(Received for publication, November 15, 1996, and in revised form, January 21, 1997)
From the § Division of Cancer Biology Research, OCI-5 encodes the rat homologue of glypican-3, a
membrane-bound heparan sulfate proteoglycan that is mutated in the
Simpson-Golabi-Behmel overgrowth syndrome. OCI-5 and glypican-3 are
95% identical. It has been recently suggested that glypican-3
interacts with insulin-like growth factor-2 (IGF-2) and that this
interaction regulates IGF-2 activity. We report here that we have
transfected OCI-5 into two different cell lines, and we have not been
able to detect an interaction between the OCI-5 proteoglycan produced
by the transfected cells and IGF-2. On the other hand, we have found
that OCI-5 interacts with FGF-2, as has already been shown for
glypican-1. This interaction is mediated by the heparan sulfate chains
of OCI-5 because it can be inhibited by heparin or by heparitinase.
Glypicans are a family of heparan sulfate proteoglycans that are
bound to the cell membrane through a glycosylphosphatidylinositol anchor (1). In mammalian cells four members of this family have been
characterized (2-5). Although the function of glypicans is still
poorly understood, David and collaborators have recently provided
convincing evidence showing that glypicans can interact with fibroblast
growth factor-2 (FGF-2)1 and stimulate
occupancy and signaling of the FGF receptor-1 (6). This is not
surprising because it is well established that many heparan sulfate
chains can interact with heparin-binding growth factors such as FGF-2
(7-9). Recently, it was reported that glypican-3 is mutated in
patients with the Simpson-Golabi-Behmel overgrowth syndrome (SGBS)
(10). This syndrome is characterized by pre- and post-natal overgrowth
with visceral and skeletal anomalies (11, 12). It was also reported in
the same paper that glypican-3 can interact with IGF-2, suggesting that
the SGBS phenotype is the result of altered regulation of IGF-2
activity. In this study we have transfected OCI-5, the rat homologue of
glypican-3 (2, 13) into WI-L2, a human lymphoblastic cell line, and
into HT-29, a human colorectal carcinoma cell line. We report that
OCI-5 produced by these cells binds to FGF-2. We have not been able, on
the other hand, to show any interaction with IGF-2.
The introduction
of the hemagglutinin A (HA) epitope into the OCI-5 cDNA has been
previously described (2). The HA-tagged OCI-5 cDNA was introduced
into the pEF-BOS vector (14). WI-L2 cells were transfected by
electroporation and selected in the presence of 800 µg/ml of
geneticin. Cells expressing OCI-5 were selected by
fluorescence-activated cell sorter after immunostaining with an anti-HA
antibody. HT-29 cells were transfected with Lipofectin (Life
Technologies, Inc.) and selected in the presence of 800 µg/ml of
geneticin. Individual clones were isolated with cloning cylinders.
WI-L2 cells were cultured in RPM1 1640, 10%
fetal calf serum, and HT-29 cells in Dulbecco's modified Eagle's
medium, 10% fetal calf serum. Both cell lines were obtained from
ATCC.
Conditioned
media of WI-L2 and HT-29 cells were collected as described below and
immunoprecipitated using the anti-HA 12CA5-I monoclonal antibody (20 µg/ml) and protein-Sepharose beads. The immunoprecipitates were then
rinsed twice with wash solution (0.1% Nonidet P-40, 5 mM
EDTA, PBS) and three times with enzyme buffer (50 mM Hepes,
pH 7.0, 10 mM NaCl, 1 mM CaCl2, 50 µg/ml BSA, 2 mM phenylmethanesulfonyl fluoride, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 1 µg/ml pepstatin, and 10 mM EDTA). Once resuspended in 100 µl of enzyme buffer,
the immunoprecipitates were incubated in the presence or the absence of
20 milliunits chondroitinase ABC (Sigma) or 2.5 milliunits of heparitinase (Seikagaku) for 3 h at 37 °C. The
digested immunoprecipitates were then transferred to reducing sample
buffer, boiled for 5 min, and run on a 4-9% polyacrylamide gradient
gel. The proteins were transferred to a PVDF membrane (DuPont NEN) and
subjected to Western blotting. Briefly, the membranes were blocked in
blocking buffer (10 mM Tris-HCI, pH 8.0, 150 mM
NaCl, 0.1% Tween 20, 5% nonfat dry milk) for 2 h at room
temperature (or overnight at 4 °C) and then incubated for 2 h
with 1 µg/ml 12CA5-I antibody at room temperature. After incubation
with the anti-mouse secondary antibody for 1 h, bands were
detected using ECL reagents (Amersham Corp.).
