Matrix metalloproteinase 13 collagenase 3 in human rheumatoid synovium.

ARTHRITIS & RHEUMATISM
Vol. 40, No. 8, August 1997, pp 1391-1399
0 1997, American College of Rheumatology
1391
MATRIX METALLOPROTEINASE 13 (COLLAGENASE 3)
IN HUMAN RHEUMATOID SYNOVIUM
OTSO LINDY, YRJO T. KONTTINEN, TIM0 SORSA, YANLI DING, SEPPO SANTAVIRTA,
ARNOLDAS kEPONIS, and CARLOS LOPEZ-OTiN
Objective. To show the eventual presence and
extent of production of matrix metalloproteinase 13
(MMP-13, or collagenase 3) in rheumatoid synovial
tissue samples and extracts, and to assess the inhibition
characteristics of recombinant MMP-13.
Methods. Immunohistochemical avidin-biotinperoxidase complex staininglmorphometry was used to
analyze MMP-13-positive cells in situ. Neutral salt
extraction of synovial tissue, electrophoresis of the
extract in different buffer systems, and Western blotting
were also used. The inhibitory properties of doxycycline,
clodronate, pamidronate, and D-penicillamine for recombinant enzyme were determined with a soluble type
I1 collagen assay.
Results. MMP-13 was detected in fibroblast- and
macrophage-like mononuclear cells in the synovial lining and stroma and in vascular endothelial cells. The
overall expression of MMP-13 in these cells in the synovial
stroma was high in rheumatoid arthritis (86 2 12%)
compared with osteoarthritis (17 f 5%) patient samples
(P = 0.0027). In a high-pH native electrophoresis gel,
immunoreactivity to anti-MMP-1 and anti-MMP-13 were
clearly separated, with anti-MMP-13-immunoreactive
material migrating faster than anti-MMP-l-immunoreactive material. Finally, in contrast to MMP-1 and
Supported by the Finnish Rheumatism Research Foundation,
the Emil Aaltonen Foundation, the Center for International Mobility
(CIMO), the Rector of the University of Helsinki, the Academy of
Finland, and an Evo Clinical Research Grant from the Helsinki
University Central Hospital, Helsinki, Finland; the European League
Against Rheumatism, Switzerland; and the Comision Interministerial
de Ciencia y Tecnologia (SAF94-0892) and Glaxo-Wellcome, Spain.
Otso Lindy, MD, Tim0 Sorsa, DDS, Yanli Ding, DDS:
University of Helsinki, Helsinki, Finland; Yrjo T. Konttinen, MD,
Seppo Santavirta, M,D: Helsinki University Central Hospital, Helsinki,
Finland; Arnoldas Ceponis, MD: Invalid Foundation, Helsinki, Finland; Carlos L6pez-Otin, PhD: Universidad de Oviedo, Oviedo, Spain.
Address reprint requests to Otso Lindy, MD, Department of
Medical Chemistry, PO Box 8 (Siltavuorenpenger 10 A), FIN-00014
University of Helsinki, Helsinki, Finland.
Submitted for publication July 23, 1996; accepted in revised
form March 17, 1997.
MMP-8, MMP-13 was found to be relatively resistant
to the inhibitory effects of doxycycline and clodronate
in vitro.
Conclusion. Due to its localization in synovial
tissue, its substrate profile, increased expression, and
relative resistance to known MMP inhibitors, MMP-13
is suggested to play a major role in the pathogenesis of
tissue destruction in rheumatoid arthritis.
The role of collagenases and other matrix metalloproteinases (MMP) in the turnover and catabolism of
connective tissue proteins has raised a lot of interest
since the initial discovery of tadpole collagenase (1).
Today, the MMP superfamily includes 4 different groups
of enzymes, namely, collagenases (fibroblast-type collagenase 1, or MMP-1, neutrophil-type collagenase 2, or
MMP-8, and collagenase 3, or MMP-13), gelatinased
type IV collagenases, stromelysins (matrilysin and macrophage metalloelastase), and membrane-type MMPs
(currently 4 known members) (2-4). Several MMPs seem
to be involved in the normal turnover of joint tissues, as
well as the pathologic destruction of the rheumatoid
joint (5). Interstitial collagenases are the only mammalian enzymes capable of cleaving the triple-helical domain of fibrillar or group 1 collagens, which include the
major interstitial collagens type I, 11, and 111 (6).
The collagenic structures of the inflamed joint
may be attacked by either MMP-8 from the neutrophils
in the synovial fluid (7) or by MMP-1 secreted by
fibroblasts and macrophage-like cells in the synovium
and/or pannus (8,9). The recent cloning of human
collagenase 3/MMP-13 from mammary carcinoma tissue
has brought up some interesting new issues (3). The
collagenase cloned from mouse and rat osteoblasts
(lO,ll), which was originally thought to be an evolutionary deviation from the MMP-1 of other mammals, was
shown to have a counterpart in human tissue. Characterization of MMP-13 showed a substrate specificity
favoring degradation of cartilage type I1 collagen and
LINDY ET AL
1392
Table 1. Clinical characteristics of the study patients*
Patient1
sexiage
Diagnosis
RA
RA
RA
RA
RA
RA
OA
OA
OA
OA
OA
OA
OA
Osteochondritis dissecans
Luxation of patella
Rupture of meniscus
Disease
duration
(years)
>5
7
14
12
10
>5
5
13
5
>10
8
>5
10
5
113
113
Site of
synovial
biopsy
Shoulder
MCP joint
Elbow
PIP joint
PIP joint
Wrist
Knee
Knee
Knee
Knee
Knee
Knee
Knee
Knee
Knee
Knee
* RA = rheumatoid arthritis; MCP = metacarpophalangeal; PIP
proximal interphalangeal; OA = osteoarthritis.
