British Journal of Rheumatology 1996;35:711718

EXPRESSION OF CYCLOOXYGENASE-2 IN HUMAN AND AN ANIMAL

MODEL OF RHEUMATOID ARTHRITIS
R. Y. KANG, J. FREERE-MOAR, E. SIGAL* and C.-Q. CHUt
Syntex Discovery Research, Institute of Immunology and *Lipid Metabolism, Palo Alto, CA 94304, USA and
^Kennedy Institute of Rheumatology, London

SUMMARY
An inducible form of cyclooxygenase-2 (COX-2) has been shown to be upregulated in vitro by various pro-inflammatory agents,
such as lipopolysaccharide, IL-1 and TNF. COX-2 appears to be responsible for the increase in prostaglandin synthesis at the
site of inflammation. To examine the involvement of COX-2 in inflammation, we analysed the expression of this gene in human
rheumatoid arthritis (RA) and in rat adjuvant-induced arthritis. Immunocytochemical studies of synovial membrane biopsies
from human RA, osteoarthritic (OA) and normal joints using a COX-2 specific antibody showed positive staining in RA, but
not in normal synovial membranes. Specifically, expression of COX-2 was detected in synovial lining cells, lymphoid aggregates
and endothelial cells of blood vessels. Although some positive staining was observed in the OA joints, the number of stained
cells was dramatically lower and the staining of the cells was less intense than in the rheumatoid tissue. By reverse transcription
and polymerase chain reaction analysis, COX-2 mRNA was detected in the rat adjuvant arthritic limb, whereas no COX-2
mRNA was detectable in the normal limb. These observations indicate that COX-2 expression is upregulated in inflammatory
joint disease and that COX-2 is a potential therapeutic target for specific inhibition.
KEY WORDS: Rheumatoid arthritis, Adjuvant-induced arthritis, Synovial membrane, Immunohistochemistry, Prostaglandin
synthase, Inflammation.

Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011

MAMMALIAN cells contain two distinct isoforms of
cyclooxygenase (COX) [1-6], the target for aspirin
and other non-steroidal anti-inflammatory drugs
(NSAIDs). Both isoforms catalyse arachidonic acid to
prostaglandin H2 (PGH3), the precursor of pro-
inflammatory prostanoids including the prostaglandins
and thromboxanes [7-9]. Prostaglandins potently
mediate diverse biological functions such as gastric
protection [10], renal integrity [11,12], vascular
homeostasis [13], as well as inflammation [14-16].
COX-1 is constitutively expressed in virtually all
mammalian tissues [17]. In contrast, COX-2 expression
appears to be restricted to the prostate gland and to
specific areas of the brain [16, 18-20]. COX-2 has been
characterized as an early inducible gene [1, 2, 5]. Unlike
COX-2 expression, COX-1 levels increase only slightly
following growth factor or hormonal stimulation
[1, 4, 16, 20-23]. Mitogens and other pro-inflammatory
stimuli induce COX-2 [2,4,21,24-31], while
corticosteroids attenuate its expression [32, 33]. There-
fore, it has been proposed that the upregulation of
COX-2 results in the increase in prostaglandin
production associated with inflammation, while
COX-1 is involved in cellular house-keeping functions
[34-36].
COX-2 expression has been shown to be upregulated
in two acute rat models of inflammation: the
carrageenan paw model and the carrageenan-induced
air pouch inflammation model [37, 38]. In rheumatoid
synovial explant cultures, COX-2 but not COX-1, has
been shown to be upregulated by IL-1)? [39]. In the
Submitted 31 July 1995; revised version accepted 2 February 1996.
Correspondence to: J. Freire-Moar, Roche Bioscience (S3-6), 3401
Hillview Avenue, Palo Alto, CA 94303, USA.
711
same study, there was also some immunohistochemical
evidence for the expression of COX-2 in rheumatoid
arthritis (RA) synovial membranes [39]. All of these
data indicate that COX-2 is directly involved in
inflammation. To further characterize the expression of
COX-2 in inflamed tissues, we examined COX-2
expression in human RA and osteoarthritis (OA), and
in rat adjuvant-induced arthritis. These are chronic
systemic conditions involving the synovial membranes
of multiple joints. As with other inflammatory diseases,
human RA is associated with high levels of PGE2
[40, 41] that are thought to be produced by COX-2 in
synovial tissue and infiltrating cells. Here, using a
specific antibody against COX-2, we have
demonstrated that high levels of COX-2 isoenzyme are
expressed in synovial tissues from RA patients, while
only low levels of COX-2 expression can be detected in
synovial tissues from OA patients. No COX-2
expression can be detected in synovial membrane from
normal joints. Using reverse transcription and
polymerase chain reaction (RT-PCR), we also
demonstrate that COX-2 mRNA is present in the joints
of adjuvant-induced arthritic rat joints, but absent in
normal rat joints.
MATERIALS AND METHODS
Antibody generation and purification
A peptide corresponding to a unique sequence of the
C-terminal region of the murine COX-2 [2] with the
sequence CNASASHSRLDDINPT was synthesized as
previously described [42], and conjugated with keyhole
limpet haemocyanin (KLH) obtained from Pierce
Chemical Co., according to the manufacturer's
protocol. White New Zealand rabbits (females) were
immunized with the peptide-KLH conjugate at the
1996 British Society for Rheumatology
712 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 8
Pocono Rabbit Farm and Laboratory, Inc., in
Canadensis, PA, USA. Sera were collected at weekly
intervals and tested for immunoreactivity against the
peptide. Positive sera were pooled and passed through
a HiTrap Protein A affinity column from Pharmacia
LKB, USA, to isolate the IgG fraction, and F(abOj
fragments were prepared using an isolation system
provided by Pierce Chemical Co., USA, according to
the manufacturer's directions. A polyclonal rabbit IgG
anti-COX-1 was obtained from Oxford Biomedical
Research Inc., MI, USA.
Antibody characterization
Recombinant human COX-1 and COX-2 were
expressed and purified in milligram quantities in a
baculovirus expression system [43]. Equivalent
concentrations of COX-1 and COX-2 were loaded onto
Mini Protean SDS gels from Biorad, USA. Gels were
run for 45 min at 200 V and electrotransferred to
nitrocellulose membranes (obtained from Schleicher &
Schuell, USA) at 4C overnight. The membranes were
blocked in Tris-buffered saline (TBS) containing 5%
non-fat milk for 1 h and probed with anti-COX-2
(1:1000) or anti-COX-1 (1:100) (obtained from Oxford
Biomedicals, USA) for 2 h in TBS. Membranes were
then washed with TBS for 30 min and incubated for 1 h
