(Received for publication, January 6, 1997, and in revised form, February 20, 1997)
From the Department of Pharmacology, ¶ Member,
Lucille P. Markey Cancer Center, University of Kentucky College of
Medicine, Lexington, Kentucky 40536 and the § Department of
Pharmacology, Duke University Medical Center,
Durham, North Carolina 27710
While transforming growth factor (TGF
)
type III receptor (RIII) is known to increase TGF
1
binding to its type II receptor (RII), the significance of this
phenomenon is not known. We used human breast cancer MCF-7 cells to
study the role of RIII in regulating autocrine TGF
1
activity because they express very little RIII and no detectable
autocrine TGF
activity. A tetracycline-repressible RIII expression
vector was stably transfected into this cell line. Expression of RIII
increased TGF
1 binding to TGF
type I receptor (RI) as
well as RII. Treatment with tetracycline suppressed RIII expression and
abolished TGF
1 binding to RI and RII. Growth of RIII-transfected cells was reduced by 40% when plated at low density on plastic. This reduction was reversed by tetracycline treatment and
was partially reversed by treatment with a TGF
1
neutralizing antibody. The activity of a TGF
-responsive promoter
construct when transiently transfected was more than 3-fold higher in
the RIII-transfected cells than in the control cells. Treating the cells with tetracycline or the TGF
1 neutralizing
antibody also significantly attenuated the increased promoter activity.
These results suggest that expression of RIII restored autocrine
TGF
1 activity in MCF-7 cells. The RIII-transfected cells
were also much less clonogenic in soft agarose than the control cells
indicating a reversion of progression. Thus, RIII may be essential for
an optimal level of the autocrine TGF
activity in some cells,
especially in the transformed cells with reduced RII expression.
Transforming growth factor (TGF
)1 isoforms are homodimer
polypeptides of 25 kDa. They are multifunctional growth factors involved in the regulation of cell proliferation, differentiation, extracellular matrix formation, and immune response (1-3). Many studies have shown that TGF
can inhibit the growth of a variety cell
types including epithelial, endothelial, lymphoid, and myeloid cells
(4). Almost all types of cells express one or more of the three
isoforms identified in mammals. Although in most systems, the majority
of TGF
s secreted is in a latent form with no biological activity, a
small percentage can be detected as mature, active TGF
s which may
act to regulate cellular functions in an autocrine fashion. For
example, neutralization of endogenous TGF
s with anti-TGF
antibodies was shown to stimulate proliferation of breast and colon
cancer cells (5-7). Repression of TGF
expression by TGF
1
antisense RNA in colon cancer cells was shown to increase clonogenicity
in soft agarose and tumorigenicity in nude mice indicating that
autocrine TGF
activity is tumor-suppressive (8, 9).
TGFs elicit their effects by binding mainly to three cell surface
proteins termed type I (RI), type II (RII), and type III (RIII)
receptors. RI and RII are serine/threonine kinase receptors of 55 and
75 kDa, respectively, that form heteromeric complex, apparently at one
to one stoichiometric ratio and necessary for TGF
signal
transduction (3). It has been shown that RI requires RII for TGF
binding, whereas RII needs RI for signaling (10, 11). Recently, several
studies showed that mutation or down-regulated expression of RI and/or
RII is associated with loss of TGF
sensitivity and progression of
human gastric, colorectal, and prostate cancers and T-cell lymphomas
(12-17). Restoration of TGF
sensitivity by replacing the mutated or
down-regulated receptor into colon and breast cancer cells reduced
their malignancy (18-20). These results again indicate that autocrine
TGF
s can act as negative growth regulators and disruption of the
autocrine TGF
loop is probably a major event contributing to
malignant progression.
TGF RIII, also called
-glycan, is a proteoglycan of 280-330 kDa.
It is the most abundant TGF
binding molecule on the cell surface of
a variety of cell types (21-23). Sequence analyses indicate that human
RIII has a relatively small cytoplasmic domain of 41 amino acid
residues that contains no consensus signaling motif (23, 24).
Therefore, it is believed not to directly transduce TGF
signal.
