Independent regulation of plasminogen activator inhibitor 2 and plasminogen activator inhibitor 1 in human synovial fibroblasts.

1526
INDEPENDENT REGULATION OF
PLASMINOGEN ACTIVATOR INHIBITOR 2 AND
PLASMINOGEN ACTIVATOR INHIBITOR 1 IN
HUMAN SYNOVIAL FIBROBLASTS
JOHN A. HAMILTON, DAISY CHEUNG, ENRICO L. FILONZI, DIANA S. PICCOLI, JOHANN WOJTA,
MARISA GALLICHIO, KATHERINE McGRATH, and KARENA LAST
Objective. To study the plasminogen activator
inhibitor(s) (PAI) produced in vitro by human synovial
fibroblast-like cells.
Methods. Human synovial cell explant cultures
were established using cells from nonrheumatoid donors. PAI-2 and PAL1 antigens were measured by
enzyme-linked immunosorbent assay, and messenger
RNA (mRNA) levels were determined by Northern blot.
Results. The synovial fibroblasts produced both
PA13 and PAI-1. Interleukin-1 (IL-1) increased PA13
but decreased PAI-1 formation, both at the protein and
the mRNA levels. Using cyclooxygenase inhibitors, evidence was obtained that an endogenous cyclooxygenase
product(s) in the IL-1-treated cultures inhibited formation of both PAIs; exogenous prostaglandin E, (10-7M)
reversed the effect of cyclooxygenase inhibition. The
to 10q7M) inhibglucocorticoid dexamethasone (
From the Department of Medicine, University of Melbourne, and the Department of Diagnostic Haematology, Royal
Melbourne Hospital, Parkville, Victoria, Australia.
Supported by a program grant from the National Health and
Medical Research Council of Australia and by the National Heart
Foundation of Australia. Dr. Wojta's work was supported by grant
7617 from the Austrian Fund for the Promotion of Scientific Research.
John A. Hamilton, PhD, DSc: Department of Medicine;
Daisy Cheung, BSc (Hons): Department of Medicine; Enrico L.
Filonzi, BSc (Hons): Department of Medicine; Diana S. Piccoli, BSc
(Hons): Department of Medicine; Johann Wojta, PhD: Department
of Diagnostic Haematology; Marisa Gallichio, BSc (Hons): Department of Diagnostic Haematology; Katherine McGrath, MBBS,
FRACP: Department of Diagnostic Haematology; Karena Last, BSc
(Hons): Department of Medicine.
Address reprint requests to John A. Hamilton, PhD, DSc,
Department of Medicine, University of Melbourne, Royal Melbourne Hospital, Parkville 3050,Victoria, Australia.
Submitted for publication March 26, 1992; accepted in
revised form August 10, 1992.
Arthritis and Rheumatism, Vol. 35, No. 12 (December 1992)
ited IL-l-stimulated PAL2 formation but reversed the
suppressive effect of IL-1 on PAI-1 production.
Conclusion. PA13 formation and PAI-1 formation can be regulated independently in human synoviocytes, illustrating the complexity of the modulation of
the net PA activity expressed by these cells.
The extent of tissue remodeling and cell migration that occurs during, for example, inflammatory
reactions may be dependent on the net cellular proteolytic activity (1). For instance, such activity in the joint
tissue of rheumatoid arthritis patients is likely to be
determined in part by the balance between the levels
of proteases and those of their corresponding inhibitors. One such proteolytic system of potential significance to these events is the plasminogen activator
(PA)-PA inhibitor (PAI) system. There are 2 mammalian PAS, urokinase-type PA (uPA) and tissue-type PA
(tPA), which are serine proteases capable of cleaving
plasminogen to generate the less specific protease,
plasmin (2). Urokinase-type PA is considered important in tissue remodeling, cell migration, and neoplasia
(3), while tPA is generally viewed as being critical in
fibrin resolution (4).
Two inhibitors of both PAS are PAI-1 and PAI-2
(5); however, PAI-2 shows a preference for uPA and
PAI-1 for tPA, leading to the view that PAI-1 is the
primary physiologic inhibitor of fibrinolysis mediated
by tPA (5). The physiologic role of PAI-2 is unclear,
although its tissue distribution suggests that it may
function locally during inflammation and wound healing (6). However, the roles of the 2 PAIs have not been
thoroughly defined as yet. Both PAIs are members of
the serine protease inhibitor (serpin) superfamily (9,
PA1 IN SYNOVIAL FIBROBLASTS
the gene structures for the 2 human molecules having
now been determined (7-10).
Activated synovial fibroblast-like cells have
been implicated in the invasive properties of the
inflamed synovium of rheumatoid lesions (11). It is
likely that their properties in such lesions are being
modified by locally generated cytokines; it has been
shown, for example, that interleukin-1 (IL-1) elicits a
number of responses in such cells (12-14). Elevated
PA levels have been found in rheumatoid tissue explant cultures and synovial fluids when compared with
their noninflamed counterparts (15,16). Also, plasmin
has been shown to degrade cartilage (17). Cytokineactivated synovial fibroblasts could be a source of uPA
activity in rheumatoid lesions, and the control of its
expression by these cells could be important in understanding the pathogenesis of such lesions, which includes prominent deposition of fibrin in the synovial
tissue (18). We have proposed that enhanced PA
activity in these cells, resulting from the action of
monocyte/macrophage cytokines, could be one manifestation of the expression of an immature, invasive,
“tumor-like’’ phenotype, which could help to explain
some of the features of the erosive synovial “pannus”
in rheumatoid joints (19).
