Research Article
2029
Anti-membrane-bound transferrin-like protein
antibodies induce cell-shape change and chondrocyte
differentiation in the presence or absence of
concanavalin A
Ryo Oda1,*, Ketut Suardita2,*, Katsumi Fujimoto2, Haiou Pan2, Weiqun Yan2, Atsushi Shimazu3,
Hideaki Shintani1 and Yukio Kato2,‡
1Department
2Department
3Department
of Operative Dentistry, Faculty of Dentistry, Hiroshima University, 1-2-3 Kasumi, Hiroshima 734-8553, Japan
of Biochemistry, Faculty of Dentistry, Hiroshima University, 1-2-3 Kasumi, Hiroshima 734-8553, Japan
of Preventive Dentistry, Faculty of Dentistry, Hiroshima University, 1-2-3 Kasumi, Hiroshima 734-8553, Japan
*These authors contributed equally to this work
‡Author for correspondence (e-mail: [email protected])
Accepted 27 January 2003
Journal of Cell Science 116, 2029-2038 © 2003 The Company of Biologists Ltd
doi:10.1242/jcs.00393
Summary
Membrane-bound transferrin-like protein (MTf), a
glycosylphosphatidylinositol-anchored
protein,
is
expressed at high levels in many tumors and in several fetal
and adult tissues including cartilage and the intestine, as
well as in the amyloid plaques of Alzheimer’s disease,
although its role remains unknown. MTf is one of the major
concanavalin A-binding proteins of the cell surface. In this
study, we examined the effects of anti-MTf antibodies and
concanavalin A on cell shape and gene expression, using
cultures of chondrocytes and MTf-overexpressing ATDC5
and C3H10T1/2 cells. In cultures expressing MTf at high
levels, concanavalin A induced cell-shape changes from
fibroblastic to spherical cells, whereas no cell-shape
changes were observed with wild-type ATDC5 or
Introduction
Because of their powerful biological actions in animal cells,
plant lectins and their derivatives are promising for various
aspects of biotechnology (Lis and Sharon, 1986). The
elucidation of lectin-glycoprotein interactions on the cell
surface will be useful for drug design and studies on the roles
of glycoproteins, because the tertiary structure of several plant
lectins and the structure of lectin-binding polysaccharides
have previously been determined. There has been a vast
accumulation of data on the actions of concanavalin A (ConA)
and other plant lectins in animal cells, whereas the functional
receptors for plant lectins have not been identified. Because
many plant lectins, with different sugar-binding properties,
induce cell-shape changes and proliferation and/or
differentiation in lymphocyte cultures, we used chondrocytes
or chondrogenic cells that responded to ConA but not to other
lymphocyte-activating lectins (Yan et al., 1990) in order to
identify a real ConA receptor. In chondrocyte cultures,
ConA induced the conversion of fibroblastic cells to welldifferentiated spherical chondrocytes and enhanced their
synthesis of cartilage-matrix proteoglycan (aggrecan) within
24 hours (Yan et al., 1990; Yan et al., 1997). This effect of
C3H10T1/2 cells expressing MTf at very low levels. The
cell-shape changes were associated with enhanced
proteoglycan synthesis and expression of cartilagecharacteristic genes, including aggrecan and type II
collagen. Some anti-MTf antibodies mimicked this action
of concanavalin A, whereas other antibodies blocked the
lectin action. The findings suggest that the crosslinking of
MTf changes the cell shape and induces chondrogenic
differentiation. MTf represents the first identification of a
plant lectin receptor involved in cell-shape changes and the
differentiation of animal cells.
Key words: Chondrocyte, Concanavalin A, GPI-anchored protein,
Membrane-bound transferrin-like protein
ConA was much greater than that of known growth factors
(Yan et al., 1990; Yan et al., 1997). However, chondrocytes
exposed to retinoic acid lost their responsiveness to ConA,
suggesting that retinoic acid decreases the expression of ConA
receptors on the cell surface (Yan et al., 1990). The ConAinduced chondrocyte differentiation can be observed in vivo:
injection of ConA induced ectopic cartilage in the
perichondrorium of mice (Wlodarski and Galus, 1992).
MTf was originally identified with the use of monoclonal
antibodies as a 97 kDa human tumor-associated antigen
(Woodbury et al., 1980; Dippold et al., 1980). Its amino acid
sequence shows ~40% identity with transferrin and lactoferrin
(Rose et al., 1986). It binds to iron (Brown et al., 1982)
and stimulates iron uptake in the absence of transferrin and
transferrin receptor (Kennard et al., 1995). Interestingly, MTf
can cross the blood-brain barrier (Demeule et al., 2002) and it
accumulates in amyloid plaques of Alzheimer’s disease, where
iron is also concentrated (Jefferies et al., 1996). Furthermore,
it has been suggested that MTf is involved in the proliferation
and differentiation of melanoma cells and eosinophils (Estin
et al., 1989; McNagny et al., 1996). However, the precise
physiological roles of MTf remain unknown.
2030
Journal of Cell Science 116 (10)
Previous studies have shown that MTf is a major ConAbinding protein on the chondrocyte surface (Kawamoto et al.,
1998). Because cartilage contains MTf at a much higher level
than other normal tissues, chondrogenic cells are a good model
for studies on MTf. In the present study, we examined the effect
of anti-MTf antibodies on cell shape and the expression of
differentiation-related genes in cultures of chondrocytes or
chondrogenic cells (ATDC5 and C3H10T1/2 cells) in the
presence or absence of ConA. The anti-MTf antibodies
markedly suppressed the effect of ConA on the cell shape and
the phenotypic expression in these cultures, or they mimicked
the action of ConA only when the cells synthesized MTf at
high levels. These effects of ConA and the anti-MTf antibodies
on MTf-expressing cells were observed within 24-48 hours.
The findings on the ConA-MTf system obtained in this study
will be useful in the understanding of the remarkable actions
of plant lectins on animal cells.
Materials and Methods
Preparation of antibodies
MTf was purified from rabbit chondrocyte plasma membrane
(Kawamoto et al., 1998). Three female BALB/c mice were immunized
by a subcutaneous injection of Ribi adjuvant solution (0.2 ml per
mouse) containing 20 µg pure MTf. The mice received two
subcutaneous injections of the same solution (0.2 ml per mouse) 21
and 35 days after the first immunization. 3 days after the last injection,
one mouse showed a high titer of anti-MTf antibodies in a serum
sample. This serum (anti-MTf-pAb1) was taken and the spleen of the
mouse was used for cell fusion with P3-X63-Ag8-U1 BALB/c
myeloma cells. A hybridoma cell clone producing a monoclonal
antibody to MTf (anti-MTf-mAb2) was obtained by limiting dilution.
