(Received for publication, October 9, 1996, and in revised form, February 10, 1997)
From the Departments of We have examined the expression of the small
extracellular chondroitin/dermatan sulfate proteoglycans (CS/DS PGs),
biglycan, decorin, and PG-100, which is the proteoglycan form of colony stimulating factor-1, in the human endothelial cell line EA.hy 926. We
have also examined whether modulation of the phenotype of EA.hy 926 cells by tumor necrosis factor- Proteoglycans (PGs)1 are complex
macromolecules that consist of a protein core and one or more
glycosaminoglycan (GAG) side chains, covalently bound to the core
protein (1). Nearly 30 individual PGs with various functions residing
in the extracellular matrix, on the cell surface, or inside the cells,
have been identified (1-4). Among the PGs of the extracellular matrix,
four small chondroitin/dermatan sulfate (CS/DS) PG species, biglycan
(5), decorin (6), PG-100 (7), and epiphycan (4, 8, 9) are currently
known. Two of these PGs, biglycan and decorin, are highly homologous
with each other in terms of their core protein structure (5), and
together with epiphycan, fibromodulin, and lumican they both belong to
the small leucine-rich PG gene family (4). With some exceptions
biglycan and decorin differ in the number of GAG chains attached to
their core proteins. Biglycan is usually substituted with two GAG
chains, whereas decorin typically has only one GAG chain (10-12). On
sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)
biglycan has variously been shown to migrate at positions of protein
molecular weight markers of ~200,000 or more, while decorin migrates
at positions of protein molecular weight markers of 87,000-180,000
(13-17). The third small extracellular CS/DS PG species, originally
isolated from the culture medium of human osteosarcoma cells and
tentatively named PG-100 (7), has recently been demonstrated to be
identical with the macrophage colony-stimulating factor CSF-1 (18).
PG-100 is therefore a PG form of this cytokine (19, 20), and it can be
called a "part-time" PG to be differentiated from biglycan,
decorin, and other PGs that usually exist only in the GAG containing
form (21). The fourth small CS/DS PG, epiphycan, has been shown to be
closely related to osteoinductive factor (22), and the expression of
this PG seems to be cartilage specific (8).
The exact functions of the small CS/DS PGs are still somewhat
controversial. Decorin, the most thoroughly investigated molecule in
this group, has been shown to bind to collagen fibrils and to regulate
fibril formation (23-26). Decorin has also been found to interact with
the cytokine transforming growth factor- We have earlier shown that the principal small extracellular CS/DS PG
species synthesized by bovine aortic endothelial (BAE) cells in
monolayer cultures is biglycan (16). This result has been shown to be
true for human umbilical vein endothelial cell monolayers as well (16).
We have also shown that BAE cell monolayer cultures do not express
detectable amounts of either decorin or type I collagen (16). However,
when BAE cells change their phenotype and form so called sprouting
cultures, the synthesis of both decorin and type I collagen is
initiated (38, 39) indicating that the synthesis of decorin and type I
collagen is associated with an in vitro angiogenesis
phenomenon.
In this study we have examined the expression of the small
extracellular CS/DS PGs in the permanent human endothelial cell line
EA.hy 926 (40). Previously this cell line has been shown to possess
several characteristics typical of endothelial cells (40-46). However,
the production of extracellular matrix molecules, including PGs, of
EA.hy 926 cells has remained unidentified. Furthermore, we have
examined whether modulation of the phenotype of EA.hy 926 cells is
associated with specific changes in the synthesis of the small
extracellular CS/DS PGs. We demonstrate that EA.hy 926 cells express
biglycan, but not decorin. We also demonstrate that these cells
synthesize PG-100, the Mr of which (~250,000) is similar to that of biglycan synthesized by these same cells. In
addition, the inhibitory effect of TNF- Cells of the permanent human endothelial cell
line EA.hy 926 (40) were grown on plastic dishes (Nunc, Roskilde,
Denmark) at 37 °C in 95% air + 5% CO2 in
bicarbonate-buffered Dulbecco's modified Eagle's medium (Flow
Laboratories, Irvine, United Kingdom) containing 100 IU/ml of
penicillin (Oriola, Finland) and 50 µg/ml streptomycin (Oriola), and
supplemented with 10% (v/v) of fetal calf serum (FCS, Life
Technologies, Inc., Paisley, Scotland), and 1 × HAT (100 µM hypoxanthine, 0.4 µM aminopterin, 16 µM thymidine; Life Technologies, Inc.). For subcultures,
the cells were detached with 0.01% (w/v) trypsin (Sigma).
Human skin fibroblasts (HSFs) that were used as control cells were
derived from a skin biopsy of a healthy person, and these cells were
subcultured as described previously (47). The culture medium used for
HSFs was the same as described above for EA.hy 926 cells with the
exception that HAT was omitted.
For identification of
sulfated PGs in the culture media, confluent cell cultures were labeled
for 24 h with 50 µCi/ml
Na2[35S]O4 (Amersham, United
Kingdom) in sulfate-free Dulbecco's modified Eagle's medium
supplemented with 10% (v/v) FCS, whereafter the media were collected
and the detached cells were removed by low speed centrifugation (2000 rpm) for 5 min. The incorporation of [35S]sulfate into
newly synthesized PGs was evaluated by the cetylpyridinium chloride
(CPC) precipitation assay (48) before further analyses of the culture
media (see below).