PAI-1 is produced by many cultured cells,
particularly those of mesodermal origin (5,6); PAI-2,
which was first isolated from placenta, appears to have
a more restricted distribution. Of particular relevance
to joints, human monocytes produce PAI-2 and PAI-1
in vitro (20,21), while PAI-1 is produced by human
chondrocytes and cartilage tissue in vitro (22,23). It
was heretofore not known whether human synovial
fibroblast-like cells produce PAI-2 and/or PAI-1. We
report herein that they can produce both PAI-2 and
PAI-1. In this particular cell type, PAI-2 levels are
increased by IL-1, but levels of PAI-1 are decreased
by the cytokine; treatment with a glucocorticoid elicited changes opposite to those brought about by IL-1.
MATERIALS AND METHODS
Synovial cell cultures. Human synovial cell explant
cultures were established from different nonrheumatoid donors as previously described (18). Passaged cells were functionally and morphologically similar to the synovial fibroblast-like cells (18). Cells were plated for 8 hours at 5 x lo4
cells/0.5 mywell in 48-well plates (Costar, Cambridge, MA)
in a-modified minimal essential medium (a-MEM; Commonwealth Serum Laboratories, Parkville, Australia) supplemented with 10% heat-inactivated (56°C for 30 minutes) fetal
calf serum (FCS). The cells were then incubated overnight in
1527
fresh medium, usually containing 1% FCS depleted of plasminogen (18) (this allows concomitant measurement of PA
activity [18]). For experiments, incubation was in 500 pl
a-MEM plus 1% FCS (or 10% FCS) in the presence or
absence of cytokine or drugs, usually for 24 hours. Supernatants and cell lysates were collected and assayed for PAI-2
and PAI- 1.
PAL2 enzyme-linked immunosorbent assay (ELISA).
PAI-2 was quantitated using an ELISA developed at Biotech
Australia Pty. Ltd. (Roseville, Australia). In this assay,
96-well plates (Nunc-Immuno Plate; Nunc, Roskilde, Denmark) were coated overnight at room temperature with
monoclonal murine anti-human PAT-2 IgG (MAI-21,2 pg/ml;
Biopool, Umea, Sweden). Samples or standard at C200
ng/ml (either native human PAI-2 [24] supplied by E.
Kruithof [Lausanne, Switzerland] or recombinant human
material [25] [supplied by Biotech Australia]) were added to
the well for 2 hours, followed by goat anti-PAI-2 IgG (2.0
pg/ml; Biotech Australia) and then by peroxidaseconjugated rabbit anti-goat IgG (1 23,000 dilution; Dakopatts,
Copenhagen, Denmark). The sensitivity of the assay was 3
ng/ml, and absorbance was linearly related to PAI-2 concentration up to 50 ng/ml. The PAI-2 assay recognizes both free
PAI-2 and PAI-2 complexed with uPA.
PAI-1 ELISA. PAI-1 antigen in the samples was
quantitated by ELISA (26,27). Briefly, a monoclonal antiPAI-1 antibody (5PAI12; obtained from B. R. Binder, University of Vienna), which recognizes active PAI-1, latent
PAI-1, and PAI-1 in complex with tPA immobilized to a
micro-ELISA plate was used to bind the PAI-1 contained in
the sample. A second peroxidase-labeled monoclonal antiPAI-1 antibody (3PAI5; obtained from B. R. Binder), which
recognizes active and latent PAI-1 as well as PAI-1 in
complex with tPA, was used to quantitate the amount of
bound PAI-1. Purified melanoma PAI-1 (26) was used as a
calibration standard (27). The sensitivity of the assay was 0.2
ng/ml, and absorbance was linearly related to PAI-1 concentration up to 40 ng/ml.
Detection of messenger RNA (mRNA). Synovial fibroblasts were plated at 5 x lo5 cells in 100 mm-diameter tissue
culture dishes (Corning Labware, Corning, NY) and treated
as described previously (28). Total RNA was prepared as
Table 1. Effect of interleukin-1 (IL-1) on human synovial fibroblast PAL2 and PAI-1 levels*
PAI-2 (ng/ml)
Addition
IL-1
Cellassociated
23
62
f2
f 3t
PAI- 1 (ng/ml)
Extracellular
Cellassociated
Extracellular
4.8 f 1
10.3 f lb
1.7 f 0.2
ND
111 2 10
64 2 5$
* Sixth-passage synovial cells from donor MC, cultured (5 X lo4
cells/well) in triplicate as described in Materials and Methods, were
incubated for 24 hours at 37°C with or without IL-la (20 pM). Cell
lysates and supernatants were collected and assayed for plasminogen activator inhibitor 2 (PAI-2) and PAI-1 content by enzymelinked immunosorbent assay. ND = not detected.
t P < 0.001 versus control.
$ P < 0.02 versus control.
HAMILTON ET AL
1528
reported earlier (28), using the method of Chirgwin et a1 (29).
Up to 10 &ane of RNA was fractionated on a formaldehydecontaining 1% agarose gel before transfer to Hybond-Nf
membrane (Amersham, Sydney, Australia). The filter was
hybridized overnight at 42°C in a standard hybridization
buffer (29) containing >2 x lo6 counts per minute of
32P-labeled PAI-2 complementary DNA (cDNA) (10). After
hybridization, the membrane was washed several times and
exposed to Kodak X A R J film (Eastman-Kodak, Rochester,
NY) at -70°C (28).