In other studies, three female BALB/c mice were immunized by four
subcutaneous injections of the antigen (10 µg MTf in 0.15 ml of Ribi
adjuvant solution per injection) on days 0, 14, 28 and 35. Sera (antiMTf-pAb2, -pAb3, and -pAb4) were taken 7 days after the last
injection.
Immunoblotting
The proteins extracted from the cultured chondrocytes (10 µg protein
per lane) were resolved by 4-20% SDS-PAGE under nonreducing
conditions. After blotting onto a PVDF membrane (Towbin et al.,
1979) and blocking with 4% nonfat milk in PBS for 2 hours at room
temperature, the membrane was incubated at 4°C with anti-MTf sera
(1:500 dilution in PBS) or anti-MTf-mAb2 (5 µg ml–1) for 14 hours
and then incubated with 125I-labeled sheep anti-mouse IgG (Fab′)2
fragment in PBS for 2 hours at room temperature.
Chondrocyte culture
The chondrocytes were isolated from the resting cartilage of the ribs
of 4-week-old male Japanese White rabbits as described previously
(Shimomura et al., 1975; Kato and Gospodarowicz, 1985). The cells
were seeded at a density of 106 cells per 150-mm tissue culture dish
and maintained in 30 ml α-modified Eagle’s medium supplemented
with 10% fetal bovine serum, 50 µg ml–1 ascorbic acid, 50 U ml–1
penicillin, 60 µg ml–1 kanamycin and 250 ng ml–1 amphotericin B
(medium A). When the cultures became subconfluent, the cells were
harvested with trypsin and EDTA, and seeded at 5×104 cells per 16mm well in 0.5 ml medium A. When the cultures again became
subconfluent, the cells were preincubated in 0.5 ml of a 1:1 mixture
of Dulbecco’s modified Eagle’s medium and Ham’s F-12 medium
(Nissui Pharmaceutical, Tokyo, Japan) supplemented with 0.5% fetal
bovine serum (medium B) for 24 hours. The medium was replaced by
Fig. 1. SDS-PAGE and immunoblotting of the chondrocyte
membrane treated with various concentrations of retinoic acid, and
the effect of retinoic acid pretreatment on MTf expression by
chondrocyte cultures. (A) Chondrocytes were exposed or not to 10–6
M retinoic acid 4 days before the end of incubation. The protein in
the crude membrane fraction (6 µg) was analysed by SDS-PAGE and
stained with silver. (B) Chondrocytes were exposed to retinoic acid at
0 M, 10–8 M, 10–7 M and 10–6 M 4 days before the end of
incubation. The protein in the ConA-bound membrane fraction
(2 µg) was analysed by SDS-PAGE and stained with silver.
(C) Chondrocytes were exposed to retinoic acid at 10–6 M 0 hours,
24 hours, 48 hours and 72 hours before the end of the incubation.
The proteins in the ConA-bound fraction (2 µg) were resolved by
SDS-PAGE and stained with silver. (D) Chondrocytes were exposed
to retinoic acid at 10–6 M 0 hours, 24 hours, 48 hours and 72 hours
before the end of the incubation. The MTf level in the chondrocyte
cultures was analyzed by immunoblotting. (E) Chondrocytes were
exposed, or not exposed, to retinoic acid at 10–6 M for 4 days (left)
and then incubated in the absence of retinoic acid for 3 days (right).
The MTf level in the chondrocyte cultures was analysed by
immunoblotting. (F) Chondrocytes in confluent cultures were
incubated with retinoic acid at 10–6 M for 4 days. The MTf and
GAPDH mRNA levels in the chondrocytes were determined by
northern blot analysis.
0.5 ml of fresh medium B in the absence or presence of anti-MTf
antibodies, ConA (5 µg ml–1) or both, and the incubation was
continued for 24 hours.
ConA affinity chromatography of plasma membrane from
retinoic-acid-exposed cultures
The cell membrane was isolated from retinoic-acid-exposed or -free
MTf is a ConA receptor
2031
Fig. 2. Effects of retinoic acid pretreatment on the shape
of cultured chondrocytes in the presence or absence of
ConA. Poorly differentiated chondrocytes were not
exposed to retinoic acid for 4 days and then incubated in
the absence (A) or presence (B) of 5 µg ml–1 ConA for 24
hours. Poorly differentiated chondrocytes were exposed
to retinoic acid at 10–6 M for 4 days and then incubated in
the absence (C) or presence (D) of 5 µg ml–1 ConA for 24
hours. Poorly differentiated chondrocytes were exposed
to retinoic acid at 10–6 for 4 days and then incubated in
the absence (E) or presence (F) of 5 µg ml–1 ConA for 72
hours. Bar, 30 µm.
chondrocyte cultures in three 150 mm dishes by the method of
Mollenhauer et al. (Mollenhauer et al., 1984) with some
modifications. The plasma membrane fraction was dissolved in 8 ml
of buffer A [10mM Tris/HCl, pH 7.4, 10 µM (p-amidinophenyl)
methanesulfonyl fluoride, 10 µM pepstatin A and 1% sodium
deoxycholic acid] and applied to a ConA-Sepharose column (1 cm ×
3 cm) (Amersham, Piscataway, NJ) equilibrated with buffer A. The
ConA-bound glycoprotein was eluted with buffer A containing 0.5 M
methyl-α-mannopyranoside as described previously (Kawamoto et
al., 1998).
Northern blotting
Total RNA was prepared from chondrocytes which had been treated
with 10–6 M retinoic acid for 4 days using the guanidine-thiocyanate
method (Smale and Sasse, 1992). RNA samples (10 µg) were
electrophoresed on a 1% agarose gel containing 2.2 M formaldehyde
and transferred to a Hybond-N membrane (Amersham). The
membrane was hybridized with a 32P-labeled rabbit MTf cDNA
probe or a 32P-labeled glyceraldehyde-3-phosphate dehydrogenase
(GAPDH) cDNA probe as described previously (Kawamoto et al.,
1998). The membrane was exposed to BioMax X-ray film at –80°C
with an intensifying screen.
