(Received for publication, December 13, 1996, and in revised form, March 21, 1997)
From the Departments of Radiation Oncology and
Pathology, University of Arizona Health Sciences Center,
Tucson, Arizona 85724
Matrilysin is a matrix metalloprotease that is
overexpressed in cancer cells of epithelial origin and in normal
tissues during events involving matrix remodeling such as the cycling
endometrium. We previously observed that inflamed ductule and acinar
epithelia in the prostate also overexpress matrilysin. The presence of
infiltrating macrophages in these areas prompted us to determine if
factors secreted from monocytes could induce matrilysin expression in a
human prostatic cell line. Conditioned media collected from the
monocyte cell line THP-1 following lipopolysaccharide treatment substantially induced matrilysin protein and mRNA expression in LNCaP prostate carcinoma cells. Matrilysin expression in LNCaP cells
was also induced by recombinant interleukin (IL)-1 (50 pM), but not by equimolar concentrations of
recombinant tumor necrosis factor- or IL-6. The matrilysin-inducing
activity of THP-1 conditioned medium was completely abrogated by
preincubation with a neutralizing antibody to IL-1
. Transient
transfection analyses with a chimeric human matrilysin
promoter-chloramphenicol acetyltransferase reporter construct
demonstrated that IL-1
activates transcription through the
matrilysin promoter in LNCaP cells. This is the first report of
matrilysin induction by an inflammatory cytokine in a cell line of
epithelial origin, and the results suggest a potential mechanism for
the overexpression of matrilysin in inflamed ducts and glands of the
prostate.
The matrix metalloproteases are a family of enzymes that degrade extracellular matrix proteins. Matrilysin (PUMP-1, MMP-7) is a relatively recently described matrix metalloprotease belonging to the stromelysin enzyme subclass (reviewed in Ref. 1). Matrilysin is capable of degrading a diverse set of extracellular matrix proteins including proteoglycans, fibronectin, entactin, laminin, gelatin, and elastin. The expression of matrilysin has been demonstrated in the cycling endometrium (2); the involuting rat prostate (3) and uterus (4); developing mononuclear phagocytes (5); and cancers of the breast (6), lung and upper respiratory tract (7), skin (8), stomach and colon (9, 10), and prostate (11). A unique feature of matrilysin expression is that it appears to predominate in epithelial cells of glandular tissue, while other matrix metalloproteases, such as stromelysin and the gelatinases, are more commonly expressed by cells in the stromal compartment (1). The expression of matrilysin during normal and pathological events that involve matrix remodeling and the cell type specificity of this expression imply an important role for matrilysin in these events.
In the normal prostate, we have observed that inflamed ductule and acinar epithelia frequently express high levels of matrilysin (11, 12). It has been clearly demonstrated by in situ hybridization that the overexpression of matrilysin mRNA is confined to the epithelial cells of these structures and is not present in the surrounding stroma or infiltrating leukocytes (12). Immunohistochemical analysis confirmed that the expression of matrilysin mRNA in the prostate epithelium correlates with the expression of matrilysin protein.
We have speculated that the high levels of matrilysin expression in
inflamed prostate epithelial cells might be due to the presence of
infiltrating macrophages within these structures. During the
inflammatory process, these cells secrete factors such as tumor
necrosis factor- (TNF-
)1 and
interleukin-1 (IL-1), which are known to induce the expression of
matrix metalloproteases in a variety of cell types such as connective
tissue cells, endothelial cells, monocytes/macrophages, neutrophils,
and tumor cells (reviewed in Ref. 13). The induction of stromelysin-1
(MMP-3) and interstitial collagenase (MMP-1) expression by IL-1 in
chondrocytes and synovial fibroblasts has been particularly well
studied (14-24). While the induction of matrilysin and other matrix
metalloproteases by inflammatory cytokines has been observed in
cultured glomerular mesangial cells (25) and some glioma cell lines
(26), there have been no reports on the effect of inflammatory
cytokines on matrilysin expression in cells of glandular epithelial
origin.
