Aliment Pharmacol Ther 1997; 11: 69–79.

NO-naproxen vs. naproxen: ulcerogenic, analgesic and

anti-inflammatory effects

N. M. DAVIES, A. G. RØSETH*, C. B. APPLEYARD, W. MCKNIGHT, P. DEL SOLDATO†,

A. CALIGNANO‡, G. CIRINO‡ & J. L. WALLACE

Intestinal Disease Research Unit, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada ; *Department of

Medicine, Aker University Hospital, Oslo, Norway ; †Nicox SA, Paris, France ; and ‡Department of Experimental

Pharmacology, University of Naples, Naples, Italy

Accepted for publication 3 September 1996

SUMMARY

Background : A novel class of nitric oxide-releasing

nonsteroidal anti-inflammatory drug (NO-NSAID)

derivatives has recently been described which exert

anti-inflammatory activities but produce significantly

less gastrointestinal injury than the parent NSAID from

which they are derived. The present studies were

performed to determine if a nitroxybutylester derivative

of naproxen was less ulcerogenic to the gastrointestinal

tract than its parent NSAID, and if it exerted

comparable analgesic and anti-inflammatory properties

to the parent NSAID.

Methods : The two drugs were compared in an acute

gastric injury model, an antral ulcer model and after

twice-daily administration for 18 days (small intestinal

damage model). Anti-inflammatory activity was

examined in the carrageenan-induced paw oedema

model, while analgesia was examined in the acetic

INTRODUCTION

NSAIDs have wide-spread clinical utility due to their

anti-inflammatory, anti-pyretic, analgesic and anti-

thrombotic properties. However, the use of these drugs

continues to be limited by their ability to induce mucosal

damage throughout the gastrointestinal tract. NSAID-

related gastrointestinal adverse events account for more

Correspondence to: Dr. J. L. Wallace Department of Pharmacology and

Therapeutics, University of Calgary, 3330 Hospital Drive NW, Calgary,

Alberta, T2N 4N1, Canada.

acid-induced writhing model. The pharmacokinetic

profiles of naproxen vs. NO-naproxen were compared

by HPLC analysis.

Results : NO-naproxen was found to produce

significantly less gastric damage despite inducing

similar increases in plasma TNFα to naproxen. With

chronic administration, small intestinal damage was

markedly less with NO-naproxen than with the parent

NSAID. However, NO-naproxen exerted superior

analgesic and comparable anti-inflammatory effects to

naproxen. NO-naproxen was not completely converted

to naproxen, but the reduced plasma levels of the latter

was not the underlying reason for reduced

gastrointestinal toxicity of NO-naproxen.

Conclusion : NO-naproxen represents a novel,

gastrointestinal-sparing NSAID derivative with superior

analgesic and comparable anti-inflammatory properties

to naproxen.

than 70 000 hospitalizations and 7000 deaths annually

in the USA." Naproxen is a stereochemically pure 2-

arylpropionic acid NSAID which has been shown to

cause mucosal damage throughout the gastrointestinal

tract.#,$ Naproxen has recently been marketed as an

over-the-counter medication in the United States. With

the increased use of over-the-counter NSAIDs, there is a

distinct possibility of an increase in the prevalence of

adverse effects induced by these drugs.

We have recently described a novel class of nitric oxide-

releasing nonsteroidal anti-inflammatory drug (NO-

NSAID) that exhibits markedly reduced gastrointestinal

# 1997 Blackwell Science Ltd 69

70 N. M. DAVIES et al.

toxicity, while retaining the anti-inflammatory and anti-

pyretic activity of the parent NSAIDs.%–( Analgesic

properties of NO-NSAIDs have not previously been

reported. The rationale behind the development of NO-

NSAIDswas that nitric oxide released from the compound

would counteract two events that occur subsequent to

suppression of prostaglandin synthesis by the NSAID:

reduced gastric mucosal blood flow and increased ad-

herence of neutrophils to the vascular endothelium of the

gastric microcirculation. Both of these events have been

suggested to be critical in the pathogenesis of exper-

imental NSAID-gastropathy.) However, it remains poss-

ible that this is not themechanism responsible for reduced

gastrointestinal toxicity of NO-NSAIDs. To further exam-

ine this question, the present study included an evalu-

ation of the effects of an NO-NSAID (NO-naproxen) vs. its

parent NSAID (naproxen) in a model of gastric ulceration

that has been reported to be neutrophil-independent*

and in NSAID-induced small intestinal injury, which

has also been suggested to occur independent of

neutrophils."! As tumour necrosis factor (TNFα) has

recently been proposed to be a mediator of NSAID-

induced gastric injury,"","# we compared the ability of

naproxen vs. NO-naproxen to cause TNFα release. We

also compared the pharmacokinetic behaviour of

naproxen vs. NO-naproxen to determine if any differences

between the two compounds in this regard would explain

differences in terms of gastrointestinal toxicity or

analgesic}anti-inflammatory efficacy.

METHODS

Animals

Male Wistar rats weighing 175–200 g and male Swiss

mice weighing 25–30 g were obtained from Charles

River Breeding Farms (Montreal, Quebec) and were

housed in polypropylene cages and fed standard lab-

oratory chow and tap water ad libitum. The mice were

used for the acetic acid-induced writhing experiments,

whereas rats were used in all other experiments. All

experimental protocols were approved by the Animal

Care Committee of the University of Calgary and in

accordance with the guidelines of the Canadian Council

on Animal Care.