=
joint tissue (12,13). MMP-13 was later cloned from
human cartilage (14,15) and quite recently, Wernicke et
a1 found that MMP-13 messenger RNA (mRNA) is also
expressed in rheumatoid arthritis (RA) and osteoarthritis (OA) synovial tissue, but not in any other tissue
examined (16). These observations suggest a synovial
tissue/articular cartilage-specific expression of MMP-13.
These observations prompted us to assess
whether RA synovial MMP-13 mRNA is translated to
the corresponding protein in situ, and in which cells and
to what extent does this occur compared with OA and
with some degenerative and/or traumatic joint changes.
Because collagenases have been considered one of the
prime targets for anticollagenolytic enzyme inhibitors,
such as tetracyclines for MMP-8 (17-19) and bisphosphonates (20), we also tested the inhibitory potential of
such compounds.
PATIENTS AND METHODS
Patients and samples. Synovial tissue specimens were
obtained from 16 consecutive patients at the time of joint
replacement surgery or diagnostic/therapeutic arthroscopy
(Table 1). Six of these patients (all women, mean age 52.5
years, age range 37-73) had RA (21), with a mean disease
duration of 8.8 years (range 5-14 years), and 7 of these patients
(6 women and 1 man, mean age 65 years, age range 48-90) had
OA (22), with a mean disease duration of 8 years (range 5-13
years). Samples were also available from 1 patient with osteochondritis dissecans, 1 with luxation of the patella, and 1 with
a ruptured meniscus. Diagnosis of the last 3 conditions were
proved by arthroscopy.
Ten synovial samples were obtained from the knee, 2
from the proximal interphalangeal joint, 1 from the meta-
carpophalangeal joint, 2 from the wrist, and l from the
shoulder joint. All samples were snap-frozen, embedded in
Tissue-Tek OCT compound (Lab-Tek Products, Elkhart, IN),
and stored at -70°C until used.
Cytospin preparation of neutrophils. To exclude crossreaction of the anti-MMP-13 antibody with MMP-8 collagenase, isolated leukocytes were used as a known MMP-8positive control. Polymorphonuclear leukocytes were isolated
from the buffy coat layer of a 20-ml blood sample by a standard
protocol in microfuge tube scale (23). A viability count was
obtained in 0.2% trypan blue in phosphate buffered saline, and
cells were adjusted to a concentration of 0.6 X 106/ml. Cytospin preparations were made at 6.0 X lo4 cells/slide. Slides
were air dried for 10 minutes and fixed in -20°C methanol for
10 minutes.
Immunohistochemical staining. Avidin-biotinperoxidase complex (ABC) staining (24) was utilized using
Vectastain Elite ABC rabbit kit (Vector, Burlingame, CA). Six
micrometer-thick cryostat sections were applied to gelatinformalin-coated slides, and fixed in acetone for 5 minutes
at 4°C.
Endogenous peroxidase activity in the tissue sections
and cytospin preparations was inhibited with 0.3% H,O, in
methanol for 30 minutes. Samples were incubated serially at
22°C in 1) normal goat serum (diluted 1:60) for 30 minutes, 2)
primary rabbit anti-human MMP-13 antibodies (diluted 1500)
for 60 minutes, 3) biotinylated goat anti-rabbit IgG (diluted
1:133) for 30 minutes, 4) AJ3C (diluted 1:lOO) for 30 minutes,
or 5) 3,3'-diaminobenzidine tetrahydrochloride (DAB; Sigma,
St. Louis, MO) in 0.006% H,O, solution in TBS (0.05M Tris,
0.15M NaC1, pH 7.4) followed by a wash in tap water.
Consecutive sections or cytospin preparations were counterstained with Harris hematoxylin or were not counterstained.
Slides with counterstaining were repeatedly washed in tap
water.
Finally, sections or cytospin preparations were dehydrated, cleared in xylene, and mounted in a synthetic mounting
medium (Diatex, Becker, Marsta, Sweden). Morphometry was
done using an ocular counting square (20 squares X 20
squares) and an oil immersion objective (1,000- magnification). At least 500 of each of synovial stromal cells, synovial
lining cells, and endothelial cells were counted to estimate the
ratio of positive cells. Lymphocytes were excluded from the cell
counts (25).
A 0.1% solution of bovine serum albumin (Sigma) in
TBS was used for dilution of the primary antibodies and sera.
Between steps (except after incubation in normal goat serum),
the slides were washed 3 times for 5 minutes in TBS. Routine
controls were omission or use of normal rabbit IgG (no. X903;
Dako, Glostrup, Denmark) instead of the primary antibody
and exposure of the tissue sections to DAB and H202 (for
endogenous peroxidase).
Double staining for MMP-13 and CD68. ABC-alkaline
phosphatase-anti-alkaline phosphatase (APAAP) double
staining was performed, and the color reactions were developed as described in detail elsewhere (24,26,27). Briefly,
consecutive 5 Fm-thick cryostat sections from 2 of the synovial
specimens used for MMP-13 staining were processed for
staining of MMP-13 according to the protocol for the ABC
technique (24) up to the step where the slides were rinsed in
tap water. Then, the slides were washed with TBS (0.05M
Tris, 0.15M NaCI, pH 7.4), and incubated at room temper-
COLLAGENASE 3 IN RHEUMATOID SYNOVIUM
ature in 1) normal rabbit serum (diluted 150) for 1 hour, 2)
mouse anti-human CD68 IgG (dilution 1 5 0 ) for 1 hour, 3)
rabbit anti-mouse IgG (dilution 1:25) for 1hour, or 4) MAAP
solution (dilution 1:25) for 1 hour. Alkaline phosphatase was
visualized after treatment with BCIP (Sigma no. B-0274) and
nitroblue tetrazolium (Sigma no. N-6876) in alkaline phosphatase buffer for 10 minutes at room temperature, covered from
light. Endogenous alkaline phosphatase activity was inhibited
by including 0.003M levamisole (Sigma no. L-9756). Finally,
the slides were dehydrated, cleared, and mounted.