Proteoglycan synthesis
The chondrocytes obtained from the rabbit primary cultures were
seeded at 104 cells per 6-mm microwell and maintained in medium
A. When the cultures became confluent, the cells were preincubated
in 0.1 ml of medium B for 24 hours. The cells were incubated with
ConA (5 µg ml–1), succinyl-ConA (sConA) (10 µg ml–1), dibutyryl
cyclic AMP (DBcAMP) (1 mM) or insulin (10 µg ml–1) in 0.1 ml of
fresh medium B in the absence or presence of anti-MTf antibodies for
24 hours. [35S]sulfate (5 µCi ml–1 final concentration) was added 6
hours before the end of the incubation (Kato et al., 1980). Uronic acid
was determined by the method of Bitter and Muir (Bitter and Muir,
1962).
DNA synthesis
The chondrocytes obtained from the rabbit primary cultures were
seeded at 104 cells per 6-mm microwell and maintained in medium
A. When the cultures became confluent, the cells were preincubated
in 0.1 ml of α-modified Eagle’s medium supplemented with
0.5% fetal bovine serum for 24 hours and then incubated with
ConA (5 µg ml–1) or sConA (10 µg ml–1) in 0.1 ml of α-modified
Eagle’s medium supplemented with 0.5% fetal bovine serum in the
absence or presence of anti-MTf-mAb2 or control mouse IgG for 24
hours. [3H]thymidine (10 µCi ml–1) was added 3 hours before the
end of the incubation (Kato et al., 1983). The lymphocytes were
isolated from the thymus of 4-week-old rabbits. The cells were
seeded at a density of 6×105 cells per 16-mm well and maintained
in 0.2 ml medium A for 72 hours. The cells were exposed to antiMTf-mAb2, control mouse IgG and/or ConA for 72 hours.
[3H]Thymidine (3 µCi ml–1) was added 6 hours before the end of
the incubation.
Plasmids, transfections and ATDC5 and C3H10T1/2 cells
Rabbit full-length MTf cDNA was inserted into the mammalian
expression vector pDNA3.1/Zeo (Invitrogen, Carlsbad, CA) at the
EcoRI-NotI site to construct pcDNA-MTf. pcDNA3.1/Zeo or pcDNAMTf was transfected into ATDC5 and C3H10T1/2 cells (RIKEN,
Tsukuba, Japan) using the SuperFect transfection reagent (Qiagen,
Valencia, CA). The transfected cells were selected with 0.2 mg ml–1
Zeocin (Invitrogen). Individual colonies were isolated, and the
expression levels of MTf were determined by immunoblotting with
anti-MTf-mAb2.
ATDC5 cells were maintained in a 1:1 mixture of Dulbecco’s
modified Eagle’s medium and Ham’s F-12 medium supplemented
with 5% fetal bovine serum, 10 µg ml–1 human transferrin
(Boehringer Mannheim, Mannheim, Germany) and 3×10–8 M sodium
selenite (Sigma, St Louis, MO) in the presence or absence of 10 µg
ml–1 bovine insulin (Sigma) as previously described (Atsumi et al.,
1990). The inoculum density of the cells was 4×104 cells per 16-mm
dish or 20×104 cells per 10 cm dish. The medium was replaced every
other day.
C3H10T1/2 cells were maintained in medium A. The inoculum
density of the cells was 8×104 cells per 35-mm dish or 20×104 cells
per 10-cm dish. The medium was replaced every other day. On day
8, the culture medium was changed to Dulbecco’s modified Eagle’s
medium and Ham’s F-12 medium (1:1) containing 1% fetal bovine
serum, insulin (6.25 µg ml–1), transferrin (6.25 µg ml–1), selenite (6.25
ng ml–1), ascorbic acid (50 µg ml–1), dexamethasone (10–10 M) and
2032
Journal of Cell Science 116 (10)
Fig. 3. Specificity of anti-MTf antibodies and
the appearance of chondrocytes not exposed
(none) or exposed to ConA or ConA plus
anti-MTf antibodies for 24 hours. (A) The
proteins in the chondrocyte cultures were
resolved by SDS-PAGE. MTf in the samples
was analysed by immunoblotting with pAb1
(lane 1), pAb2 (lane 2), pAb3 (lane 3), pAb4
(lane 4) or mAb2 (lane 5). (B) Poorly
differentiated chondrocytes were incubated
with 1% control serum in the absence
(control) or presence of 5 µg ml–1 ConA for
24 hours. Alternatively, these cells were
incubated with 1% pAb1-4 serum in the
presence of 5 µg ml–1 ConA for 24 hours
(ConA+pAb1-4). In the other studies, these
cells were incubated with control mouse IgG
(40 µg ml–1) or mAb2 (10-40 µg ml–1) in the
presence of 5 µg ml–1 ConA for 24 hours.
(C) Chondrocytes were incubated with the
F(ab′)2 fragment of mAb2 (40 µg ml–1) in the
presence of 5 µg ml–1 ConA for 24 hours.
Bar, 30 µm.
transforming growth factor β1 (TGF-β1, 5 ng ml–1). The medium was
replaced every other day.
Number of spherical/polygonal/spindle-like cells
Spherical/polygonal/spindle-like cells and spread cells were counted
separately under a phase-contrast microscope. At least 300 cells were
evaluated and the proportion of spherical/polygonal/spindle-like cells
among total cells was calculated.
RT-PCR and Southern blot analysis
The first-strand cDNA was synthesized with 1 µg ml–1 total RNA
from ATDC5 cells. PCR was performed using an aliquot of the firststrand cDNA as a template under standard conditions with Klentaq
Polymerase (Clonetech, Palo Alto, CA) for 18, 23 and 18 cycles for
aggrecan, type II collagen and GAPDH, respectively. These cycles
were optimal for comparison between the amplified products. The
PCR products were separated on 1.5% agarose gels and transferred to
NytranN membranes (Schleicher & Schuell, Dassel, Germany). The
membranes were hybridized with 32P-labeled mouse aggrecan, 32Plabeled mouse-type II collagen or 32P-labeled mouse GAPDH cDNA
(Nakamasu et al., 1999).
Results
Close relationship between MTf levels and
responsiveness of chondrocytes to ConA
To identify the retinoic-acid-sensitive membrane proteins that
were potential ConA receptors (Yan et al., 1990), we incubated
chondrocytes with retinoic acid at various concentrations for
1, 2 and 4 days, and examined retinoic-acid-induced changes
in the membrane fraction by SDS-PAGE. The membrane
fraction was resolved with over 40 bands on SDS-PAGE, but
retinoic acid at 10–6 M for 4 days had little effect on the
electrophoresis profile of the membrane proteins (Fig. 1A). The
proteins were purified by ConA-affinity chromatography.