with goat anti-rabbit-alkaline phosphatase (AP)
(1:500) from Boehringer Mannheim, USA. After the
1 h incubation, the membranes were washed and
allowed to soak in BM-Purple alkaline phosphatase
substrate from Boehringer Mannheim, USA. Colour
detection of immunoreactive bands was apparent after
20-30 min of incubation at room temperature. The
membranes were washed in ddH2O and dried.
Detection of COX-1 and COX-2 mRNA in rat
adjuvant-induced arthritic joints by RT-PCR
All animal studies were performed according to
institutional and federal guidelines. Mineral oil was
obtained from Sigma, USA. Heat-killed and dried
Mycobacterium butyricum was obtained from Difco,
USA. Female Sprague-Dawley rats from Charles
River, Portage, MI, were injected intradermally in the
proximal portion of the tail with a suspension of
bacteria and mineral oil (1 mg/0.1 ml) ground using
mortar and pestle. Seventeen days later, the rats were
euthanized and hind paws were amputated at the
tarsocrural joint. The hind paws were homogenized in
the presence of RNAzol from Biotecx Laboratories,
USA, with a Brinkman tissue homogenizer, and total
RNA was isolated according to the manufacturer's
protocol. One microgram aliquots of total RNA were
reverse transcribed using the Geneamp RNA PCR
system from Perkin-Elmer Cetus, Foster City, CA. The
cDNA was then diluted 1:5 with 1 x TE buffer from
Digene Diagnostics, MD. All PCR reagents and
equipment were from Perkin-Elmer Cetus, Foster City,
CA. One microlitre aliquots of the cDNA were
amplified in 50 /il of PCR mixture containing 5 /il of
10 x PCR buffer, 8/il of 1.25 mM dNTP cocktail,
1.0 JUM final concentration each of sense and antisense
primers, and 2.5 U of Taq polymerase. The sequences
of the sense primers were 5'-ACTCACTCAGTTT-
GTTGAGTCATTC-3' and 5-TGCATGTGGCT-
GTGGATGTCATCAA-3' for COX-2 and COX-1,
respectively. The sequences for the antisense primers
were 5'-TTT GATTAGTACTGTAGGGTTAATG-3'
and 5'-CACTAAGACAGACCCGTCATCTCCA-3'
for COX-2 and COX-1, respectively [29]. Rat G3PDH
primers from Clontech, Palo Alto, CA, were used in
separate PCR reactions to control for the efficiency of
cDNA synthesis in each sample. The COX-1 primer set
yields a PCR product of 450 bp and the COX-2 primer
set yields a 583 bp product [29]. The amplification
conditions were 94C for 1 min (denaturation), 55C
for 2 min (annealing) and 72C for 3 min (extension)
for 55 cycles. Aliquots of the PCR products were run
on a 1.2% TBE agarose gel for analysis.
Immunohistochemical localization of COX-2
Synovial membrane biopsies were taken from 10
patients (three males and seven females) with RA
fulfilling the revised American Rheumatism
Association criteria [44]. The age of the patients ranged
Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011
from 30 to 68 yr and disease duration from 2 to 21 yr.
All patients had active disease when the samples were
taken and were receiving drug therapy: NSAIDs
(naproxen, n = 3; ibuprofen, n = 1), D-penicillamine
(n = 3) and corticosteroids (n = 4).
The synovial biopsies were snap frozen in iso-
pentene cooled in a liquid nitrogen bath and stored at
70C. Six micrometre sections were fixed in
acetone/methanol (1:1) at -20C for 10 min and
treated with 0.3% hydrogen peroxide in methanol for
20 min. The sections were blocked with 20% normal
goat serum for 30 min at room temperature. Bio-
tinylated rabbit F(ab0i anti-COX-2 (15 /ig/ml) or bio-
tinylated normal rabbit F(abO2 antibodies (15pg/ml)
from a pre-immunization bleed were incubated on the
sections overnight at 4C. The sections were then
washed with 1 x phosphate-buffered saline (PBS) and
incubated with streptavidin-linked horseradish peroxi-
dase from Amersham International, Aylesbury, for
60 min at room temperature. Diaminobenzidine
(0.25 mg/ml) from Sigma Chemical Co., Poole, was
dissolved in 1 x PBS containing 0.03% hydrogen
peroxide, applied to sections and allowed to develop
for 5 min. The sections were counterstained with
haemotoxylin and examined with a light microscope.
In order to identify vascular endothelial cells,
parallel staining on sequential sections was performed
with anti-COX-2 and mouse monoclonal anti-von
Willebrand Factor antibody (Dako, High Wycombe).
The binding of anti-von Willebrand Factor antibody
was developed by horseradish peroxidase-conjugated
goat anti-mouse Ig and followed by diaminobenzidine.
Frozen sections of knee joint biopsies from OA
patients (n = 5) and normal synovial membrane
(/i = 3) from amputation knee joints of patients with
osteogenic sarcoma were also included in this study.
Prior to staining, biotinylated anti-COX-2 F(abO2
antibodies were incubated with a range of concen-
 KANG ET AL:. PGHS-2 EXPRESSION IN RA 713