Reagents for the double-staining protocol were from
Dako A& unless otherwise specified. Between each step, the
sections were washed 3 times for 5 minutes with buffer.
Preparation of the antiserum. The immunization procedure is described elsewhere (3). Briefly, recombinant human
MMP-13 was produced in Escherichia coli, and 8M urea
was used to solubilize the insoluble fraction of the cell lysate.
Then, 1 ml of the 8M urea extract was electrophoresed through
a 12% polyacrylamide gel, and the portion of the gel containing the recombinant protein was excised, ground, and incubated with 2 ml of deionized water at 37°C for 20 hours with
sporadic vortexing. Next, 1 ml of sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE)-purified
protein was used to immunize a New Zealand white rabbit
according to the method described by Vaitukaitis (28). The
rabbit was bled 6 weeks after the injection, and IgG were
purified by chromatography through DEAE-cellulose (Whatman DE52; Whatman, Kent, UK), equilibrated, and eluted
with 0.02M phosphate buffer, pH 7.2. The IgG concentration
was 1 pglpl. Preliminary experiments examining the possible
cross-reactivity with other MMPs, including MMP-1, MMP-8,
and stromelysins, failed to reveal any significant crossreactivity.
Absorption control tests. The specificity of immunoreaction was tested in 1 RA and 1 OA sample by preabsorbing
the anti-MMP-13 with an excess of antigen, as described
previously (29). Briefly, the maximum reproducible dilution
(1:1,000) of anti-MMP-13 for the tissue was determined, and
this working dilution was incubated with 0.1 nmoles of the
antigen overnight at 4°C. This presaturated antibody dilution
was then used in place of the primary antibody.
Extraction of proteolytic enzymes. Synovial tissue
specimens were stored at -70°C until used. No cartilage was
included in the samples to be used for tissue extraction. Tissue
pieces were homogenized in 3 volumes of TBS-1M MgCI,,
extracted on ice for 1 hour, and clarified by centrifugation.
The supernatant was dialyzed against 3 changes of TNC
(0.005M CaCI, in TBS, pH 7.5) and stored at -70°C until
analyzed (30).
Calculation of the isoelectric point (PI). Calculations
were performed using pK, values for amino acids. The amino
acid sequences for MMP-1 and MMP-13 were retrieved from
the Swiss-Prot protein databank, signal peptides were removed
with a text editor program, and charges were calculated for
procollagenases and collagenases using the Microsoft Excel 5.0
spreadsheet program. Possible sialic acid residues in glycosylated forms of MMP-l and MMP-13 were not taken into
account in the calculations of the apparent PI (31-33). The
charges calculated were used to choose the pH for the
nondenaturing electrophoresis buffer, since molecules of approximately the same size but different charge would show
different electrophoretic mobilities.
1393
Electrophoresis. Three buffer systems were used for
separating MMP-1 and MMP-13 in the extracts. First, SDSPAGE was run as described by Laemmli (34), without reduction. Second, low pH Triton-acid-urea gels (TAU gels, 10%)
were run essentially as described by Smith (35). Third, high-pH
gels were run in am Ornstein-Davis buffer system using 7.5%
gel (36,37). A minigel apparatus from Bio-Rad (Richmond,
CA) was used for all gels.
Protein blotting and immunostaining. Synovial tissue
extracts were prepared for the different buffer systems, electrophoresed, and blotted onto nitrocellulose filter. Proteins
from SDS-PAGE gels were blotted in routinely used Trisglycine buffer in a Bio-Rad Mini-TransBlot system (34,38).
From the TAU gels, proteins were capillary blotted overnight
at room temperature. After a brief soak in 2.3% (weighti
volume) SDS-5% (v/v) 2-mercaptoethanol, blotting from
Ornstein-Davis native gels was carried out with 0.1% SDS
added to the transfer buffer (39).
Nitrocellulose filters were blocked with 3% (wiv) gelatin, washed in TBS, and incubated overnight with rabbit
anti-human MMP-13 (1500) or rabbit anti-human MMP-1
(1500). Goat anti-rabbit IgG horseradish peroxidase conjugate was diluted to 1500, and the blots were developed using
4-chloro-1-naphthol (Sigma) as a chromogen.
Collagenase assay. Soluble rat tail tendon type I and
bovine cartilage type I1 collagens were purified by pepsin
extraction and selective salt precipitation at acidic and neutral
pH, as described in detail elsewhere (40). Recombinant human
MMP-13 was activated with 0.001M aminophenyl mercuric
acetate (APMA) for 30 minutes and incubated with type I
and type I1 collagen at a concentration of 1.5 pM for 3 hours
at 22°C. The enzyme reaction was terminated by boiling for
5 minutes after addition of sample buffer without 2mercaptoethanol. Intact collagen a chains and their 3/4 (aA)
and ?h(aB) degradation products were separated by 11%
SDS-PAGE (41). The percentage of collagen degraded was
estimated by image analysis (Molecular Analyst; Bio-Rad) of
Coomassie blue-stained gels. Specifically, the gel was scanned
on the hard disk of the computer, intact monomeric collagen
bands were marked, and the marked areas were analyzed for
pixel intensity. The scores of 3h fragments were multiplied by
Y3 to compensate for the ?A fragment not scanned, and the
percentage from the total (intact band score plus degraded
band score) was calculated.