In the experiments examining the effect of TNF- Cell number calculations were
performed with a hemocytometer. Before the calculations, the cells were
fixed with a formaldehyde containing buffer solution (10%, v/v, of
37% formaldehyde in 0.085 M NaCl and 0.1 M
Na2 SO4). Two aliquots of each cell suspension were taken for the cell counting.
The viability of the cells in
the control and TNF- For the identification of
CS/DS and heparan sulfate (HS) PGs in the culture media, equal volumes
of [35S]sulfate-labeled media were ethanol-precipitated,
dried, dissolved into appropriate buffers, and digested with 0.1 IU/ml
of chondroitin ABC lyase from Proteus vulgaris
(Seikagaku Co., Tokyo, Japan) or 6 milliunits/ml of heparitinase
from Flavobacterium heparinum (Seikagaku Co.), respectively
(50, 51).
The immunoprecipitation of biglycan and
PG-100 from the culture media with polyclonal antisera against these
two PGs (see below) was performed as described previously (52).
Aliquots of [35S]sulfate-labeled media corresponding to
500,000 cpm as determined by the CPC precipitation assay (48) were
taken for the analysis. The immunoprecipitants were run on a gradient
SDS-PAGE as described (see below).
Partial purification of PGs by ion exchange
chromatography was performed by applying equal aliquots of
[35S]sulfate-labeled media to a DEAE-Sephacel (Pharmacia,
Uppsala, Sweden) column equilibrated with 0.2 M NaCl in the
buffer containing 50 mM sodium acetate, 2.0 M
urea, 1 mM phenylmethylsulfonyl fluoride, and 0.1% Triton
X-100. The column was eluted with a linear gradient of 0.2-1.0
M NaCl in the same buffer as above. The amount of
radioactivity in sulfated PGs in each fraction was determined by the
CPC precipitation assay (48). All fractions that contained
[35S]sulfate-labeled PGs were pooled before further
analyses.
SDS-PAGE of the
samples was carried out essentially as described by Laemmli (53) on a
4-12% linear gradient gel with a 3% stacking gel (16). The positions
of the radioactive bands were visualized by fluorography of dried gels
previously treated with 2,5-diphenyloxazole (Fisons, UK).
14C-Methylated protein molecular weight standards
(Amersham) were used to estimate the average sizes of
[35S]sulfate-labeled macromolecules.
[35S]Sulfate-labeled
medium samples or partially purified PGs from similar medium samples
were lyophilized, whereafter they were digested with chondroitin ABC
lyase and heparitinase to remove all GAG chains (see above). Next,
SDS-PAGE was carried out (see above) under reducing conditions using a
12.5% gel, whereafter the proteins were electroblotted to a
nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany).
Specific bands were detected by immunostaining (18, 54) using
polyclonal antisera against biglycan and PG-100 (see below).
Biglycan antiserum was a polyclonal antiserum made
in a rabbit against a synthetic peptide of human biglycan (amino acids 11-24) that was conjugated to bovine serum albumin before injections (5). This antiserum, called LF-51, was kindly provided by Dr. L. Fisher
(National Institute of Dental Research, Bethesda, MD). PG-100 antiserum
was a polyclonal antiserum that was raised in a rabbit against PG-100
core protein as described (7). This antiserum was kindly provided by
Dr. H. Kresse (University of Münster, Münster,
Germany).
[35S]Sulfate-labeled
medium samples from control and TNF- Total cellular RNA
was isolated from the cultures using the single-step method (58).
Northern blot analyses using GeneScreen Plus membranes (DuPont NEN)
were performed essentially as described previously (16). Random priming
(Multiprime DNA Labeling System, Amersham, UK) was used for labeling of
the cDNAs with 5 The following cDNA probes were used: a
biglycan cDNA (5) for the full-length human biglycan core protein;
a 1338-base pair CSF-1 cDNA (18) for the PG-100 core protein; a
decorin cDNA (6) for the full-length human decorin core protein; a
1.1-kilobase perlecan cDNA, called HS-1 (60); and pHCAL1U (61) for
pro- The following cytokines were used: TNF- Student's unpaired t test
was used to describe the significance of the differences between the
results of the control and cytokine-treated EA.hy 926 cell
cultures.
SDS-PAGE of [35S]sulfate-labeled medium
sample from a confluent monolayer culture of EA.hy 926 cells
demonstrated that these cells synthesize and secrete into the medium
two size classes of [35S]sulfate-labeled macromolecules,
one that remains on the top of the separating gel (Fig.
1, lane 1, band III), and another one with
the Mr of ~250,000 (Fig. 1, lane 1, band
II). Chondroitin ABC lyase and heparitinase digestions of
[35S]sulfate-labeled medium samples prior to SDS-PAGE
indicated that the macromolecules of band II are CS/DS PGs, while those
of band III represent HS containing PGs (Fig. 1, lanes
2-4), e.g. perlecan (60), since these cells express
abundant mRNA for this PG (data not shown). The composition of
sulfated PGs in the culture medium of EA.hy 926 endothelial cell
monolayers differed from that of sulfated PGs present in the culture
medium of HSFs used as control cells. Besides containing PGs migrating
in band II and band III (Fig. 1, lane 5), the culture medium
of HSFs contained two additional size classes of sulfated PGs, one with
the Mr of ~120,000 (Fig. 1, lane 5, band
I) and another one that does not enter the separating gel (Fig. 1,
lane 5, band IV). Both of these additional PG size classes
were sensitive to chondroitin ABC lyase digestion (Fig. 1, lane
6) indicating that they represent CS/DS PGs. Based on the results
of previous studies, the band I PGs (Mr of
~120,000) most probably represent decorin, while the PGs of band IV
mainly represent versican (6, 13, 62-65).
