Design, synthesis and biological testing of a novel series of anti-inflammatory drugs

Article abstract of DOI:10.1211/0022357011778043

Many of the non-steroidal anti-inflammatory drugs (NSAIDs) currently marketed produce severe gastro-toxic side effects. The benefits of producing NSAIDs without these side effects are obvious, particularly for patients requiring long-term therapy. The aim of this investigation was to produce novel NSAIDs, based on paracetamol, that exhibit little or no gastro-toxicity. The work covers design, synthesis and testing of 13 drug candidates. The analgesic and anti-inflammatory potencies of the drug candidates were measured using the mouse abdominal constriction assay and the carrageenan-induced rat paw oedema assay, respectively. The stomachs of the rats were examined post-mortem, to assess the gastro-toxicity of the drugs. Of the 13 compounds described herein, 11 were shown to possess analgesic activity at 2-10 times the potency of aspirin, while 8 demonstrated anti-inflammatory activity at 3-10 times the potency of aspirin. Significantly, all of the compounds showed very low gastro-toxicity when compared with aspirin. The results of this study indicate that it is possible to develop novel, potent NSAIDs based on the structure of paracetamol. These compounds have the advantage of demonstrating much lower gastro-toxicity than NSAIDs currently available. Drugs of this type may, in future, provide effective treatments for inflammatory disorders.

Full text of DOI:10.1211/0022357011778043

Journal of
Pharmacy and Pharmacology
JPP 2001, 53: 1505–1514
# 2001 The Authors
Received February 15, 2001
Accepted July 16, 2001
ISSN 0022-3573
Design, synthesis and biological testing of a novel
series of anti-inflammatory drugs
Judith C. Duffy, John C. Dearden and Chris Rostron
Abstract
Many of the non-steroidal anti-inflammatory drugs (NSAIDs) currently marketed produce
severe gastro-toxic side effects. The benefits of producing NSAIDs without these side effects are
obvious, particularly for patients requiring long-term therapy. The aim of this investigation was
to produce novel NSAIDs, based on paracetamol, that exhibit little or no gastro-toxicity. The
work covers design, synthesis and testing of 13 drug candidates. The analgesic and anti-
inflammatory potencies of the drug candidates were measured using the mouse abdominal
constriction assay and the carrageenan-induced rat paw oedema assay, respectively. The
stomachs of the rats were examined post-mortem, to assess the gastro-toxicity of the drugs. Of
the 13 compounds described herein, 11 were shown to possess analgesic activity at 2–10 times
the potency of aspirin, while 8 demonstrated anti-inflammatory activity at 3–10 times the
potency of aspirin. Significantly, all of the compounds showed very low gastro-toxicity when
compared with aspirin. The results of this study indicate that it is possible to develop novel,
potent NSAIDs based on the structure of paracetamol. These compounds have the advantage
of demonstrating much lower gastro-toxicity than NSAIDs currently available. Drugs of this
type may, in future, provide effective treatments for inflammatory disorders.
Introduction
Non-steroidal anti-inflammatory analgesic drugs (NSAIDs) are used to alleviate
mild to moderate pain. They are also used for the treatment of inflammatory
disorders such as arthritis and spondylitis. Although there are more than 20 drugs
of this type regularly prescribed in the UK, there are several reasons why the
production of new drugs is desirable.
Patients show considerable variation in their response to NSAID therapy. Only
60% of patients will respond to any particular NSAID (Mehta 1997) and in the
treatment of specific disease states one drug may be more effective than another.
Thus, it is to the advantage of the physician to have a wide range of drugs from
which to choose. Another consideration is the incidence of gastro-intestinal side
effects, termed NSAID gastropathy, which result from treatment with standard
NSAIDs. These side effects include discomfort, nausea, diarrhoea, vomiting,
petechial lesion and ulcer formation. Gastro-intestinal intolerance is the main
reason for cessation of anti-inflammatory therapy. Apart from the distress suffered
by the patient, the cost of treating NSAID-induced side effects amounts to billions
of dollars annually (Wallace & Granger 1992). For these reasons, this study was
undertaken to design novel NSAIDs that would show little or no gastro-intestinal
toxicity.
School of Pharmacy & Chemistry,
Liverpool John Moores
University, Byrom St., Liverpool,
L3 3AF, UK
Judith C. Duffy, John C.
Dearden, Chris Rostron
Correspondence: J. C. Duffy,
School of Pharmacy and
Chemistry, Liverpool John
Moores University, Byrom St.,
Liverpool, L3 3AF, UK.
1505

1506
Judith C. Duffy et al
Table 1 Structures of compounds synthesised.
Paracetamol itself was selected as the parent structure
for all of the drugs in this study. The potential use of
paracetamol derivatives presented an interesting op-
portunity for drug development. Paracetamol is not an
anti-inflammatory drug, but its derivatives have been
found here to possess anti-inflammatory activity. Para-
cetamol does not exhibit gastro-toxicity, and this
favourable characteristic has been maintained in the
drugs of this study. This represents an advantage over
currently available anti-inflammatory drugs. In this
study a maximum of three new drugs was synthesised
and tested simultaneously. This ensured that any alter-
ation in structure, which led to an increase or a decrease
in activity, could readily be identified. This structural
modification could then be incorporated in, or excluded
from, the next series of drugs as appropriate. This
iterative process allowed for rapid structural opti-
misation, with the minimum number of drugs requiring
synthesis and testing. The structures of the compounds
synthesised are shown in Table 1.