Production and purification of human recombinant
MMP-13. Human MMP-13 was produced using a vaccinia
virus expression system as previously described (3). Vaccinia
virus-expressing human MMP-13 was obtained using a plaque
selection system (42). Plasmid pRB-col3 was obtained by
inserting an Eco RIIHin dIII fragment containing the gene
downstream of a vaccinia virus synthetic earlyilate promoter, in
plasmid pBR21. Confluent monolayers of CV-1 cells in T25
flasks were infected with 1 plaque forming unit (PFU) per cell
of vaccinia virus vRB12 clone a4 and transfected with 10 pg of
plasmid pRB-col3. At 2 days postinfection, the progeny virus
was harvested. The recombinant, termed VV-col3, was selected by 2 consecutive rounds of plaque purification on BSC-1
cell monolayers. For production of the recombinant protein,
preconfluent BSC-1 cells in 900-cm2 roller bottles were infected with wild-type vaccinia virus (strain WR) or VV-col3, at
a multiplicity of infection of about 5 PFUicell. The medium
was harvested at 24 hours postinfection.
1394
LINDY ET AL
Figure 1. Immunostaining for matrix metalloproteinase 13 (MMP-13) in rheumatoid arthritis (RA) and osteoarthritis (OA) synovial
membranes. A, RA synovium, showing the lining (arrows). B, Consecutive section from A, representing a negative-staining control in
which the primary antibodies were omitted from the staining sequence. C, Expression of MMP-13 in RA synovial stromal cells (arrows) and
vascular endothelium (open circle). D, RA sublining tissue stained for MMP-13 (arrows). E, OA synovium,showing the lining (arrows). (Frozen
sections; avidin-biotin-peroxidase complex stained, without counterstain; original magnification X 160 in A and E, X400 in B, C, and D).
The recombinant protein was purified from the culture
medium by chromatography onto an S-Sepharose fast flow
column (2.5 X 10 cm; Pharmacia, Uppsala, Sweden) equilibrated in 0.02M Tris HC1, pH 7.2,0.005M CaCl,, 0.05% NaN,.
The matrix was washed in the same buffer to background
readings, followed by another wash with the same buffer with
0.2M NaCl to remove impurities. Finally, the column was
eluted with the same buffer with 0.5M NaCl, and the fractions
containing pro-MMP-13 were identified by analysis in
SDS-PAGE.
Assessment of the inhibition characteristics of the
recombinant MMP-13. Briefly, autoactive MMP-13 was incubated with type I and type I1 collagen and analyzed as
described above. Autoactivation refers to an apparently
spontaneous conversion of pro-MMP-13 into its enzymatically active form in the absence of proteolytic activators or
organomercurials (12). Collagenase activity was determined
in the presence of doxycycline, clodronate, pamidronate,
and D-penicillamine, at a wide range of concentrations. The
concentrations were chosen to also cover those attained
COLLAGENASE 3 IN RHEUMATOID SYNOVIUM
1395
Figure 2. Double-labeling for MMP-13 and CD68 (marker for monocyteimacrophages) in A, rheumatoid arthritis and
B, osteoarthritis synovial membranes. Note that some stromal and lining cells (arrows) are positive for both
anti-MMP-13 and anti-CD68, whereas other, apparently fibroblast-like cells (arrowheads) are positive only for
MMP-13. (Frozen sections, avidin-biotin-peroxidase complex and alkaline phosphatase-anti-alkaline phosphatase
double-stained, without counterstain; original magnification X 400). See Figure 1 for definitions.
in vivo in patients receiving regular treatment with these
drugs.
Statistical analysis. BMDP 7.01 software was used for
the statistical calculations. P values were estimated using the
Mann-Whitney test for unpaired groups.
RESULTS
Immunohistochemistry results. All the synovial
samples used in this study showed immunoreactivity to
anti-MMP-13. Staining for MMP-13 was observed in
synovial lining, endothelial, and stromal cells (Figure 1).
Double labeling for MMP-13 and CD68 confirmed that
MMP-13 was expressed in both synovial fibroblasts and
monocyte macrophages, as well as in both A and B
synoviocytes (Figures 2A and B). In addition, in RA,
accumulations of mononuclear inflammatory cells, apparently lymphocytes, in perivascular and sublining
areas were negative for both MMP-13 and CD68 (data
not shown).
The results of morphometric assessments are
presented in Table 2. Although the majority of the
lining, synovial stromal, and endothelial cells from both
RA and OA patients were found to express MMP-13,
the ratio of MMP-13-positive synovial lining and stroma1 cells was significantly higher in RA than in OA
patients (for lining cells P = 0.0262; for stromal cells P =
0.0027). The ratio of MMP-13-positive endothelial cells
in RA and OA patients did not differ statistically sig-
nificantly (P = 0.3491). The control samples from patients with other joint derangements were excluded from
statistical comparison due to the small number of cases.
Cytospin preparations of neutrophils (known to
be MMP-8-positive) did not stain with anti-MMP-13.
Positive staining for MMP-13 was found in the ductal
epithelium of the mammalian carcinoma specimens (not
shown). Absorption control with purified recombinant
MMP-13 abolished staining from strongly positive cells,
leaving a light brown background. The antigen t o antibody
ratio in the absorption test was calculated to give at least a
20-fold antigen excess over the binding sites.
Detection of synovial tissue extract collagenases
blotted onto nitrocellulose. Regular SDS-PAGE did not
allow unequivocal separation of the 2 MMPs (data not
shown). After calculation of the isoelectric points, 2
buffer systems were chosen for separation, both with a
reduction step added in sample preparation. At low pH,
where pro-MMP-1, MMP-1, pro-MMP-13, and MMP13, according to calculations, have different charges, the
recovery of protein on nitrocellulose was low. In contrast, at p H 9.5, protein recovery was better and a clear
separation of MMP-1 and MMP-13 could be achieved:
the relative mobility for the fastest-migrating MMP-1
and MMP-13 bands was 0.4 and 0.56, respectively (Figure 3). A control sample of 300 ng of MMP-1 blotted
from a high-pH native gel was not detected by antiMMP-13 antibody (data not shown).