About 5% of the membrane proteins were recovered in the
ConA-bound fraction. When this fraction was analyzed by
SDS-PAGE, the levels of 76 kDa (p76) and 140 kDa proteins
(p140) were markedly affected (Fig. 1B,C): retinoic acid
decreased the level of p76, whereas it increased the level of
p140 dose-dependently. The p76 level decreased 24-48 hours
after the retinoic acid was added (Fig. 1C and data not shown),
and scarcely any of the p76 remained in cultures exposed to
retinoic acid at 10–7 M or 10–6 M for 4 days (Fig. 1B). Because
MTf is a ConA receptor
2033
decreased after the addition of retinoic acid (Fig.
1F).
In monolayer cultures in a low-serum medium,
the chondrocytes took a fibroblastic configuration
(Fig. 2A) and the addition of ConA induced a cell
shape change from fibroblastic to spherical
chondrocytes within 24 hours (Fig. 2B) (Yan et
al., 1990). However, pretreatment with retinoic
acid for 4 days prevented MTf expression, as
indicated by immunoblotting (Fig. 1D), and the
responsiveness of the chondrocytes to ConA
(Fig. 2D), and removal of retinoic acid for 3
days restored the MTf level (Fig. 1E) and
responsiveness to ConA (Fig. 2F). These findings
indicate a close relationship between MTf level
and responsiveness to ConA.
Inhibition of ConA-induced chondrocyte
phenotypic expression by pAb1 and mAb2
and ConA-like actions of pAb4
If MTf is a receptor for ConA, antibodies to
MTf should modulate the lectin action on
chondrocytes. To test this hypothesis, we purified
MTf from rabbit chondrocyte plasma membrane
using HPLC and lectin-affinity chromatography
(Kawamoto et al., 1998). Using the purified MTf,
we prepared four polyclonal antisera (antiMTf-pAb1, -pAb2, -pAb3 and -pAb4) and a
monoclonal antibody (anti-MTf-mAb2) that had
proven to be specific for MTf in immunoblot
analysis with chondrocyte extracts (Fig. 3A).
Fig. 4. Effects of mAb2 on
Rabbit chondrocytes were incubated with antiproteoglycan synthesis and
MTf-pAb1 serum or the control serum in the
DNA synthesis in chondrocyte
absence or presence of ConA for 24 hours. The
cultures, and on DNA synthesis
in lymphocyte cultures in the
chondrocytes adopted a fibroblastic configuration
presence or absence of ConA or
in monolayer cultures at a low serum
sConA. (A-C) Poorly
concentration (Fig. 3B, control) (Yan et al., 1990).
differentiated chondrocytes
The addition of ConA to the cultures altered the
were incubated in the absence or
cell shape within 12 hours, with almost all of the
–1
presence of 5 µg ml ConA or
lectin-exposed cells becoming spherical at 24
–1
10 µg ml sConA with or
hours (Fig. 3B, ConA) (Yan et al., 1990). This
without mAb2 or control mouse
effect of ConA was eradicated by the anti-MTfIgG for 24 hours. Alternatively,
pAb1 serum (Fig. 3B, ConA+pAb1). Anti-MTfthese cells were exposed to
pAb2 and -pAb3 suppressed the action of ConA
DBcAMP or insulin in the
presence or absence of mAb2 at
on the cell shape to a lesser extent, whereas anti40 µg ml–1. (D) Lymphocytes
MTf-pAb4 had little effect on the cell-shape
were incubated in the presence
change (Fig. 3B, ConA+pAb2, ConA+pAb3 and
or absence of 3 µg ml–1 ConA with or without mAb2 or control IgG at 40 µg
ConA+pAb4). The anti-MTf-mAb2 suppressed
–1
ml for 72 hours. The values are averages ± s.d. for the four cultures. (A:
the ConA-induced cell-shape change doseConA/sConA versus ConA/sConA + mAb2; *P<0.01; **P<0.005;
dependently (Fig. 3B, ConA+mAb2), and the
***P<0.0001. B: *P<0.01; **P<0.001.)
F(ab′)2 fragment of mAb2 also suppressed the
cell-shape change (Fig. 3C, ConA+ F(ab′)2). The
p76 had been identified as rabbit MTf by N-terminal
anti-MTf-mAb2 was prepared from the mouse that produced
sequencing (Kawamoto et al., 1998), the changes in p76 levels
anti-MTf-pAb1.
were examined by immunoblotting with anti-MTf antibody:
The anti-MTf-mAb2 suppressed the ConA-stimulation of
Immunoblotting showed that incubation with 10–6 M retinoic
the incorporation of [35S]sulfate into glycosaminoglycans
acid for 24-72 hours markedly decreased the expression of
synthesized by chondrocytes (Fig. 4A, left panel). Under these
p76/MTf (Fig. 1D). However, p76/MTf expression was
conditions, the majority of 35S-labeled glycosaminoglycans
recovered 3 days after the removal of retinoic acid (Fig. 1E).
were associated with cartilage-characteristic proteoglycan
Northern blot analysis showed that the MTf mRNA level also
(aggrecan) (Yan et al., 1990).
2034
Journal of Cell Science 116 (10)
Fig. 5. Effects of anti-MTf-antibodies on
cell shape and proteoglycan synthesis by
chondrocytes in the absence of ConA.
(A) Poorly differentiated chondrocytes
were incubated with 1% control serum, 1%
pAb2-4, 100 µg ml–1 control mouse IgG or
100 µg ml–1 IgG purified from the pAb4
serum for 24 hours. (B) Poorly
differentiated chondrocytes were incubated
with the control serum or pAb4 serum at
concentrations of 0.05-1% for 24 hours.
(C) Poorly differentiated chondrocytes
were incubated with 1% control serum or
1% pAb4 for 7 days. Bar, 30 µm. The
values are averages ± s.d. for the four
cultures. *P<0.01.
Previous studies have shown that a divalent succinylated
derivative of ConA (sConA) enhances chondrocyte phenotypic
expression without inducing rapid changes in cell shape,
although native tetravalent ConA enhances both rapid cellshape change and chondrocyte phenotypic expression
(proteoglycan synthesis) (Yan et al., 1997). ConA- but not
sConA-induces extensive crosslinking of the cell surface
glycoproteins and patch/cap formation (clustering of lectinbinding membrane proteins) (Gunther et al., 1973). In our
study, the anti-MTf-mAb2 also suppressed the sConAstimulation of proteoglycan synthesis by the chondrocytes
(Fig. 4A, right panel).