Reagents. Recombinant human IL-1 a (specific activity 5.7 x lo7 unitshg) was obtained as a gift from P.
Lomedico (Hoffmann-La Roche, Nutley, NJ). Dexamethasone, indomethacin, and prostaglandin E, were from Sigma
(St. Louis, MO). Human cDNA probes were obtained from
the following sources: PAI-2 from T-C. Wun (Monsanto,
St. Louis, MO), PAI-1 from D. Loskutoff (Scripps Clinic
and Research Foundation, La Jolla, CA), and GAPDH
from 0. Bernard (Walter and Eliza Hall Institute, Parkville,
Australia).
Statistical analysis. Measurements in supernatants
and cell lysates are presented as the mean 2 SEM values
from triplicate cultures. The significance of differences was
assessed using Student's 2-tailed t-test; P values less than
0.05 were considered significant.
RESULTS
Production of PAL2 and PAI-1 by synovial fibroblasts. The conditioned media and cell lysates of
human synovial fibroblast-like cells were examined for
both PAI-2 and PAI-1, by ELISA. In Table 1, it can be
seen that the cells produced both PAIs, with PAI-2
being predominantly found in the cell lysates while
PAI-1 was mostly in the conditioned medium. This
relative distribution was observed in a total of 30
human synovial fibroblast populations, irrespective of
passage number.
Table 2. Effect of FCS concentration on IL-1 down-modulation of
PAI-1 levels*
Extracellular PAL1 (ng/ml)
Experiment 2
Experiment 1
Addition
1% FCS
10% FCS
1% FCS
10% FCS
IL-1
73 f 2
64 +. 2t
133 +. 6
88 2 2t
66 +. 5
205 +. 11
115 2 41
72
2
3
* Seventh-passage synovial cells from donor GE, cultured (5 x lo4
cells/well) in triplicate in a-modified minimal essential medium
containing 1% fetal calf serum (FCS) or 10% FCS as described in
Materials and Methods, were incubated for 24 hours at 37°C with or
without IL-la (20 pM). Supernatants were collected and assayed for
PAI-1 content by enzyme-linked immunosorbent assay. Values are
the mean 2 SEM. See Table 1 for other definitions.
t P < 0.05 versus control.
t P < 0.01 versus control.
§ P < 0.002 versus control.
160{
z
;;a
c
v
c;'
140 -
120100-
"1 7r""/
60
0
0.02
0.2
2
IL-1 (pM)
I
20
Figure 1. Human synovial fibroblast plasminogen activator inhibitor 2 (PAI-2) levels after treatment with interleukin-1 (IL-1). Fifthpassage synovial cells from donor JT, cultured (5 x lo4 celldwell) in
triplicate as described in Materials and Methods, were incubated for
24 hours at 37°C with increasing concentrations of IL-la. Cell
lysates were collected and assayed for PAI-2 content by enzymelinked immunosorbent assay. Values are the mean f SEM.
Effect of IL-1 on synovial fibroblast PAI-2 and
PAI-1 levels. IL-1 has been found to stimulate the
formation of a number of synovial products (for examples, see refs. 12-14). As demonstrated in Table 1,
IL-1 was shown in the present study to raise levels of
PAI-2. The stimulation of PAI-2 by IL-1 was observed
in 15 synovial cell lines studied.
In contrast, Table 1 shows that there was a
reduction in levels of PAI-1 in the presence of IL-1.
Under the same culture conditions (i.e., in 1% FCS),a
slight reduction in PAI-1 levels induced by IL-1 was
noted with 14 other synovial lines but not observed
with 4; this small effect of IL-1 was not due to a
redistribution of PAI-1 between extracellular and cellassociated compartments.
The data in Table 1 were obtained from cultures
performed in low serum concentrations, in order to
reduce the contribution of serum. It was reasoned that
raising the FCS level to 10% might stimulate the cells
to produce PAI-1 and the inhibitory action of IL-1
might therefore be more readily apparent, since it
presumably can suppress only newly formed PAI-1.
As can be observed in 2 separate experiments (Table
2), the basal PAI-1 level was higher for cultures in 10%
FCS than for cultures in 1% FCS, and the inhibitory
action of IL-1 could be demonstrated more easily.
The dose-responsiveness of the effect of IL-1
on synoviocyte PAI-2 levels in cell lysates is presented
PA1 IN SYNOVIAL FIBROBLASTS
1529
h
3
200
\
bn
c
v
z
100
U
0
4
8
12
24
Time (h)
Figure 2. Kinetics of the IL-1 effect on human synovial fibroblast
PAI-2 levels. Second-passage synovial cells from donor AB, cultured (5 X lo4 cells/well) in triplicate as described in Materials and
Methods, were incubated at 37°C for various time periods in the
absence (0)or presence (0)of IL-la (20 pM). Cell lysates were
collected and assayed for PAL2 by content by enzyme-linked
immunosorbent assay. Values are the mean f SEM (where not
shown, error bars are smaller than the symbol). See Figure 1 for
definitions.
in Figure 1; in this and other experiments, the IL-1 was
active at concentrations of -0.2 pM. The kinetics with
IL-1 at an optimal concentration are depicted in Figure
2. The first changes in the cell lysates were measurable
between 2 and 4 hours after addition of the cytokine.
Under all conditions, the majority of the PAI-2 remained associated with the cell. The kinetics of the
suppressive effect of IL-1 at an optimal concentration
on supernatant PAI-1 levels under the same culture
conditions are presented in Figure 3. Differences were
observed within 3 hours.