96-microwell plates were coated a
day in advance. Briefly, 20 µg/ml stock solutions of BSA, IGF-2
(Vector Biosystems), and basic FGF (ID Labs) were serially diluted 1:2
in PBS. Wells were left overnight at 4 °C, and then the unbound
proteins were removed by aspiration. Each well was then washed twice
with PBS and blocked with 100 µl of blocking solution (2.5 mg/ml BSA
in serum-free RPMI 1640) for at least 2 h at 37 °C. In the
meantime, cells were harvested, washed three times in the blocking
solution, and then diluted in blocking solution at 8 × 105 cells/ml. 100 µl of the cell suspension was added to
each well, and cells were allowed to attach to the coated wells for
1 h at 37 °C. For glycosaminoglycan inhibition assays, blocking
was allowed to progress for another hour in the presence of 100 µg/ml
heparin, dermatan sulfate, or chondroitin sulfate (all from
Sigma) prior to cell attachment. Unattached cells were
removed by gently aspirating the blocking solution with a
micropipettetor, and wells were washed twice with 100 µl of PBS.
Attached cells were fixed with 100 µl of 3% paraformaldehyde in PBS
for 20 min at 37 °C and then stained with 100 µl of 3% toluidine
blue overnight at room temperature. The following day the dye was
dumped out, and the wells were washed with water. The plates were dried
and then read spectrophotometrically at 570 nm.
Conditioned media were
collected from cells that were grown at high density overnight under
serum-free conditions and concentrated to desired volumes using
Centricon-10 filters (Amicon) at 4 °C. Proteinase inhibitors and
Nonidet P-40 (final concentration of 0.1%) were added to retentates
prior to immunoprecipitation. For ligand blots, nonreducing sample
buffer was added, and the samples were stored at Concentrated conditioned media were run on
a 4-9% polyacrylamide gel gradient, and proteins were then
transferred to a PVDF membrane. The transferred membrane was
immediately immersed in TBS (10 mM Tris-HCI, pH 8.0, 150 mM NaCl) with 3% Nonidet P-40 for 30 min at 4 °C. The
membrane was then blocked for 3 h with 1% BSA in TBS. After a
brief rinse with TBS-T (TBS with 0.1% Tween 20), the membrane was
incubated with 1 µCi/ml of [125I]IGF-II (ICN, 214 µCi/µg) in TBS-T overnight at 4 °C. The membrane was then washed
twice with TBS-T for 20 min each and then three times with TBS for 30 min each. After drying the membrane was exposed to a PhosphorImager
screen.
Conditioned media, collected as described
above, were immunoprecipitated with a 1:10 v/v ratio of a
12CA5-I/protein A-Sepharose slurry (12CA5-I antibody prebound to
protein A beads) overnight at 4 °C. When indicated,
immunoprecipitates were then subjected to heparitinase digestion as
described above. Digested beads were washed three times with binding
buffer (0.5% BSA in PBS) and resuspended in a final volume of 500 µl. 50 ng/ml of FGF-2 or 1 µg/ml of IGF-2 were then added, and
binding was permitted for an additional 2 h at 4 °C on an
orbital shaker. The immunoprecipitates were then washed three times in
PBS, reducing sample buffer was added, the samples were boiled for 5 min, and electrophoresis was performed with a 15% polyacrylamide gel.
Proteins were transferred to a PVDF membrane and Western blotted as
described above with either a primary mouse anti-FGF-2 or anti-IGF-2
antibody (both from Upstate Biotechnology Inc.).
WI-L2 and HT-29, two cell lines that do not express
OCI-5/glypican-3, were transfected with an expression vector containing an HA-tagged OCI-5 cDNA. WI-L2 is a human lymphoblastic cell line that does not bind significantly to FGF-2-coated cultured dishes (15).
Based on this property this cell line was previously used to clone
syndecan-1 with a ligand affinity cloning procedure designed to
identify molecules that are able to interact with FGF-2. We have
previously reported that OCI-5-transfected COS-1 cells secrete detectable amounts of this proteoglycan into the medium (2). Fig.
1 shows that this is also the case with
OCI-5-transfected WI-L2 and HT-29 cells. It can also be seen in this
figure that heparitinase, but not chondroitinase, is able to reduce the
molecular weight of the secreted OCI-5, indicating that, like in COS-1
cells (2), OCI-5 expressed by WI-L2 and HT-29 cells display only heparan sulfate chains. It is important to note that because reducing conditions were used for the electrophoresis, Fig. 1 shows that the
conditioned media from WI-L2 and HT-29 cells contain significant amounts of a 41-kDa band recognized specifically by the anti-HA antibody. Under nonreducing conditions, however, the amount of the
41-kDa band is significantly reduced, and a parallel increase in the
glycanated OCI-5 is observed (compare Fig. 1 with Fig. 4A).