The above results showed that EA.hy 926 cells in a monolayer culture
synthesize and secrete into the medium only one size class (band II) of
small CS/DS PGs. The average size of these CS/DS PG molecules is of the
same range as that of HSF-derived small extracellular CS/DS PGs,
previously identified as biglycan (64, 65). To clarify whether the
EA.hy 926 cell derived small CS/DS PGs that migrate on SDS-PAGE in the
band II also represent biglycan, immunoprecipitation of
[35S]sulfate-labeled medium sample from these cells using
an antiserum against biglycan (LF-51) was performed. As shown in Fig.
2 (panel A, lanes 1 and 3), the
biglycan antiserum LF-51 was able to immunoprecipitate [35S]sulfate-labeled macromolecules that migrate on
SDS-PAGE gel in band II. This result was true for both the EA.hy 926 endothelial cell and HSF cultures. No cross-reactivity to band I CS/DS
PGs that were present only in the culture medium of HSFs and that have
previously been identified as decorin (64, 65) was observed (Fig. 2,
panel A, lane 3). These results confirmed that the CS/DS PGs
of band II synthesized by EA.hy 926 cells contain biglycan molecules.
As not all the CS/DS PGs of band II in the culture medium of EA.hy 926 cells could be immunoprecipitated with the antiserum LF-51 (data not
shown), the possibility remained that there are additional species of
small CS/DS PGs migrating in the position of band II. A few years ago,
evidence for the existence of a new member of the small extracellular
CS/DS PGs of this size range, tentatively named PG-100, has been
presented (7). Recently, this PG was shown to represent a GAG-linked
form of the macrophage colony stimulating factor CSF-1 (18). The result
from immunoprecipitation using an antiserum against PG-100 (66)
demonstrated that EA.hy 926 cells synthesize and secrete into the
medium PG-100 and that the Mr of these molecules
is similar to that of biglycan synthesized by the same cells (Fig. 2,
panel B, lane 1).
We also examined whether monolayer cultures of EA.hy 926 endothelial
cells contain mRNA for small extracellular CS/DS PGs, biglycan,
decorin, and PG-100. Northern blot analysis using a full-length
cDNA probe for human biglycan core protein demonstrated that EA.hy
926 cells express mRNA that hybridizes to a biglycan cDNA probe
(Fig. 3). Furthermore, mRNA for PG-100 was detected (Fig. 3). In contrast, mRNA that hybridizes to a cDNA probe for decorin core protein was not expressed in detectable amounts by monolayer cultures of EA.hy 926 cells. Human skin fibroblasts used as
control cells expressed abundant mRNA for the core proteins of all
three PGs (Fig. 3).
Previously, we have demonstrated that monolayer cultures of BAE cells
that do not express detectable amounts of decorin also fail to express
type I collagen (16). To examine whether EA.hy 926 endothelial cells
are similar in this respect, expression of type I collagen by these
cells was examined. Northern blot analysis of total cellular RNA from
monolayer cultures of EA.hy 926 cells demonstrated that these cells do
not express detectable amounts of mRNA for The functions of the small extracellular CS/DS
PGs are still somewhat controversial. Evidence exists that these PGs,
especially biglycan and decorin, interfere with the cell adhesion
process (34, 67, 68). The small extracellular CS/DS PGs have also been
shown to be involved in an in vitro phenomenon of
angiogenesis (39). Therefore, we examined whether potentially
angiogenically active cytokines modulate both the phenotype of EA.hy
926 cells and the synthesis of the small extracellular CS/DS PGs of
these cells. Several cytokines, including TNF-
Proteoglycan synthesis of EA.hy 926 cells was concomitantly affected.
The net incorporation of [35S]sulfate into PGs secreted
into the culture medium by the cells was slightly (31-34%), but
significantly increased in response to TNF-
Table I.
Effect of TNF-
Earlier a permanent human endothelial cell line EA.hy 926, that is
capable of expressing von Willebrand factor, has been established (40).