PA1 mRNA levels. Northern blot analysis (Figure 4A) indicated that IL-1 could raise the level of
PAI-2 mRNA, possibly reflecting an increase in PAI-2
gene transcription. Consistent with this, we have
found that the RNA synthesis inhibitor actinomycin D
(2 pglml) blocked the increased formation of PAI-2
due to IL-1 in an 8-hour experiment (results not
shown). Protein synthesis also appeared to be required, since cycloheximide (10 pglml) inhibited the
increase in antigen expresshn (not shown). The reduction in PAI-1 transcripts by IL-1 is presented in Figure
4B; PAI-1 mRNA is usually found as a doublet (79,23).
Effect of a cyclooxygenase inhibitor. Endogenous cyclooxygenase products, including prostaglandin E, (PGE,), are produced in IL-1-treated
synovial cell lines and can, in turn, regulate the levels
of certain products under the control of IL-1
(12,13,30). As can be seen in Table 3, the cyclooxygenase inhibitor indomethacin potentiated the action of
IL-1 in enhancing PAI-2 formation and exogenous
PGE, reversed the effect of indomethacin on PAI-2
expression. The effect of indomethacin was noted in 8
other experiments and was confirmed using another
cyclooxygenase inhibitor, naproxen (lO-'A4). These
data suggest that an endogenous cyclooxygenase product(s) is quite critical in determining the increase in
PAI-2 levels brought about by IL-1.
Likewise, the inhibitory effect of IL-1 on PAI-1
levels could be reversed, at least partially, by indomethacin and restored by exogenous PGE,; data from
an experiment with cells cultured in both 1% FCS and
10% FCS are presented in Table 4. This effect of
indomethacin was found for 9 of 12 synovial cell lines
cultured in 1% FCS and 5 of 5 lines cultured in 10%
FCS. With some cell lines, indomethacin (lO-'A4)
raised the basal PAI-1 levels.
Effect of glucocorticoids. A glucocorticoid, such
as dexamethasone, can often have an effect opposite
to that of IL-1 in many cell types, including synoviocytes (29-31). In Table 5, it is shown that dexamethasone could suppress the action of IL-1 in enhancing
h
2
-.
bn
cl
v
c
o r * ' . ' . ' . ' . ' . ' . ' "
0
3
12
6
24
Time (h)
Figure 3. Kinetics of the IL-1 effect on human synovial fibroblast
PAL1 levels. Sixth-passage synovial cells from donor PL, cultured
(5 x lo4 cells/well) in triplicate as described in Materials and
Methods, were incubated at 37°C for various time periods in the
absence (0)or presence (0)of IL-1a (20 pM) in a-modified minimal
essential medium containing 1% fetal calf serum. Supernatants were
collected and assayed for PAI- 1 by enzyme-linked immunosorbent
assay. Values are the mean SEM (where not shown, error bars are
smaller than the symbol). See Figure 1 for definitions.
*
1530
HAMILTON E T AL
Figure 4. Effect of IL-1 on steady-state PA1 messenger RNA levels. Cytoplasmic RNA was isolated from &hour cultures of human
synoviocytes, and Northern blot analysis was performed as described in Materials and Methods. A, Studies using fifth-passage
cells from donor LL. The blot was probed with 32P-labeled PAI-2
complementary DNA (cDNA). Lane 1, Control; lane 2, IL-la (20
pM). Ethidium bromide-stained ribosomal RNA bands indicated
even loading in the lanes. B, Studies using second-passage cells from
donor FH. The blot was probed with "P-labeled PAI-1 cDNA and
GAPDH cDNA (the latter as a housekeeping gene to check RNA
loading). Lane 1, Control; lane 2, IL-la (20 pM). See Figure 1 for
other definitions.
PAI-2 formation, an observation confirmed under the
same culture conditions with 5 other lines using lop610p7M dexamethasone. With 3 of 5 lines, treatment
with the steroid slightly reduced basal PAI-2 levels
(Table 5 ) . Dexamethasone (lO-'it4) also suppressed
the increase in PAI-2 mRNA levels resulting from IL-I
action, as measured by Northern blot analysis (results
not shown).
When PAI-1 levels were monitored in the presence of glucocorticoid, it was found that the steroid
did not suppress these levels and reversed the inhibitory action of IL-1 (Table 6). When this experiment
was carried out with 12 additional lines cultured in 1%
FCS or 10% FCS, dexamethasone (lop6 to 10-7M)
enhanced basal PAI-1 levels in 5 cases but had no
significant effect in 7; the opposing action of the
steroid on the IL-1-mediated changes was demonstrated in 11 of 12 lines.
DISCUSSION
This is the first report on the nature of the PAIs
produced by cultured human synovial fibroblast-like
cells. We have shown that this cell population can
PA1 IN SYNOVIAL FIBROBLASTS
1531
Table 5. Effect of glucocorticoid on human synovial fibroblast
PAL2 levels*
Table 3. Cyclooxygenase inhibition and PAI-2 regulation*
PAI-2 (ng/ml)
Addition
Iedo
Indo
+ PGE,
IL-1
IL-1
IL-1
+ Indo
+ Indo + PGE,
Cell-associated
Extracellular
23 f 2
16 t 2
10 t 2
521
ND
ND
62 t 3
165 f 8 t
52 f 4
10 f 1
27 f 2f
12 f 1
* Sixth-passage synovial cells from donor MC, cultured
lo4
cells/well) in triplicate in 1% fetal calf serum as described in
Materials and Methods, were incubated for 24 hours at 37°C with or
without IL-la (20 pM). Some cultures also contained indomethacin
(Indo) (W5M) in the presence or absence of prostaglandin E,
(PGE,) (10-7M). Cell lysates and supernatants were collected and
assayed for PAL2 content by enzyme-linked immunosorbent assay.