LINDY ET AL
1396
Table 2. Quantitative evaluation of matrix metalloproteinase 13
(MMP-13) staining*
MMP-13-positive synovial cells (%)
Patient
Lining cells
Endothelial cells
Stromal cells
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
98
90
95
89
95
83
80
82
88
75
82
92
89
80
84
85
97
90
98
92
95
88
85
90
93
80
90
98
93
80
89
85
98
72
94
80
97
74
11
15
20
16
17
26
16
4
15
10
* Ratios were estimated for 500 cells per type.
Inhibition characteristics of the recombinant
MMP-13. Clodronate, pamidronate, and D-penicillamine
were assessed over a range of concentrations that can be
reached with regular treatment and/or that have been
shown to inhibit other collagenases (17,18,20). Activity
was measured using a soluble type I1 collagen assay
(Figure 4).
Clodronate, pamidronate, and D-penicillamine
did not inhibit recombinant MMP-13 (Figure 5). Doxycycline had inhibitory potential starting at approximately
200 f l ,but the 50% inhibition concentration (IC5,J
values were not yet reached at 500 p M (Figure 5). This
differs from the IC,, values for MMP-8 (26
and for
MMP-1 (280 rJ.M) (18).
synovial lining and stroma in vivo. The results presented
in this paper are in clear accordance with those reported
by Wernicke et al, who cloned MMP-13 from an RA
synovial sample and demonstrated expression of its
mRNA in RA and OA synovial tissue, but did not find it
in several normal human tissues (16). Our results demonstrate protein level expression of MMP-13 in RA and
OA synovium. MMP-13 seems thus to be specifically
expressed in synovial tissue, which makes it an enzyme
of potential pathogenetic importance in the arthritides.
The combination of high tensile strength provided by type I1 collagen and high swelling pressure of
the proteoglycan matrix is important for the biomechanical function of cartilage (44,45). Different strategies have been used in various attempts to overcome the
poor inherent healing capacity of articular cartilage:
heterologous chondrocytes, periosteum, perichondrium,
and osteochondral grafts, and, more recently, autologous chondrocytes (46) have been transplanted to treat
focal articular cartilage defects. Because of the importance of type I1 collagen to the biomechanical properties
of the cartilage and the poor inherent healing capacity of
the tissue, the initial and specific cleavage of type I1
collagen must be a highly regulated event.
Collagenases of the MMP family are the only
mammalian enzymes which are able to cleave the al(I1)
M),
40
M
0
D
r
20
DISCUSSION
In the original work by Freije and coworkers (3),
MMP-13 mRNA was found in breast carcinomas, but
not in normal human tissues. However, articular cartilage was not included in their normal tissue panel. Later,
Flannery and Sandy reported MMP-13 mRNA expression in chondrocytes (43), a finding later confirmed by
Mitchell et a1 (15) and by Reboul et a1 (14). Reboul and
coworkers did not find MMP-13 mRNA in interleukinlp-stimulated synoviocyte cultures or in 1 OA synovial
tissue sample. In light of the data reported by Reboul et
al, the present data and those recently published by
other investigators (16), it is possible that MMP-1 and
MMP-13 are differently regulated under conditions used
in synoviocyte cultures. Alternatively, in vitro cell cultures do not reproduce the conditions that prevail in the
40
w
80 I
0
2
4
6
8
10
12
14
PH
Figure 3. Charge-pH plot of MMP-1 ( 0 ) and MMP-13 (W), demonstrating the charge difference. In native gels, MMP-1 and MMP-13
would migrate according to their intrinsic charge. For maximum
separation, pH 3 and Ornstein-Davis buffer systems were chosen. At
low pH, the recovery of protein on nitrocellulose was low (not shown).
The blotted Ornstein-Davis gel shown demonstrates neutral salt
extracts of 3 representative rheumatoid arthritis samples stained for
MMP-1 (first 3 lanes) and MMP-13 (second 3 lanes). Note that the
different net molecular charges allowed separation of the 2 MMP
superfamily members at pH 9.5, which was used to demonstrate the
specificity of the antisera. For clarity, pro-MMPs are not shown. See
Figure 1 for definitions.
COLLAGENASE 3 IN RHEUMATOID SYNOVIUM
triple helix at the 775Gly-776Lxu
peptide bond (47). Hitherto, only MMP-1 and MMP-8 have been considered to be
responsible for thc degradation of type I1 collagen.
MMP-8, which cleaves type I collagen faster than type I1
collagen, can be found in synovial fluid neutrophils (7,23).
The rheumatoid changes are characterized by development of peripheral erosions associated with fibroblast- and
macrophagc-like cells at the advancing edge of the pannus
(48). These cells have been suggested to contain fibroblasttypc collagenase 1 (MMP-I) (8,9), which, however, is not
particularly effective against monomeric type I1 collagen
(30,49,50). Therefore, the occurrence of a third type of
collagenasc, collagenase 3/MMP-13, in synovial tissue in
both RA and OA is of particular interest. MMP-13 was
observed in fibroblast- and macrophage-lke cclls both in
thc synovial stroma and the lining. According to recent
data, MMP-13 differs from MMP-I in that it clcavcs type
I1 collagen more efficiently than other fibrillar collagens
(12). In addition, MMP-13 is active against aggrecan, the
major proteoglycan of the hyaline articular cartilage (13).