Acute gastric Damage

Rats were deprived of food, but not water, for 18 h and

were then given naproxen orally at doses of 40 or

80 mg}kg or equimolar doses of NO-naproxen (58 or

116 mg}kg). A further group of rats was treated with an

equal volume (1 mL}kg) of the vehicle. Each group

consisted of 10 rats. The compounds were initially

dissolved in dimethylsulphoxide (DMSO), then diluted in

0.5% carboxymethylcellulose (final concentration of

DMSOwas 5%). The rats were anaesthetized with sodium

pentobarbital 3 h after drug administration and the

stomach was excised and opened by an incision along the

greater curvature. The extent of macroscopic damage

was determined by a observer unaware of the treatment

the rats had received, as described previously."$ Briefly,

this method involved measuring the length of the lesions

in mm, then summing the lengths of all lesions observed

in each stomach. After scoring, the tissue was fixed in

neutral buffered formalin and processed by routine

techniques prior to embedding in paraffin, sectioning and

staining with haematoxylin & eosin. Blood samples were

taken from the descending aorta and transferred into

plastic Eppendorf tubes and the rats were sacrificed by

cervical dislocation. The blood samples were centrifuged

for 2 min (14000 g), and the plasma was removed and

stored at ®70 °C for subsequent determination of TNFα

concentration.

Plasma TNFα concentrations

Plasma concentrations of biologically active TNFα were

determined using a modified version"# of a previously

described cytotoxicity assay."% Briefly, the murine fibro-

blasts cell line WEH 164 (5 10& cells}well ; American

Type Culture Collection, Rockville, MD) were incubated

in 96-well plates at 37 °C with 5% CO#. The culture

medium (50 µL}well) consisted of RPMI 1640 medium

containing 10% foetal bovine serum, 1% glutamine, and

1% penicillin–streptomycin (1:1) solution. Actinomycin

D (25 µL; 30 mg}mL) was added to each well, followed

by 50 µL samples of various dilutions of filtered rat

plasma or purified recombinant human TNFα as stan-

dard. The cells were incubated for 24 h at 37 °C before

the reaction was stopped by adding lysis buffer (100 µL).

This assay utilizes 3-(4,5-dimethylthiazol-2-yl)-2,5-

diphenyl-tetrazolium bromide; thiazolyl blue (MTT) as a

marker of cell viability. Live cells can metabolize MTT,

yielding a purple colour, the intensity of which can be

quantified by measuring absorbance at 620 nm using a

spectrophotometer. One international unit of TNFα

activity was defined as the concentration that would kill

50% of the cells.

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

71GASTROINTESTINAL-SPARING NAPROXEN DERIVATIVE

Refeeding model of gastric antral ulcer

The refeeding model of antral ulceration described by

Satoh et al."& was used to determine if NO-naproxen was

ulcerogenic in a model in which neutrophil adherence}activation has been reported not to be important.* Rats

(n¯6 each) were deprived of food for 20 h, fed for 2 h,

and then given naproxen (80 mg}kg p.o.) or an equi-

molar dose of NO-naproxen (116 mg}kg p.o.) before

fasting for a further 24 h. A group of control rats (n¯6)

was treated with vehicle. The rats were sacrificed

(random order) by cervical dislocation and the stomach

was excised and pinned to a wax platform, after which it

was photographed. The area of antral ulceration was

determined by computerized planimetry by an observer

unaware of the treatments the rats had received. The

stomach was then fixed in neutral buffered formalin and

processed by routine techniques for subsequent exam-

ination by light microscopy.

Small intestinal injury

Rats were divided into two groups (n¯6 each) and

placed in individual metabolic cages to facilitate complete

collection of faeces. The rats were orally dosed with

naproxen (20 mg}kg) or an equimolar dose of NO-

naproxen (29 mg}kg) twice-daily for 1 week. At the end

of this period, the doses of naproxen and NO-naproxen

were increased by 50% (to 30 and 44 mg}kg, respect-

ively) and treatment was continued on a twice-daily

basis for 7 days. At the end of the second week of

treatment, the doses of naproxen and NO-naproxen were

further increased by 50% (to 45 mg}kg and 66 mg}kg,

respectively). Originally it was our intention to continue

twice-daily dosing for 1 week. However, as the rats

treated with naproxen appeared moribund after 3 days of

treatmentwith the 45 mg}kg dose, we opted to terminate

the experiment at that time. The rats were anaesthetized

with ether 3 h after the final dose of NSAID or NO-NSAID

and blood was removed by cardiac puncture for measure-

ment of haematocrit. The rats were then killed by cervical

dislocation. An additional six untreated rats were also

killed at this time in order to provide ‘baseline ’ data on

haematocrit and small intestinal damage. A 10 cm

segment of ileum extending proximally from a point

10 cm proximal to the ileo-caecal valve was excised and

was rinsed with 10 mL of phosphate buffered saline. The

segments were ligated and filled with 2 mL of Carnoy’s

fixative, then submerged in Carnoy’s fixative for 1 h. The

segments were then opened along the anti-mesenteric

side, stapled to cardboard sheets and photographed for

computerized planimetric determination of ulcer area.

This analysis was performed by an individual unaware of

the treatments the rats had received. Data were expressed

as percentage of the total area of the segment that

exhibited ulcers. To ensure that damage being assessed

were indeed ulcers, sections of tissue were obtained from

damaged areas for blind histological evaluation. Tissue

sections were embedded in plastic using a commercially

available kit (JB-4 embedding kit, Polysciences. Inc.