The purpose of the work reported here was to determine if factors
secreted from monocytic cells could induce matrilysin expression in a
prostatic cell line and, if so, by what mechanism. We report that
IL-1 secreted from monocytic cells induces matrilysin expression in
LNCaP cells and that the mechanism of this induction involves an
increase in matrilysin gene transcription.
LNCaP and THP-1 cells were obtained from
American Type Culture Collection (Rockville, MD) and were maintained in
RPMI 1640 medium supplemented with 10% fetal calf serum and penicillin
(100 units/ml)/streptomycin (100 µg/ml). All cells were maintained in
a humidified incubator at 37 °C and 5% CO2. For all
experiments, LNCaP cells were seeded in full medium and allowed to
attach for 20-24 h prior to treatment. To generate THP-1 conditioned
medium (THP-1 CM), THP-1 cells at a density of 1 × 106 cells/ml were incubated for 4 h in serum-free RPMI
1640 medium containing 5 µg/ml lipopolysaccharide (LPS) (Sigma) and
then clarified by centrifugation and passed through a 0.2-µm filter.
Recombinant human IL-1, IL-1
, and IL-6 and neutralizing
antibodies against the IL-1 proteins were purchased from Genzyme
Diagnostics (Cambridge, MA); recombinant TNF-
was from Boehringer
Mannheim; and antibody against mouse IL-2 was from Collaborative
Biomedical Products (Bedford, MA).
An antibody sandwich assay was developed for detection and quantification of matrilysin. The capture antibody (10D2, a mouse monoclonal antibody produced in the laboratory of Dr. Raymond Nagle using purified promatrilysin from Dr. Mark Navre, Syntex, Palo Alto, CA) was coated onto 96-well EIA plates (Costar, Cambridge, MA). This antibody is specific for human promatrilysin, and therefore, the ELISA does not detect active matrilysin. The detection antibodies included antibody Rb2, a rabbit polyclonal antibody to human matrilysin (12), followed by a horseradish peroxidase-conjugated goat anti-rabbit antibody (Pierce) used to detect bound Rb2. Horseradish peroxidase activity was quantitated using a hydrogen peroxide/o-phenyldiamine (Sigma) colorimetric system. Purified promatrilysin was used to generate a standard curve for each assay. No cross-reactivity was observed with either purified gelatinase A or gelatinase B (a gift from Dr. Eric Howard, University of Oklahoma). The assay was linear in the range of 0.2-12.5 ng/ml. Samples were diluted prior to analysis until the readings fell within the linear range of the assay, and the results were multiplied by the dilution factor.
Northern AnalysisTotal RNA was isolated from cells by the
acid guanidinium thiocyanate/phenol/chloroform method (27). Equal
amounts of RNA (20 µg) were electrophoresed on a
MOPS/formaldehyde-agarose (1%) gel. The RNA was transblotted onto a
nylon membrane (GeneScreenTM, DuPont NEN) and cross-linked with
ultraviolet light (GS GenelinkerTM, Bio-Rad). The membranes were then
hybridized with a 32P-labeled random-primed probe generated
from a full-length matrilysin cDNA (28) using the RTS RadPrime DNA
labeling system (Life Technologies, Inc.) and washed according to the
manufacturer's instructions. The membranes were then exposed to a
storage phosphor screen (Molecular Dynamics, Inc., Sunnyvale, CA). A
Molecular Dynamics PhosphorImager equipped with the ImageQuant software
package was used for obtaining and analyzing digital images from the
screens. As a control for loading and transfer of RNA, all membranes
were stripped and reprobed for glyceraldehyde-3-phosphate
dehydrogenase as described above with a probe generated from an
800-base pair XbaI-PstI fragment from pHcGAP
(American Type Culture Collection).
Detection of matrilysin protein in media by Western analysis was done as described previously (29). Rb2 was used as the primary antibody, and horseradish peroxidase-conjugated goat anti-rabbit antibody was used as the secondary antibody. Bands were visualized by treating the membranes with Western blotting ECL detection reagent (Amersham International, Buckinghamshire, United Kingdom) and exposing them to Kodak autoradiographic film.