Values are the mean f SEM. See Table 1 for other definitions.
t P < 0.001 versus treatment with IL-1 alone.
t. P < 0.002 versus treatment with IL-1 alone.
(5
X
produce both PAI-2 and PAI-1. As has been generally
reported for other cell types (6,32), PAI-2 was mainly
cell associated and presumably in a nonglycosylated
form (33) (although when monocyte/macrophages are
treated in certain ways, they can secrete PAI-2 that is
mostly in a glycosylated form [33]). The converse is
true for the distribution of PAI-1 in the synovial cell
cultures, as is usually the case for this PA1 (5,6).
It was also shown that IL-la stimulates levels
of PAI-2 protein and mRNA, but reduces those for
PAI-1. Purified recombinant human IL-lp gave similar
results (data not shown). These observations indicate
PAL2 (ng/ml)
Addition
Dex
IL- 1
IL-1
+ Dex
Cell-associated
Extracellular
18 2 0.2
13 2 1
521
422
83 2 2
32f I t
30 t 2
12 f 1f
* Sixth-passage synovial cells from donor GD, cultured (5 x lo4
cells/well) in triplicate as described in Materials and Methods, were
incubated for 24 hours at 37°C with or without IL-la (20 pM) in the
presence or absence of dexamethasone (Dex) (10-7M). Cell lysates
and supernatants were collected and assayed for PAI-2 content by
enzyme-linked immunosorbent assay. Values are the mean f SEM.
See Table 1 for other definitions.
t P < 0.001 versus treatment with IL-1 alone.
t. P < 0.002 versus treatment with IL-1 alone.
that the synthesis of these 2 PAIs can be modulated
independently, as has been found in other cells (6).
IL-1 has been shown in many studies to enhance PAI-1 synthesis (3,although in human chondrocytes we showed for the first time that it could have
the reverse effect (23). It is not that the synovial cells
cannot produce PAL1 in response to a cytokine, since
transforming growth factor p can stimulate them to do
so (Hamilton JA et al: submitted for publication). In
human umbilical vein endothelial cells, IL-1 did not
affect PAI-2 production, but increased the amount of
PAI-1 produced by the cells (6). Others have shown
Table 6. Effect of glucocorticoid on human synovial fibroblast
PAI-1 levels*
Table 4. Cyclooxygenase inhibition and PAI- 1 regulation*
PAI-I (ng/ml)
Extracellular PAI- 1 (ng/ml)
Addition
-
1% FCS
10% FCS
7.0 f 0.5
8.5 f 0.4
1.8 t 0.2
1.9 +. 0.2
2.9
5.7
2.5
Indo
Indo
+ PGE,
4.1 t 0.2
4.8 t 0.7
0.6 2 0.1
IL-1
IL-1
IL-1
+ Indo
+ Indo + PGE,
3.2
0.4
f 0.3t
f 0.1
f 0.1
f 0.3.t
f 0.1
~~
* Seventh-passage synovial cells from donor FH, cultured (5 x lo4
cells/well) in triplicate in a-modified minimal essential medium
containing 1% fetal calf serum (FCS) or 10% FCS as described in
Materials and Methods, were incubated for 24 hours at 37°C with or
without IL-la (20 pM). Some cultures also contained indomethacin
(Indo) (10-5M) in the presence or absence of prostaglandin E,
(PGE,)
Supernatants were collected and assayed for PAI-1
content by enzyme-linked immunosorbent assay. Values are the
mean f SEM. See Table 1 for other definitions.
t P < 0.05 versus treatment with IL-1 alone.
$ P < 0.001 versus treatment with IL-1 alone.
1% FCS
10% FCS
Dex
PGE,
Dex + PGE,
5.5 f 0.6
7.0 t 0.4
2.2 t 0.5
2.9 f 0.4
11.3 f 0.2
13.8 f 0.4
6.5 f 0.1
8.5 f 0.9
IL- 1
IL-1
IL-1
IL-1
3.2 2 0.3
4.4 f 0.2t
1.7 f 0.4
2.5 f 0.2
7.1 ? 0.3
10.1 2 0.4f
5.2 2 0.1
7.1 f 0.5
Addition
+ Dex
+ PGE,
+ Dex + PGE,
* Third-passage synovial cells from donor WD, cultured (5 x lo4
celldwell) in triplicate in 1% fetal calf serum (FCS) or 10% FCS as
described in Materials and Methods, were incubated for 24 hours at
37°C with or without IL-la (20 pM) in the presence or absence of
dexamethasone (Dex) ( 10-7M), prostaglandin E, (PGE,) (10-7M),
or both. Supernatants were collected and assayed for PAI-1 content
by enzyme-linked immunosorbent assay. Values are the mean f
SEM. See Table 1 for other definitions.
t P < 0.05 versus treatment with IL-1 alone.
f P < 0.01 versus treatment with IL-1 alone.
1532
that IL-1 stimulates PAI-2 mRNA and PAI-1 mRNA
expression in human MRC-5 fetal lung fibroblasts, but
induces an increase in only the latter mRNA species in
an SV40-transformed human fibroblast cell line,
XP12RO (34). Thus, the cellular PA1 response to IL-1
is cell type specific, as can occur for cytokine effects in
a number of vitro systems. The diverse findings for
PA1 regulation with the various human cell populations, including fibroblasts, highlight the importance of
using fibroblasts isolated from synovial tissue if one is
to be able to relate the findings to arthritic disease.