trations (5-50/ig/ml) of synthetic COX-2 peptide
overnight in order to confirm the specificity of
anti-COX-2 antibody staining in the tissue.
The number of COX-2-positive cells was determined
by counting up to 500 cells in each area of the tissue
examined and the results were expressed as a
percentage of positive cells over the total cells.
Percentages of positive cells in RA and OA synovial
membrane were compared using the Wilcoxon rank
sum test (two-tailed).
RESULTS
Antibody characterization
The specificity of the anti-COX-2 antibody was
determined by Western blot analysis using recombinant
human COX-1 and COX-2 expressed in baculovirus
[43], as indicated in Materials and methods. When
membranes containing equivalent amounts of human
COX-1 and COX-2 were probed with the anti-COX-2
antibody, a 70 kDa immunoreactive band was detected
only in the sample containing human COX-2 protein,
but not from the sample containing the human COX-1
an anti-COX-1 antibody was used (Fig. lb). In several
other experiments, the anti-COX-1 antibody showed
cross-reactivity with human COX-2. Because of the
lack of specificity, anti-COX-1 antibody was not used
in later experiments for immunohistochemical staining
of synovial membranes.
No immunoreactive bands were detected when
membrane blots were probed with a control rabbit IgG
antibody.
The anti-COX-2 antibody was also able to
immunostain lipopolysaccharide (LPS)-treated human
mononuclear cells (data not shown). No immuno-
staining was detected on untreated human mono-
nuclear cells.
COX-2 mRNA expression in rat adjuvant-induced
arthritis
Initial attempts to detect COX-2 expression in
inflamed rat joints gave erratic results because of the
sample processing necessary for calcium removal from
the bone. As an alternative method, COX expression in
normal and inflamed rat joints was analysed by

Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011

protein (Fig. la). However, a 70 kDa immunoreactive RT-PCR using specific primers for COX-1 and COX-2
band in the human COX-1 sample was detected when as indicated in Materials and methods. Oligonucleotide

(a)

tn tn
O
->
CO

CO iS
CM C\l
1 3 *7 3
X X O X X o
o O .2 O O .2
o o o O O o

70 kD- 70 kD *

FIG. 1.Western blot characterization of anti-COX-2 antibody. Inununoblots containing equivalent amounts (100 ng) of human recombinant
COX-1 (hCOX-1) and COX-2 (hCOX-2) were probed with anti-COX-2 antibody (a) and anti-COX-1 antibody (b). Detection of immunoreactive
bands was carried out using a goat anti-rabbit antibody conjugated to alkaline phosphatase as indicated in Materials and methods.
714 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 8

All the sections from the 10 RA patients showed

Arthritic joint Normal joint positive staining with the COX-2 antibody (Fig. 3a).
The specificity of staining detected with the COX-2
I I antibody was evaluated by competition studies using
C\J C\J
X Q X X
Q
X different concentrations of the COX-2 antigenic
Q_ Q_
O CO O O CO O peptide, and by immunostaining tissues with an
o o o o o o irrelevant antibody of the same isotype. Biotinylated
normal F(abO2 fragments derived from pre-immuniz-
ation rabbit serum did not stain any cells at 15 /ig/ml
in RA tissue (Fig. 3b). The staining of the COX-2
antibody on RA synovial membranes was inhibited by
1000bp COX-2 peptide in a concentration-dependent manner,
and staining was abrogated at 50 /xg/ml of COX-2
peptide (Fig. 3c).
COX-2-specific staining was also observed in some
cells in the synovial membrane sections from OA
patients. However, the staining intensity was much
lower than in RA tissue and the number of positive
cells was fewer (Fig. 3d). None of the normal synovial
membrane sections were stained with the COX-2
antibody (Fig. 3e). A quantitative analysis of the
staining of RA and OA synovial membranes with

Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011

FIG. 2.Induction of COX-2 mRNA expression in rat adjuvant-in-
duced arthritis. COX-1 and COX-2 mRNA expression in the hind
paws of normal and adjuvant-induced arthritic rats were analysed by
RT-PCR as indicated in Materials and methods using specific
primers for COX-1, COX-2 and G3PDH. The PCR products
generated after 55 cycles of amplification were analysed by
electrophoresis in a 1.2% TBE agarose gel. A 1 kb DNA ladder was
used as a molecular weight standard. The sizes of the amplified DNA
fragments are indicated on the side in base pairs.
primers specific for COX-1 amplified a 450 bp DNA
fragment in the normal as well as inflamed joints.
However, primers specific for COX-2 yielded a 583 bp
fragment only in the inflamed joint (Fig. 2). To control
for the efficiency of cDNA synthesis among the
different samples and to ensure the use of equivalent
amounts of cDNA in the PCR reactions, aliquots of
cDNA from the various samples were amplified using
G3PDH primers. In all of these experiments, negative
controls were included to ensure that positive results of
the RT-PCR were not caused by contamination.
Detection of COX-2 in rheumatoid synovia! tissue
COX expression in inflamed and normal joints was
analysed by immunohistochemistry using the anti-
COX-2 specific antibody as described in Materials and
methods.
TABLE I
Expression of cyclooxygenase-2 in RA and OA synovia! membrane*
Lymphoid
Lining layer aggregate Inter-aggregate
RA (n - 10) 85 6.0 20 3.0 15 2.0
OA (n = 5) 24 4.0 3 0.5 7 2.0
'Results are expressed as the percentage of positive staining cells.
anti-COX-2 antibody was carried out by counting the
number of COX-2-positive cells in each sample, as
indicated in Materials and methods. The results of
these analyses are summarized in Table I, demonstrat-
ing that most of the cells in the RA synovial
membrane express COX-2, while only a small
percentage of cells are expressing COX-2 in OA
synovial membranes.
The COX-2-positive cells showed a cytoplasmic
staining pattern. COX-2 antibody stained most of the
lining layer cells (Figs 3a and 4a). In the sublining
layers, stained cells were scattered throughout the
tissue section with an accumulation in perivascular
areas and within lymphoid aggregates (Fig. 4b).
Morphologically, these positive cells were monocytes/
macrophages. Most of the vascular endothelial cells
were also stained (Fig. 4c and d). In the cartilage/
pannus junction, some cells in the synovial pannus were
stained, but chondrocytes were not (Fig. 4e). In all
experiments, a concentration of 15 Mg/ml of COX-2
antibody was used.
DISCUSSION
Previous studies have suggested that COX-2 is
responsible for increased prostaglandin production in
inflamed tissue, while COX-1 is involved in the
production of prostaglandins necessary to maintain
homeostasis, such as the integrity of the gastro-
intestinal mucosa, and other normal physiological
functions. COX-1 expression has been detected in
multiple tissues [17], while COX-2 expression appears
to be restricted mainly to inflamed tissues. Induction of
COX-2 expression has been reported in two rat models
of inflammation [37, 38]. In these models, COX-2-
specific inhibitors reduced the oedema in the inflamed
areas without any of the side-effects associated with
traditional NSAIDs that are not selective for COX-2.
 KANG ET AL.: PGHS-2 EXPRESSION IN RA 715