Based on its substrate profile against the 2 major components of the hyaline articular cartilage and its now-
Figure 4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
of type I (first 5 lanes) and type I1 (second 5 lanes) collagen. Lane 1,
Intact d(1) and u2(I) chains of type I collagen; lanes 2 and 4,
time-dependent (4 hours and 12 hours) degradation of type I collagen
by autoactivated recombinant MMP-13; lanes 3 and 5, the degradation
was not increased by the presence of 0.001Morganomercurial aminophenyl mercuric acetate (APMA)-type pro-MMP activator; lane 6 ,
intact rrl(I1) chains; lanes 7 and 9, time-dependent (4 hours and 12
hours) degradation by autoactivated recombinant MMP-13; lanes 8
and 10, the degradation was the same in the presence of APMA.
Arrowheads indicate intact monomeric collagen chains. The bar graph
shows the percentage degradation of collagen substrate. See Figure 1
for other definitions.
1397
A &
b
B
CC
D
b
Figure 5. Inhibition of MMP-13-mediated degradstion of typc I1 collagen by doxycycline, clodronate, pamidronate, and D-penicillamine.
The sodium dodecyl sulfate-polyacrylamide gels represent a range of
roncuntrations that were selected to cover thc concentrations attained
in vivo. as follows. A, For doxycycline, lane 1 contains 1.5 pM type I1
collagen alonc: lanes 2-10 contain type I I collagen with 160 ng 0 1
recornbinant MMP-I3 enzyme protein without and with 20.40, ti). 100,
200,300,100, and 5(W1 pA1 doxycycline, respectively. H, For clodronate,
lane 1 contains type I1 collagen alone; lanes 2-8 contain type I1
collagen with recombinant MMP-13 without and with 50, IN).750.500,
750. and 1,WO JLM clodronate, respectivcly. C, For paniidronate. lane
1 contains type I1 collagen alone; lanes 2-8 contain type I1 collagcn
with recombinant MMP-13 without and with 50, 100,250,500,750, and
1,000 phf pamidronate, respectively. D, For D-prnirillarninc, lane 1
contains type I1 collagen alone; lanes 2-6 contain type I1 collagen with
recombinant MMP-13 without and with 50, 100, 200, and 500 JLM
D-penicillamine, respectively. Arrowheads indicate intact monomeric
collagen chains.
demonstrated presence in synovial membrane, we suggest
that MMP-13 may represent a particularly significant mediator of tissue destruction in the arthritides.
Our immunohistochemical findings of the presence of MMP-13 in RA and OA synovial membranes
were confirmed by demonstrating MMP-13 in synovial
tissue extracts. Notably, no cartilage was included in our
samples. Electrophoretic separation in native gels was
efficiently used to separate collagenases that displayed
similar apparent molecular weights in SDS-PAGE. Estimates of charge and isoelectric points are naturally
only apparent, since the calculations cannot include the
effects of the conformation or sialic acid content of
MMP-1 and MMP-13. Because of the significant degree
of glycosylation of MMP-1 and MMP-13 (10% and 25%,
respectively), possible sialic acid end groups may contribute to the total charge and mobility of the collagenases. According to our results after electroblotting of
1398
the Ornstein-Davis gels, RA and OA synovial membranes contained both MMP-13 and MMP-1, which
were clearly separated from each other under the conditions used. Separation in different gel systems also
served in the characterization of the MMP-13 and
MMP-1 antisera. Selection of the electrophoresis system, therefore, helped to show the presence of MMP-13
protein in RA and OA synovial tissue.
MMPs have a great inherent tissue destruction
potential. Therefore, many MMPs, including MMP-1,
are not constitutively expressed (or only at a low level),
but appear first upon appropriate stimulation of the
producer cells with, for example, certain cytokines (5).
Furthermore, they are usually not stored in the producer
cells, but are rapidly secreted after their synthesis (5).
This type of regulation may be important for the expression of these enzymes. Therefore, localization of
MMP-13-positive cells in the vicinity of subsynovial
lymphocyte infiltrates may reflect increased MMP-13
production due to as-yet-unidentified inflammatory cytokines produced by the mononuclear cells in rheumatoid synovitis in situ. Interestingly, there was no difference in MMP-1 and MMP-13 mRNA expression as
analyzed by densitometry from Northern blots standardized for GAPDH expression (16). This suggests that, in
addition to transcriptional events mediating the regulation of MMP-13 expression, posttranscriptional events
can also be important.
Because of the potential significance of MMPs in
joint-destructive events, inhibitors of MMPs have received extensive interest (see ref. 51). In particular,
drugs already widely used, including tetracyclines (51)
and bisphosphonates (20), would seem to offer certain
practical advantages over as-yet-unregistered drugs. Tetracyclines and bisphosphonates have been previously
reported to inhibit several members of the MMP superfamily (51,20) and may share a common type of mechanism of action. They are now reported to be relatively
ineffective against MMP-13 compared with their potency against MMP-1 and MMP-8. This type of inhibitory profile, in conjunction with its tissue-specific localization, substrate profile, and increased expression in
RA, may further contribute to the pathogenetic significance of MMP-13.
In conclusion, our findings suggest that MMP-13
is present in RA synovial tissue and that MMP-1 and
MMP-8 are not the only collagenases involved. In fact,
some of the earlier reports on synovial collagenase may
have to be reassessed due to the possible presence of
collagenase 3/MMP-13 activity. Because of its cellular
localization in the fibroblast-like cells, apparent synovial
LINDY ET AL
tissue-specific localization, and substrate specificity, it
may play a major role in local tissue-destructive events.
Increased expression in RA may be a reflection of the
catabolic state in the joint and could be explained by the
more obvious inflammatory nature of RA compared
with OA. Finally, the relative resistance of MMP-13 to
several known collagenase inhibitors further adds to its
pathogenetic significance.
ACKNOWLEDGMENT
Rabbit polyclonal anti-MMP-1 was kindly provided by
Prof. H. Birkedal-Hansen,Department of Oral Biology, School of
Dentistry, The University of Alabama at Birmingham.