Evidence was also given above for downmodulation of the IL-1-mediated stimulation of PAI-2
formation by an endogenous cyclooxygenase product(s), since cyclooxygenase inhibitors could potentiate the action of IL-1. Exogenous PGE, reversed the
effect of indomethacin. Likewise, the inhibition of
PAI-1 formation by IL-1 was partially due to the
formation of a cyclooxygenase product(s). Regulation
of the metabolism of IL- I-treated synovial fibroblasts
by an endogenous cyclooxygenase product(s) has been
noted before (12,13,30,35) and demonstrates that this
type of product can influence the degree of response
elicited by IL-1.
We and others have previously shown that
enhancement of uPA activity in human synoviocytes
by IL- 1 and mononuclear cell-conditioned medium
was partially dependent on the generation of a cyclooxygenase product (13,35,36). It has been postulated that the potential down-regulation of synoviocyte
uPA by cyclooxygenase inhibitors could represent a
significant portion of the clinical action of these drugs
in the treatment of inflammatory arthritides (13); in
light of the findings reported herein, it could be that
this modulation of uPA activity may be occurring in
part by prevention of suppression of PA1 levels produced by endogenous prostanoids in the IL-1-treated
synoviocyte cultures. In other words, part of the
action of this class of nonsteroidal antiinflammatory
drug in joint disease may be to maintain PA1 levels by
blocking the formation of endogenous PGE,.
Given the evidence that IL-1 induction of synovial cell PA activity is somewhat dependent on the
generation of endogenous PGE, and, subsequently, on
intracellular elevation of CAMP (13,36), it is likely that
endogenous PGE, is functioning via this mechanism in
down-regulating PAL2 and PAI-1 expression. In support of this contention, there is evidence for 2 putative
binding sites for a cAMP response element in the
promoter region of the human PAI-2 gene (37), and
cholera toxin, an activator of adenylate cyclase, in-
HAMILTON ET AL
duced PAL2 in cultured murine macrophages (38);
treatment of UCT/gl-1 or endothelial cells with exogenous cAMP or forskolin caused a decrease in PAI-I
activity (3p-41).
The glucocorticoid, dexamethasone, while usually having slight and varied effects on basal PA1 levels
in the synoviocytes, inhibited IL- 1-stimulated PAL2
formation but reversed the suppressive effect of IL-1
on PAI-1 levels. These results provide yet another
example of the contrasting pattern of regulation of the
2 PAIs and also provide further examples of the
opposing actions of glucocorticoids and IL-1 (31). It
has been reported that, in other cells, dexamethasone
can reduce PAL2 formation while potentiating that of
PAI-1 (32). Dexamethasone has been shown to reduce
PAI-2 activity in murine macrophages (38), while 1
study showed down-regulation of secreted PAI-2 antigen and mRNA expression in the human fibrosarcoma
line HT1080 (32) whereas another showed no reduction in the amounts of secreted PAL2 in this line (6).
Dexamethasone can increase PAI-1 activity in a number of cell types, but fails to do so in a variety of other
cell lines (5).
Net uPA activity and mRNA expression in
human synovial cells have previously been shown to
be down-regulated by glucocorticoids (28,42,43); part
of the loss of activity could be due to enhanced PAI-I
expression. This down-regulation of uPA activity has
been proposed as providing a partial explanation for
the antiinflammatory action of glucocorticoids in rheumatic disease (28,42,43). The enhancement of PAI-I
formation by this class of drug in these cells would
provide extra protection against any putative neighboring matrix breakdown in inflamed joints arising
from the uPA activity of the synoviocytes; however,
the inhibition of PAI-2 formation in the same cell type
would favor the expression of net uPA activity.
We have reported previously that IL-1 stimulates low and transient uPA activity and uPA mRNA
expression in the same synoviocytes (13,28,44). The
varied regulation of the 2 PAIs described above,
including the influence of endogenous cyclooxygenase
products, indicates the complexity in the regulation of
this protease activity by the synovial cells. It is intriguing that this cell type, which is believed to play a part
in some of the destructive processes evident in inflammatory joint disease (ll), should be producing both
types of PAI, one (PAI-1) often believed to be more
important in fibrinolysis and the other (PAI-2) in
inflammation, pregnancy, and extravascular events
(6). However, it is possible that the functions of these
PA1 IN SYNOVIAL FIBROBLASTS
2 PAIs may not be as clearly delineated as is generally
assumed and it may be, for example, that PAI-1 has a
role in controlling inflammation and tissue destruction.
The suppression by IL-1 of synoviocyte PAI-1 formation may therefore favor the possibility of local PAmediated proteolysis in inflamed joints.
Human monocytes produce PAI-1 and PAI-2
(20), and chondrocytes produce PAI-1 (22,23). Perhaps the prominent fibrin deposition noted in rheumatoid synovial tissue (45) owes its presence in part to
PAIs in joints (presumably PAI-I [5]), and based on
our findings above, the synoviocyte could be making a
contribution. PAI-2 and PAI-1 have been found in
rheumatoid synovial fluids (Hamilton JA et al: unpublished observations) and could be derived in part from
the synoviocytes.
ACKNOWLEDGMENTS
The authors thank Biotech Australia Pty. Ltd. for the
supply of recombinant human PAI-2 and goat anti-PAI-2
IgG and for support and contributions to this program, J.
