Expression and physiological role of CCN4⁄ Wnt-induced
secreted protein 1 mRNA splicing variants in chondrocytes
Takeshi Yanagita
1,2
, Satoshi Kubota
1
, Harumi Kawaki
1
, Kazumi Kawata
1
, Seiji Kondo
1
,
Teruko Takano-Yamamoto
3
, Shinji Tanaka
4
and Masaharu Takigawa
1
1 Department of Biochemistry & Molecular Dentistry, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical
Sciences, Japan
2 Department of Orthodontics and Dentofacial Orthopedics, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical
Sciences, Japan
3 Division of Orthodontics and Dentofacial Orthopedics, Tohoku University Graduate School of Dentistry, Miyagi, Japan
4 Department of Hepato-Biliary-Pancreatic Surgery, Graduate School of Medicine and Dentistry, Tokyo Medical and Dental University, Japan
The Wnt1-induced secreted proteins (WISPs) were
identified together in a human study reported in 1998
and then later classified as members of the CCN
family, which also includes cysteine-rich 61 (Cyr61 ⁄
CCN1), connective tissue growth factor (CTGF ⁄
CCN2) and nephroblastoma overexpressed (Nov ⁄

have been described in various pathological conditions. Several previous
reports described a CCN4 ⁄ WISP1 variant (WISP1v) lacking the second
module in a few malignancies, but no information concerning the produc-
tion of WISP1 variants in normal tissue is currently available. The expres-
sion of CCN4 ⁄ WISP1 mRNA and its variants were analyzed in a human
chondrosarcoma-derived chondrocytic cell line, HCS-2 ⁄ 8, and primary rab-
bit growth cartilage (RGC) chondrocytes. First, we found WISP1v and a
novel variant of WISP1 (WISP1vx) to be expressed in HCS-2 ⁄ 8, as well as
full-length WISP1 mRNA. This new variant was lacking the coding regions
for the second and third modules and a small part of the first module. To
monitor the expression of CCN4 ⁄ WISP1 mRNA along chondrocyte differ-
entiation, RGC cells were cultured and sampled until they were mineral-
ized. As a result, we identified a WISP1v ortholog in normal RGC cells.
Interestingly, the WISP1v mRNA level increased dramatically along with
terminal differentiation. Furthermore, overexpression of WISP1v provoked
expression of an alkaline phosphatase gene that is a marker of terminal dif-
ferentiation in HCS-2 ⁄ 8 cells. These findings indicate that WISP1v thus
plays a critical role in chondrocyte differentiation toward endochondral
ossification, whereas HCS-2 ⁄ 8-specific WISP1vx may be associated with the
transformed phenotypes of chondrosarcomas.
Abbreviations
ALP, alkaline phosphatase; CT, C-terminal cysteine knot; HUVEC, human umbilical vein endothelial cell; IGFBP, insulin-like growth factor
binding protein-like; RGC, rabbit growth cartilage; TSP1, thrombospondin type 1; VWC, von Willebrand factor type C; WISP, Wnt-induced
secreted protein.
FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS 1655
described as ‘Expression in low-metastatic cells type 1
(Elm1)’ gene [6].
Currently, all six members tend to be referred to as
CCN1–6, based on the unified nomenclature [1,7].
Members have been described as being involved in a