O
NH
C
R1
R5
R3
O
R4
Compound
R1
R3
R5
R4
1
2
n-C4H9
CH3
H
H
H
H
H
n-C4H9
n-C4H9
C2H5
C2H5
C2H5
C2H5
CH3
CH3
Benzyl
H
3
n-C4H9
CH3
H
H
4
CH3
C2H5
iPr
CH3
C2H5
iPr
Cl
5
CH3
6
CH3
7
CH3
Cl
8
C6H6
CH3
H
H
9
C2H5
CH3
Benzyl
C6H6
H
C2H5
CH3
H
10
11
12
13
14
CH3
CH3
Materials and Methods
CH3
C6H6
H
tBu
H
tBu
CH3
Materials
CH3
tBu
H
4-Amino-2,6-dichlorophenol, 4-aminophenol, benzyl
bromide, 2,6-diphenylphenol, 2,6-diethylaniline, iodo-
ethane, 4-methoxyaniline, N-(4-methyl phenyl) aceta-
mide and pentanoic anhydride were obtained from
Aldrich Chemical Company. Carrageenan was supplied
by Bhanda Chemicals and Lambda Pharmaceuticals.
Benzyl chloride, tertiary butyl bromide, calcium chlor-
ide, ethanoic anhydride, hydrochloric acid, nitric acid,
petroleum ether (boiling range 30–60mC), silica gel,
sodium carbonate, sodium bicarbonate, sodium dithio-
nite, metallic sodium, sodium nitrite, sodium sulfate,
sulfanilic acid, sulfuric acid and granulated tin were
obtained from BDH Ltd. N-(3,5-ditertiary butyl-4-
hydroxyphenyl) acetamide, N-(3,5-diethyl-4-hydroxy-
phenyl) acetamide, N-(3,5-diisopropyl-4-hydroxyphe-
nyl acetamide and N-(3,5-dimethyl-4-hydroxyphenyl)
acetamide were available in our laboratory, synthesised
previously by O’Hara (1976).
Previous research by Tomlinson (1971) showed that a
parabolic relationship existed between the analgesic
activity of substituted N-phenylacetamides and their 1-
octanol\water partition coefficients (log P), for com-
pounds where log P was less than 1.7. Continuing this
work, O’Hara (1976) showed that a parabolic relation-
ship existed between analgesic activity and log P of ring-
substituted alkyl derivatives of paracetamol, where log
P was less than 1.8. Dearden et al (1980) later extended
this range of compounds to include derivatives in which
the log P was increased to 4.4. In this extended series, a
second parabolic relationship was apparent in the log P
range 1.8–4.4. This unexpected second parabola led to
speculation that the drugs with the higher log P values
were acting to produce an anti-inflammatory effect. The
analgesic activity had been measured using the mouse
abdominalconstrictionassay, forwhichthereisevidence
of an inflammatory component (Collier et al 1968).
Chemical synthesis
Since many commercially available NSAIDs have log P Compounds were characterised using infra-red spec-
values within this range, it was considered possible that troscopy (Pye-Unicam 100 SP3). Liquids were analysed
a paracetamol derivative with potent analgesic activity as films, solids as nujol mulls; air was used as reference
could be synthesised which would also possess anti- beam. Melting points were determined using
inflammatory activity.
Gallenkamp melting point apparatus, and are uncor-

1507
Novel anti-inflammatory drugs
rected; literature values are quoted where available. N-(3,5-dimethyl-4-ethoxyphenyl)acetamide (4)
Molecular masses were obtained using electrospray Metallic sodium (0.01 mol) was dissolved in 4 mL an-
mass spectrometry (Micromass VG Platform II instru- hydrous ethanol as above. A suspension of 0.01 mol N-
ment, MassLynx software, positive ion mode).
(3,5-dimethyl-4-hydroxyphenyl)acetamide in 3 mL an-
hydrous ethanol was added, followed by 0.01 mol iodo-
ethane. After stirring for 18 h and heating for 3 h in
darkness, the sodium iodide precipitate was removed by
N-(4-hydroxyphenyl)pentanamide (Compound 1)
Pentanoic anhydride (0.04 mol) was added to a sus- filtration. Ethanol was removed by rotary evaporation
pension of 4-aminophenol (0.04 mol) in 12 mL distilled and the crude product was washed with 10 mL of 1.25 
water. Themixturewas heated onwater-bathfor 10 min. sodium hydroxide solution, then distilled water until
After cooling, the precipitate was filtered through a washings were neutral. The product was recrystallised
Buchner funnel and dried. Traces of pentanoic acid from 50% ethanol.
were removed by stirring with 100 mL of 1.2  sodium
bicarbonate for 20 min. The mixture was filtered and (Noguchi et al (1984)). IR (cm ) 1210 and 1280 (aral-
Yield 1.6 g; mp 133.5mC (literature mp 132–133mC
1
−
the product was recrystallised from water.
1
kylether); measured mass 207.34; calculated mass
−
Yield 3.5 g; mp 99.5mC; IR (cm ) 1650 (CꢀO amide); 207.27.
measured mass 193.27; calculated mass 193.25.
N-(3,5-diethyl-4-ethoxyphenyl)acetamide (5)
Theprocedureforsynthesisofcompound4wasfollowed
using 0.01 mol N-(3,5-diethyl-4-hydroxyphenyl)aceta-
mide in place of dimethyl derivative.
N-(4-butoxyphenyl)acetamide (2)
The method was adapted from that of Cizmarik et al
(1976).
1
−
Yield 1.15 g; mp 96mC; IR (cm ) 1200 and 1270
(aralkylether); measured mass 235.45; calculated mass
235.32.
Metallic sodium (0.1 mol) was dissolved in 40 mL
anhydrous ethanol under reflux with a calcium chloride
closure. 4-Acetamidophenol (0.1 mol) suspended in
10 mL anhydrous ethanol was added, followed by
0.1 mol 1-chlorobutane. The resulting solution was
stirred magnetically for 18 h at room temperature, then
heated under reflux for 3 h. The mixture was filtered to
remove sodium chloride precipitate. Ethanol was re-
moved by rotary evaporation. Cold distilled water
(200 mL) was added to the residue. The crude product
was washed with 2.5  sodium hydroxide (50 mL), then
with distilled water until washings were neutral. The
product was recrystallised from distilled water.