* *

mat,

Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011

(C) (d)

(e)
FlO. 3.Immunohistochemical staining of synovial membranes from patients with RA or OA and normal joints. Six micrometre sections of
synovial membranes from RA, OA and normal joints were stained using biotinylated F(ab')j fragments (15 //g/ml) from anti-COX-2 antibody
or an irrelevant antibody as indicated in Materials and methods. Immunoreactivity was detected using streptavidin-linked horseradish peroxidase
(brownish staining). Sections were counterstained with haemotoxylin (blue staining), (a) Synovial membrane from an RA patient stained with
anti-COX-2 antibody, (b) Control staining of RA synovial membrane stained with normal rabbit IgG. (c) Control staining of RA synovial
membranes with anti-COX-2 IgG pre-absorbed with 50 iig/ml of COX-2 peptide. (d) OA synovial membrane stained with anti-COX-2 antibody,
(e) Normal synovial membrane stained with anti-COX-2 antibody. Original magnifications: (a-d) x 100; (e) x40.
716 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 8

> '* /

(a) (b)

Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011

(c) (d)

64
FIG. 4.Localization of COX-2 antigen in RA synovial membranes. Synovia! membrane sections from RA patients were stained with a
biotinylated F(ab')j fragment anti-COX-2 antibody as described in Materials and methods. Positive staining was seen as brown deposits and
haematoxylin counterstaining was blue, (a) Staining of the lining layer from a RA synovhim. (b) Staining of lymphoid aggregates from RA
synovium. (c) von WiUebrand Factor staining indicating vascular endothelium. (d) COX-2 staining on a sequential section to that of (c), showing
vascular endothelial cells stained for COX-2. (e) Staining of the cartilage (Q/pannus (P) junction in RA synovium. Original magnifications x 100.
 KANG ET AL.: PGHS-2 EXPRESSION IN RA
NSAIDs inhibit COX effectively, yet use of these
drugs leads to detrimental side-effects. Published reports
to date support the hypothesis that the side-effects of
NSAIDs are because of the inhibition of COX-1, while
the anti-inflammatory effects are the result of COX-2
inhibition [16]. Unlike NSAIDs which act on the COX
enzyme, corticosteroids specifically inhibit COX-2
expression at the transcriptional level and are potent
agents in the treatment of inflammatory disease [32].
Two previous studies have examined cyclooxygenase
expression in human RA [39,45]. However, these
studies did not address the specific expression of
COX-2 in relation to non-inflammatory joint disease.
In order to further characterize the expression of
COX-2 in inflamed tissue, we analysed the expression
of COX-2 in human RA and in OA (a non-
inflammatory joint disease), as well as in normal joints.
In our studies, we found that cells expressing COX-2
displayed a cytoplasmic staining pattern consistent with
findings from in vitro studies [7]. Staining was observed
throughout the lining layer and scattered staining was
seen in sublining layers of the RA synovium. The
scattered staining was concentrated in perivascular
areas, vascular endothelial cells, and lymphoid
aggregates consisting of monocytes and macrophages.
In contrast to RA, inflammation is a minor
component of OA [46]. Staining of OA synovial
membranes with the anti-COX-2 antibody showed very
little reactivity. This suggests that COX-2 may be a
contributor to the inflammatory component of RA.
The causative agents of COX-2 upregulation in RA
have not been identified. However, in vitro studies have
demonstrated that cytokines such as IL-1 and TNFa,
which are present at high levels in synovial fluid of RA
patients, induce COX-2 expression [4,6,21]. It is
possible that these cytokines may modulate the in vivo
expression of COX-2 in RA.
We extended our studies on COX-2 expression by
examining rodent models of inflammation. Rat
adjuvant-induced arthritis is a well-characterized model
of human RA. Symptomatically, peripheral and axial
joints are particularly affected. Histological sections of
these joints show pannus formation with mononuclear
and polymorphonuclear cell infiltration similar in
appearance to human RA [47]. By RT-PCR, we found
mRNA expression of COX-2 in the affected joints, but
not in normal joints. Also, COX-2 mRNA was
detectable in the carrageenan-induced oedema in rat
and in freshly isolated human synoviocytes (R. Y. Kang
et al., unpublished observation). These results agree
with previously published data showing induction of
COX-2 by carrageenan in the rat paw, the air-pouch
model of inflammation [37, 38] and in synovial cells
isolated from RA joints [39]. All of these data strongly
support the hypothesis that COX-2 is involved in
inflammation, and that the correlative increase in PG
production associated with inflammation is very likely