Plasmid ConstructsTo generate the heterologous human
matrilysin promoter (HMP) construct used in these studies, 1170 base
pairs of the HMP located directly upstream of the TATA box were
amplified by polymerase chain reaction and subcloned into pBLCAT2 (30).
Polymerase chain reaction amplification of the HMP (kindly provided by
Dr. Lynn Matrisian, Vanderbilt University) was carried out using the
following heterologous primers, which contained either
HindIII (upstream primer) or XbaI (downstream
primer) restriction site sequences linked to matrilysin
promoter-specific sequences (matrilysin-specific regions of each primer
are underlined): upstream sense primer, 5-CCCAAGCTTAGCTCCAGCATATTT-3
; and downstream antisense
primer, 5
-TGCTCTAGAGCTTCTCAGCCTCG-3
. The resultant
1217-base pair amplification product was digested with
HindIII/XbaI, gel-purified, and directionally cloned into HindIII/XbaI-digested pBLCAT2
immediately upstream of the thymidine kinase minimal promoter. The
resulting plasmid, named pHMPCAT, was confirmed by DNA sequencing.
LNCaP cells, grown in 12.5-cm flasks, were transfected (10 µg of plasmid/flask) using Lipofectin (5 µl/ml; Life Technologies, Inc.) in 1 ml of serum-free RPMI 1640 medium according to the manufacturer's instructions. Plasmids were isolated using QiaPrep columns (QIAGEN Inc., Chatsworth, CA) according to the manufacturer's instructions. Following transfection, the cells were washed once with serum-free RPMI 1640 medium and then incubated in treatment medium. Cell lysates were collected and analyzed for chloramphenicol acetyltransferase (CAT) enzyme activity as described previously (31). The cell lysate added to each CAT enzyme assay was normalized for equal protein as determined by the Bio-Rad protein assay. The percent butylation of [14C]chloramphenicol by cell lysates was determined by TLC, followed by exposure of the TLC plates to a storage phosphor screen and analysis of the digital images obtained as described above for Northern analyses.
The monocyte cell line THP-1 (32) was used as a
source of monocyte-derived factors to test the hypothesis that these
factors induce matrilysin expression in prostate epithelial cells. The human prostate carcinoma cell line LNCaP was used because these cells
express matrilysin protein in cell culture and have retained some
features of normal prostate epithelial cells including androgen responsiveness and secretion of prostate-specific antigen. LNCaP cells
were incubated in serum-free RPMI 1640 medium alone, in serum-free RPMI
1640 medium treated with LPS, or in THP-1 CM. After a 48-h incubation,
the amount of matrilysin protein secreted into the medium was
determined. Western analysis for matrilysin protein revealed that THP-1
CM dramatically induced the expression of matrilysin protein by LNCaP
cells (Fig. 1A). Nearly all of the matrilysin
protein present was in the 28-kDa proenzyme form, with only a slight
band corresponding to the 19-kDa activated form (lane D).
THP-1 cells did not secrete any matrilysin protein in response to LPS
activation (lane B), and LPS alone did not induce matrilysin
protein expression in LNCaP cells (lane C). An ELISA was
used to better quantitate and follow the induction of LNCaP matrilysin
protein expression by THP-1 cell-secreted factors. Quantitation of
promatrilysin by ELISA demonstrated an ~70-fold increase in secreted
matrilysin protein in media from LNCaP cells treated with THP-1 CM
compared with untreated controls (Fig. 1B).