Yield 11.4 g; mp 109.4mC (literature mp 109mC
N-(3,5-di-isopropyl-4-ethoxyphenyl)acetamide (6)
Theprocedureforsynthesisofcompound4wasfollowed
using 0.01 mol N-(3,5-di-isopropyl-4-hydroxyphenyl)-
acetamide in place of dimethyl derivative.
Recrystallisation from 50% aqueous ethanol pro-
duced impure product. The product was recrystallised,
with difficulty, from distilled water, hence only sufficient
was recrystallised to perform animal testing.
1
−
Yield – not determined; mp 112mC; IR (cm ) 1260
(aralkylether); measured mass 263.48; calculated mass
263.38.
1
−
(Cizmarik et al (1976)); IR (cm ) 1240 and 1260 (aral-
kylether); measured mass 207.33; calculated mass
207.27.
N-(3,5-dichloro-4-ethoxyphenyl)acetamide (7)
Part 1. Synthesis of N-(3,5-dichloro-4-hydroxyphenyl)-
acetamide. The procedure for synthesis of compound 1
was followed, using 0.05 mol 4-amino-2,6-dichloro-
phenol suspended in 24 mL distilled water and adding
0.1 mol ethanoic anhydride.
N-(4-butoxyphenyl)pentanamide (3)
Metallic sodium (0.03 mol) was dissolved in 12 mL
anhydrous ethanol as above. Compound 1 (0.03 mol,
prepared as above) suspended in 3 mL anhydrous etha-
nol was added, followed by 0.03 mol 1-chlorobutane.
The procedure continued as for the synthesis of com-
pound 2, and the product was recrystallised from 50%
ethanol.
1
−
Yield 5.7 g; IR (cm ) 1660 (CꢀO amide).
Part 2. Synthesis of N-(3,5-dichloro-4-ethoxyphenyl)-
1
−
Yield 1.9 g; mp 87.1mC; IR (cm ) 1250 and 1290 acetamide. Metallic sodium (0.026 mol) was dissolved in
(aralkylether); measured mass 249.46; calculated mass 12 mL anhydrous ethanol as above. A suspension of
249.35.
0.026 mol of N-(3,5-dichloro-4-hydroxyphenyl)aceta-

1508
mide in 6 mL ethanol was added. The procedure con- with stirring and cooling in ice to 25–30mC. The mixture
Judith C. Duffy et al
tinued as for synthesis of compound 4.
1
was left for 5 h, then 220 mL distilled water was added.
−
Yield 3.2 g; mp 173mC; IR (cm ) showed phenolic OH The organic layer was removed and washed with water
peak at 2750–3380 diminished; measured mass 247.29; until neutral. Chloroform was removed by rotary evap-
calculated mass 247.14.
oration and the resultant oil was returned to the reaction
vessel. A further 0.5 mol 70% nitric acid was added and
the mixture was stirred for 5 h. Distilled water (220 mL)
was added and the chloroform layer was separated,
washed and chloroform removed by rotary evaporation.
N-(4-methoxyphenyl)benzamide (8)
4-Methoxyaniline (0.05 mol) was suspended in 45 mL
2.5  sodium hydroxide. Benzoyl chloride (0.05 mol)
was added slowly, the mixture was heated on a water-
bath for 10 min and then cooled. The precipitate was
filtered and impurities were removed by mixing product
with warm water, filtering and drying.
1
−
IR (cm ) 1340 and 1510–1530 (aromatic NO2).
Part 4. Synthesis of 3,5-diethyl-4-methoxyaniline. 2,6-
Diethyl-4-nitroanisole (0.33 mol) was added to 0.5 mol
granulated tin in a round-bottomed flask fitted with 2
reflux condensers. Concentrated hydrochloric acid
(4.3 mol) was added slowly and the mixture was warmed
on a steam bath. After the reaction subsided, the mixture
was cooled and 2.5 mol sodium hydroxide in 166 mL
water was added slowly. Chloroform (200 mL) and
water (200 mL) were added and the mixture was filtered.
The product was repeatedly extracted into chloroform
until the chloroform layer was colourless. Combined
chloroform extractions were washed with 2.5  sodium
hydroxide (100 mL), then water until neutral. Chloro-
form was removed by rotary evaporation.
1
−
Yield 7.8 g; mp 159mC; IR (cm ) 1640 (CꢀO amide);
measured mass 227.35; calculated mass 227.26.
N-(3,5-diethyl-4-methoxyphenyl)acetamide (9)
Part 1. Synthesis of 2,6-diethylphenol. Concentrated
sulfuric acid (3 mol) was added to 1200 mL distilled
water. 2,6-Diethylaniline (1.5 mol) was added and the
mixture was heated on a steam bath for 15 min, followed
by cooling in ice-water to 5mC. Sodium nitrite (1.5 mol)
in 200 mL water was slowly added, and once the reaction
had subsided, the mixture was heated to 50mC on a
steam bath for 1 h. After standing overnight the re-
sultant precipitate was filtered and dried. Sample was
recrystallised from 50% aqueous ethanol.
1
−
IR (cm ) 1610 (NH2), also disappearance of NO2
peaks at 1340 and 1510–1530.
1
−
IR (cm ) 3200–3600 (phenolic OH).