Anti-MTf-mAb2 had little effect on the stimulation of
proteoglycan synthesis by either a permeable analogue of
cyclic AMP or insulin at the pharmacological level (Fig. 4B).
The anti-MTf-mAb2 did not interfere with the effect of ConA
on DNA synthesis in the chondrocytes (Fig. 4C). In addition,
the anti-MTf-mAb2 had little effect on the ConA stimulation
of lymphocyte proliferation (Fig. 4D), because MTf is barely
expressed in lymphocytes (Kawamoto et al., 1998; Nakamasu
et al., 1999).
Interestingly enough, the addition of the anti-MTf-pAb4
serum, as in the case of ConA, induced the expression of the
spherical phenotype (Fig. 5A, pAb4). The anti-MTf-pAb2 and
-pAb3 sera also induced cell-shape change, although their
effect was far less than that of the anti-MTf-pAb4 (Fig. 5A,
pAb2 and pAb3). IgG purified from the anti-MTf-pAb4
serum using a Protein-A-affinity gel column also elicited the
ConA-like action at 24 hours (Fig. 5A, pAb4 IgG), with the
ConA-like action of the anti-MTf-pAb4 being indicated
by the increase in incorporation of [35S]sulfate into
glycosaminoglycans (Fig. 5B). In cultures exposed to the anti-
MTf-pAb4 for 7 days, all of the cells were eventually
surrounded by an abundant refractile matrix (Fig. 5C). All of
the mice that were immunized with MTf – but none of the
control mice – produced neutralizing and/or mimicking
antibodies for the ConA action on chondrocytes.
Effects of ConA on chondrocyte phenotypic expression
in ATDC5 cultures overexpressing MTf
MTf is upregulated in the mouse embryonic carcinoma-derived
ATDC5 cells during chondrogenic differentiation (Nakamasu
et al., 1999). These cells initiate chondrogenic differentiation
only after the addition of some growth factor, such as insulin/
insulin-like growth factor-I, TGF-β or bone morphogenic
proteins (Shukunami et al., 1996; Atsumi et al., 1990; Fujii et
al., 1999). Using this model, we examined the role of MTf
in ConA-induced chondrogenic differentiation. Rabbit MTf
cDNA was expressed under the control of the CMV promoter
in stably transfected ATDC5 cells, and two MTf-expressing
clones were isolated (ATDC5-MTf1 and ATDC5-MTf5);
immunoblotting confirmed the expression of rabbit MTf in
these clones (Fig. 6A). In the absence of added growth factors,
parental ATDC5 cells and empty vector-integrated cells (Pc1
and Pc2) did not differentiate chondrogenic cells, regardless of
the presence or absence of ConA (Fig. 6B), because of an
absence of MTf expression (Nakamasu et al., 1999). By
contrast, ~30% of the MTf-overexpressing cells became
spherical chondrocytes even in the absence of ConA, and
ConA further increased the number of spherical chondrocytes
dose-dependently (Fig. 6B), with the cell-shape change being
accompanied by an increase in the uronic-acid-containing
macromolecule (proteoglycan) (Fig. 6C). Even in the absence
MTf is a ConA receptor
2035
control cells (Fig. 6D), indicating
the involvement of MTf in
chondrogenic
differentiation.
Furthermore, sConA increased
these mRNA levels in the MTfoverexpressing cells but not in the
control cells. And, within 48
hours, the anti-MTf-mAb2 or antiMTf-pAb1 suppressed expression
of the spherical phenotype
induced
by
the
MTfoverexpressing ATDC5 cells in
response to ConA (Fig. 6E).
Effect of ConA on cell shape in
cultures of C3H10T1/2 cells
overexpressing MTf
We isolated the C3H10T1/2 cells
(T4) overexpressing rabbit MTf to
examine whether ConA would
induce cell shape change in the
mouse pluripotent mesenchymal
cell line expressing MTf at high
levels. Wild-type C3H10T1/2 and
T4 cells showed MTf mRNA
expression at low and high levels,
respectively (data not shown), and
immunoblotting confirmed the
expression of rabbit MTf at a high
level in T4 cells, and its absence
in the wild-type C3H10T1/2 cells
(Fig. 7A). ConA induced cell
shape change from fibroblastic to
spherical or spindle-like cells
within 48 hours in T4 cells, but not
in the wild-type or empty vectorintegrated cells (Fig. 7B,C).
Furthermore,
the
cell-shape
change was suppressed in the
presence of the anti-MTf-mAb2
(Fig. 7D,E).
Discussion
ConA
enhances
phenotypic
expression by chondrocytes (Yan
et al., 1990). Because this effect of
ConA is greater than that of
growth factors, identification of
ConA receptors might give us
insight into the process of
chondrocyte differentiation. MTf
is a major ConA-binding protein
on the chondrocyte surface, has a
potential N-glycosylation site
and binds to ConA-Sepharose
(Kawamoto et al., 1998). Digestion of MTf of 76 kDa with
N-glycosidase F for 5-15 hours yielded a product of 66 kDa
(R.O., E. Usui and Y.K., unpublished). Accordingly, MTf
elutes from a ConA-Sepharose column with methyl-α-
Fig. 6. Effects of MTf overexpression on chondrogenic differentiation of ATDC5 cells in response to
ConA. ConA (5-20 µg ml–1) or sConA (35 µg ml–1) was added to the cultures in the presence of 5%
serum, transferrin and selenite with no other growth factors from day 10, every four days or every
other day, respectively. (A) Rabbit MTf levels in the cell layer of Pc1 (lane 1), Pc2 (lane 2), MTf1
(lane 3) and MTf5 (lane4) were analysed by immunoblotting. The proportion of spherical/polygonal
chondrocytes among total cells (B) and the uronic acid content (C) were determined on day 23.