to be the result of the enzymatic activity of the COX-2
isoform. Conversely, the homeostatic functions of
cyclooxygenase are likely to be due to COX-1 activity.
Because common NSAIDs display no specificity for
717
one isoform, a COX-2-specific inhibitor may be an
attractive therapeutic allowing anti-inflammatory
activity devoid of side-effects such as NSAID
gastropathy.
ACKNOWLEDGEMENTS
We thank Dr J. Krstenansky for the synthesis of the
COX-2 peptide, Dr J. Barnett and J. Chow for the
human COX-1 and COX-2 proteins, and Dr B. H.
Devens, Dr D. Webb and Dr P. Whiteley for then-
review of this manuscript.
REFERENCES
1. Kujubu DA, Fletcher BS, Varnum BC, Lim RW,
Herschman HR. TIS10, a phorbol ester tumor promoter-
inducible mRNA from Swiss 3T3 cells, encodes a novel
prostaglandin synthase/cyclooxygenase homologue.
J Biol Chem 1991;266:12866-72.
2. Fletcher BS, Kujubu DA, Perrin DM, Herschman HR.
Structure of the mitogen-inducible TIS10 gene and
demonstration that the TlSlO-encoded protein is a
functional prostaglandin G/H synthase. / Biol Chem
1992;267:4338-44.
Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011
3. Regier MK, DeWitt DL, Schindler MS, Smith WL.
Subcellular localization of prostaglandin endoperoxide
synthasc-2 in murinc 3T3 cells. Arch Biochem Biophys
1993;301:439-44.
4. Hla T, Neilson K. Human cyclooxygenase-2 cDNA. Proc
Natl Acad Sci USA 1992;89:7384-8.
5. Xie W, Chapman JG, Robertson DL, Erikson RL,
Simmons DL. Expression of a mitogen-responsive gene
encoding prostaglandin synthase is regulated by mRNA
splicing. Proc Natl Acad Sci USA 1991;88:2692-6.
6. Feng L, Sun W, Xia Y et al. Cloning two isofonns of rat
cyclooxygenase: differential regulation of their ex-
pression. Arch Biochem Biophys 1993;307:361-8.
7. Needleman P, Turk J, Jakshichk BA, Morrison AR,
Lefkowith JB. Arachidonic acid metabolism. Annu Rev
Biochem 1986;55:69-102.
8. Yamamoto S. Biology and chemistry of prostaglandins
and related eicosanoids. In: Curtis-Prior PB, ed. Prosta-
glandins: biology and chemistry of prostaglandins and
related eicosanoids. London: Churchill Livingstone,
1988:37-45.
9. Smith WL, DeWitt DL, Karaemer SA et al. Structure-
function relationships in sheep, mouse and human
prostaglandin endoperoxide G/H synthases. Adv Prosta-
glandin Thromboxane Leukotriene Res 1990;20:14-21.
10. Whittle BJR, Higgs GA, Eakins KE, Moncada S, Vane
JR. Selective inhibition of prostaglandin production in
inflammatory exudates and gastric mucosa. Nature
1980;284:271-3.
11. Moore PK. Prostanoids and the kidney. In: Moore PK,
ed. Prostanoids: pharmacological, physiological, and
clinical relevance. New York: Cambridge University
Press, 1985:173-99.
12. Quilley J, McGiff JC. Renal arachidonic acid
metabolism. In: Hushka PV, Mais DE, eds. Eicosanoids
in the cardiovascular and renal systems. Norwell, MA:
MTP Press Limited, 1988:16-33.
13. Weksler BB, Eldor A, Falcone D, Levin RI, Jaffe EA,
Minick CR. Prostaglandins and vascular endothelium.
In: Herman AG, Vanhoutte PM, Denolin H, Goossens
A, eds. Cardiovascular pharmacology of the prosta-
glandins. New York: Raven Press, 1982:137-48.
14. Cromwell O. Contribution of eicosanoids to the im-
mediate and late inflammatory responses. In: Church M,
718 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 8
Robinson C, eds. Eicosanoids in inflammatory conditions
of the lung, skin and joints. Norwell, MA: MTP Press
Limited, 1988:21-42.
15. Zurier RB. Prostaglandins and inflammation. In:
Curtis-Prior PB, ed. Prostaglandins: biology and
chemistry of prostaglandins and related eicosanoids. New
York: Churchill Livingstone, 1988:595-607.
16. DeWitt DL, Meade EA, Smith WL. PGH synthase
isoenzyme selectivity: the potential for safer nonsteroidal
antiinflammatory drugs. Am J Med 1993;95:40-6.
17. DeWitt DL. Prostaglandin endoperoxidase synthase:
regulation of enzyme expression. Biochim Biophys Acta
1991;1083:121-34.
18. Tippetts MT, Varnum BC, Lim RW, Herschman HR.
Tumor promoter-inducible genes are differentially
expressed in the developing mouse. Mol Cell Biol
1988;8:4570-2.
19. Smith WL, Gutekunst DI, Lyons RH Jr. Immunocyto-
chemical localization of the prostaglandin-forming
cyclooxygenase in the cerebral cortex. Prostaglandins
1980; 19:61-9.
20. Yamagata K, Andreasson KI, Kaufmann WE, Barnes
CA, Worley PF. Expression of a mitogen-inducible
cyclooxygenase in brain neurons: regulation by synaptic
activity and glucocorticoids. Neuron 1993;ll:371-86.
21. Jones DA, Carlton CP, Mclntyre TM, Zimmerman GA,
Prescott SM. Molecular cloning of human prostaglandin
endoperoxidase synthase type II and demonstration of
expression in response to cytokines. J Biol Chem
1993^68:9049-54.
22. Tohkin M, Kishino J, Ishizaki J, Arita H. Pancreatic-
type phospholipase A2 stimulates prostaglandin syn-