Part 5. Synthesis of N-(3,5-diethyl-4-methoxyphenyl)-
acetamide. 3,5-Diethyl-4-methoxyaniline (0.18 mol) was
suspended in 45 mL distilled water and 0.2 mol ethanoic
anhydride was added. The mixture was stirred vigor-
ously, heated for 10 min, then cooled. Sodium bicar-
bonate (200 mL, 0.6 ) was added and the mixture was
stirred for a further 30 min. Chloroform (250 mL) was
added. The chloroform layer was separated and washed
3 times with 200 mL distilled water. Chloroform was
removed by rotary evaporation, leaving a red oil. This
was purified, with difficulty, from 70% aqueous ethanol
solution, hence only sufficient was recrystallised to allow
for animal testing.
Part2. Synthesisof2,6-diethylanisole. 2,6-Diethylphenol
(1 mol) was dissolved in chloroform and dried using
anhydrous sodium sulfate. Chloroform was removed in
a rotary evaporator. Metallic sodium (0.27 mol) was
dissolved in 108 mL anhydrous ethanol as above. Dried
2,6-diethylphenol (0.67 mol) was added to the sodium
ethylate, followed by 0.27 mol iodomethane, and syn-
thesis continued as for product 2. Distilled water
(50 mL) was added to remove water-soluble impurities,
and the product was extracted into 75 mL chloroform,
which was subsequently removed by rotary evaporation.
1
−
IR (cm ) 1200 and 1260 (aralkylether), also dis-
1
−
Yield – not determined; mp 94.5–95mC; IR (cm )
appearance of phenolic OH peak at 3200–3600.
1660 (CꢀO amide); measured mass 221.30; calculated
mass 221.30.
Part 3. Synthesis of 2,6-diethyl-4-nitroanisole. The
method was adapted from that of Bisarbi-Ershadi &
Rigterink (1989).
N-(3,5-dimethyl-4-benzyloxyphenyl)acetamide (10)
2,6-Diethylanisole (0.66 mol) was added to 660 mL Metallic sodium (0.02 mol) was dissolved in 10 mL
chloroform in a 3-necked round-bottomed flask fitted anhydrous ethanol as above. A suspension of 0.02 mol
with thermometer, dropping funnel and magnetic N-(3,5-dimethyl-4-hydroxyphenyl)acetamide in 10 mL
stirrer. Nitric acid (0.5 mol, 70%) was added over 1.5 h ethanol was added, followed by 0.001 mol ascorbic acid

1509
Novel anti-inflammatory drugs
and 0.02 mol benzyl bromide. The mixture was stirred lised, with difficulty, from 70% ethanol solution, hence
overnight, heated under reflux for 3 h, then cooled. The only sufficient was purified to allow for animal testing.
1
−
product was extracted into 30 mL chloroform. Chloro-
Yield – not determined; mp 101–103mC; IR (cm )
form was removed by rotary evaporation and the com- 1620 (CꢀO amide); measured mass 303.35; calculated
pound was recrystallised from 50% aqueous ethanol.
1
mass 303.36.
−
Yield 2.4 g; mp 148mC; IR (cm ) absence of phenolic
1
OH peak at 3300 cm ; measured mass 269.34; calcu-
−
N-(4-tertiarybutoxyphenyl)acetamide (13)
lated mass 269.34.
The procedure for synthesis of compound 2 was fol-
lowed. Tertiarybutylbromide was used in place of 1-
chlorobutane. Residual 4-acetamidophenol was re-
moved by washing with toluene, which was subsequently
removed by rotary evaporation. The product was mixed
with 0.6  sodium bicarbonate (100 mL) for 30 min,
washed with distilled water and recrystallised, with
difficulty, from 0.12  sodium bicarbonate in 10%
aqueous ethanol. Only sufficient compound was purified
to allow for animal testing.
N-(3-benzyl-4-hydroxyphenyl)acetamide (11)
Part 1. Synthesis of 3-benzyl-4-aminophenol. The method
was adapted from that of O’Hara (1976).
Sulfanilic acid (0.25 mol) and sodium carbonate
(0.12 mol) were dissolved with heating and stirring in
250 mL distilled water. Sodium nitrite (0.27 mol) was
dissolved in 80 mL distilled water and the two solutions
were cooled in ice to 15mC. Both solutions were added to
1.7 mol concentrated hydrochloric acid in 300 g crushed
ice, giving mixture A, which was allowed to stand for
20 min.
Yield – not determined; mp 126–127mC (literature mp
1
−
130mC (Cadogan et al)). IR (cm ) 1220 and 1240 (aral-
kylether); measured mass 207.29; calculated mass
207.27.
2-Benzylphenol (0.25 mol) was dissolved in a solution
of 1.38 mol sodium hydroxide in 300 mL distilled water
and cooled by the addition of 200 g ice. This solution
was added to mixture A and allowed to stand for 3 h,
after which it was heated to 45mC and 0.6 mol sodium
dithionite was added. The mixture was heated to 60–
65mC and the temperature was maintained until product
appeared. The product was filtered and washed with
0.03  sodium dithionite solution.
N-(3,5-ditertiarybutyl-4-hydroxyphenyl)acetamide
(14)
This drug had previously been synthesised by O’Hara
(1976).
Pharmacological evaluation
1
−
IR (cm ) 1500 and 1510 (aromatic NH2), 3300 (phen-
All animal testing was performed in accordance with
Home Office regulations and with ethical approval from
the University.
olic OH).
Part 2. Synthesis of N-(3-benzyl-4-hydroxyphenyl)-
acetamide. 3-Benzyl-4-aminophenol (0.16 mol) was sus-
pended in 60 mL distilled water. Ethanoic anhydride
(0.16 mol) was added and the mixture was heated for
10 min. The resulting solid was filtered and mixed with
1.2  sodium bicarbonate (200 mL), then filtered and
washed with water. The product was recrystallised, with
difficulty, from 0.03  sodium dithionite solution, hence
only sufficient was purified to allow for animal testing.