Warrington, PA). Thin sections (1–1.5 µm) were stained

with Lee’s methylene blue-basic fuchsin and were

examined by light microscopy.

Body weights were recorded daily. Complete stool

collections were performed on the day prior to beginning

drug administration (control period) and on the second

day of administration of each of the three doses of each

drug. The samples were weighed, then frozen in plastic

containers at ®70 °C until the assay for granulocyte

marker protein (GMP) was performed. For this analysis,

the faecal samples were thawed and homogenized for

20–30 s on ice in 4 mL of buffer containing of 0.25 m

Thimerosal and 10 m CaCl#

at pH 8.4 per gram of

faeces. The samples were then centrifuged at 4 °C for

20 min at 10 000 g and the supernatant was removed

and stored at ®70 °C until analysis was performed. GMP

was quantified by enzyme-linked immunosorbent assay

as previously described."',"( GMP is a stable 70-kDa, iron-

binding protein isolated from rat neutrophils, and is

stable in stools for at least 3 days. It has previously been

suggested to be a sensitive, non-invasive marker of

gastrointestinal inflammation."',"(

Pharmacokinetic studies

Rats were anaesthetized with halothane and the left

jugular vein was cannulated with polyethylene tubing

tipped with silastic tubing. The rats were allowed to

recover overnight before the pharmacokinetic studies

were performed. Groups of rats (n¯4 in each) were

orally dosed with either naproxen (3 or 30 mg}kg) or

NO-naproxen (4.4 or 44 mg}kg). Whole blood samples

(0.25 mL) were withdrawn from the cannula before and

0.5–48 h after drug administration. The whole blood

samples were immediately centrifuged, and the plasma

transferred to new 1 mL polypropylene vials and stored

at ®20 °C until analysis.

The plasma samples were analysed using a reverse-

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

72 N. M. DAVIES et al.

phase HPLC method. Briefly, to a 100 µL aliquot of serum

were added 25 µL of mefenamic acid (internal standard;

0.1 mg}mL in acetonitrile) and 1 mL of ice-cold aceto-

nitrile. The mixture was vortex-mixed for 30 s, then

centrifuged for 10 min at 1000 g. The supernatant

was then transferred to clean glass culture tubes and

evaporated to dryness. The residue was reconstituted

in 500 µL of mobile phase and aliquots of 50 µL were

injected into the HPLC system. The HPLC system was a

Hewlett Packard (HP) 1050 series with a variable

wavelength detector. A 5 µm C18 analytical column (HP

Spherisorb ODS2 250¬4 mm) was used. The mobile

phase consisted of acetonitrile}3% acetic acid (60}40) at

a flow rate of 0.8 mL}min. All analyses were performed

at ambient temperature. The detection wavelength used

was 254 nm.

A calibration curve was prepared over a range of

0.5–100 µg}mL in rat plasma for each analysis.

Naproxen and NO-naproxen concentrations in the

samples were determined by interpolation from cali-

bration curves. Plots of the area ratios of the peaks

corresponding to naproxen and NO-naproxen and

internal standard were linear from 0.5–100 µg}mL

naproxen or NO-naproxen and the interday variation of

the calibration curve slope was less than 2%. The peaks

corresponding to naproxen, mefenamic acid and NO-

naproxen eluted at 4.9 min, 10.2 min and 15.1 min,

respectively.

The area under the plasma concentration vs. time

(AUC!–¢), peak plasma concentration (C

max), the time of

its attainment (tmax

) and terminal elimination half-life

(t"/#

) were estimated by non-compartmental analysis and

linear trapezoidal approximation of the experimental

data points using WN (Scientific Consulting Inc.,

Apex, NC).

Analgesic effects

Mice were given vehicle (DMSO), naproxen (0.3–

10 mg}kg) or equimolar doses of NO-naproxen (n¯12

for vehicle, 5–7 for each dose of the test compounds)

orally 1 h prior to intraperitoneal injection of 0.5 mL of

0.6% acetic acid. The number of writhes or abdominal

constriction responses displayed by each mouse was

counted over a 20-min period beginning 5 min after

acetic acid administration. The individual recording

these responses was unaware of the treatments each

mouse had received.

Anti-inflammatory effects

Groups of five rats each were orally treated with vehicle,

naproxen (3 or 30 mg}kg) or equimolar doses of NO-

naproxen (4.4 or 44 mg}kg) and 1 h later were anaes-

thetized with ether. A 0.1% solution of lambda

carrageenan (0.1 mL) was injected into the right hind

foot pad, as described previously."( The volume of the

paw was measured by hydroplethysmometry (Ugo Basile,

Italy) prior to carrageenan injection and every hour

thereafter for 5 h by an observer unaware of the

treatment. Changes in paw volume relative to the

measurement taken before carrageenan administration

were calculated.

Statistical analysis

All data are expressed as the mean³S.E.M. Groups of

data were compared using an analysis of variance and

the Student–Neuman–Keuls test, except for the GMP

excretion data, which were compared using a paired t-

test with Bonferroni correction. With all analyses,

differences were considered to be significant when

P!0.05.

Materials

Naproxen, mefenamic acid, Thimerosal and MTT were

obtained from Sigma Chemical Company (St. Louis, MO).

Acetonitrile and carboxymethylcellulose and CaCl#were

obtained from BDH Chemicals Ltd. (Edmonton, AB,

Canada). The RPMI 1640 medium was obtained from

Canadian Life Technologies (Burlington, ON, Canada).