Bartlett for the supply of synovial tissue, and S. Wong for
typing the manuscript.
REFERENCES
1. Reich E: Activation of plasminogen: a general mechanism for producing localised extracellular proteolysis,
Molecular Basis of Biological Degradative Processes.
Edited by RD Berlin, M Herrman, IH Lepow. New
York, Academic Press, 1978
2. Saksela 0: Plasminogen activation and regulation of
pericellular proteolysis. Biochim Biophys Acta 823:3565, 1985
3. Blasi F, Vassalli J-D, Dan0 K: Urokinase-type plasminogen activator: proenzyme, receptor and inhibitors. J
Cell Biol 104:801-804, 1987
4. Loscalzo J , Braunwald E: Tissue plasminogen activator.
N Engl J Med 319:925-931, 1988
5. Andreasen PA, Georg B, Lund LR, Riccio A, Stacey
SN: Plasminogen activator inhibitors: hormonally regulated serpins. Mol Cell Endocrinol 68:l-19, 1990
6. Schleef RR, Wagner NV, Loskutoff DJ: Detection of
both type 1 and type 2 plasminogen activator inhibitors
in human cells. J Cell Physiol 134:26%274, 1988
7. Loskutoff D, Linders M, Keijer J, Veerman H, van
Heerikhuizen H, Pannekoek H: Structure of the human
plasminogen activator inhibitor 1 gene: nonrandom distribution of introns. Biochemistry 26:3763-3768, 1987
8. Bosma PJ, van den Berg EA, Kooistra T, Siemiencak
DR, Slightom JL: Human plasminogen activator inhibitor-1 gene: promoter and structural gene nucleotide
sequences. J Biol Chem 263:9129-9141, 1988
1533
9. Strandberg L , Lawrence D, Ny T: The organization of
the human plasminogen activator-inhibitor-1 gene. Eur J
Biochem 176:609-616, 1988
10. Ye RD, Ahern SM, LeBeau MM, Lebo RV, Sadler JE:
Structure of the gene for human plasminogen activator
inhibitor-2: the nearest mammalian homologue of
chicken ovalbumin. J Biol Chem 2645495-5502, 1989
11. Harris ED Jr: Pathogenesis of rheumatoid arthritis,
Textbook of Rheumatology. Second edition. Edited by
WN Kelley, ED Harris Jr, S Ruddy, CB Sledge. Philadelphia, WB Saunders, 1985
12. Dayer J-M, de Rochemonteix B, Burrus B, Demczuk S,
Dinarello C: Human recombinant interleukin-1 stimulates collagenase and prostaglandin E, production by
human synovial cells. J Clin Invest 77:645-648, 1986
13. Leizer T , Clarris BJ, Ash PE, van Damme J, Saklatvala
J, Hamilton JA: Interleukin-lp and interleukin-la stimulate the plasminogen activator activity and prostaglandin E, levels of human synovial cells. Arthritis
Rheum 30562-566, 1987
14. Guerne P, Zuraw B, Vaughan J, Lotz M: Synovium as a
source of IL-6 in vitro: contribution to local and systemic manifestations of arthritis. J Clin Invest 83585592, 1989
15. Werb Z, Mainardi CL, Vater CA, H a m s ED Jr: Endogenous activation of latent collagenase by rheumatoid
synovial cells: evidence for a role of plasminogen activator. N Engl J Med 296:1017-1023, 1977
16. Mochan E, Uhl J: Elevations in synovial fluid plasminogen activator in patients with rheumatoid arthritis. J
Rheumatol 11: 123-128, 1984
17. Lack CJ, Rogers HJ: Action of plasmin on cartilage.
Nature 182:948-949, 1958
18. Hamilton JA, Slywka J: Stimulation of human synovial
fibroblast plasminogen activator production by mononuclear cell supernatants. J Immunol 1262351-855, 1981
19. Hamilton JA: Hypothesis: in vitro evidence for the
invasive and tumor-like properties of the rheumatoid
pannus. J Rheumatol 102345-851, 1983
20. Castellote JC, Grau E, Linde MA, Piejol-Max N , Rutlant ML: Detection of both type 1 and type 2 plasminogen activator inhibitors in human monocytes. Thromb
Haemost 63:67-71, 1990
21. Vassalli J-D, Dayer J-M, Wohlwend A, Belin D: Concomitant secretion of prourokinase and of a plasminogen
activator-specific inhibitor by cultured human monocytes-macrophages. J Exp Med 159: 1653-1668, 1984
22. Yamada H , Stephens RW, Nakagawa T , McNicol D:
Detection and partial characterisation of a specific plasminogen activator inhibitor in human chondrocyte cultures. J Biochem 104:96&967, 1988
23. Campbell IK, Last K, Novak U, Lund LR, Hamilton
JA: Recombinant human interleukin-1 inhibits plasminogen activator inhibitor- 1 (PAI- 1) production by human
1534
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
articular cartilage and chondrocytes. Biochem Biophys
Res Commun 174:251-257, 1991
Kruithof EKO, Vassalli J-D, Schleuning WD, Mattaliano RJ, Bachmann F: Purification and characterization of a plasminogen activator inhibitor from the histiocytic lymphoma cell line U-937. J Biol Chem 261:
11207-11213, 1986
Antalis TM, Clark MA, Barnes T, Lehrbach PR, Devine
PL, Schevzov G, Goss NH, Stephens RW, Tolstoshev
P: Cloning and expression of a cDNA coding for a
human monocyte-derived plasminogen activator inhibitor. Proc Natl Acad Sci U S A 85:955-989, 1985