Analgesic activity
Analgesic activity was tested using the abdominal con-
striction assay (Collier et al 1968). Male white CD-1
mice, weighing 25–30 g (bred in-house or obtained from
Charles River suppliers), were fasted for 16 h and al-
lowed free access to water. Homogeneous drug suspen-
sions were made in distilled water with 1 drop of Tween
80. Mice were dosed by gavage, each receiving 0.5 mL
drug suspension (0.5 mL vehicle for control group).
Thirty minutes post dose, mice were injected intra-
peritoneally with 0.1 mL of 0.7% v\v acetic acid sol-
ution and placed in a deep-sided, open-top plastic box.
The number of full abdominal constrictions (comprising
1
−
Yield – not determined; mp 128.5–129.5mC. IR (cm )
1620–1640 (CꢀO amide); measured mass 241.30; cal-
culated mass 241.29.
N-(3,5-diphenyl-4-hydroxyphenyl)acetamide (12)
The procedure for synthesis of compound 11 was fol- a stretching of the hind limbs to full extent allowing the
lowed. 2,6-Diphenylphenol replaced 2-benzylphenol as abdomen to touch the cage floor) performed by mice
the starting material and was suspended in a solution of were counted for a 15-min period.
1.1 mol sodium hydroxide in 300 mL distilled water Five mice were used for each of the four dose levels of
containing 25 mL ethanol. The product was recrystal- drug (doses selected from prior knowledge of similar

1510
Judith C. Duffy et al
Table 2 Scoring system used to determine gastro-toxicity.
compounds). A control group of five mice was used on
each day of testing, as the test is very susceptible to
external conditions. The total number of constrictions
was summed for the five mice in each group. Analgesic
activity was recorded as the percentage inhibition of
abdominal constrictions when drug was present com-
pared with control group. Results were calculated using
equation 1
Description of lesion
Lesion score
0
0
1–5
1–150 minute
150–200j minute
reddened areas
slight sloughing
severe sloughing
1–20 small
5–10
5–10
5–10
10–20
10–30
30–35
35–40
40–50
50–70
70–90
90–100
% Inhibition l 100k(100i(average drug
1–5 medium
5–10j medium
1–5 severe
response\average control response))
(1)
ED50 (the effective dose to inhibit the biological
response by 50%) values were obtained from dose–
response plots.
5–20 severe
1–10 streak
10j streak
Further details of the scoring system are given in the text.
Anti-inflammatory activity
Anti-inflammatory activity was measured using the car-
rageenan-induced rat paw oedema assay (Winter et al
1962). White male Wistar rats from a breeding stock
originally supplied by Charles River were all bred in-
house. As facilities allowed for only limited breeding,
overall weights ranged from 90–400g. Within individual
studies, however, all rats were of comparable weight.
Drug suspensions were prepared as described pre-
viously. Carrageenan solution (2%) was prepared using
0.1 g carrageenan from Lambda Pharmaceuticals and
0.1 g carrageenan from Bhanda chemicals in 0.9%
sodiumchloridesolution(10 mL). Carrageenansolution
was prepared24 hbefore use and stored in a refrigerator,
whereupon it became gelatinous.
Activity l 100k(100i(average volume drug
treated\average volume for control))
(3)
Four dose levels were used for each drug, from which
dose–response graphs were produced. ED30 (the effec-
tive dose to inhibit the biological response by 30%)
values were obtained visually from these graphs. ED30
values were used rather than ED50 values so that
activities of the less potent compounds could be quan-
tified.
Gastro-toxicity
Gastro-toxicity was measured using a visual lesion as-
sessment method, adapted from Dearden & Nicholson
(1984). After paw volumes at 4 h were measured, rats
were killed either by injection with 0.25 mL euthatal or
by being placed in a CO2 chamber. Stomachs were
removed, rinsed with tap water and stretched over an
inverted test-tube. They were examined by eye for signs
of damage and scores were recorded for each rat ac-
cording to the scale shown in Table 2.
Rats were fasted for 16 h before use and allowed free
access to water. Rats were dosed by gavage with drug
suspension (0.5–2 mL) or vehicle (control group). Thirty
minutes post dose each rat was injected with 2% carra-
geenan solution (0.1 mL) into the subplantar surface of
the right hind paw. A black mark was made at the knee
joint for reference. Paw volume, up to the reference
mark, was measured plethysmographically immediately
following injection and each hour thereafter. Paw
volume at the 3-h time point was used for subsequent
calculations. Three groups of drug-treated rats and one
control group were used each test day. Five rats were
used in each group, the mean paw oedema value for the
test group being compared with the mean value for the
control group for that day:
Examination of stomachs by eye provided a simple
andreliablemethodtoquantifygastricirritancy. Lesions
produced by anti-inflammatory drugs were clearly vis-
ible and severity could be readily assessed. There are
several scales recorded in the literature for estimating
gastro-toxicity; however it is difficult to compare results
between different groups as the manner of interpreting
damage will vary considerably. The ulcerogenic activi-
ties of the compounds in this study were therefore
assessed relative to each other only and a comparison
was made with aspirin, which is commercially available.
Oedema level l (paw volume at time tkvolume at
time 0)\volume at time 0
(2)
Anti-inflammatory activity was measured as the per- As the compounds showed very low gastro-toxicity, a
centage reduction in oedema level when drug was pres- scale was devised so that even limited damage could be
ent, relative to control:
quantified. Asthe ratswerefastedovernight, they tended

1511
Novel anti-inflammatory drugs
Table 3 Results of the analgesic, anti-inflammatory and gastro-toxicity tests.