Human recombinant TNFα was obtained from R&D

Systems (Minneapolis, MN). NO-naproxen was

synthesized by Nicox SA (Paris, France). Halothane was

obtained from Halocarbon Laboratories (River Edge, NJ).

All other reagents were obtained from Fisher Scientific

Ltd. (Edmonton, AB, Canada).

RESULTS

Acute gastric damage

Oral administration of naproxen caused extensive haem-

orrhagic damage in the corpus region of the stomach at

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

73GASTROINTESTINAL-SPARING NAPROXEN DERIVATIVE

Figure 1. Severity of gastric damage (upper panel) and plasma

tumour necrosis factor-α (TNFα) levels (lower panel) following

oral administration of naproxen or NO-naproxen. The doses of

NO-naproxen (58 and 116 mg}kg) represent equimolar doses to

those of naproxen (40 and 80 mg}kg, respectively). Blood

samples were taken and damage was assessed 3 h after

administration of the drugs or vehicle. Bars are means³S.E.M. of

10 rats per group. *P!0.05, **P!0.01 compared to the

vehicle-treated group.

each of the doses tested (Figure 1). The lesions were

usually linear and were often located on the crests of the

rugal folds. Histological examination of the tissue con-

firmed that the injury induced by naproxen was limited

to the mucosal layer, not penetrating the muscularis

mucosae. Oral administration of NO-naproxen caused

only marginal damage in the corpus region of the

stomach at each of the doses tested, with average gastric

damage scores of 1.6³0.8 and 4.8³3.2 mm for the 58

and 116 mg}kg doses, respectively, not significantly

different from that observed in vehicle-treated rats. The

lack of erosion formation following administration of NO-

naproxen or vehicle was confirmed by histology.

Figure 2. Effects of twice-daily oral administration of NO-

naproxen or naproxen on body weight. The dose of NO-naproxen

given was equimolar to that of naproxen. The dose of naproxen

was 20 mg}kg for the first week, 30 mg}kg for the second week

and 45 mg}kg thereafter. The arrows indicate days on which

faecal samples were processed for measurement of granulocyte

marker protein excretion. Each group consisted of six rats. The

mean body weights in the two groups did not differ significantly

on day 1, but did differ significantly (P!0.05) on all days

thereafter.

Plasma TNFα concentrations

Oral administration of naproxen or NO-naproxen

resulted in significant increases in plasma TNFα concen-

trations compared to vehicle-treated rats (Figure 1).

There were no significant differences in plasma TNFα

concentrations between the naproxen and NO-naproxen

groups.

Refeeding model of antral ulcer

Administration of naproxen to rats in this refeeding

paradigm resulted in the development of discrete, spheri-

cal ulcers in the antrum, which were histologically

confirmed to penetrate the muscularis mucosae. The

mean size of these ulcers was 3.9³1.0 mm#, significantly

greater than the small amount of damage observed in

vehicle-treated rats (0.1³0.1 mm# ; P 0.05). In rats

treated with NO-naproxen, the extent of antral injury

(0.2³0.2 mm#) did not differ significantly from that

observed in vehicle-treated rats.

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

74 N. M. DAVIES et al.

Figure 3. Effects of chronic twice daily administration of NO-

naproxen and naproxen on faecal granulocyte marker protein

(GMP) excretion. Results are expressed as percentage of the

excretion levels on the day prior to administration of the test

drugs (control). Each group consisted of six rats. NO-naproxen

was given at equimolar doses to those of naproxen. See Figure 2

for treatment regimen. *P!0.05, **P!0.01 compared to the

corresponding control data using a paired t-test with Bonferroni

correction.

Small intestinal injury

The body weights of rats treated daily with naproxen or

NO-naproxen did not differ significantly at the outset of

the study (Figure 2). However, within a day of beginning

treatment and on all days thereafter, body weights in the

naproxen-treated ratswere significantly lower than those

in rats treated with NO-naproxen (P!0.05).

Prior to beginning drug administration, faecal GMP

concentration and faecal wet weight did not differ

significantly between the two groups of rats. In the rats

that received naproxen, basal mean fecal GMP con-

centration was 27.7³6.8 µg}g and mean faecal wet

weight was 20.6 (0.8 g (total faecal GMP excretion of

570.6³140.1 µg). In rats in the NO-naproxen group,

basal mean faecal GMP concentration was

29.5³4.2 µg}g and mean faecal wet weight was

22.4³0.8 g (total faecal GMP excretion of

660.8³94.1 µg). As shown on Figure 3, faecal GMP

excretion significantly increased over basal levels in the

rats treated with naproxen, but not in rats treated with

NO-naproxen. The increase in faecal calprotecin ex-

cretion in the naproxen-treated rats was due to a

significant increase both in faecal GMP concentration

and faecal wet weight.

Figure 4. Effects of repeated oral administration of naproxen, NO-

naproxen or vehicle on intestinal ulcer area (upper panel) and

haematocrit (lower panel). The rats were treated with equimolar

doses of the drugs for a total of 18 days, with the doses

increasing by 50% on days 8 and 15 (see Figure 2). **P!0.01,

***P!0.001 compared to the control group (no drug

treatment).

As the rats treated with naproxen appeared moribund

on day 18 of the study, we decided not to proceed with

further administration of the test drugs (in accordance

with the policy of our institutional Animal Care Com-

mittee). Examination of the ileum of rats treated with

naproxen revealed penetrating ulcers along the mes-

enteric border. These lesions were histologically con-

firmed to extend to the depth of the muscularis propria.