(D) The mRNA levels of aggrecan, collagen type IIA, collagen type IIB and GAPDH were
determined on day 23 by RT-PCR and Southern-blot analysis. (E) The appearance of ATDC5-MTf5
cells exposed to 15 µg ml–1 ConA in the presence or absence of mAb2 (50 µg ml–1) (Experiment A)
or pAb1 serum (1%) (Experiment B) on day 12. The lectin and/or antibodies were added to the
cultures on day 10. Bar, 30 µm.
of ConA, the mRNA levels of aggrecan and collagen type IIB
(chondrocyte specific), as well as the collagen type IIA
(prechondrogenic stage characteristic) mRNA level, were
much higher in the MTf-overexpressing cells than in the
2036
Journal of Cell Science 116 (10)
Fig. 7. Effects of MTf overexpression on chondrogenic differentiation of C3H10T1/2 cells in response to ConA. C3H10T1/2 cells were
transfected with MTf-expression or empty vector. (A) MTf levels in the cells transfected with the empty vector (lane 1) or MTf-expression
vector (lane 2) were analyzed by immunoblotting. (B) ConA (5 µg ml–1) was added to confluent cultures of these cells for 48 hours. (C) The
proportion of spherical/polygonal/spindle-like cells among total cells was calculated. (D) The antibody mAb2 or control IgG (100 µg ml–1) was
added to confluent cultures of MTf-overexpressing C3H10T1/2 cells for 4 days. ConA (5 µg ml–1) was added to these cultures for 48 hours.
(E) The proportion of spherical/polygonal/spindle-like cells among total cells was calculated. Bar, 30 µm.
mannopyranoside (Kawamoto et al., 1998). ConA actions on
chondrocytes can be abolished by anti-MTf antibodies (Fig.
3B,C) or methyl-α-mannopyranoside (Yan et al., 1990; Yan et
al., 1997). These findings suggest that sugar is needed for the
interaction between ConA and MTf. When retinoic acid
decreased the MTf level, the chondrocytes lost responsiveness
to ConA. The responsiveness of the chondrocytes, ATDC5
cells and C3H10T1/2 cells to the lectin depended upon the
expression of MTf. Furthermore, some anti-MTf antibodies
mimicked the ConA actions. These findings strongly suggest
that MTf is a ConA receptor in chondrocytes, prechondrogenic
cells and mesenchymal cells.
ConA alters the shape, proliferation and/or differentiation of
various animal cells. MTf is not always expressed in these
cells, suggesting that other ConA receptors are present in nonchondrogenic cells including lymphocytes. However, it is
noteworthy that MTf is expressed at high or moderate levels in
many tumors and fetal tissues (Brown et al., 1981; Danielsen
and van Deurs, 1995), as well as several adult tissues, including
the capillaries (Rothenberger et al., 1996), salivary glands
(Richardson, 2000) and eosinophil precursors (McNagny et al.,
1996). In these cells, MTf might work, at least partially, as a
ConA receptor.
ConA modulated both cell shape and the expression of
aggrecan and type II collagen genes in chondrocytes and
ATDC5 cells, via crosslinking of MTf and/or simple binding
to MTf. Crosslinking of MTf by ConA appears to be essential
for rapid cell-shape change but not for the expression of the
differentiation-related genes, because simple binding of sConA
induced the gene expression in chondrocytes (Fig. 4A) but not
the rapid cell-shape change (Yan et al., 1997). Unlike ConA,
sConA cannot induce clustering of cell-surface proteins. The
lack of effect of sConA on cell shape suggests that the ConAinduced cell-shape change from fibroblastic cells to spherical
MTf is a ConA receptor
cells is not directly linked with the ConA-induced, aggrecan
and type II collagen expressions. This conclusion was
unexpected because the cell shape has a great effect on the
differentiation of prechondrogenic cells and dedifferentiated
cartilage cells in some experimental systems (Zanetti and
Solursh, 1984; Brown and Benya, 1988). However, in cultures
of MTf-overexpressing C3H10T1/2 cells, ConA altered cell
shape but did not enhance the expression of aggrecan or type
II collagen (data not shown), which also suggests that MTf has
two distinct roles: modulating cell shape and stimulating
chondrocyte phenotypic expression.
Cell shape is determined by extracellular proteins (e.g.
adhesion factors, matrix proteins and proteases) and
membrane-bound proteins (e.g. integrins, cytoskeletonassociated proteins, small G proteins and tyrosine kinases), and
the crosslinking of MTf might affect the activity of these
proteins. Signals for the cell-shape change induced by ConA
remain unknown. However, in subconfluent cultures of rabbit
chondrocytes, SB203580 [an inhibitor of p38 mitogenactivated protein (MAP) kinase] at 10 µM suppressed the
ConA-induced cell-shape changes, whereas U0126 [an
inhibitor of MAP kinase kinase (MEK)] or SP600125 [an
inhibitor of c-Jun N-terminal kinase (JNK)] had little effect at
the same concentrations (data not shown). These findings
suggest that p38 MAP kinase is involved in the cell shape
change induced by the ConA-MTf system.
MTf is a glycosylphosphatidylinositol (GPI)-anchored
protein (Food et al., 1994) and, because MTf does not have a
cytoplasmic domain, the effect of ConA might be mediated by
the binding of a ConA-MTf complex to a transmembrane
receptor. One GPI-anchored protein [CD14, which binds to
lipopolysaccharide (LPS)] forms a complex with a
transmembrane Toll-like receptor-4 to induce inflammatory
responses (Chow et al., 1999). The ciliary neurotrophic factor
receptor is also a GPI-anchored protein that binds to a
transmembrane gp130 protein, the signaling component of the
IL-6 receptor (Davis et al., 1993). Similarly, the GPI-anchored
glial-cell-derived neurotrophic factor receptor α associates,
after ligand binding, with a transmembrane tyrosine kinase
receptor Ret (Jing et al., 1996). A more likely mechanism for
ConA actions is activation of signaling molecules in lipid rafts
via crosslinking of MTf or simple binding to MTf. GPIanchored proteins are found on the outer surface of lipid rafts,
and signalling molecules, such as the Src kinase family and G
proteins, are located in the inner surface of lipid rafts (Rodgers
et al., 1994; Mumby, 1997; Simons and Ikonen, 1997). Binding
of specific antibodies to some GPI-anchored proteins (e.g.
CD14, Thy-1, Ly-6 and Qa-2) elicits striking biological
reactions, including tyrosine phosphorylation, increase of
cytoplasmic calcium, cell aggregation, phagocytosis, IL-2
production and/or DNA synthesis in various cells (Horejsi et
al., 1998).
It is clear that MTf is not a ConA receptor for lymphocyte
activations. We speculate, however, that some GPI-anchored
proteins might work as a ConA receptor in lymphocytes,
because ConA and antibodies to some GPI-anchored proteins
have similar effects on lymphocytes, including increases in
cytoplasmic calcium, cell-shape change, DNA synthesis and
IL-2 production (Robinson, 1991; Horejsi et al., 1998).