Compound Analgesic
ED50 (mol kg
Anti-inflammatory
1
Gastro-toxicity
score
1
−
−
)
ED30 (mol kg
)
4
4
4
4
4
4
4
4
4
−
−
−
−
−
−
−
−
−
1
2.4i10
3.5i10
2.1i10
1.9i10
5.1i10
4.2i10
1.1i10
3.9i10
3.7i10
Inactive
1.4
4
−
2
3.6i10
Inactive
1.2
3
1.7
4
−
4
1.16i10
1.2
4
−
5
1.7i10
1.9i10
3.3i10
1.6
4
−
6
2.3
4
−
7
2.1
8
Inactive
1.1
4
−
9
4.0i10
Inactive
1.8
10
Inactive
1.3
4
4
−
−
11
1.4i10
Incalculable
3.3i10
Incalculable
1.3
12a
13b
14
1.5
4
4
−
−
49% inhibition at 2.4i10
48% inhibition at 3.8i10
Incalculable
2.6
5
−
4
−
4.9i10
9.4i10
2.1i10
1.1i10
4
−
3
−
Aspirin
9.5
aAlthough analgesic and anti-inflammatory activity was seen the responses were not uniform and therefore
estimates of ED50 and ED30 could not be made. bSufficient compound was available to allow testing at only
one dose level.
to eat sawdust which could lead to minute lesions on the
In-vivo experiments were performed as this gives a
surface of the stomach. As these were apparent also in better indication of the efficacy and toxicity of the drugs.
the control groups they were not considered to be Experiments performed in-vitro cannot account for
significant. Consequently, 1 or 2 medium lesions scored pharmacokinetic effects, which may render a drug in-
much higher than 100 of these minute lesions.
active in-vivo, despite apparent potency in-vitro.
The terms minute, small, medium and severe refer to
the diameter of the circular lesions. A streak lesion
appears as a furrow along the stomach lining, where a
line of lesions has formed, leading to severe breakdown
of the mucosa and bleeding into the stomach. Aspirin
was the only drug tested that produced this effect.
The average lesion score per rat, for all twenty rats
used in producing the anti-inflammatory data, was used
irrespective of dose of drug given. These were then
Results and Discussion
The results of the biological tests are shown in Table 3.
Rationale for compound selection
scaled to give relative toxicity values (that of control In this study a series of paracetamol derivatives was
being 1). The scaled gastro-toxicity scores are recorded designed and tested for analgesic and anti-inflammatory
in Table 3. The incidence of lesions has been shown not activity. The work follows on from the findings of
to be directly related to the amount of drug given but O’Hara (1976) and Dearden et al (1980) vide supra. One
rather to show a dose–response pattern independent of important consideration in the design of anti-inflam-
the dose–response curve for anti-inflammatory activity matory drugs is the problem of gastric irritancy, the
(Elliott 1979). Therefore, to obtain absolute values for most common side effect associated with NSAIDs. As
toxicity a separate dose–response study would have to paracetamol is devoid of gastro-toxic effects, the possi-
have been undertaken, doubling the number of rats bility of capitalising on this property, by using para-
used. The results produced by this study give an in- cetamol derivatives, presented the opportunity to im-
dication of the toxicity score obtained at the doses prove on current therapies.
required to give an anti-inflammatory effect. We believe
that this gives a most realistic indication of gastro- cated that substitutions in positions 2 and 6 reduced
toxicity. analgesic activity, whereas substitution in positions 3
The results from O’Hara’s work (O’Hara 1976) indi-

1512
and 5 enhanced activity. No attempt had been made by
Judith C. Duffy et al
As compound 8 showed no anti-inflammatory ac-
O’Hara to substitute the acetamide or hydroxy func- tivity, no further substitutions were made at the aceta-
tions. In this study, therefore, the first structural alter- mide function and as compound 10 was the only com-
ations were to convert the acetamide function to a pound devoid of both analgesic and anti-inflammatory
pentanamide (compound 1). The hydroxy group was activity, the use of such a bulky group in position 4 was
converted to a butoxy group in compound 2 and both of avoided in the next design–test iteration.
the modifications were included in compound 3. The log
We had carried out a separate study, using the Cata-
P values calculated for compounds 1, 2 and 3 were 2.1, lyst Molecular Modelling package, to design an an-
2.8 and 4.4, respectively, (MedChem software, version algesic drug (Duffy et al 1995). That study resulted in
3.52, Medicinal Chemistry Project, Pomona College, the design of compound 11, which incorporated a bulky,
USA). These values all fall within the range in which the hydrophobic group in position 3. Progressing from
second parabolic relationship had been demonstrated compound 11, compound 12 was designed, in which a
by Dearden et al (1980). The compounds were therefore bulky group was incorporated in both positions 3 and 5.
predicted to possess anti-inflammatory activity. Al- Calculated log P values were 2.6 and 3.7 for compounds
though compound 2 (of similar log P to compound 1) 11 and 12, respectively.
possessed anti-inflammatory activity, compounds 1 and
Compound 13 was designed to test whether replacing
3 did not. This was attributed to the substitution of the the n-butoxy function with a tertiary butoxy function
acetamide function with the pentanamide group. In the would lead to a significant change in activity. The
next iteration of the design cycle therefore the acetamide calculated log P for this compound was 2.5, similar to
group was not substituted.
that of compound 2 (log P, 2.8). Comparing the results
As 3- and 5-substitution had enhanced the activity of for compounds 2 and 13 should distinguish the effects of
O’Hara’s compounds, substitutions were made in these the topological differences in the substituents in position
positions for compounds 4, 5 and 6 using methyl, ethyl 4. As there were problems in the synthesis of compound
and isopropyl groups, respectively. We wished to main- 13, sufficient could be made only to allow testing at one
tain an alkoxy function in place of the hydroxy, as this dosage level.
had proved successful in compound 2. The butoxy
The calculated log P values for the compounds in this
function was not used in position 4 for these compounds, study all fell within the range 2.1–4.6, similar to the
as the log P values would have been too high. An ethoxy values investigated by Dearden et al (1980). However,
group was substituted in position 4 giving calculated log unlike the previous study, no overall correlation was
P values of 2.8, 3.8 and 4.6 for compounds 4, 5 and 6, found between log P and in-vivo analgesic or anti-
respectively. This made synthesis easier as the butoxy inflammatory activity. This suggested that factors other
group would have created more steric hindrance. The than hydrophobicity, such as steric or electronic effects,
influence of halogen substituents in place of alkyl groups were more significant in determining activity within this
in the 3 and 5 positions was tested in compound 7. This series.
compound had a calculated log P of 3.1, again falling
within the range of the second parabola. Compounds
4–7 were found to possess superior efficacy to compound
2.