Besides von Willebrand factor, this cell line has been shown to produce
several other molecules characteristic of endothelial cells, such as
thrombomodulin (41, 71), prostacyclin (43), tissue-type plasminogen
activator (44), type I plasminogen activator inhibitor (44),
endothelin-1 (45, 46), endothelin converting enzyme (72), and vascular
adhesion protein-1 (73). EA.hy 926 cells have also been shown to
release an endothelial specific compound, platelet activating factor
(42). However, the synthesis of the extracellular matrix molecules,
including PGs, of EA.hy 926 cells has remained largely unidentified. In
the present study we have examined PG expression of EA.hy 926 cells. In
particular, we have focused on the expression of the small
extracellular CS/DS PG species, except the cartilage specific epiphycan
(4, 8), by these cells. As the small extracellular CS/DS PGs have been suggested to be involved in the regulation of endothelial cell behavior
(35, 39, 67, 74), we have also examined whether modulation of the
phenotype of EA.hy 926 cells is associated with specific changes in the
synthesis of the small CS/DS PGs. We have demonstrated that EA.hy 926 cells, when forming monolayer cultures typical of macrovascular
endothelial cells, express and synthesize detectable amounts of two
small extracellular CS/DS PG species, namely biglycan and PG-100. In
contrast, EA.hy 926 cells failed to express detectable amounts of
decorin. The exact amounts of biglycan and PG-100 synthesized by EA.hy
926 cells were not determined. However, according to our experience
(see also Figs. 8 and 9) EA.hy 926 cells normally synthesize more
biglycan than PG-100 whereas the opposite is true when these cells are
exposed to TNF- The importance of decorin in the fibrillogenesis of type I collagen
(23, 25) and in the regulation of the activity of TGF- In the present study we have shown that EA.hy 926 cells do not express
decorin or type I collagen even after having changed their phenotype
from a polygonal shape into a spindle shape. Previously Iruela-Arispe
et al. (38) and Järveläinen et al.
(39) have demonstrated that BAE cells initiate the synthesis of both
type I collagen and decorin during their morphological transition from a polygonal monolayer to a sprouting phenotype (an in vitro
model for angiogenesis). Although EA.hy 926 cells of this study changed their phenotype, the cells still grew in a monolayer culture and formed
no cords or tube-like structures as did the BAE cells in the studies
cited above. Therefore, it is likely that decorin and type I collagen
are needed in later stages of angiogenesis displayed by endothelial
cells in vitro. Indeed, recently we have been able to
demonstrate that EA.hy 926 cells, when they are grown on floating
collagen type lattices together with rat fibroblasts, start to
build up cords and after 6 days in culture human decorin can be
detected by immunostaining.3
In summary, this study has shown that EA.hy 926 cells express and
synthesize a set of small extracellular CS/DS PGs, characteristic of
macrovascular endothelial cells of various origin. This study has also
confirmed that different regulatory mechanisms are involved in the gene
expression of the small extracellular CS/DS PGs. Furthermore, additional indirect evidence has been presented suggesting that biglycan plays a role in the cell adhesion process. Thus, EA.hy 926 cells represent an excellent in vitro model system to be
used in studies on various aspects of PG expression in endothelial cells in man.
We thank Terttu Jompero, Taina Kalevo, and
Liisa Peltonen for excellent technical assistance. We are also grateful
to Dr. Larry Fisher for providing us with the antiserum LF-51 against biglycan and the biglycan cDNA probe, and Drs. Tom Krusius, Renato Iozzo, and Eero Vuorio for the cDNA probes for decorin, perlecan, and pro-
Medical Biochemistry and
** Internal Medicine,
Department of Anatomy, University of Kuopio, FIN-70211
Kuopio, Finland
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
(TNF-
) is associated with
specific changes in the synthesis of these PGs. We demonstrate that
EA.hy 926 cells, when they form monolayer cultures typical of
macrovascular endothelial cells, express and synthesize detectable amounts of biglycan and PG-100, but not decorin. On SDS-polyacrylamide gel electrophoresis both PGs behave like proteins of the relative molecular weight of ~250,000. TNF-
that changed the morphology of
the cells from a polygonal shape into a spindle shape and that also
stimulated the detachment of the cells from culture dish, markedly
decreased the net synthesis of biglycan, whereas the net synthesis of
PG-100 was increased. These changes were parallel with those observed
at the mRNA level of the corresponding PGs. The proportions of the
different sulfated CS/DS disaccharide units of PGs were not affected by
TNF-
. Several other growth factors/cytokines, such as
interferon-
, fibroblast growth factors-2 (FGF-2) and -7 (FGF-7),
interleukin-1
, and transforming growth factor-
, unlike TNF-
,
modulated neither the morphology nor the biglycan expression of EA.hy
926 cells under the conditions used in the experiments. However, PG-100
expression was increased also in response to FGF-2 and -7 and
transforming growth factor-
. None of the above cytokines, including
TNF-
, was able to induce decorin expression in the cells. Our
results indicate that the regulatory elements controlling the
expression of the small extracellular CS/DS PGs in human endothelial
cells are different.
(TGF-
) and to be
involved in cell proliferation and differentiation (27-29). The role
of biglycan in these processes is obscure, although it has been shown
to bind to these same molecules (30-32). The pericellular localization
of biglycan (33) as well as its ability to bind to fibronectin (34)
suggest that biglycan is likely to interfere with cell adhesion (34).
There is also recent evidence to suggest that biglycan is associated
with cell migration (35). CSF-1 is known to control the growth and
differentiation of mononuclear phagocytes (36) as well as the
growth of several other cell types, including endothelial cells (37).
However, PG-100 exhibits less than 1% of the biological activity of
mature CSF-1 (18) and should perhaps be considered an extracellular
storage form of this cytokine (19).
on biglycan expression and
its stimulatory effect on PG-100 expression by EA.hy 926 cells, associated with the altered morphology and adhesion of the cells, are
shown.