However, RIII binds all three TGF
isoforms with high affinity. The
Kd ranges from 50 to 300 pM depending on
cell types (21, 25, 26). Expression of RIII in the cells that lacks
RIII significantly increased TGF
1 binding to RII by 2.5-fold or more (24, 27). It was shown that in the absence of RIII,
only a small population of RII in myoblasts could bind TGF
1 with high affinity, while a larger population of
RII had a much lower affinity for TGF
1. However,
expression of RIII converted all RII receptors into one population with
high affinity for TGF
1 (27). In the presence of ligands,
RIII has been shown to form a heteromeric complex with RII suggesting
that RIII enhances TGF
binding to RII by directly presenting the
ligands to RII (27, 28).
While it is clear that RIII can enhance TGF1 binding to
RII, the significance of this phenomenon remains to be elucidated. Since TGF
isoforms are produced mainly in a latent form and mature, activated TGF
levels are very low, if detectable, in most cell types, the majority of RII receptors with low affinity would probably not be occupied by the ligands in the absence of RIII. On the basis of
what is known about the stoichiometric interaction between TGF
RI
and RII, it appears that RIII should enhance autocrine TGF
1 activity if a cell expresses a similar or lower
level of RII than RI. Under this condition, autocrine TGF
activity
is minimized without RIII because only a small percentage of RII is
ligand-bound and can activate a small percentage of RI to transduce the
signal. Thus, expression of RIII should increase the percentage of
ligand-bound RII and consequently the percentage of activated RI. To
test this hypothesis, we expressed RIII in human breast cancer MCF-7
cells which express a very low level of RII with no detectable
autocrine TGF
activity. We show that expression of RIII inhibited
both anchorage-dependent and anchorage-independent growth
of MCF-7 cells and the inhibition could be reversed partially by a
TGF
1 neutralizing antibody, indicating that RIII-induced growth inhibition is at least partially due to restored autocrine negative activity of TGF
1.
MCF-7 cells were originally obtained from the Michigan Cancer Foundation. The cell line was adapted to McCoy's 5A medium with 10% fetal bovine serum (FBS), pyruvate, vitamins, amino acids, and antibiotics (29). Working cultures were maintained at 37 °C in a humidified incubator with 5% CO2 and routinely checked for mycoplasma contamination. MCF-7 limiting dilution clones were obtained by plating parental cells into 96-well culture plates at 0.5 cell/well.
RIII Expression Vector Construction and TransfectionThe
full-length cDNA of rat TGF RIII (24) was subcloned by
blunt-ended ligation into a tetracycline-repressible expression system
as described previously (20). The sense orientation of the RIII
cDNA was confirmed by restriction digestion and agarose gel
electrophoresis.
The expression vectors were linearized and transfected into one of MCF-7 limiting dilution clones with a BTX Electro Cell Manipulator at 250 V and 950 microfarads. The control cells were transfected with the empty vectors. The transfected cells were plated in 10-cm culture plates and maintained in the 10% FBS medium for 2 days. Selection of stable transfectants were accomplished by adding Geneticin (G418 sulfate; Life Technologies, Inc.) to the culture medium at 600 µg/ml. G418-resistant clones were ring-cloned and expanded for screening of RIII expression. Control clones were pooled and designated as Neo cell.
RNA AnalysisTotal RNA from G418-resistant cells was
isolated by guanidine thiocyanate homogenization and acidic phenol
extraction (30). To measure the transfected rat RIII mRNA levels in
the clones, we constructed a rat RIII riboprobe by inserting a 470-base
pair BamHI fragment of the rat RIII cDNA into pBSK()
plasmid (Stratagene Cloning Systems). The recombinant plasmid was then
linearized with EcoRI and T3 RNA polymerase was
used to synthesize a radioactive antisense RIII probe. RNase protection
assays were performed using this radioactive antisense riboprobe to
measure RIII mRNA levels in the transfected clones as described
previously (20).