Anti-inflammatory activity
To test if a larger group in place of the pentanamide, The presence of an acetamide function appeared es-
used in compounds 1 and 3, would produce an anti- sential for anti-inflammatory activity. In compounds 1,
inflammatory effect, compound 8 was synthesised 3 and 8, where the acetamide had been substituted, anti-
whereina phenyl amide replaced the acetamide function. inflammatory activity was abolished. Replacement of
Attempts were also made to determine whether an the hydroxy group with a butoxy function enhanced the
alkoxy or aralkoxy replacement for the hydroxy func- compound’s anti-inflammatory activity. There was little
tion in position 4 would improve activity. In compound difference in potency between the n-butoxy and the
8 a methoxy group was used, the calculated log P value tertiary-butoxy substitutions (compounds 2 and 13,
being 2.7. Compound 9 represented a modification of respectively). In compound 10, where the benzyloxy
compound 5, possessing a 3,5-diethyl substitution in function was used, activity was again abolished. The
addition to a methoxy group in position 4 (calculated most effective anti-inflammatory agents were those
log P, 3.3). In compound 10 a benzyloxy function was where the hydroxy had been substituted by an ethoxy
used in position 4 with 3,5 dimethyl substitution (cal- group (compounds 4–7). Anti-inflammatory potency
culated log P 4.0).
within this series decreased as the size and hydro-

1513
Novel anti-inflammatory drugs
phobicity of the alkyl substituent increases. This may be stituted phenols in the body. Substituted phenols, by
due to the optimum log P value having been exceeded or acting as free-radical scavengers, have been shown to
to increasing steric hindrance at the receptor site. Re- act as anti-inflammatory agents (Swingle et al 1985).
placement of the alkyl groups in positions 3 and 5 with This may account for the activity of these compounds.
a chloro group (compound 7) reduced anti-inflam-
The ability of certain drugs to insert into the lipid
matory activity, as did replacing the ethoxy in position bilayer of cells which mediate inflammation has also
4 (compound 5), with the smaller methoxy group (com- been proposed as a mechanism for anti-inflammatory
pound 9). In compound 11, substitution was made in activity (Hwang & Shen 1981). Once inserted into the
position 3, but not in position 5. This substitution membrane, the drugs interrupt cell signalling processes
pattern also produced an effective drug. Although com- and inhibit activation of inflammatory cells. Para-
pound 12 did possess some anti-inflammatory and an- cetamol itself is not sufficiently hydrophobic, but the
algesic activity, the responses obtained were not uniform increasedhydrophobicityofthe compoundsinthisstudy
and therefore could not be quantified. The reason for may enable them to act via such a mechanism.
this is unclear.
Analgesic activity
Toxicity
With the exception of compound 14 (synthesised pre- Prostaglandins play a role in mediating inflammation
viously by O’Hara), the di-chloro derivative (compound and also in protecting the gastric mucosa. Two major
7) was the most effective analgesic. Compound 11, enzyme systems involved in the production of prosta-
designed using the Catalyst program, was the second glandins are cyclooxygenase-1 (COX-1) and cyclo-
most effective analgesic. Unlike anti-inflammatory ac- oxygenase-2 (COX-2). COX-1 is responsible for main-
tivity, analgesic activity was not abolished by substi- tenance of the gastric mucosa (Wallace & Cirino 1984),
tution of the acetamide function.
with COX-2 responsible for mediating inflammation.
All compounds which possessed anti-inflammatory Currently available NSAIDs are believed to owe their
activity possessed analgesic activity, but not vice versa. anti-inflammatory activity and gastro-toxicity to the
This may have been because the receptor which mediates inhibition of COX-2 and COX-1, respectively. The lack
analgesic activity is more accommodating than the of gastro-toxicity apparent in the drugs synthesised in
receptor which mediates anti-inflammatory activity. It this study may have been due to their selective inhibition
may also have been due to the analgesic assay being of COX-2, or to the fact that they act by a method other
more susceptible to false-positive results than the anti- than inhibition of prostaglandin synthesis. Gastro-tox-
inflammatory assay.
icity of NSAIDs may be caused by a direct irritant effect
The pathways via which pain and inflammation are of the acidic drugs on the gastric mucosa. The lack of
mediated are particularly complex. Interfering with dif- gastro-toxicity of these new drugs may be due, in part,
ferent mediators of these pathways may result in a more to their non-acidic nature. As NSAID gastropathy is the
potent analgesic or anti-inflammatory effect. This re- main reason for cessation of anti-inflammatory therapy,
lationship is also seen in the fenamates, where flufenamic these drugs present a possible alternative to those cur-
acid is a more potent anti-inflammatory and mefenamic rently available. Although some commercial NSAIDs
acid is a more potent analgesic (Lembo et al 1983).
are more potent, the lack of gastro-toxicity of the drugs
For all the compounds where the hydroxy group had described herein may enable higher doses of them to be
been replaced with an alkoxy function, metabolism in given more safely.