On average, these ulcers occupiedE3.7% of the mucosal

surface of the segment examined (Figure 4). In rats

treated with NO-naproxen, ulcers were not observed, as

was the case in untreated control rats sacrificed at the

same time.

As shown in the lower panel of Figure 4, rats treated

with naproxen had a significantly lower haematocrit

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

75GASTROINTESTINAL-SPARING NAPROXEN DERIVATIVE

Figure 5. Mean plasma concentrations of naproxen following

oral administration of naproxen 3 mg}kg or NO-naproxen

4.4 mg}kg. NO-naproxen itself was not detectable in plasma

when administered at this dose.

than untreated controls, whereas the haematocrit in rats

treated with NO-naproxen was unchanged from controls.

Pharmacokinetic studies

Figure 5 illustrates the plasma naproxen levels following

oral administration of naproxen vs. NO-naproxen, with

further pharmacokinetic parameters listed in Table 1. At

the lower doses tested, NO-naproxen delivered signifi-

cantly less naproxen into the plasma than an equimolar

dose of naproxen itself (note reduced AUC!–¢ and C

max).

At the higher doses tested, the significant difference in

maximal concentration of naproxen achieved was still

evident (Table 1).

As markedly less naproxen was ‘delivered’ to the

plasma following administration of NO-naproxen than

with an equimolar dose of naproxen, we speculated that

Drug AUC!–¢ C

maxtmax

t"#

(mg}kg) (µg.h}mL) (µg}mL) (h) (h)

Naproxen (3) 108.3³13.7 10.5³0.9 1.7³0.3 7.60³2.70

NO-naproxen (4.4) 39.3³6.7* 5.8³2.6* 1.5³0.5 7.84³2.43

Naproxen (30) 584.4³67.1 64.9³5.7 1.6³0.8 4.56³0.87

NO-naproxen (44) 354.2³84.4 24.9³3.8** 2.3³0.6 6.72³1.47

Blood samples were taken from 0.5–48 h after a single oral administration of the test

drugs and the levels of naproxen in plasma were determined by HPLC. AUC!–¢ : area

under the plasma concentration vs. time. Cmax

¯ peak plasma concentration, tmax

¯ the

time of attainment of Cmax

, t"#

¯ terminal elimination half-life. At the doses shown in

this table, intact NO-naproxen was never detectable in plasma. *P!0.05,**P!0.01

compared to the corresponding naproxen group.

Table 1. Pharmacokinetic parameters for

naproxen and NO-naproxen

this might be the underlying reason for the relative lack

of gastrointestinal damage with NO-naproxen. In order

to test this hypothesis, we performed pharmacokinetic

studies in a group of rats (n¯4) given NO-naproxen at a

dose of 116 mg}kg, a dose which was found not to cause

significant damage to the stomach (see Figure 1).

Administration of this dose of NO-naproxen resulted in

high plasma levels of naproxen. Indeed, the maximal

plasma concentration (Cmax

) achieved with this dose of

NO-naproxen was 1164.2³175.0 µg}mL, far greater

than that achieved after administration of naproxen at a

dose of 30 mg}kg, which produced significant gastric

injury (Figure 1). Thus, reduced delivery of naproxen to

the serum by NO-naproxen did not appear to account for

the reduced gastrointestinal toxicity of the latter com-

pound. While intact NO-naproxen was not detectable in

plasma following administration of the 4.4 or 44 mg}kg

doses, it was detectable for several hours following

administration of the 116 mg}kg dose (Cmax

of

409.1³108.4 µg}mL).

Analgesic activity

Vehicle-treated control mice given acetic acid intra-

peritoneally exhibited extensive writhing and abdominal

contractions (mean of 46.5³3.0 per 20 min). NO-

naproxen significantly reduced the incidence of writhing

at the dose of 0.44 mg}kg, while a dose of naproxen of

3 mg}kg was required to observe significant inhibition

(Figure 6). Additional experiments were performed to

determine the lowest effective dose of NO-naproxen. At

0.15 mg}kg, NO-naproxen inhibited writhing by

38.0³2.2% (P!0.05), while at 0.05 mg}kg, the com-

pound had no significant effect (13.4³4.2% inhibition).

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

76 N. M. DAVIES et al.

Figure 6. Inhibition of acetic acid-induced writhing in mice by

naproxen and NO-naproxen. Asterisks denote significant

reductions in the number of writhing movements observed over

a 20-min period following intraperitoneal administration of

acetic acid. NO-naproxen was given at doses equimolar to those

of naproxen. **P!0.01 compared to vehicle-treated controls.

Anti-inflammatory activity

As shown in Figure 7, both naproxen (3 mg}kg) and NO-

naproxen (4.4 mg}kg) significantly reduced the paw

oedema stimulated by carrageenan compared to the

vehicle-treated control group. There were no significant

differences between the two drugs at any of the time

points examined. When higher doses of naproxen

(30 mg}kg) and NO-naproxen (44 mg}kg) were tested, a

similar reduction of oedema was observed and again,

there were no significant differences between the effects

observed with the two drugs (data not shown).

DISCUSSION

Consistent with our findings for other NSAIDs,%–' the

addition of a nitric oxide-releasing moiety to naproxen,

through an ester linkage, greatly reduced the ability of

this compound to produce acute gastric mucosal damage.