Simian recombinant
TGF1 was purified from conditioned media of transfected
Chinese hamster ovary cells as described previously (31). Purified
TGF
1 was iodinated by the chloramine T method as
described by Ruff and Rizzino (32). To measure the expression of cell
surface TGF
receptors, [125I]TGF
1 (200 pM) was incubated with a cell monolayer in 35-mm tissue
culture wells and then cross-linked to its receptors as described by
Segarini et al. (33). Labeled cell monolayers were solubilized in 200 µl of 1% Triton X-100 containing 1 mM
phenylmethylsulfonyl fluoride. Equal amounts of cell lysate protein
were electrophoresed in 4-10% gradient SDS-polyacrylamide gel
electrophoresis under reducing conditions and exposed for
autoradiography.
To
measure the amount of active TGF1 in the media
conditioned by control or RIII-transfected cells, monolayer cells in
6-well culture plates were cultured in the 10% FBS medium until
confluence. The cells were washed twice with a serum-free McCoy's 5A
medium and incubated in 1.5 ml of this serum-free medium for additional 72 h. The conditioned medium was then collected in siliconized microcentrifuge tubes and the cell number was counted with a
hemocytometer. The amount of mature, activated TGF
1 in
the conditioned medium was determined with a TGF
1 ELISA
kit from Promega (Madison, WI) according to the manufacturer's
instructions.
To study the effect of RIII
expression on cell proliferation at low seeding densities, the control
and RIII-transfected cells were plated at 200 or 400 cells per well in
24-well culture plates. After 9 days of incubation, relative cell
number was determined with an MTT assay as described previously (20).
To determine whether autocrine TGF1 activity was
enhanced after RIII expression, a TGF
1 neutralizing
antibody (R&D Systems) was added to the cells after plating at a final
concentration of 30 µg/ml.
To determine
whether RIII expression enhanced autocrine TGF1
activity, we measured a TGF
-responsive promoter activity using a
plasmid called p3TP-Lux from Dr. J. Massague. The promoter activity is
reported by luciferase (10). The p3TP-Lux (20 µg) and a
-galactosidase expression plasmid (5 µg) were transiently
co-transfected into the MCF-7 Neo control or RIII-expressing cells
(107 cells) by electroporation in the same manner as the
stable transfection described above. The cells were then equally
divided into replicate culture dishes, part of which were treated with
tetracycline (0.1 µg/ml) or the TGF
1 neutralizing
antibody (30 µg/ml) for 48 h. The treated and control cells were
lysed in luciferase buffer (100 mM
K2HPO4, pH 7.8, 1 mM
dithiothreitol) using three cycles of freeze-thaw. The activities of
luciferase and
-galactosidase in the cell extract were assayed using
published procedures (34, 35). Luciferase activity was normalized to
-galactosidase activity and expressed as relative luciferase
activity.
To compare the clonogenic potential of the control and RIII-transfected cells in a semisolid medium, soft agarose assays were performed as described previously (20). Briefly, 6 × 103 cells were suspended in 1 ml of 0.4% low melting point agarose (Life Technologies) dissolved in the 10% FBS medium and plated on the top of a 1-ml underlayer of 0.8% agarose in the same medium in 6-well culture plates. For tetracycline treatment, we added tetracycline in soft agarose at 0.1 µg/ml as well as 200 µl of the medium containing 0.2 µg/ml tetracycline on the top of the solidified agarose. Control wells received 200 µl of the medium without tetracycline. The media were replenished every 3 days. After 3 weeks of incubation in the humidified incubator with 5% CO2 at 37 °C, the cell colonies were visualized by staining with 1 ml of p-iodonitrotetrazolium violet staining (Sigma). To quantitate the clonogenicity, the colonies in each well were counted under a magnifying lens.
MCF-7 cells express very low levels of
RIII and undetectable RII on the cell surface using the receptor
cross-linking method (20). However, they express low levels of RII
mRNA and their growth can be slightly inhibited by high
concentrations of exogenous TGF1 suggesting that they do
have functional TGF
receptors and intracellular signaling components
(20). In addition, this cell line expresses mainly TGF
1
with no detectable TGF
2 and TGF
3 (36).2 Therefore, this cell line is ideal
for the examination of whether TGF
RIII plays a role in regulating
autocrine TGF
1 activity and, consequently, cell
growth.