Cell Culture
on the synthesis of
the small extracellular CS/DS PGs, the cells were grown to confluence
in the serum containing culture medium. Next, the cells were kept in
the serum-free culture medium for 4 h, whereafter the cells were
labeled as described above in serum-free culture medium containing
TNF-
at various concentrations. After removing the detached cells
from the culture media by low speed centrifugation (see above), the
amount of [35S]sulfate-labeled PGs in the samples was
determined by the CPC precipitation assay (48) before further
analyses.
treated cultures was measured using the trypan
blue exclusion test as described (49).
stimulated EA.hy 926 cell
cultures were digested with 0.1 IU/ml of chondroitin ABC lyase in a
buffer containing 200 mM Tris, 60 mM sodium
acetate, pH 8.0, and 5 µg of chondroitin sulfate A (Seikagaku Co.) as
a carrier. The digestions were performed for 6 h at 37 °C. After ethanol precipitation overnight at 4 °C the supernatants and
the wash solutions (70% ethanol) were combined, ethanol was removed by
evaporation, and the sulfated disaccharides released by chondroitin ABC
lyase treatment were separated by ion exchange chromatography (55). The
average length of the GAG chains was estimated using a Sephacryl S-300
(Pharmacia) gel filtration (56, 57).
-[
-32P]dCTP (Amersham) to a high
specific activity. For the quantitation of the hybridization signals,
the membranes were probed with a rat glyceraldehyde-3-phosphate
dehydrogenase cDNA (59).
1(I) collagen. These cDNA probes were kindly provided by
Drs. Fisher (National Institute of Dental Research), Kresse (University
of Münster, Münster, Germany), Krusius (The Finnish Red
Cross, Helsinki, Finland), Iozzo (Thomas Jefferson University,
Philadelphia, PA), and Vuorio (University of Turku, Turku, Finland),
respectively.
and
fibroblast growth factor-7 (FGF-7), purchased from PeproTech;
fibroblast growth factor-2 (FGF-2), purchased from PeproTech or
Boehringer Mannheim; interleukin-1
(IL-1
), interferon-
(IFN-
), and TGF-
, all purchased from Boehringer Mannheim. TNF-
was used at concentrations of 5, 25, and 50 ng/ml. FGF-7, FGF-2,
IL-1
, IFN-
, and TGF-
were used at concentrations of 50 ng/ml,
10 ng/ml, 5 IU/ml, 1000 IU/ml, and 20 ng/ml, respectively.
Identification of Small Extracellular Chondroitin/Dermatan Sulfate
Proteoglycans in Monolayer Cultures of EA.hy 926 Human Endothelial
Cells
Fig. 1.
Comparison of size classes of newly
synthesized, [35S]sulfate-labeled macromolecules in the
culture medium of EA.hy 926 human endothelial cells and human skin
fibroblasts. [35S]Sulfate-labeled medium samples
from a monolayer culture of EA.hy 926 cells and HSF control culture
were run under reducing conditions on a 4-12% gradient SDS-PAGE with
a 3% stacking gel before and after chondroitin ABC lyase and/or
heparitinase digestions. The gel was treated with 2,5-diphenyloxazole
and dried, and the radioactive bands were visualized by fluorography.
Lanes 1-4 represent [35S]sulfate-labeled
medium samples from EA.hy 926 cell culture before GAG degrading enzyme
treatments (lane 1), and after digestions with chondroitin
ABC lyase (lane 2), heparitinase (lane 3), and chondroitin ABC lyase and heparitinase (lane 4). Lanes
5-8 represent [35S]sulfate-labeled medium samples
from HSF culture before GAG degrading enzyme treatments (lane
5), and after digestions with chondroitin ABC lyase (lane
6), heparitinase (lane 7), and chondroitin ABC lyase
and heparitinase (lane 8). The arrow indicates
the upper limit of the separating gel. Roman numerals I-IV
indicate the different size classes of
[35S]sulfate-labeled macromolecules.
[View Larger Version of this Image (42K GIF file)]
Fig. 2.
Immunoprecipitation of biglycan and PG-100
from [35S]sulfate-labeled culture medium of EA.hy 926 cell monolayers. Antisera against biglycan (panel A)
and PG-100 (panel B) were used for the immunoprecipitation
of [35S]sulfate-labeled medium samples from a monolayer
culture of EA.hy 926 endothelial cells. Immunoprecipitations of
identical medium samples with normal rabbit serum were used as negative
controls. Aliquots corresponding to 500,000 cpm as determined by CPC
precipitation assay were used for immunoprecipitations. SDS-PAGE
analyses identical to that described in the legend for Fig. 1 were
performed. The arrows in panels A and
B indicate the upper limit of the separating gel.
Lanes 1 and 2 in panel A represent
[35S]sulfate-labeled medium samples from EA.hy 926 cell
culture after immunoprecipitation with the biglycan antiserum LF-51
(lane 1) or NRS (lane 2), and lanes 3 and 4 represent [35S]sulfate-labeled medium
samples from HSF culture after immunoprecipitation with the biglycan
antiserum LF-51 (lane 3) or NRS (lane 4).
Lanes 1 and 2 in panel B represent
[35S]sulfate-labeled medium samples from EA.hy 926 cell
culture after immunoprecipitation with the antiserum against PG-100
(lane 1) or NRS (lane 2).