This type of damage has been shown to be neutrophil-

dependent"*,#! and to be associated with marked

increases in plasma levels of TNFα."","# Moreover, TNFα

appears to contribute to the pathogenesis of the acute

gastric damage caused by NSAIDs, since inhibitors of

TNFα synthesis and an antibody against TNFα signifi-

cantly reduced the damaging effects of NSAIDs in

rats."","# Interestingly, despite the greatly reduced gastric

damage observed with NO-naproxen vs. naproxen, these

compounds caused similar increases in plasma TNFα

release. This suggests that whatever the mechanism

through which the NO-releasing moiety reduces ‘ulcero-

genicity’, it does so independent of the effects of the

compound on TNFα release. TNFα release caused by

NSAIDs is likely linked to the ability of these drugs to

block prostaglandin synthesis.#" Since previous studies

have shown that addition of the NO-releasing moiety to

NSAIDs does not interfere with their ability to block

prostaglandin synthesis,%,&,## it is perhaps not surprising

that similar elevations in plasma TNFα occurred fol-

lowing administration of NO-naproxen vs. naproxen.

The potential pro-ulcerogenic effects of NO-naproxen

were also examined in a model of chronic-type ulcers in

the antrum region of the stomach. This model, unlike the

acute injury caused by NSAIDs, has been suggested to

occur through neutrophil-independent mechanisms.* As

one of the mechanisms suggested to account for the

reduced gastric toxicity of NO-NSAIDs is an inhibition of

neutrophil adherence to the vascular endothelium,% it

might be anticipated that these agents would not exhibit

reduced toxicity in a model of ulceration that is not

neutrophil-dependent. However, while naproxen con-

sistently produced antral ulcers, NO-naproxen was found

not to produce significant injury.

Similar differences between NO-naproxen and its parent

NSAID, in terms of gastrointestinal damage, were

observed when the drugs were administered repeatedly

over an 18-day period. Daily dosing with naproxen

resulted in reduced body weight gain, a significant

decrease in haematocrit and the development of pen-

etrating ulcers in the ileum. The decrease in haematocrit

was presumably due to bleeding from the gastrointestinal

tract. Naproxen also caused a significant increase in

faecal excretion of GMP, a protein found mainly in

neutrophils. Faecal excretion of the neutrophil markers

GMP and calprotectin has previously been used to non-

invasively detect gastrointestinal mucosal inflammation

in healthy volunteers and animals given a number of

NSAIDs, including naproxen."',#$ In sharp contrast to

our findings with naproxen, daily administration of

equimolar doses of NO-naproxen had no effect on

haematocrit, did not cause ileal ulcer development and

did not significantly affect faecal GMP excretion. While it

has been suggested that NSAID-induced enteropathy

occurs via a neutrophil-independent mechanism,"!

others have proposed a key role for neutrophil influx into

the mucosa in the pathogenesis of this disorder.#% It is

entirely possible that one of the reasons for reduced

intestinal damage with NO-naproxen vs. naproxen is

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

77GASTROINTESTINAL-SPARING NAPROXEN DERIVATIVE

Figure 7. Reduction of carrageenan-

induced paw oedema by orally

administered naproxen (3 mg}kg) or an

equimolar dose of NO-naproxen

(4.4 mg}kg). Both drugs significantly

reduced oedema (P!0.05) compared to

the vehicle-treated controls at each time

point examined. There were no significant

differences between the effects of the two

drugs at any time.

related to a diminution of neutrophil influx into the

intestinal tissue, as manifest by reduced faecal GMP

excretion. As mentioned above, NO-NSAIDs have pre-

viously been shown not to induce neutrophil adherence

to the vascular endothelium, in contrast to the parent

NSAID.% Nitric oxide has well characterized inhibitory

effects on neutrophil activation}adherence.#&,#'

While NO-naproxen exhibited markedly reduced gas-

trointestinal toxicity compared to naproxen, the anti-

inflammatory and analgesic properties of the former

were comparable or better than the latter. In the

carrageenan-induced paw oedema model, NO-naproxen

and naproxen exhibited comparable effects. NO-

naproxen has also been shown to be effective in a rat

model of chronic inflammation.#( On the other hand, in

the analgesia model, NO-naproxen was found to produce

significant effects at doses much lower than the effective

doses of naproxen. It is possible that the enhanced

analgesic effects of this compound were attributable to

anti-nociceptive effects of the NO-releasing moiety. Nitric

oxide has been suggested to exert anti-nociceptive effects.