Tetracycline-repressible RIII expression vectors were transfected into
one of the MCF-7 limiting dilution clones. Using RNase protection
assays and a rat RIII riboprobe, we identified two RIII-expressing
clones. Both clones showed similar levels of RIII expression and growth
properties on plastic and in soft agarose (data not shown). Therefore,
we used one clone for the study. The mRNA level of the transfected
RIII in this clone, designated as RIII cell, is shown in Fig.
1. Consistent with the mRNA level, RIII cells also
expressed a higher level of cell surface RIII protein than the Neo
cells by the receptor cross-linking assay (Fig. 2). The
specificity of [125I]TGF1 binding to the
receptor was confirmed by competition with a 100-fold excess of
unlabeled TGF
1 (Fig. 2, third lane).
Treatment with tetracycline at 0.1 µg/ml for 4 days prior to the
receptor cross-linking assay almost completely suppressed the
expression of transfected RIII (Fig. 2, fourth lane). This
reversible expression of RIII by tetracycline was particularly useful
in the subsequent experiments for the demonstration of the specific
effect of the RIII on the cell growth and autocrine TGF
activity.
Consistent with a previous observation (20), TGF RI and RII were not
detectable by the receptor cross-linking assay on the Neo cell surface
(Fig. 2). However, two bands corresponding to RI and RII were detected
in the RIII cell samples. Since they were absent in the Neo cells and
in the tetracycline-treated RIII cells (Fig. 2), the increased
TGF
1 binding to RI and RII appears to be specifically
due to RIII expression. This is consistent with what was observed in
myoblasts (24, 27).
While RIII cells
showed increased TGF1 binding to RI and RII, their
sensitivity to exogenous TGF
1 remained relatively
similar to that of the Neo cells in a growth inhibition assay (20)
(data not shown). Therefore, we hypothesized that MCF-7 cells may
express enough activated TGF
1 such that the expression
of RIII regenerated a maximal TGF
growth inhibition in an autocrine
manner. To test this hypothesis, we first examined whether the cells
produced mature, activated TGF
1 in our system. Using a
TGF
1 ELISA kit we were able to detect a modest amount of
activated TGF
1 (200 pg/106 cells/72 h) in
the medium conditioned by the Neo cells. The amount of activated
TGF
1 in the medium conditioned by the RIII cells was
reduced by more than half (Fig. 3). This reduction could
be reversed by treating the RIII cells with tetracycline, suggesting that the reduced amount of free activated TGF
1 in the
medium was due to its binding to RIII. This result led us to further examine whether the cell growth was inhibited by the autocrine TGF
1 after RIII expression.
Plating Efficiency Assay
It is well known that TGF
isoforms suppress cell cycle progression of newly plated, non-dividing
cells more effectively than that of exponentially growing cells.
Plating cells at low density is one way to ensure that the cells are
maintained for a period of non- or slow-dividing state before entering
exponential growth. This approach has been successfully employed
previously to demonstrate the autocrine inhibitory activity of TGF
(18, 19). When Neo and RIII cells were plated in 24-well tissue culture
plates at 200 and 400 cells/well, the RIII cells proliferated
significantly slower than the Neo cells by about 40% after nine days
of incubation (Fig. 4A). The growth reduction
was due to reduced number of colonies as well as number of cells per
colony. Treatment of the cells with tetracycline at 0.1 µg/ml prior
to and during the plating efficiency assay completely reversed the
growth property of the RIII cells back to that of the Neo cells as
shown in Fig. 4B suggesting that the growth inhibition was
specifically due to RIII expression. To assess whether the growth
reduction was due to RIII-enhanced autocrine TGF
1
activity, we treated the cells with a TGF
1 neutralizing antibody or a control antibody after plating. While the growth of Neo
cells was not affected by the antibody, the growth of the RIII cells
was stimulated (Fig. 5). Even though the stimulation was
only 15%, it is statistically significant (p < 0.05)
by Student's t test. Since RIII expression reduced the
growth by 40%, 15% growth stimulation represents a 37.5% recovery of
the growth. The reason that tetracycline was more effective than the
TGF
1 neutralizing antibody to stimulate the growth of
the RIII cells was probably due to the fact that tetracycline could
completely suppress RIII expression whereas the antibody had to compete
with the RIII for endogenous TGF
1.