[35S]Sulfate-labeled macromolecules representing biglycan
and PG-100 are indicated by the asterisk (*).
[View Larger Version of this Image (45K GIF file)]
Fig. 3.
The expression of mRNAs for the core
proteins of the small extracellular CS/DS PGs and 1(I) collagen by
EA.hy 926 cell monolayer and HSF control cultures. Northern blot
analysis of total cellular RNA (10 µg/lane) from a monolayer culture
of EA.hy 926 cells and from a HSF culture was performed. The blot was
probed with 32P-labeled cDNAs for biglycan, decorin,
and PG-100 core proteins, and for pro-
1(I) collagen and
glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as
indicated. The average sizes of the mRNAs for the small CS/DS PGs
and pro-
1(I) collagen are shown. kb, kilobase.
[View Larger Version of this Image (20K GIF file)]
1(I) collagen (Fig.
3). The opposite was true for HSFs (Fig. 3).
, IFN-
, FGF-2 and
-7, IL-1
, and TGF-
, were tested. Of these cytokines TNF-
was
found to be the only one that was able to modulate the phenotype of EA.hy 926 cells under the conditions used in the experiments. Specifically, this cytokine changed the morphology of EA.hy 926 cells
from a polygonal shape into a spindle shape (Fig. 4).
Furthermore, TNF-
stimulated the detachment of the cells from the
culture dish which was indicated by an increase in the number of cells floating free in the culture medium (Fig. 5). As TNF-
is known to induce apoptosis and general cytotoxic effects on a variety of cells including endothelial cells (69, 70), we performed the trypan
blue exclusion test to examine whether this was true also in the
present study. The results demonstrated that almost all cells that
remained attached to the culture dish were negative for trypan blue
staining. This was evident for the control (96% of the cells trypan
blue negative) as well as for the TNF-
(50 ng/ml) treated (98% of
the cells trypan blue negative) cultures. In contrast, most of the
cells floating free in the medium were positive for trypan blue
staining, the proportion of positively stained cells being greater in
the TNF-
(50 ng/ml) treated (80% of the cells trypan blue positive)
than in the control cultures (63% of the cells trypan blue positive).
These results collectively indicate that, although TNF-
induced some cytotoxicity on EA.hy 926 cells in this study, the changes
in cell morphology and the increase in the detachment of the cells
could be only partially accounted for the cytotoxic effects of
TNF-
.
Fig. 4.
The effect of TNF- on the shape of EA.hy
926 cells in culture. The cells were grown to confluence in the
culture medium containing 10% (v/v) FCS. Next, the cells were kept in
the serum-free culture medium for 4 h, whereafter TNF-
at
concentrations of 5, 25, and 50 ng/ml was added on the cells. The cells
were grown for additional 24 h, after which phase-contrast
micrographs were taken to visualize the morphology of the cells.
Panel A represents the cells of the control culture (TNF-
not added), whereas panel B represents the cells exposed to
50 ng/ml TNF-
. A similar morphological change of the cells as shown
in panel B was evident in the cultures exposed to 5 and 25 ng/ml TNF-
(data not shown).
[View Larger Version of this Image (137K GIF file)]
Fig. 5.
The effect of TNF- on the number of EA.hy
926 cells on plastic and in the culture medium. The cells were
grown and exposed to TNF-
as described in the legend for Fig. 4.
After an incubation time of 24 h, the number of cells on plastic
and that of cells floating free in the culture medium were calculated using a hemocytometer. Each symbol represents the mean ± S.D. of
duplicate cell number calculations of three separate cultures. The
significance of the differences between the control and TNF-
treated
cultures was tested using Student's unpaired t test (*, p < 0.05; **, p < 0.01).
[View Larger Version of this Image (21K GIF file)]
stimulation (Fig.
6). The increased [35S]sulfate
incorporation observed at all TNF-
concentrations (5, 25, and 50 ng/ml) seemed to be most prominent in band II CS/DS PGs (Fig.
7). As the degree of sulfation of the CS/DS disaccharide isomers (Table I) or the average GAG chain length (data
not shown) were not altered in TNF-
treated cultures, we examined
whether the increase in [35S]sulfate incorporation in
response to TNF-
was due to an increase in the net synthesis of
biglycan or PG-100, or both of them. Western blot analyses of
chondroitin ABC lyase digested medium samples from TNF-
-treated
EA.hy 926 cell cultures using antisera against biglycan and PG-100
demonstrated that the net synthesis of PG-100 markedly increases in
response to TNF-
, while the net synthesis of biglycan decreases
(Fig. 8). Northern blot analyses of total cellular RNA
from these cultures using cDNA probes for biglycan and PG-100 core
proteins demonstrated that the changes observed at the product
level were parallel with those observed at the mRNA level of the
corresponding PGs (Fig. 9). The other cytokines used in
this study neither modulated the phenotype of EA.hy 926 cells nor
influenced significantly biglycan expression of the cells under the
conditions used in the experiments (Fig. 10). In contrast, the expression of PG-100 by EA.hy 926 cells was found to be
increased in response to FGF-2 and -7 and TGF-
stimulation of the
cells, but not in response to IFN-
or IL-1
stimulation (Fig. 10).