Wagner OF, Binder BR: Purification of an active plasminogen activator inhibitor immunologically related to
the endothelial type plasminogen activator inhibitor
from the conditioned media of a human melanoma cell
line. J Biol Chem 261:1447+14481, 1986
Wojta J, Holzer M, Hufnagl P, Christ G, Hoover RL,
Binder BR: Hyperthermia stimulates plasminogen activator inhibitor type-1 (PAI-1) expression in human umbilical vein endothelial cells (HUVEC) in vitro. Am J
Pathol 139:911-919, 1991
Vitti G, Hamilton JA: Modulation of urokinase-type
plasminogen activator messenger RNA levels in human
synovial fibroblasts by interleukin-1 , retinoic acid, and a
glucocorticoid. Arthritis Rheum 31: 1046-1051, 1988
Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ:
Isolation of biologically active ribonucleic acid from
sources enriched in ribonuclease. Biochemistry 18:
5292-5299, 1979
Hamilton JA, Piccoli DS, Cebon J, Layton JE, Rathanaswani P, McColl SR, Leizer T: Cytokine regulation of
colony-stimulating factor (CSF) production in cultured
human synovial fibroblasts. 11. Similarities and differences in the control of interleukin-1 induction of granulocyte-macrophage CSF and granulocyte-CSF production. Blood 79:1413-1419, 1992
Allison AC, Lee SW: The mode of action of antirheumatic drugs: anti-inflammatory and immunosuppressive effects of glucocorticoids. Prog Drug Res 33:
63-81, 1989
Medcalf RL, van den Berg E , Schleuning W-D: Glucocorticoid-modulated gene expression of tissue- and
urinary-type plasminogen activator and plasminogen activator inhibitor 1 and 2. J Cell Biol 106:971-978, 1988
Belin D, Wohlwend A, Schleuning W-D, Kruithof EKO,
Vassalli J-D: Facultative polypeptide translocation allows a single mRNA to encode the secreted and cytosolic forms of plasminogen activators inhibitor 2. EMBO
J 8:3287-3294, 1989
Michel JB, Quertermous T: Modulation of mRNA levels
for urinary- and tissue-type plasminogen activator and
plasminogen activator inhibitors 1 and 2 in human fibroblasts by interleukin-1. J Immunol 143:89&895, 1989
HAMILTON ET AL
35. Hamilton JA, Leizer T, Lingelbach SR: The stimulation
of human synovial fibroblast plasminogen activator activity: involvement of cyclic AMP and cyclooxygenase
products. Biochim Biophys Acta 886: 195-202, 1986
36. Mochan E, Uhl J, Newton R: Evidence that interleukin-1 induction of synovial cell plasminogen activator is
mediated via prostaglandin E, and CAMP. Arthritis
Rheum 29:1078-1084, 1986
37. Kruithof EKO, Cousin E: Plasminogen activator inhibitor 2: isolation and characterisation of the promoter
region of the gene. Biochem Biophys Res Commun
156:383-388, 1988
38. Wohlwend A, Belin D, Vassalli J-D: Plasminogen activator-specific inhibitors in mouse macrophages: in vivo
and in vitro modulation of their synthesis and secretion.
J Immunol 139:1278-1284, 1987
39. Andreasen PA, Christensen TH, Huang J-Y, Nielsen
LS, Wilson EL, Dan0 K: Hormonal regulation of extracellular plasminogen activators and Mr 54000 plasminogen activator inhibitor in human neoplastic cell lines,
studied with monoclonal antibodies. Mol Cell Endocrinol45:137-147, 1986
40. Santell L, Levin EG: Cyclic AMP potentiates phorbol
ester stimulation of tissue plasminogen activator release
and inhibits secretion of plasminogen activator inhibitor-1 from human endothelial cells. J Biol Chem 263:
16802-16808, 1988
41. Kooistra T, Bosma PJ, Toet K , Cohen LH, Griffioen M,
van den Berg E, le Clercq L, van Hinsbergh VWM: Role
of protein kinase C and cyclic adenosine monophosphate in the regulation of tissue-type plasminogen activator, plasminogen activator- 1, and platelet derived
growth factor mRNA in human endothelial cells: possible involvement of proto-oncogenes c-jun and c-fos.
Arteriosclerosis Thromb 11:1042-1052, 1991
42. Hamilton JA, Bootes AJ, Phillips PE, Slywka J: Human
synovial fibroblast plasminogen activator: modulation of
enzyme activity by antiinflammatory steroids. Arthritis
Rheum 24:129&1303, 1981
43. Medcalf RL, Richards RI, Crawford RJ, Hamilton JA:
Suppression of urokinase-type plasminogen activator
mRNA levels in human fibrosarcoma cells and synovial
fibroblasts by anti-inflammatory glucocorticoids. EMBO
J 5:2217-2222, 1986
44. Hamilton JA, Piccoli DS, Leizer T , Butler DM, Croatto
M, Royston AKM: Transforming growth factor p stimulates urokinase-type plasminogen activator and DNA
synthesis, but not prostaglandin E, production, in human synovial fibroblasts. Proc Natl Acad Sci U S A
88:7 180-7 184, 1991
45. Clemmensen I, Helund B, Andersen RB: Fibrin and
fibronectin in rheumatoid synovial membrane and rheumatoid synovial fluid. Arthritis Rheum 26:479-485, 1983