Activity of a TGF
To confirm our
observation that RIII expression enhanced autocrine TGF1
activity, we measured the activity of a TGF
-responsive promoter/luciferase construct after being transiently transfected into
the Neo and RIII cells. This construct has been widely used to measure
the efficacy of TGF
in activating its signaling receptors and
stimulating gene expression (10). While RIII expression inhibited the
growth of the MCF-7 cells, it stimulated TGF
-responsive promoter
activity as indicated by luciferase activity by more than 3-fold (Fig.
6A). Treatment of the transiently transfected cells with tetracycline for 48 h reduced the luciferase activity almost to the level of the Neo cells, again demonstrating the specificity of the RIII effect. Similar to the growth assay, the stimulated promoter activity could be partially, but significantly (p < 0.05) reduced by treating the transfected RIII
cells with the TGF
1 neutralizing antibody for 48 h.
Thus, with two different approaches, we have showed that RIII can
restore autocrine TGF
1 activity in the MCF-7 cells.
Clonogenicity in Soft Agarose
The growth inhibitory activity
of autocrine TGF is known to maintain cancer cells at a less
malignant state as previously shown by its suppressive effect on the
clonogenicity in soft agarose and the tumorigenicity in nude mice (8,
18, 20). Since RIII expression restored autocrine TGF
1
activity, we compared the ability of Neo and RIII cells to form
colonies in a soft agarose assay. As shown in Fig.
7A, the number of colonies formed by RIII cells were considerably less than that formed by Neo cells.
Tetracycline had no effect on Neo cells, but significantly increased
the number of colonies formed by RIII cells (Fig. 7B). These
results suggest that by restoring autocrine TGF
1
activity, RIII can also reduce the anchorage-independent growth of the
MCF-7 cells in the same manner as RII (20).
Although RIII has been shown to enhance TGF1
binding to RII, this is the first report demonstrating an important
role of RIII in regulating autocrine TGF
1 activity.
Expression of RIII in MCF-7 cells increased affinity labeling of both
RI and RII by [125I]TGF
1 to a detectable
level. Increased TGF
1 binding to RII after RIII
expression is consistent with what was observed in myoblasts (24, 27).
However, RIII expression in myoblasts had no effect on
TGF
1 binding to RI. The authors suspected that the
amount of RII with high affinity for TGF
1 in the
myoblasts without RIII was sufficiently high enough to support
TGF
1 binding to all cell surface RI molecules (27). In
contrast, MCF-7 cells expressed a very low level of RII mRNA and no
detectable amount of cell surface RII protein with the affinity
labeling (20). As a result, even though they express a relatively high
level of RI mRNA (20), little RI could be affinity-labeled with
TGF
1 since RI requires TGF
-bound RII for ligand
binding. RIII expression in MCF-7 cells had no effect on RII mRNA
expression,2 suggesting that the increased
TGF
1 binding to RII in the RIII cells was due to
increased binding affinity as observed in the myoblasts (27). It
appears that the increased TGF
1 binding to RI resulted
from the increased TGF
1 binding to RII. Therefore, whether RIII is necessary for maintaining an optimal autocrine TGF
1 activity in a given cell type appears to be
dependent on the ratio of RII to RI.
In the myoblasts that lack RIII expression, expression of RIII was
shown to convert RII molecules into a single population with high
affinity (27). Due to its low expression level, we do not know whether
the two populations of RII also exist in MCF-7 cells. The fact that
their growth can only be inhibited by high concentrations of exogenous
TGF1 (20) suggests that they contain the low affinity
RII. Since the TGF
1 neutralizing antibody had no effect
on the growth of the Neo cells, it appears that they contain no or
little high affinity RII which does not confer a significant amount of
autocrine TGF
1 activity.
TGF RI, RII, and RIII are widely expressed in many types of normal
and transformed cells (37). Whether RIII plays a role in regulating
autocrine TGF
activity in certain types of normal cells remains to
be elucidated. However, our data would suggest that if a normal cell
expresses less RII than RI, RIII could be essential to confer TGF
sensitivity. This could also be the case during development of a cell
or an organ when the ratio of RII to RI may vary at different stages.