Similar findings have been reported for K-glypican (5). These results
suggest that there is a proteolytic cleavage site within the OCI-5 core
protein and that at least a pair of cysteine residues form a disulfide
linkage between the NH2- and COOH-terminal sides of the
cleavage site. Because the HA epitope is located at the NH2
terminus, only this part of OCI-5 can be detected under reducing
conditions with the anti-HA antibody. This 41-kDa cleavage product,
however, does not contain the potential glycosaminoglycan attachment
sites.
To investigate the interaction of OCI-5 with FGF-2 and IGF-2,
conditioned media from OCI-5-transfected WI-L2 cells were collected, and OCI-5 was immunoprecipitated by using the anti-HA antibody bound to
protein A beads. FGF-2 or IGF-2 were then added, and binding of these
growth factors to the immobilized OCI-5 was determined by Western blot
using antibodies against FGF-2 and IGF-2. Fig. 2 shows
that although significant amounts of FGF-2 bound to the immunoprecipitated OCI-5, no detectable IGF-2 is found in the precipitates. If OCI-5 is previously incubated with heparitinase, the
amount of bound FGF-2 is significantly reduced, indicating that the
interaction of FGF-2 with OCI-5 is mediated by the heparan sulfate
chains (Fig. 2).
We and others have shown that glypicans can be found both in the
conditioned medium and on the cell surface (2-5). It was decided,
therefore, to investigate whether cell-bound OCI-5 is able to interact
with FGF-2 and IGF-2. A procedure previously used by others to
demonstrate that syndecans interact with FGF-2 was used (15). WI-L2
cells transfected with OCI-5 were tested for attachment on microwells
covered with FGF-2 or IGF-2. Although a significant
dose-dependent attachment of WI-L2 cells to FGF-2 was
observed (Fig. 3B), concentrations of IGF-2
ranging between 2.5 and 20 µg/ml failed to produce cell attachment
above the nonspecific levels seen with BSA. The attachment to FGF-2 was
completely inhibited if wells were preincubated with heparin but not
with chondroitin sulfate or dermatan sulfate (Fig. 3A).
These results indicate that, like in the binding assay, the interaction
between OCI-5 and FGF-2 is mediated by the heparan sulfate chains.
Because we were unable to detect any interaction between IGF-2 and
OCI-5 in our cell-attachment and binding assays, we decided to
determine if this interaction could be observed by ligand blot analysis. This is the experimental approach that was used by Pilia et al. to show an interaction between glypican-3 and IGF-2
(10). For this assay we collected conditioned media from WI-L2 and
HT-29 cells transfected with OCI-5. Fig. 4A
shows a Western blot analysis under nonreducing conditions of the
solubilized form of OCI-5 found in the conditioned media of transfected
WI-L2 and HT-29 cells. An IGF-2 ligand blot performed in parallel using
the same conditioned media is shown in Fig. 4B. It can be
seen that despite the presence of large amounts of glycanated and
nonglycanated OCI-5 bands in the Western blot, the only bands detected
in the ligand blot displayed molecular masses similar to isoforms of IGFBP-4 (33-29 kDa), an IGF binding protein known to be produced by
HT-29 cells (16) (Fig. 4B).
Here we show that OCI-5, the highly conserved rat homologue of
glypican-3, can interact with FGF-2 and that this interaction is
mediated by the heparan sulfate chains of OCI-5. In this respect, therefore, OCI-5/glypican-3 behaves like glypican-1 (6). We have been
unable, on the other hand, to show an interaction between OCI-5 and
IGF-2. This result differs from a previous report that showed by ligand
blot analysis that glypican-3 can interact with IGF-2 (10). Besides
failing to detect OCI-5/IGF-2 interaction by ligand blot analysis, we
have also shown here that immobilized OCI-5 does not bind to IGF-2 and
that it does not promote cell attachment to IGF-2-covered microwells.
Currently we do not have an explanation for the different results
obtained by us and Pilia et al. (10). These investigators
have shown an interaction between glypican-3 and IGF-2 using
conditioned medium from CaCo-2, a human colorectal carcinoma cell line.
Because the type of GAG chains attached to proteoglycans can be cell
type-specific (17), it could be argued that the difference in the
results is due to the fact that we have not used the same cell line.
However, the ligand blot shown by Pilia et al. (10) seems to
indicate that IGF-2 interacts with the protein core of glypican-3,
because the binding increases after digestion with chondroitinase AC.
This would exclude cell specificity of GAG chains as an explanation for
the different results. We have also performed ligand blot analysis with
IEC-18, a cell line that expresses OCI-5 endogenously, but, like with WI-L2 and HT-29 cells, we were unable to detect any interaction between
IGF-2 and OCI-5 (data not shown). It is also important to note that
whereas we have been able to detect IGFBP-4 in our ligand blot from
HT-29 cells, Pilia et al. (10) did not detect any of the
IGFBPs that are reported to be produced by CaCo-2 cells (18).