Furthermore, as in control cultures of EA.hy 926 cells, there were no
detectable amounts of mRNA for decorin core protein or
1(I)
collagen in any of the cytokine-stimulated cultures (data not
shown).
Fig. 6.
The effect of TNF- on the net
incorporation of [35S]sulfate into newly synthesized PGs
secreted into the culture medium by EA.hy 926 cells. The cells
were grown to confluence in the culture medium containing 10% (v/v)
FCS, whereafter the cells were kept in the serum-free culture medium
for 4 h. Next, the serum- and sulfate-free culture medium
containing 50 µCi/ml [35S]sulfate and TNF-
at
various concentrations (0, 5, 25, and 50 ng/ml) were changed. The cells
were grown for an additional 24 h, after which the media were
collected. The amount of [35S]sulfate-labeled PGs in the
samples was determined by the CPC precipitation assay. Each
symbol represents the mean ± S.D. of duplicate
determinations of three separate cultures. The significance of the
differences between the control and TNF-
treated cultures was tested
using Student's unpaired t test (*, p < 0.05; **, p < 0.01).
[View Larger Version of this Image (15K GIF file)]
Fig. 7.
The effect of TNF- on the net
incorporation of [35S]sulfate into various size classes
of newly synthesized PGs secreted into the culture medium by EA.hy 926 cells. EA.hy 926 cells were cultured, treated with TNF-
, and
labeled with [35S]sulfate as described in the legend for
Fig. 6. Equal volumes of the medium from each culture condition were
taken and analyzed by SDS-PAGE before and after chondroitin ABC lyase
and heparitinase digestions as described in the legend for Fig. 1.
Panel A represents medium samples that were not digested
with GAG degrading enzymes, and panels B-D represent medium
samples after digestions with chondroitin ABC lyase (panel
B), heparitinase (panel C), or chondroitin ABC lyase
and heparitinase (panel D).
[View Larger Version of this Image (37K GIF file)]
treatment on the molar proportions of sulfated
chondroitin/dermatan sulfate disaccharide isomers in the secreted proteoglycans, analyzed by CarboPac PA1 HPLC column after chondroitin ABC lyase digestion
TNF-
Disaccharide isomer
di-4S
di-6S
Disulfated
Trisulfated
ng/ml
0
11.2
2.8
85.7
0.3
5
16.7
3.6
79.1
0.6
25
15.7
4.4
79.2
0.7
50
20.9
5.1
73.3
0.7
Fig. 8.
Western blotting for the biglycan and PG-100
core proteins in the culture media of TNF- stimulated EA.hy 926 cells and their non-stimulated counterparts. The cells were grown
and stimulated with TNF-
as described in the legend for Fig. 6. An equal volume of the medium from each culture condition was run through
a DEAE-Sephacel column to partially purify the PGs secreted into the
medium by the cells. Next, the samples were dialyzed against tap water,
lyophilized, and dissolved in 100 µl of double distilled water. Equal
aliquots were taken for digestion with chondroitin ABC lyase and
heparitinase, whereafter the samples were run on a 12.5% SDS-PAGE. The
proteins were transferred to nitrocellulose filter and the blots were
incubated in the presence of antisera against biglycan (panel
A) and PG-100 (panel B) to determine the amounts of the
core proteins in different samples. The lanes in panels A
and B contain medium samples from EA.hy 926 endothelial cell
control culture (0) and from cultures of EA.hy 926 cells stimulated
with 5, 25, and 50 ng/ml TNF-
as indicated.
[View Larger Version of this Image (46K GIF file)]
Fig. 9.
The effect of TNF- on the mRNA level
of biglycan and PG-100 in EA.hy 926 cell monolayers. The cells
were grown as described in the legend for Fig. 6. Total cellular RNA
was isolated and 10 µg of RNA from each culture condition was taken
for Northern blot analysis. The blot was probed with
32P-labeled cDNAs for the biglycan and PG-100 core
proteins, and for glyceraldehyde-3-phosphate dehydrogenase
(GAPDH). The average sizes of the biglycan and PG-100 core
protein mRNAs are shown.
[View Larger Version of this Image (29K GIF file)]
Fig. 10.
The effect of various cytokines on the
mRNA level of biglycan and PG-100 in EA.hy 926 cell
monolayers. The cells were grown to confluence in the culture
medium containing 10% (v/v) FCS, whereafter the cells were exposed for
24 h to 1000 IU/ml of IFN-, 10 ng/ml of FGF-2, 5 IU/ml of
IL-1
, 20 ng/ml of TGF-
, and 50 ng/ml of FGF-7 under the
serum-free culture medium. Total cellular RNA was isolated and 10 µg
of total cellular RNA from each culture condition was taken for
Northern blot analysis. The blot was probed with
32P-labeled cDNAs for the biglycan and PG-100 core
proteins, and for glyceraldehyde-3-phosphate dehydrogenase
(GAPDH). The average sizes of the biglycan and PG-100 core
protein mRNAs are shown. 1, control; 2,
IFN-
; 3, FGF-2; 4, IL-1
; 5,
TGF-
; 6, FGF-7.