REFERENCES
1. Gross 3, Lapibre CM: Collagenolytic activity in amphibian tissues:
a tissue culture assay. Proc Natl Acad Sci U S A 48:1014-1022,
1962
2. Nagase H, Barrett AJ,Woessner J F Jr: Nomenclature and glossary
of the matrix metalloproteinases. Matrix Suppl 1:421-424, 1992
3. Freije JMP, Diez-Itza I, Balbin M, SBnchez LM, Blasco R,
Tolivia J, M p e z - 0 t h C Molecular cloning and expression of
collagenase-3, a novel human matrix metalloproteinase produced
by breast carcinomas. J Biol Chem 269:16766-16773, 1994
4. Puente XS, Pendas AM, Llano E, Velasco G, L6pez-0th C:
Molecular cloning of a novel membrane-type matrix metalloproteinase from a human breast carcinoma. Cancer Res 56:944-949,
1996
5. Werb Z, Alexander CM: Proteinases and matrix degradation. In,
Textbook of Rheumatology. Fourth edition. Edited by WN Kelley,
ED Harris Jr, S Ruddy, CB Sledge. Philadelphia, WB Saunders,
1993
6. Prockop DJ, Kivirikko KI: Collagens: molecular biology, diseases,
and potentials for therapy. Annu Rev Biochem 64:403-434, 1995
7. Harris ED Jr, DiBona DR, Krane SM: Collagenase in synovial
fluid. J Clin Invest 48:2104-2113, 1969
8. Evanson JM, Jeffrey JJ, Krane SM: Human collagenase: identification and characterization of an enzyme from rheumatoid synovium in culture. Science 158:499-502, 1967
9. Dayer JM, Krane SM, Russell RGG, Robinson DR. Production of
collagenase and prostaglandins by isolated adherent rheumatoid
synovial cells. Proc Natl Acad Sci U S A 73:945-949, 1976
10. Quinn CO, Scott DK, Brinckerhoff CE, Matrisian LM, Jeffrey JJ,
Partridge NC: Rat collagenase: cloning, amino acid sequence
comparison, and parathyroid hormone regulation in osteoblastic
cells. J Biol Chem 265:22342-22347, 1990
11. Henriet P, Rousseau GG, Eeckhout Y: Cloning and sequencing of
mouse collagenase cDNA divergence of mouse and rat collagenases from the other mammalian collagenases. FEBS Lett 310:
175-178, 1992
12. Knauper V, Lbpez-0th C, Smith B, Knight G, Murphy G:
Biochemical characterization of human collagenase-3. J Biol
Chem 271:1544-1550, 1996
13. Fosang AJ, Last K, Knauper V, Murphy G, Neame PJ: Degradation of cartilage aggrecan by collagenase-3 (MMP-13). FEBS Lett
380:17-20, 1996
14. Reboul P, Pelletier J-P, Tardif G, Cloutier J-M, Martel-Pelletier J:
The new collagenase, collagenase-3, is expressed and synthesized
by human chondrocytes but not by synoviocytes: a role in osteoarthritis. J Clin Invest 97:2011-2019, 1996
15. Mitchell PG, Magna HA, Reeves LM, Loprestimorrow LL,
COLLAGENASE 3 IN RHEUMATOID SYNOVIUM
Yocum SA, Rosner PJ, Geoghegan KF, Hambor JE: Cloning,
expression, and type I1 collagenolytic activity of matrix
metalloproteinase-13 from human osteoarthritic cartilage. J Clin
Invest 97:761-768, 1996
16. Wernicke D, Seyfert C, Hinzmann B, Gromnica-Ihle E: Cloning of
collagenase 3 from the synovial membrane and its expression in
rheumatoid arthritis and osteoarthritis. J Rheumatol 23:590-595,
1996
17. Golub LM, Ramamurthy NS, McNamara TF, Gomes B, Wolff M,
Casino A, Kapoor A, Zambon J, Ciancio S, Schneir M Tetracyclines
inhibit tissue collagenase activity: a new mechanism in the treatment
of periodontal disease. J Periodontal Res 19651-655, 1984
18. Suomalainen K, Sorsa T, Golub LM, Ramamurthy N, Lee H-M,
Uitto V-J, Saari H, Konttinen Y T Specificity of the anti-collagenase action of tetracyclines: relevance to their anti-inflammatory
potential. Antimicrob Agents Chemother 36227-229, 1992
19. Smith GN, Brandt KD, Hasty KA: Activation of recombinant
human neutrophil procollagenase in the presence of doxycycline
results in fragmentation of the enzyme and loss of enzyme activity.