Some GPI-anchored proteins, including Thy-1, have been
shown physiologically to be involved in signaling via
2037
immunoreceptors (Hueber et al., 1997; Romagnoli and Bron,
1997). We showed here that even in the absence of ConA, the
overexpression of MTf moderately altered the shape of ATDC5
cells and enhanced the expression of cartilage-characteristic
genes, which suggests that GPI-anchored MTf plays a specific
physiological role in chondrocyte differentiation.
We are now using ConA to enhance chondrogenic
differentiation of human bone marrow mesenchymal cells and
to enhance the phenotypic expression of cultured chondrocytes
in vitro. The cartilage-like tissue formed in the presence of
ConA in vitro might be useful for cell therapy. The results
obtained in this study should be useful in promoting the
application of ConA to tissue engineering.
We thank T. Kawamoto and M. Noshiro for their contributions. We
also thank the Research Center for Molecular Medicine, Hiroshima
University School of Medicine, for the use of their facilities.
References
Atsumi, T., Miwa, Y., Kimata, K. and Ikawa, Y. (1990). A chondrogenic
cell line derived from a differentiating culture of AT805 teratocartinoma
cells. Cell Differ Dev. 30, 109-116.
Bitter, T. and Muir, H. M. (1962). Modified uronic acid carbazole reaction.
Anal. Biochem. 4, 330-334.
Brown, P. D. and Benya, P. D. (1988). Alterations in chondrocyte cytoskeletal
architecture during phenotypic modulation by retinoic acid and
dihydrocytochalasin B-induced reexpression. J. Cell Biol. 106, 171-179.
Brown, J. P., Woodbury, R. G., Hart, C. E., Hellstrom, I. and Hellstrom,
K. E. (1981). Quantitative analysis of melanoma-associated antigen p97 in
normal and neoplastic tissues. Proc. Natl Acad. Sci. USA 78, 539-543.
Brown, J. P., Hewick, R. M., Hellstrom, I., Hellstrom, K. E., Doolittle, R.
F. and Dreyer, W. J. (1982). Human melanoma-associated antigen p97 is
structurally and functionally related to transferrin. Nature 296, 171-173.
Chow, J. C., Young, D. W., Golenbock, D. T., Christ, W. J. and Gusovsky,
F. (1999). Toll-like receptor-4 mediates lipopolysaccharide-induced signal
transduction. J. Biol. Chem. 274, 10689-10692.
Danielsen, E. M. and van Deurs, B. (1995). A transferrin-like GPI-linked
iron-binding protein in detergent-insoluble noncaveolar microdomains at
the apical surface of fetal intestinal epithelial cells. J. Cell Biol. 131, 939950.
Davis, S., Aldrich, T. H., Stahl, N., Pan, L., Taga, T., Kishimoto, T., Ip, N.
Y. and Yancopoulos, G. D. (1993). LIFR beta and gp130 as
heterodimerizing signal transducers of the tripartite CNTF receptor. Science
260, 1805-1808.
Demeule, M., Poirier, J., Jodoin, J., Bertrand, Y., Desrosiers, R. R.,
Dagenais, C., Nguyen, T., Lanthier, J., Gabathuler, R., Kennard, M. et
al. (2002). High transcytosis of melanotransferrin (P97) across the bloodbrain barrier. J. Neurochem. 83, 924-933.
Dippold, W. G., Lloyd, K. O., Li, L. T., Ikeda, H., Oettgen, H. F. and Old,
L. J. (1980). Cell surface antigens of human malignant melanoma: definition
of six antigenic systems with mouse monoclonal antibodies. Proc. Natl.
Acad. Sci. USA 77, 6114-6118.
Estin, C. D., Stevenson, U., Kahn, M., Hellstrom, I. and Hellstrom, K. E.
(1989). Transfected mouse melanoma lines that express various levels of
human melanoma-associated antigen p97. J. Natl. Cancer Inst. 81, 445-448.
Food, M. R., Rothenberger, S., Gabathuler, R., Haidl, I. D., Reid, G. and
Jefferies, W. A. (1994). Transport and expression in human melanomas of
a transferrin-like glycosylphosphatidylinositol anchored protein. J. Biol.
Chem. 269, 3034-3040.
Fujii, M., Takeda, K., Imamura, T., Aoki, H., Sampath, T. K., Enomoto,
S., Kawabata, M., Kato, M., Ichijo, H. and Miyazono, K. (1999). Roles
of bone morphogenetic protein type I receptors and Smad proteins
in osteoblast and chondroblast differentiation. Mol. Biol. Cell 10, 38013813.
Gunther, G. R., Wang, J. L., Yahara, I., Cunningham, B. A. and Edelman,
G. M. (1973). Concanavalin A derivatives with altered biological activities.
Proc. Natl. Acad. Sci. USA 70, 1012-1016.
Horejsi, V., Cebecauer, M., Cerny, J., Brdicka, T., Angelisova, P. and
Drbal, K. (1998). Signal transduction in leucocytes via GPI-anchored
2038
Journal of Cell Science 116 (10)
proteins: an experimental artefact or an aspect of immunoreceptor function?
Immunol. Lett. 63, 63-73.
Hueber, A. O., Bernard, A. M., Battari, C. L., Marguet, D., Massol, P.,
Foa, C., Brun, N., Garcia, S., Stewart, C., Pierres, M. and He, H. T.
(1997). Thymocytes in Thy-1–/– mice show augmented TCR signaling and
impaired differentiation. Curr. Biol. 7, 705-708.
Jefferies, W. A., Food, M. R., Gabathuler, R., Rothenberger, S., Yamada,
T., Yasuhara, O. and McGeer, P. L. (1996). Reactive microglia specifically
associated with amyloid plaques in Alzheimer’s disease brain tissue express
melanotransferrin. Brain Res. 712, 122-126.
Jing, S., Wen, D., Yu, Y., Holst, P. L., Luo, Y., Fang, M., Tamir, R., Antonio,
L., Hu, Z., Cupples, R. et al. (1996). GDNF-induced activation of the ret
protein tyrosine kinase is mediated by GDNFR-alpha, a novel receptor for
GDNF. Cell 85, 1113-1124.
Kato, Y. and Gospodarowicz, D. (1985). Effect of exogenous extracellular
matrices on proteoglycan synthesis by cultured rabbit costal chondrocytes.