[View Larger Version of this Image (21K GIF file)]
. The apparent discrepancy between this and the
result shown in Fig. 2 (see panel A, lane 1, and panel
B, lane 1) is most probably due to the different precipitation
capability of the two antisera used and the different exposure times of
the x-ray films. In this study we have also demonstrated that the same
EA.hy 926 endothelial cell monolayers that do not express decorin, fail
to express type I collagen. These results are in a full agreement with
those of our previous studies demonstrating that monolayer cultures of bovine aortic and human umbilical vein endothelial cells express and
synthesize biglycan, but not decorin or type I collagen (16). BAE cell
monolayers are also capable of producing
PG-100.2 Thus, EA.hy 926 cells seem to
express and synthesize a set of small extracellular CS/DS PGs typical
of monolayer cultures of macrovascular endothelial cells. As the
characteristic properties of human endothelial cells tend to change
during culture and as the life span of human endothelial cells is
highly limited, earlier studies concerning, e.g. synthesis
of PGs by endothelial cells, have mainly been performed using cells of
animal origin. This study has proved that EA.hy 926 cells represent an
excellent in vitro human model system to be used for studies
examining the factors and the regulatory mechanisms underlying the
differential expression of the small extracellular CS/DS PGs. These
cells are also very useful when the importance of the small
extracellular CS/DS PGs in the control of endothelial cell behavior in
man will be examined.
(28) has been
demonstrated. In contrast, the functions of biglycan and PG-100 are
not exactly known. Indirect evidence exists suggesting that biglycan is
involved in the regulation of cell adhesion. This evidence is based,
e.g. on the notion that biglycan has a pericellular
localization (33), and on the fact that biglycan, like decorin (75,
76), interferes with the molecules such as fibronectin (34), important
in the cell adhesion. In this study, we have shown that the
morphological change of EA.hy 926 cells from a polygonal shape into a
spindle shape and the increase in the number of detached cells in
response to TNF-
treatment are associated with a significant
decrease in the net synthesis of biglycan by the cells. We have also
shown that several other cytokines, such as IFN-
, FGF-2 and -7, IL-1
, and TGF-
, that did not change the phenotype of EA.hy 926 cells did also not influence biglycan expression of these cells under
the conditions used in the experiments. Since the above change in the
cell morphology and the increase of the cell detachment could be only
partially accounted for the cytotoxic effects of TNF-
(see
"Results"), the results of this study support the finding by
Bidanset et al. (34) that biglycan plays a role in cell
adhesion. We are currently trying to overexpress biglycan in EA.hy 926 cells to examine whether the above changes in the morphology and
adhesion of EA.hy 926 cells in response to TNF-
can be prevented. In
regard with PG-100, the net synthesis of this part-time PG was
increased in EA.hy 926 cells in response to TNF-
. The expression of
PG-100 was also found to be increased in response to FGF-2 and -7 and
TGF-
. Thus, unlike the expression of biglycan, PG-100 expression did
not show a strict correlation with the morphological change of EA.hy
926 cells. This indicates that the functional role of PG-100 for
endothelial cells is different from that of biglycan. It can also be
concluded that the regulatory elements controlling the expression of
PG-100 and biglycan differ from each other. As regards biglycan, in a recent study with T47D cells it was shown that TNF-
is capable of
altering biglycan promoter activity (77). The finding that the
expression of PG-100 is up-regulated in response to TNF-
excludes
the possibility that the effect of TNF-
on PG synthesis by
endothelial cells and other cell types is merely inhibitory as
demonstrated in previous studies (78-82). However, the obvious discrepancy in this study that the net incorporation of
[35S]sulfate into all extracellular PGs was increased
only by 31-34% in response to TNF-
(Fig. 6), while the net
incorporation of [35S]sulfate into small extracellular
CS/DS PGs was increased by severalfold (Fig. 7) suggests that the
synthesis of most extracellular PG species has to decrease in response
to TNF-
. Therefore, PG-100 may be one of the few PGs, the synthesis
of which is increased in response to TNF-
(Fig. 8).
*
This study was supported by the Turku University Foundation,
Finnish Foundation for Cardiovascular Research, The Finnish Cultural Foundation, The Academy of Finland, and The Orion-Farmos Research Foundation.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: Dept. of Medical
Biochemistry, University of Turku, Kiinamyllynkatu 10, FIN-20520 Turku,
Finland. Tel.: 358-2-3337444; Fax: 358-2-3337229; E-mail: hannu.jarvelainen{at}utu.fi.
1
The abbreviations used are: PG(s),
proteoglycan(s); BAE, bovine aortic endothelial; CPC, cetylpyridinium
chloride; CS, chondroitin sulfate; CSF-1, colony stimulating factor-1;
DS, dermatan sulfate; FCS, fetal calf serum; FGF, fibroblast growth
factor; GAG, glycosaminoglycan; HS, heparan sulfate; HSF, human skin
fibroblast; IFN-, interferon-
; IL-1
, interleukin-1
; PAGE,
polyacrylamide gel electrophoresis; TNF-
, tumor necrosis factor-
;
TGF-
, transforming growth factor-
.
2
E. Schönherr, unpublished result.
3
E. Schönherr, L. Nelimarkka, and H. Järveläinen, unpublished results.
1(I) collagen, respectively. The antibody against PG-100 and
the PG-100 cDNA probe were kindly provided by Dr. Hans Kresse. We
thank Dr. Thomas Hering (Case Western Reserve University, Cleveland, OH) for providing valuable comments of the manuscript.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.