The Matrilysin-inducing Factor Secreted by THP-1 Cells Is IL-1
Monocytic cells are known to secrete the cytokines
TNF-, IL-1, and IL-6 in response to activation by LPS. To determine
if one of these cytokines might be responsible for THP-1 CM induction of matrilysin expression in LNCaP cells, we treated LNCaP cells with
equimolar amounts of purified recombinant forms of these cytokines. As
shown in Fig. 2, treatment with IL-1 (50 pM)
induced an ~100-fold increase in promatrilysin expression by LNCaP
cells compared with untreated cells, while neither TNF-
nor IL-6 was capable of inducing a significant change in matrilysin expression. These data strongly implicate IL-1 as being a factor in THP-1 CM that
is at least partially responsible for the induction of matrilysin
expression in LNCaP cells.
Two forms of IL-1 receptor agonists are known, IL-1 and IL-1
.
Neutralizing antibodies specific for IL-1
or IL-1
were therefore used to determine if either of these proteins was responsible for THP-1
CM induction of matrilysin expression in LNCaP cells. As shown in Fig.
3, preincubation of THP-1 CM with anti-IL-1
antibody
prior to incubation with LNCaP cells completely abrogated the ability
of THP-1 CM to induce matrilysin expression in LNCaP cells, while
preincubation with anti-IL-1
antibody had no effect. The
neutralization observed with anti-IL-1
antibody was specific as
preincubation with anti-IL-1
or a monoclonal antibody to mouse IL-2
did not alter the ability of THP-1 CM to induce LNCaP matrilysin expression. Anti-IL-1
antibody was able to neutralize the induction of LNCaP matrilysin expression by a 50 pM dose of
recombinant IL-1
, demonstrating that the antibody was active (data
not shown).
The effect of THP-1 CM on steady-state matrilysin mRNA levels in
LNCaP cells was also studied. LNCaP cells were treated for 20 h
with THP-1 CM with or without antibody preincubation. At the end of the
treatment period, total RNA was collected from LNCaP cells and analyzed
by Northern hybridization. Treatment with THP-1 CM resulted in a
substantial increase in steady-state matrilysin mRNA (Fig.
4A, lane B) over that observed in
untreated control cells (lane A). Quantitation by digital
image analysis revealed that, following glyceraldehyde-3-phosphate
dehydrogenase correction for loading and transfer, this increase was
~50-fold over untreated LNCaP cells (Fig. 4B, column
1). Preincubation of THP-1 CM with IL-1
neutralizing antibody
completely blocked its capacity to induce steady-state matrilysin
mRNA levels (lane C and column 2), while the
irrelevant antibody against mouse IL-2 had no effect (lane D
and column 3).
Transcriptional Enhancer Elements Present in the Matrilysin Promoter Region Are Responsive to IL-1
To determine if
transcriptional activation of the matrilysin gene promoter could be
involved in the IL-1-induced increase of steady-state matrilysin
mRNA, we conducted transient transfection experiments in LNCaP
cells with a human matrilysin promoter-TK-CAT reporter construct
(pHMPCAT). Treatment of pHMPCAT-transfected cells with 200 pM IL-1
for 48 h resulted in a 3.7-fold increase in
CAT enzyme activity compared with untreated cells (Fig.
5), whereas no induction was seen in cells transfected
with the parental vector (pBLCAT2).
We have demonstrated that IL-1 secreted from the THP-1 monocyte
cell line induces expression of the matrix metalloprotease matrilysin
by the prostate carcinoma cell line LNCaP. This is the first time to
our knowledge that IL-1 has been demonstrated to induce matrilysin in
an epithelial cell line. THP-1 cells are induced to secrete
inflammatory cytokines including TNF-
, IL-1, and IL-6 by treatment
with LPS (33). Although LNCaP cells have been reported to express
receptors for all three of these cytokines (34-36), and TNF-
has
been shown to be an important regulator of matrix metalloprotease
expression in non-epithelial cells (37-39), only IL-1 induced
matrilysin protein expression in LNCaP cells. Both IL-1
and IL-1
induced matrilysin expression in LNCaP cells; however, IL-1
antibody
treatment did not affect matrilysin induction by THP-1 CM. We can
therefore conclude that THP-1 cells secrete little or no IL-1
, which
is in agreement with published observations of IL-1 expression by these
cells (40). Unlike IL-1
, IL-1
exists primarily as an
intracellular and membrane-bound protein and is not normally secreted
(41).