For example, nitric oxide derived from -arginine was

shown to potentiate the anti-nociceptive effects of β-

endorphin in the mouse,#) and to inhibit electrically or

chemically evoked nociceptive signals in the rat.#*

Pharmacokinetic studies were performed to determine

the extent to which NO-naproxen ‘delivered’ naproxen

into the plasma. If substantially less naproxen reached

the systemic circulation from NO-naproxen than when

equimolar doses of the parent NSAID itself were given,

this might explain in part the reduced gastrointestinal

toxicity. Indeed, we observed that the amount of

naproxen reaching the plasma from NO-naproxen was

only about half that achieved with an equimolar dose of

naproxen. However, this could not account for the

differences in gastrointestinal tolerability. In one series of

experiments, a much higher dose of NO-naproxen

(116 mg}kg) was administered. This dose of NO-

naproxen did not cause significant gastric damage, but

was found to produce very high plasma levels of

naproxen. Indeed, these plasma levels of naproxen far

exceeded those achieved by a smaller, but gastric

damage-inducing dose (30 mg}kg) of the parent com-

pound. It is noteworthy that NO-naproxen itself was not

detectable in plasma when the drug was given at doses in

the range that produced anti-inflammatory and analgesic

effects. However, when administered at a dose of

116 mg}kg, NO-naproxen could be detected in plasma. It

is possible that intactNO-naproxenwas present in plasma

at levels below the limits of detection when lower doses of

the drug were administered. It is also possible that any

intact NO-naproxen in plasma contributed to the anti-

inflammatory and analgesic effects observed with this

compound. In previous studies, we found that intact NO-

NSAIDs were capable of inhibiting cyclo-oxygenase in

vitro.##

Why did NO-naproxen exhibit greatly reduced gas-

trointestinal toxicity despite exerting comparable anti-

inflammatory effects and superior analgesic effects to

naproxen? There are several possible mechanisms which

could account for this apparent discrepancy. As discussed

above, nitric oxide is a potent inhibitor of leukocyte

activation and adherence,#&,#' processes which may

contribute to the pathogenesis of gastrointestinal injury

in at least some of the models utilized in this study."*,#!,#%

Nitric oxide is also capable of scavenging oxygen-derived

free radicals, such as superoxide anion.$! There is

evidence that reactive oxygen metabolites, including

superoxide anion, play an important role in the patho-

genesis of experimental NSAID gastropathy.$" It is poss-

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

78 N. M. DAVIES et al.

ible that nitric oxide release from NO-naproxen, as has

been demonstrated with other NO-NSAIDs possessing

the same NO-releasing moiety,%,&,$# results in mainten-

ance of gastrointestinal blood flow during the period of

luminal exposure to the NSAID. NO-releasing com-

pounds can also increase mucus gel thickness in the rat

stomach$$ and the secretion ofmucus by gastric epithelial

cells,$% which could conceivable reduce the topical

irritant effects of naproxen on the gastrointestinal epi-

thelium and therefore reduce the severity of injury. Nitric

oxide has also been suggested to directly modulate the

permeability of the intestinal epithelium.$& As luminal

bacterial products may provide the chemotactic drive to

recruit neutrophils into the intestinal mucosa and

eventually to the lumen, a reduction of intestinal

epithelial permeability by nitric oxide could influence

susceptibility to NSAID-induced injury. Further studies

are required to determine if these mechanisms, or others,

underlie the reduced gastrointestinal toxicity of NO-

naproxen and other NO-NSAIDs. In conclusion, these

studies suggest that NO-naproxen is an anti-inflam-

matory and analgesic drug with a much more favourable

gastrointestinal toxicity profile than naproxen.

ACKNOWLEDGEMENTS

This work was supported by a grant from the Medical

Research Council of Canada (MRC). Dr J. L. Wallace is an

MRC Senior Scientist and an Alberta Heritage Foun-

dation for Medical Research Scientist. Dr N. M. Davies is

supported by a Fellowship co-sponsored by the Canadian

Association of Gastroenterology, Astra Canada and the

MRC.

REFERENCES

1 Fries JF. NSAID gastropathy: epidemiology. J Musculoskeletal

Med 1991; 8: 21–8.

2 Aabakken L, Osnes M. Gastroduodenal lesions induced by

naproxen. An endoscopic evaluation of regional differences

and a natural course. Scand J Gastroenterol 1990; 25:

1215–22.

3 Uribe A, Johansson C, Slezak P, Rubio C. Ulcerations of the

colon associated with naproxen and acetylsalicylic acid treat-

ment. Gastrointest Endosc 1986; 32: 242–4.

4 Wallace JL, Reuter B, Cicala C, McKnight W, Grisham MB,

Cirino G. Novel nonsteroidal anti-inflammatory drug deriva-

tives with markedly reduced ulcerogenic properties. Gastro-

enterology 1994; 107: 173–9.

5 Wallace JL, Reuter B, Cicala C, McKnight W, Grisham MB,

Cirino G. A diclofenac derivative without ulcerogenic proper-

ties. Eur J Pharmacol 1994; 257: 249–55.

6 Reuter B, Cirino G, Wallace JL. Markedly reduced intestinal

toxicity of a diclofenac derivative. Life Sci 1994; 55: PL1–8.

7 Wallace JL, Pittman QJ, CirinoG. Nitric oxide releasing NSAIDs:

a novel class of GI-sparing anti-inflammatory drugs. In:

Proznansky W, ed. New Molecular Approaches to Anti-

Inflammatory Therapy. Basle : Birkauser Verlag, 1995: 121–9.

8 Wallace JL. Mechanisms of NSAID-induced gastrointestinal

damage—potential for development of GI-safe NSAIDs. Can J

Physiol Pharmacol 1994; 71: 98–102.

9 Trevethick MA, Clayton NM, Bahl AK, Sanjar S, Strong P.

Neutrophil depletion does not prevent indomethacin-induced

ulceration of the rat gastric antrum. Agents Actions 1994; 41:

C226–7.

10 Yamada T, Deitch E, Specian RD, Perry MA, Sartor RB, Grisham

MB. Mechanisms of acute and chronic intestinal inflammation

induced by indomethacin. Inflammation 1993; 17: 641–62.

11 Santucci L, Fiorucci S, Giansanti M, Brunori PM, Di Matteo FM,

Morelli A. Pentoxifylline prevents indomethacin induced acute

gastric mucosal damage in rats : role of tumour necrosis factor-

α. Gut 1994; 35: 909–15.

12 Appleyard CB, McCafferty D-M, Tigley AW, Swain MG, Wallace

JL. Tumor necrosis factor mediation of NSAID-induced gastric

damage: role of leukocyte adherence. Am J Physiol 1996; 270:

G42–8.