thesis in mouse osteoblastic cells (MC3T3-E1) via a
specific binding site. J Biol Chem 1993;268:2865-71.
23. Smith CJ, Morrow JD, Roberts LJ II, Marnett LJ.
Differentiation of monocytoid THP-1 cells with phorbol
ester induces expression of prostaglandin endoperoxidase
synthase-1 (Cox-1). Biochem Biophys Res Commun
1993;192:787-93.
24. Ryseck RP, Raynoschek C, Macdonald-Bravoy H,
Dorfman K, Mattei M-G, Bravo R. Identification of an
immediate early gene, pghs-B, whose protein product has
prostaglandin synthase/cyclooxygenase activity. Cell
Growth Diff 1992;3:443-50.
25. Hempel SL, Monick MM, Hunninghake GW. Lipo-
polysaccharide induces prostaglandin H synthase-2
protein and mRNA in human alveolar macrophages and
blood monocytes. J Clin Invest 1994;93:391-6.
26. DuBois RN, Awad J, MOITOW J, Roberts LJ II, Bishop
PR. Regulation of eicosanoid production and mitogene-
sis in rat intestinal epithelial cells by transforming growth
factor-alpha and phorbol ester. J Clin Invest
1994;93:493-8.
27. Sirois J, Levy L, Simmons DL, Richards JS. Character-
ization and hormonal regulation of the promoter of the
rat prostaglandin endoperoxidase synthase 2 gene in
granulosa cells. J Biol Chem 1993;268:12199-206.
28. Sirois J, Simmons DL, Richards JS. Hormonal
regulation of messenger ribonucleic acid encoding a
novel isoform of prostaglandin endoperoxidase H
synthase in rat preovulatory follicles. J Biol Chem
1993;267:11586-92.
29. Lee SH, Soyoola E, Chanmugam P et al. Selective
expression of mitogen-inducible cyclooxygenase in
macrophages stimulated with lipopolysaccharide. / Biol
Chem 1992;267:25934-8.
30. O'Sullivan MG, Chilton FH, Huggins EM Jr, McCall
CE. Lipopolysaccharide priming of alveolar macro-
phages for enhanced synthesis of prostanoids involves
induction of a novel prostaglandin H synthase. J Biol
Chem 1992;267:14547-50.
31. Kennedy BP, Chan CC, Culp SA, Cromlish WA.
Cloning and expression of rat prostaglandin endoperox-
ide synthase (cyclooxygenase)-! cDNA. Biochem Biophys
Res Commun 1993;197:494-500.
32. O'Banion MK, Sadowski HB, Winn V, Young DA. A
serum- and glucocorticoid-regulated 4-kilobase mRNA
encodes a cyclooxygenase-related protein. J Biol Chem
1991;266:23261-7.
33. Kujubu DA, Herschman HR. Dexamethasone inhibits
mitogen induction of TSI10 prostaglandin synthase/
cyclooxygenase gene. J Biol Chem 1992;267:7991-4.
34. Vane J. Towards a better aspirin. Nature 1994;367:215-6.
35. Meade EA, Smith WL, DeWitt DL. Differential
inhibition of prostaglandin endoperoxidase synthase
(cyclooxygenase) isoenzyme by aspirin and other
non-steroidal anti-inflammatory drugs. J Biol Chem
1993;268:6610-4.
36. Mitchell JA, Akarasereenont P, Thiemermann C, Flower
RJ, Vane JR. Selectivity of nonsteroidal antiinflam-
matory drugs as inhibitors of constitutive and inducible
cyclooxygenase. Proc Natl Acad Sci USA 1994;90:
Downloaded from rheumatology.oxfordjournals.org by guest on July 2, 2011
11693-7.
37. Seibert K, Zhang Y, Leahy K et al. Pharmacological and
biochemical demonstration of the role of cyclooxygenase
2 in inflammation. Proc Natl Acad Sci USA 1994;91:
12013-7.
38. Masferrer JL, Zweifel BS, Manning PT et al. Selective
inhibition of inducible cyclooxygenase 2 in vivo is
antiinflammatory and nonulcerogenic. Proc Natl Acad
Sci USA 1994^1:3228-32.
39. Crofford LJ, Wilder RL, Ristimaki AP et al. Cyclooxy-
genase-1 and -2 expression in rheumatoid synovial
tissues. J Clin Invest 1994;93:1095101.
40. Cush JJ, Lipsky PE. The immunopathogenesis of
rheumatoid arthritis: the role of cytokines in chronic
inflammation. Clin Aspects Autoimmun 1987;1:2-13.
41. Salmon JA, Higgs GA, Vane JR et al. Synthesis of
arachidonate cyclooxygenase products by rheumatoid
and nonrheumatoid synovial lining in nonproliferative
organ culture. Ann Rheum Dis 1983;42:36-9.
42. Krstenansky JL, Owen TJ, Yates MT, Mao SJT. The
C-terminal binding domain of hirullin PI8: antithrombin
activity and comparison to hirudin peptides. FEBS Lett
1990;26<h425-9.
43. Barnett J, Chow J, Ives D et al. Purification,
characterization and selective inhibition of human
prostaglandin G/H synthase 1 and 2 expressed in the
baculovirus system. Biochim Biophys Acta 1994;1209:
130-9.
44. Arnett FC, Edworthy SM, Bloch DA et al. The
American Rheumatism Association 1987 revised criteria
for the classification of rheumatoid arthritis. Arthritis
Rheum 1988;31:315-24.
45. Sano H, Hla T, Maier JAM et al. In vivo cyclooxygenase
expression in synovialtissuesof patients with rheumatoid
arthritis and osteoarthritis and rats with adjuvant and
streptococcal cell wall arthritis. / Clin Invest 1992;89:
97-108.
46. Dieppe P. Strategies for the prevention of osteoarthritis.
Int J Tiss Reac 1993;15:93-7.
47. Brahn E. Animal models of rheumatoid arthritis. Clues
to etiology and treatment Clin Orthopaed Relat Res
1991;265:42-53.