In transformed cells including breast, colon, stomach, prostate cancer
cells, retinoblastoma cells, and lymphoma cells, RII is often
down-regulated (12-14, 17, 38-42). While in some cancers such as
colon and gastric cancers, the down-regulation of RII can be due to
gene mutation (12, 14, 15), in other cases, it may be due to reduced
gene expression as in the MCF-7 cells. If the latter is true for a
given type of transformed cells, RIII level will be critical to the
maintenance of the autocrine TGF
activity and consequently a less
malignant phenotype of the cells. The fact that the expression of RIII
shown in this report can be as effective as the expression of a
moderate level of RII (20) in suppressing the anchorage-independent
growth of the MCF-7 cells demonstrates this important role of RIII in keeping the autocrine TGF
1 activity at an optimal level.
Since TGF
3 behaves similarly to TGF
1 in
receptor binding and TGF
2 requires RIII for binding to
the signaling receptors (27), it is conceivable that RIII can also be
critical to the maintenance of the autocrine activities of
TGF
2 and TGF
3 in the cells that express
these two isoforms.
While exogenous TGF can inhibit the growth of many types of
transformed cells (4), studies have shown that in some systems, autocrine TGF
can render cells insensitive to the growth inhibitory activity of exogenous TGF
1 (18). This is likely due to
the fact that the cells not only produce high levels of one or more TGF
isoforms, but also can activate them. In
TGF
1-transfected fibroblasts and fibrosarcoma cells, it
was found that a significant amount of cell-activated
TGF
1 was bound to the cell surface even though very
little activated TGF
1 could be detected in the
conditioned medium (43). The MCF-7 cells we used produced a readily
detectable amount of mature, activated TGF
1 in the
conditioned medium as shown in Fig. 3. In addition, the cells express a
low level of RII which should be easily saturable by the ligand.
Therefore, we hypothesized that although RIII-enhanced
TGF
1 binding to the signaling receptor RII, the reason
that the sensitivity of the MCF-7 cells to the growth inhibitory
activity of the exogenous TGF
1 was not increased after
RIII expression was due to the fact that the autocrine
TGF
1 had generated a maximal level of growth inhibition
by saturating RII and/or the intracellular signaling pathway. Since
RIII is known to form a heteromeric complex with RII in the presence of
activated TGF
(27, 28), its main role is probably to make the RII
more accessible by the ligand. In addition, by capturing and retaining
ligand, RIII may protect the ligand from cellular degradation and/or
inactivation, rendering more ligand available for binding to the
signaling receptors. The RIII-enhanced autocrine TGF
1
activity was demonstrated in the MCF-7 cells by the fact that a
TGF
1 neutralizing antibody could partially reverse the
RIII-induced growth inhibition and gene expression.
Because of its short cytoplasmic domain with no consensus signaling
motif and its absence in some of the TGF-sensitive cells, RIII has
been regarded as a non-signaling, accessory TGF
receptor. As a
result, the RIII-generated growth inhibition we observed in MCF-7 cells
was solely attributed to the increased autocrine TGF
1
binding to RII and, consequently the activation of RI·RII receptor
complex even though TGF
1 neutralizing antibody only abrogated about one-third of the RIII-induced growth inhibition as well
as gene expression. However, our study cannot rule out the possibility
that RIII may transduce signal by itself or in association with RII. We
are currently attempting to address this possibility.
In summary, our study showed that the expression of RIII can lead to
the inhibition of both anchorage-dependent and
anchorage-independent growth in MCF-7 cells. This is at least in part
mediated by RIII-restored autocrine inhibitory activity of the
endogenous TGF1. RIII may be essential for an optimal
level of autocrine TGF
activity in various types of normal cells.
Its expression is likely to be critical for the maintenance of the
growth inhibitory activity of autocrine TGF
isoforms and,
consequently of a less malignant phenotype in other adenocarcinoma
cells with reduced RII expression.
We thank Dr. H. Bujard at the University of Heidelberg, Germany, for the tetracycline-repressible expression plasmids and Dr. J. Massague at Memorial Sloan-Kettering Cancer Center, New York, for the p3TP-Lux plasmid. We also thank Dr. David Kaetzel at the University of Kentucky College of Medicine for critical reading of this manuscript.