13 Wallace JL, Whittle BJR. Role of prostanoids in the protective

actions of BW755C on the gastric mucosa. Eur J Pharmacol

1985; 115: 45–52.

14 Mosman T. Rapid colorimetric assay for cellular growth and

survival : application to proliferation and cytotoxic assays. J

Immunol Meth 1983; 65: 55–63.

15 SatohH, GuthPH, GrossmanMI. Role of food in gastrointestinal

ulceration produced by indomethacin in the rat. Gastro-

enterology 1982; 83: 210–5.

16 Foster R, Røseth AG, Fagerhol MK, Peters TJ, Somadsundaram

S, Bjarnason I. Indomethacin enteropathy in rats : assessed by

a novel granulocytic marker protein (GMP). Proc World

Congress Gastroenterol 1995; 1958 (Abstract).

17 Kristinsson J, Røseth AG, Sundset A, Nygaard K, Aadland E,

Løberg EM, Paulsen JE, Fagerhol MK. Granulocyte marker

protein (GMP) is elevated in stools from rats with azoxymethane

induced colon cancer and aberrant crypts. Gastroenterology

1996; 110: A546 (Abstract).

18 Cirino G, Peers SH, Flower RJ, Browning JL, Pepinsky RB.

Human recombinant lipocortin 1 has acute local anti-

inflammatory properties in the rat paw edema test. Proc Natl

Acad Sci USA 1989; 86: 3428–32.

19 Wallace JL, Keenan CM, Granger DN. Gastric ulceration

induced by nonsteroidal anti-inflammatory drugs is a

neutrophil-dependent process. Am J Physiol 1990; 259:

G462–7.

20 Wallace JL, Arfors KE, McKnight GW. A monoclonal antibody

against the CD18 leukocyte adhesion molecule prevents

indomethacin-induced gastric damage in the rabbit. Gastro-

enterology 1991; 100: 878–83.

21 Kunkel SL, Remick DG, Strieter RM, Larrick JW. Mechanisms

that regulate the production and effects of tumour necrosis

factor-α. Crit Rev Immunol 1989; 9: 93–117.

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79

79GASTROINTESTINAL-SPARING NAPROXEN DERIVATIVE

22 Mitchell JA, Cirino G, Akarasereenont P, Wallace JL, Flower RJ,

Vane JR. Flurbinitroxybutylester : a novel anti-inflammatory

drug devoid of ulcerogenic activity, inhibits cyclo-oxygenase-1

and cyclo-oxgenase-2. Can J Physiol Pharmacol 1994; 72:

270 (Abstract).

23 Meling TR, Aabakken L, Roseth A, Osnes M. Faecal calprotectin

shedding after short-term treatment with non-steroidal anti-

inflammatory drugs. Scand J Gastroenterol 1996; 31: 339–44.

24 Bjarnason I, Hayllar J, Smethurst P, Price A, Gumpel MJ.

Metronidazole reduces intestinal inflammation and blood loss

in non-steroidal anti-inflammatory drug induced enteropathy.

Gut 1992; 33: 1204–8.

25 May GR, Crook P, Moore PK, Page CP. The role of nitric oxide

as an endogenous regulator of platelet and neutrophil ac-

tivation within the pulmonary circulation of the rabbit. Br J

Pharmacol 1991; 102: 759–63.

26 Kubes P, Suzuki M, Granger DN. Nitric oxide : an endogenous

modulator of leukocyte adhesion. Proc Natl Acad Sci USA

1991; 88: 4651–5.

27 Cuzzolin L, Conforti A, Adami A, Lussignoli S, Memestrina F,

Del Soldato P, Benoni G. Anti-inflammatory potency and

gastrointestinal toxicity of a new compound, NO-naproxen.

Pharmacol Res 1995; 31: 61–5.

28 Tseng LF, Xu JY, Pieper GM. Increase of nitric oxide production

by -arginine potentiates i.c.v. administered β-endorphin-

induced antinociception in the mouse. Eur J Pharmacol 1992;

212: 301–3.

29 Haley JE, Dickenson AH, Schachter M. Electrophysiological

evidence for a role of nitric oxide in prolonged chemical

nociception in the rat. Neuropharmacology 1992; 31: 251–8.

30 Rubanyi GM, Ho EH, Cantor EH, Lumma WC, Botelho LH.

Cytoprotective function of nitric oxide : inactivation of super-

oxide radicals produced by human leukocytes. Biochem

Biophys Res Commun 1991; 181: 1392–7.

31 Vaananen PM, Meddings JB, Wallace JL. Role of oxygen-

derived free radicals in indomethacin-induced gastric injury.

Am J Physiol 1991; 261: G470–5.

32 Wallace JL, McKnight W, Del Soldato P, Cirino G. Anti-

thrombotic properties of a nitric oxide-releasing, gastric sparing

aspirin derivative. J Clin Invest 1995; 96: 2711–8.

33 Brown JF, Hanson PJ, Whittle BJR. Nitric oxide donors increase

mucus gel thickness in rat stomach. Eur J Pharmacol 1992;

223: 101–4.

34 Brown JF, Keates AC, Hanson PJ, Whittle BJR. Nitric oxide

generators and cGMP stimulate mucus secretion by rat gastric

mucosal cells. Am J Physiol 1993; 265: G418–22.

35 Kubes P. Nitric oxide modulates epithelial permeability in the

feline small intestine. Am J Physiol 1992; 262: G1138–42.

# 1997 Blackwell Science Ltd, Aliment Pharmacol Ther 11, 69–79