The reason that prompted the investigation on the potential interaction
of glypican-3 with IGF-2 is that there is considerable clinical overlap
between SGBS patients and patients with the Beckwith-Wiedemann syndrome, an overgrowth syndrome in which the overexpression of IGF-2
is thought to play an important role (19). However, it also evident
that there are numerous distinguishing features between both syndromes
(11, 20).
Certainly the fact that OCI-5 can interact with FGF-2 does not rule out
the possibility of OCI-5 interacting with other proteins, including
molecules that regulate IGF-2 activity. In this respect, it is also
important to note that the core proteins of glypicans are highly
conserved during evolution (5, 10) and that the position of 14 cysteine
residues is conserved in all glypicans. This raises the possibility
that the core proteins of glypicans have functions other than just
acting as anchors for GAG chains.
We thank Lynda Woodcock and Cassandra
Cheng for assistance in the preparation of this manuscript and Drs.
Yaacov Ben-David and Michael Pollack for critically reviewing this
manuscript.
OCI-5 Expression Vectors and Transfections
20 °C.
Fig. 1.
Characterization of OCI-5 secreted by WI-L2
and HT-29 cells. OCI-5 from conditioned media were
immunoprecipitated with 12CA5 antibody, incubated with buffer alone
(lanes 1, 4, 7, and 10),
heparitinase (lanes 2, 5, 8, and
11), or chondroitinase (lanes 3, 6,
9, 12) and analyzed by Western blot with the
12CA5 antibody. WI-L2 wild type (lanes 1, 2, and
3), WI-L2 + OCI-5 (lanes 4, 5, and
6), HT-29 wild type (lanes 7, 8, and
9), and HT-29 + OCI-5 (lanes 10, 11,
and 12) are shown. The 70-kDa OCI-5 core protein (filled arrowhead) and 41-kDa degradation product
(open arrowhead) are indicated. The numbers on
the right represent molecular mass markers.
[View Larger Version of this Image (64K GIF file)]
Fig. 4.
IGF-2 ligand blot. A, Western blot
analysis of OCI-5 with the anti-HA antibody under nonreducing
conditions of conditioned media from WI-L2 (lanes 1 and
2) and HT-29 cells (lanes 3 and 4).
Lanes 1 and 3, OCI-5 transfected cells;
lanes 2 and 4, wild type cells. The 41-kDa
(open arrowhead) and other (open triangles) OCI-5
core protein cleavage products are indicated. B, parallel IGF-2 ligand blot analysis of the same media. Bands
corresponding to IGFBP4 (filled arrowheads) are indicated.
The numbers in the middle represent molecular
mass markers.
[View Larger Version of this Image (90K GIF file)]
Fig. 2.
Binding assays. Conditioned media from
WI-L2 + OCI-5 (lanes 2-5) and WI-L2 wild type (lanes
6-9) were immunoprecipitated with the anti-HA antibody, incubated
at 37 °C for 3 h in the presence (lanes 2,
4, 6, and 8) or the absence
(lanes 3, 5, 7, and 9) of heparitinase, and then 1 µg/ml of IGF-2 (lanes 2,
3, 6, and 7) or 50 ng/ml of FGF-2
(lanes 2, 3, 6, and 7) were
added at 4 °C for 2 h. Immunoprecipitates were analyzed by
Western blot with anti-FGF-2 and anti-IGF-2 antibodies. Lane
1 contains recombinant FGF-2 (5 ng), and lane 10 contains recombinant IGF-2 (25 ng). The 70-kDa OCI-5 core protein
(filled arrowhead) and 41-kDa cleavage product (open
arrowhead) are indicated. The numbers on the
right represent molecular mass markers.
[View Larger Version of this Image (38K GIF file)]
Fig. 3.
Attachment of WI-L2 cells to IGF-2 and FGF-2.
A, attachment of WI-L2 cells to 96-well plates coated with
BSA (20 µg/ml), IGF-2 (20 µg/ml), or FGF-2 (5 µg/ml).
B, attachment of WI-L2 cells to 96-well plates coated with
increasing concentrations of FGF-2 and IGF-2. The result represent
averages ± S.E. of duplicates. CS, chondroitin
sulfate; DS, dermatan sulfate. Wild type W1-L2 attachment to
BSA and IGF-2 are identical to those observed for W1-L2 + OCI-5 (data
not shown; refer to panel A).
[View Larger Version of this Image (29K GIF file)]
*
This work was supported by a grant from the Medical Research
Council of Canada.The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
¶
To whom correspondence should be addressed.
1
The abbreviations used are: FGF, fibroblast
growth factor; SGBS, Simpson-Golabi-Behmel overgrowth syndrome; IGF,
insulin-like growth factor; HA, hemagglutinin A; PBS,
phosphate-buffered saline; BSA, bovine serum albumin; PVDF,
polyvinylidene difluoride.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.