J. Cell Biol. 100, 486-495.
Kato, Y., Nomura, Y., Daikuhara, Y., Nasu, N., Tsuji, M., Asada, A. and
Suzuki, F. (1980). Cartilage-derived factor (CDF) I. Stimulation of
proteoglycan synthesis in rat and rabbit costal chondrocytes in culture. Exp.
Cell Res. 130, 73-81.
Kato, Y., Hiraki, Y., Inoue, H., Kinoshita, M., Yutani, Y. and Suzuki, F.
(1983). Differential and synergistic actions of somatomedin-like growth
factors, fibroblast growth factor and epidermal growth factor in rabbit costal
chondrocytes. Eur. J. Biochem. 129, 685-690.
Kawamoto, T., Pan, H., Yan, W., Ishida, H., Usui, E., Oda, R., Nakamasu,
K., Noshiro, M., Kawashima-Ohya, Y., Fujii, M. et al. (1998). Expression
of membrane-bound transferrin-like protein p97 on the cell surface of
chondrocytes. Eur. J. Biochem. 256, 503-509.
Kennard, M. L., Richardson, D. R., Gabathuler, R., Ponka, P. and
Jefferies, W. A. (1995). A novel iron uptake mechanism mediated by GPIanchored human p97. EMBO J. 14, 4178-4186.
Lis, H. and Sharon, N. (1986). Biological properties of lectins. In The Lectins
(ed. I. E. Liener, N. Sharon and I. J. Goldstein), pp. 265-291. Orlando, FL:
Academic Press
McNagny, K. M., Rossi, F., Smith, G. and Graf, T. (1996). The eosinophilspecific cell surface antigen, EOS47, is a chicken homologue of the
oncofetal antigen melanotransferrin. Blood 87, 1343-1352.
Mollenhauer, J., Bee, J. A., Lizarbe, M. A. and von der Mark, K. (1984).
Role of anchorin CII, a 31,000-mol-wt membrane protein, in the interaction
of chondrocytes with type II collagen. J. Cell Biol. 98, 1572-1579.
Mumby, S. M. (1997). Reversible palmitoylation of signaling proteins. Curr.
Opin. Cell Biol. 9, 148-154.
Nakamasu, K., Kawamoto, T., Shen, M., Gotoh, O., Teramoto, M.,
Noshiro, M. and Kato, Y. (1999). Membrane-bound transferrin-like protein
(MTf): structure, evolution and selective expression during chondrogenic
differentiation of mouse embryonic cells. Biochim. Biophys. Acta 1447, 258264.
Richardson, D. R. (2000). The role of the membrane-bound tumour antigen,
melanotransferrin (p97), in iron uptake by the human malignant melanoma
cell. Eur. J. Biochem. 267, 1290-1298.
Robinson, P. J. (1991). Phosphatidylinositol membrane anchors and T-cell
activation. Immunol. Today 12, 35-41.
Rodgers, W., Crise, B. and Rose, J. K. (1994). Signals determining protein
tyrosine kinase and glycosyl-phosphatidylinositol-anchored protein
targeting to a glycolipid-enriched membrane fraction. Mol. Cell. Biol. 14,
5384-5391.
Romagnoli, P. and Bron, C. (1997). Phosphatidylinositol-based glycolipidanchored proteins enhance proximal TCR signaling events. J. Immunol. 158,
5757-5764.
Rose, T. M., Plowman, G. D., Teplow, D. B., Dreyer, W. J., Hellstrom, K.
E. and Brown, J. P. (1986). Primary structure of the human melanomaassociated antigen p97 (melanotransferrin) deduced from the mRNA
sequence. Proc. Natl. Acad. Sci. USA 83, 1261-1265.
Rothenberger, S., Food, M. R., Gabathuler, R., Kennard, M. L., Yamada,
T., Yasuhara, O., McGeer, P. L. and Jefferies, W. A. (1996). Coincident
expression and distribution of melanotransferrin and transferrin receptor in
human brain capillary endothelium. Brain Res. 712, 117-121.
Shimomura, Y., Yoneda, T. and Suzuki, F. (1975). Osteogenesis by
chondrocytes from growth cartilage of rat rib. Calcif. Tissue Res. 19, 179187.
Shukunami, C., Shigeno, C., Atsumi, T., Ishizeki, K., Suzuki, F. and
Hiraki, Y. (1996). Chondrogenic differentiation of clonal mouse embryonic
cell line ATDC5 in vitro: differentiation-dependent gene expression of
parathyroid hormone (PTH)/PTH-related peptide receptor. J. Cell Biol. 133,
457-468.
Simons, K. and Ikonen, E. (1997). Functional rafts in cell membranes. Nature
387, 569-572.
Smale, G. and Sasse, J. (1992). RNA isolation from cartilage using density
gradient centrifugation in cesium trifluoroacetate: an RNA preparation
technique effective in the presence of high proteoglycan content. Anal.
Biochem. 203, 352-356.
Towbin, H., Staehelin, T. and Gordon, J. (1979). Electrophoretic transfer of
proteins from polyacrylamide gels to nitrocellulose sheets: procedure and
some applications. Proc. Natl. Acad. Sci. USA 76, 4350-4354.
Wlodarski, K. H. and Galus, K. (1992). Osteoblastic and chondroblastic
response to a variety of locally administered immunomodulators in mice.
Folia Biol. (Praha) 38, 284-292.
Woodbury, R. G., Brown, J. P., Yeh, M. Y., Hellstrom, I. and Hellstrom,
K. E. (1980). Identification of a cell surface protein, p97, in human
melanomas and certain other neoplasms. Proc. Natl. Acad. Sci. USA 77,
2183-2187.
Yan, W., Nakashima, K., Iwamoto, M. and Kato, Y. (1990). Stimulation by
concanavalin A of cartilage-matrix proteoglycan synthesis in chondrocyte
cultures. J. Biol. Chem. 265, 10125-10131.
Yan, W., Pan, H., Ishida, H., Nakashima, K., Suzuki, F., Nishimura, M.,
Jikko, A., Oda, R. and Kato, Y. (1997). Effects of concanavalin A on
chondrocyte hypertrophy and matrix calcification. J. Biol. Chem. 272, 78337840.
Zanetti, N. C. and Solursh, M. (1986). Epithelial effects on limb
chondrogenesis involve extracellular matrix and cell shape. Dev. Biol. 113,
110-118.