Arthritis Rheum 39:235-244, 1996
20. Teronen 0,Konttinen YT, Salo T, Ingman T, Sorsa T Clodronate
inhibits MMP-1 and MMP-8 in GCF and PISF (abstract 31). Bone
17:605, 1995
21. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS,
Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller
JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 1987 revised criteria for the classification of
rheumatoid arthritis. Arthritis Rheum 31:315-324, 1988
22. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K,
Christy W, Cooke TD, Greenwald R, Hochberg M, Howell D,
Kaplan D, Koopman W, Longley S 111, Mankin H, McShane DJ,
Medsger T Jr, Meenan R, Mikkelsen W, Moskowitz R, Murphy W,
Rothschild B, Segal M, Sokoloff L, Wolfe F Development of
criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 29:10391049, 1986
23. Konttinen YT, Lindy 0, Kemppinen P, Saari H, Suomalainen K,
Vauhkonen M, Lindy S, Sorsa T: Collagenase reserves in polymorphonuclear neutrophil leukocytes from synovial fluid and peripheral blood of patients with rheumatoid arthritis. Matrix 11:296301, 1991
24. Hsu S-M, Raine L, Fanger H: Use of avidin-biotin-peroxidase
complex (ABC) in immunoperoxidase techniques: a comparison
between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 29:577-580, 1981
25. Konttinen YT, Reitamo S, Ranki A, Hayry P, Kankaanpaa U,
Wegelius 0: Characterization of the immunocompetent cells of
rheumatoid synovium from tissue sections and eluates. Arthritis
Rheum 24:7i-79, 1981
26. Cordell JL. Falini B. Erber WN. Ghosh AK. Abdulaziz Z. MacDonald S, Pulford &,Stein H, Mason DY: Immunoenzymatic
labeling of monoclonal antibodies using immune complexes of
alkaline phosphatase and monoclonal anti-alkaline phosphatase
(APAAP complexes). J Histochem Cytochem 32:219-229, 1984
27. Konttinen YT, Hukkanen M, Segerberg M, Rees R, Kemppinen P,
Sorsa T, Saari H, Polak JM, Santavirta S: Relationship between
neuropeptide immunoreactive nerves and inflammatory cells in
adjuvant arthritic rats. Scand J Rheumatol 21:55-59, 1992
28. Vaitukaitis JL: Production of antisera with small doses of immunogen: multiple intradermal injections. Methods Enzymol 73:46-52,
1981
29. Van Noorden S: Tissue preparation and immunostaining techniques for light microscopy. In, Immunocytochemistry: Modern
Methods and Applications. Second edition. Edited by JM Polak, S
van Noorden. Bristol, Wright & Sons, 1986
30. Konttinen YT, Lindy 0, Suomalainen K, Ritchlin C, Saari H,
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
1399
Vauhkonen M, Lauhio A, Santavirta S, Sorsa T: Substrate specificity and activation mechanisms of collagenase from human
rheumatoid synovium. Matrix 11:395-403, 1991
Segel IH: Biochemical Calculations. New York, Wiley and Sons,
1968
Ribeiro JM, Sillero A An algorithm for the computer calculation
of the coefficients of a polynomial that allows determination of
isoelectric points of proteins and other macromolecules. Comput
Biol Med 20:235-242, 1990
Henriksson G, Englund AK, Johansson G, Lundahl P: Calculation
of the isoelectric points of native proteins with spreading of pK,
values. Electrophoresis 16:1377-1380, 1995
Laemmli UK: Cleavage of structural proteins during the assembly
of the head of bacteriophage T4. Nature 227:680-685, 1970
Smith BJ: Acetic acid-urea polyacrylamide gel electrophoresis of
proteins. Methods Mol Biol 32:39-47, 1994
Davis BJ: Disc electrophoresis. 11. Method and application to
human serum proteins. Ann N Y Acad Sci 121:404-427, 1964
Ornstein L: Disc electrophoresis. I. Background and theory. Ann
N Y Acad Sci 121:321-349, 1964
Towbin H, Staehelin T, Gordon J: Electrophoretic transfer of
proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A 76:43504354, 1979
Waterborg JH, Harrington RE: Western blotting of histones from
acid-urea-Triton- and sodium dodecyl sulfate-polyacrylamide gels.
Anal Biochem 162:430-433, 1987
Miller EJ, Rhodes KR: Preparation and characterization of the
different types of collagen. In, Structural and Pontractile Proteins:
Part A. Methods in Enzymology. Edited by LW Cunningham, DW
Fredriksen. New York, Academic Press, 1982
Turto H, Lindy S, Uitto V-J, Wegelius 0, Uitto J: Human
leukocyte collagenase: characterization of enzyme kinetics by a
new method. Anal Biochem 83557-569, 1977
Blasco R, Moss B: Selection of recombinant vaccinia viruses on the
basis of plaque formation. Gene 158:157-162, 1995
Flannery CR, Sandy JD: Identification of MMP-13 (collagenase-3)
as the major matrix metalloproteinase expressed by chondrocytes
during retinoic acid-induced matrix catabolism (abstract), FortyFirst Annual Meeting of the Orthopedic Research Society, Orlando, FL, February 1995
Brandt KD, Mankin HJ: Osteoarthritis and polychondritis. In,
Textbook of Rheumatology. Fourth edition. Edited by WN Kelley,
ED Harris Jr, S Ruddy, CB Sledge. Philadelphia, WB Saunders,
1993
Poole AR: Cartilage in health and disease. In, Arthritis and Allied
Conditions. Twelfth edition. Edited by DJ McCarty, WJ Koopman. Philadelphia, Lea and Febiger, 1993
Brittberg M, Lindahl A, Nilsson A, Ohlsson C, Isaksson 0,
Peterson L Treatment of deep cartilage defects in the knee with
autologous chondrocyte transplantation. N Engl J Med 3312389895, 1994
Krane SM: Mechanisms of tissue destruction in rheumatoid arthritis.
In, Arthritis and Allied Conditions. Twelfth edition. Edited by DJ
McCarty, WJ Koopman. Philadelphia, Lea and Febiger, 1993
Shiozawa S, Shiozawa K, Fujita T: Morphologic observations in
the early phase of the cartilage-pannus junction: light and electron
microscopic studies of active cellular pannus. Arthritis Rheum
26:472-478, 1983
Welgus HG, Jeffrey JJ, Eisen AZ: The collagen substrate specificity of human skin fibroblast collagenase. J Biol Chem 256:95119515, 1981
Docherty AJP, Murphy G: The tissue metalloproteinase family
and the inhibitor TIMP: a study using cDNA and recombinant
proteins. Ann Rheum Dis 49:469-479, 1990
Inhibition of matrix metalloproteinases: therapeutic potential.
Edited by RA Greenwald, LM Golub. Ann N Y Acad Sci 732,1994