the liver may occur via O-dealkylation, resulting in the
As these new drugs are based on paracetamol, another
formation of the corresponding substituted phenol important consideration is that of hepatotoxicity. Al-
(phase I transformation). Ninety percent of paracetamol though no estimation of hepatotoxicity was made in this
is metabolised by conversion of the hydroxy to a glucu- work, the conclusions drawn by Van de Straat et al
ronide or sulfate conjugate, which is readily excreted in (1986) are relevant. Their work on similar substituted
the urine. Theoretically the substituted phenol deriva- paracetamol derivatives showed that although para-
tives, formed via phase I transformation, may also cetamol and the 3-mono-substituted derivatives did
undergo conjugation. However 3,5-disubstitution induce hepatotoxicity, 3,5-disubstituted derivatives did
would hinder access to the hydroxy group, and hence not. From these results it is predicted that the 3,5-
slow metabolism of the drugs. If these compounds do disubstituted derivatives synthesised in the present work
undergo O-dealkylation, they would then exist as sub- would not be hepatotoxic.

1514
Judith C. Duffy et al
Duffy, J. C., Dearden, J. C., Green, D. V. S. (1995) Use of catalyst in
the design of novel non-steroidal anti-inflammatory analgesics. In:
Proceedings of the 10th European Symposium on SAR Barcelona.
Prous Scientific Publishers, Barcelona
Conclusion
The results show that 11 of the drugs designed in this
study possess an analgesic potency 2–10 times that of
aspirin, while 8 possess an anti-inflammatory activity
3–10 times that of aspirin. The most significant results,
however, relate to the gastro-toxicity profile of these
compounds. The toxicity score is below 2.6 for every
compound designed. This compares favourably with
the value of 9.5 obtained for aspirin. Of all the com-
pounds tested, aspirin was the only drug to produce
characteristic streak lesions. Although the drugs synthe-
sised in this study proved potent in the assays used, this
is no guarantee of their efficacy in man. Interspecies
differences in pharmacokinetics and disease states may
render the drugs ineffective in treating human disorders.
More detailed pharmacokinetic and toxicity studies on
these drugs are essential.
Elliott, P. N. C. (1979) The effect of combination of aspirin and
sodium salicylate on the rat stomach. Br. J. Pharmacol. 67: 484P
Hwang, S. B., Shen, T. Y. (1981) Membrane effects of anti-
inflammatory agents 2. Interaction of non-steroidal anti-
inflammatory drugs with liposome and purple membranes. J. Med.
Chem. 24: 1202–1211
Lembo, S., Sasso, V., Silipo, C., Vittoria, A. (1983) Physicochemical
properties and drug design in a set of analgesics. Il Farmaco. Ed.
Sci. 38: 750–761
Mehta, D. K. (ed) (1997) British National Formulary. London: Phar-
maceutical Press.
Noguchi, Kato, Takahashi, Ishiguri, Yamamoto, Hirata, N. (1984)
European Patent No: 0 100 615. Chem. Abs. (1984) 100: 209360v
O’Hara, J. H. (1976) Physico-chemical parameters and biological ac-
tivities of a series of alkyl substituted 4-hydroxy acetanilides. PhD
Thesis, Liverpool Polytechnic
Swingle, K. F., Bell, R. L., Moore, G. G. I. (1985) Anti-inflammatory
activity of antioxidants. In: Rainsford, K. D. (ed) Anti-inflam-
matory and anti-rheumatic drugs. CRC Press, Florida. Vol III, pp
105–126
References
Tomlinson, E. (1971) Physico-chemical parameters and analgesic ac-
tivities in a series of substituted acetanilides. PhD Thesis, Liverpool
Polytechnic
Bisarbi-Ershadi, B., Rigterink R. H. (1989) United States Patent No:
4 868 215. Chem. Abs. (1990) 112: 98225f
Cadogan J. I. G., Raphael R. A., Rees, C. W. (eds) (1982) Dictionary
of organic compounds. 5th Edn, Chapman & Hall, New York
Cizmarik, J., Borvansky, A., Svec, P. (1976) Study of local anaes-
thetics. LII. Piperidinoethyl esters of alkoxyphenylcarbamic acids.
Acta Facultatis Pharmaceuticae Tom XXIX: 53–75
Van de Straat, R., de Vries, J., Kulkens, T., Debets, J. J. J., Vermeulen,
N. P. E. (1986) Paracetamol, 3-monoalkyl- and 3,5-dialkyl deriva-
tives. Comparison of their microsomal P-450 dependent oxidation
and toxicity in freshly isolated hepatocytes. Biochem. Pharmacol.
35: 3693–3699
Collier, H. O. J., Dineen, L. C., Johnson, C. A., Schneider, C. (1968)
The abdominal constriction response and its suppression by an-
algesic drugs. Br. J. Pharmacol. 32: 295–310
Wallace, J. L., Cirino, G. (1994) The development of gastro-intestinal
sparing non-steroidal anti-inflammatory drugs. Trends Pharmacol.
Sci. 15: 405–406
Dearden, J. C., Nicholson, R. M. (1984) Correlation between gastric
irritancy and anti-inflammatory activity of non-steroidal anti-
inflammatory drugs. J. Pharm. Pharmacol. 36: 713–715
Dearden, J. C., O’Hara, J. H., Townend, M. S. (1980) A double-
peaked quantitative structure-activity relationship (QSAR) in a
series of paracetamol derivatives. J. Pharm. Pharmacol. 32 (Suppl.):
102
Wallace, J. L., Granger, D. N. (1992) Pathogenesis of NSAID gastro-
pathy – are neutrophils the culprits? Trends Pharmacol. Sci. 13:
129–131
Winter, C. A., Risley, E. A., Nuss, G. W. (1962) Carrageenan-induced
oedema in the hind paw of rat as an assay for anti-inflammatory
activity. Proc. Soc. Exp. Biol. Ther. III: 544–547