Monocytes have been reported to express matrilysin in vitro (5), and this expression is reportedly enhanced by LPS treatment. Therefore, one might expect that the THP-1 cells would express matrilysin. However, we observed no matrilysin expression by THP-1 cells even after LPS treatment. The expression of matrilysin by monocytes has been determined to be developmentally regulated as fully differentiated macrophages do not express matrilysin (5). It is possible then that THP-1 cells, which are a leukemic cell line of unknown passage number, have not retained the ability to express matrilysin.
The increase of steady-state matrilysin mRNA in LNCaP cells induced
by THP-1 cell-secreted IL-1 was comparable to that observed for
matrilysin protein induction, suggesting that regulation of the
matrilysin mRNA level is the major mechanism by which IL-1
treatment increased matrilysin protein expression by LNCaP cells. This
result is in agreement with reports that IL-1
is a transcriptional regulator of other matrix metalloprotease genes including stromelysin-1 (16, 42, 43). Transient transfection experiments with the reporter
plasmid pHMPCAT in LNCaP cells revealed a significant IL-1
-induced
increase in CAT transcription, although the magnitude of this increase
does not account for the total increase in steady-state matrilysin
mRNA observed in IL-1
-treated LNCaP cells. This could be due to
problems inherent to transient transfection analysis, or the matrilysin
gene may contain additional IL-1
response elements not present in
pHMPCAT. The known enhancer elements in the portion of the matrilysin
promoter present in pHMPCAT include an AP-1-binding site, two PEA-3
elements, and a number of NF-IL-6 consensus sequences (T(T/G)NNGNAA(T/G)) (1). There are also three putative transforming growth factor-
inhibitory elements. We are now performing deletion analysis of the human matrilysin promoter in pHMPCAT to determine the
regions responsible for IL-1
transcriptional regulation in LNCaP
cells. IL-1
may also have additional effects on matrilysin mRNA
stability in LNCaP cells, and we are conducting experiments to address
this possibility.
Although we cannot rule out the possibility that other factors could be
necessary to permit a similar matrilysin induction in prostate
epithelial cells in vivo, our finding that IL-1 induces the expression of matrilysin in a prostatic cell line offers a potential explanation for the matrilysin overexpression observed in
inflamed dilated ducts and atrophic glands of the normal prostate. Saarialho-Kere et al. (44) reported that normal prostate
glands constitutively express matrilysin; however, this basal
expression must be very low as we have not observed strong matrilysin
expression in normal, non-inflamed prostate glands and ducts (11, 12). We propose that infiltrating macrophages present near inflamed ducts
and glands secrete IL-1
, which induces transcription of the
matrilysin gene in ductule and glandular epithelial cells. This
hypothesis is supported by results from a limited number of experiments
with normal human prostate epithelial cells. These cells secrete very
low levels of promatrilysin protein in vitro, and treatment
with THP-1 CM or IL-1
(50 pM) results in a 2-3-fold increase in promatrilysin protein expression (data not shown).
The pathological significance of dilated ducts and atrophic glands in
the human prostate is not known; however, it is likely that a
substantial amount of extracellular matrix remodeling occurs at these
sites. The presence of macrophage infiltration near areas of epithelial
matrilysin expression is not limited to the prostate. For example, the
number of macrophages present in endometrial tissue increases
substantially in the late secretory/premenstrual endometrium (45),
correlating with a reported increase in matrilysin expression in the
late secretory and menstrual endometrial epithelium (2). During
inflammation, macrophages secrete a number of matrix-degrading enzymes
and are capable of inducing stromal fibroblasts to express matrix
metalloproteases. IL-1 induction of matrilysin expression by
epithelial cells may be another important component of
inflammation-associated tissue remodeling in the prostate as well as in
other glandular epithelial tissues.
We thank Kathy McDaniel for technical assistance in developing the ELISA and J. David Knox for helpful discussions and critical reading of the manuscript.