In this study, we analyzed the expression of full-
length WISP1 mRNA and its splicing variants in nor-
mal and malignant-transformed chondrocytes. For the
first time, we describe the expression of WISP1v in nor-
mal growth plate chondrocytes, which was regulated
along with the terminal differentiation of those cells.
Furthermore, we also identified a novel WISP1 mRNA
variant in malignant-transformed human chondrocytes.
The roles of these variants in chondrocyte biology have
been analyzed and are now discussed.
Results
CCN4 ⁄ WISP1 mRNA splicing variants in human
chondrocytic HCS-2/8 cells
To characterize mRNAs from ccn4 ⁄ wisp1, RT-PCR
was performed using total RNAs from HCS-2 ⁄ 8,
human umbilical vein endothelial cells (HUVEC),
HeLa, MDA231, SaOS2 and HEK293 cells with speci-
fic primers that recognize the IGFBP (sense) and CT
(antisense) coding exons. These primers were designed
to amplify the splicing variant lacking the VWC mod-
ule area (WISP1v), as well as full-length WISP1
(Fig. 1A). As a result, no ccn4 ⁄ wisp1 transcripts were
Fig. 1. Analysis of CCN4 ⁄ WISP1 mRNA in several cell lines as evaluated by RT-PCR. (A) The sense and antisense primers were designed to
recognize the sequences in IGFBP module-encoding exons and CT module-encoding exons, respectively. The length of the PCR products
from the full-length CCN4 ⁄ WISP1 amplified by these primers was expected to be 606 bp. (B) The results of the RT-PCR analysis.
CCN4 ⁄ WISP1 and WISP1v-cloned vectors were also amplified as controls. According to the DNA size markers analyzed together, the size of
the PCR product I was  600 bp, whereas those of II and III were  400 and 200 bp, respectively. Experiments were repeated at least three
times and representative results are shown. Equal sample load was confirmed by the RT-PCR analysis of gapdh mRNA.
CCN4 splicing variants in chondrocytes T. Yanagita et al.
1656 FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS

Importantly, the nucleotide sequences of exon–
intron boundaries of WISP1vx strictly conserve the
general rule of mRNA splicing sites in higher eukaryo-
tes (Fig. 2B). These findings indicate that WISP1vx is
not an artifact of PCR amplification, but a splicing
variant uniquely present in HCS-2 ⁄ 8 chondrocytic
cells.
Detection of CCN4 ⁄ WISP1 proteins in HCS-2/8
cells
In order to examine whether CCN4 ⁄ WISP1 and
WISP1v mRNAs yielded corresponding proteins in
HCS-2 ⁄ 8 cells, we analyzed the HCS-2 ⁄ 8 cell lysate
using western blotting with an anti-(CCN4 ⁄ WISP1)
serum targeting the CT module. First, mammalian
expression vectors for FLAG epitope-tagged CCN4 ⁄
WISP1, WISP1v and WISP1vx proteins were con-
structed, and the proteins were overexpressed by
DNA transfection of these expression vectors
(Fig. 3A) into HCS-2 ⁄ 8 cells. The overexpressed
CCN4 ⁄ WISP1, WISP1v and WISP1vx served as pos-
itive controls for western blotting. Because the FLAG
epitope itself is a small octapeptide, the exogenesis
CCN4 ⁄ WISP1 and its variants are indistinguishable
from endogenous ones on SDS⁄ PAGE. As shown in
Fig. 3B, analysis of transfectants with the anti-
(CCN4 ⁄ WISP1) clearly showed strong enhancement
of the CCN4 ⁄ WISP1 and WISP1v signals that were
also present in HCS-2 ⁄ 8 cells without the overexpres-
sion. As expected, no WISP1vx signal was detected,
because it lacked the CT module in which the epitope

expression of type II and type X collagen genes. To
assess whether WISP1v and WISP1vx are associated
with malignant phenotypes, or with normal chondro-
cytic phenotypes, we investigated whether CCN4 ⁄
WISP1 mRNA and its variants are also expressed in
normal primary chondrocytes. Rabbit growth carti-
lage (RGC) cells were isolated, and chondrocytic
differentiation was induced in vitro. Thereafter, RT-
PCR was performed to analyze the ccn4 ⁄ wisp1
mRNA expression pattern in RGC cells. Interestingly,
WISP1v, as well full-length CCN4 ⁄ WISP1 mRNA,
was distinctly detected in normal RGC, whereas
WISP1vx was not. As summarized in Fig. 4, nucleo-
tide sequence analysis indicated that the splicing sites
used to generate rabbit WISP1v are at the same posi-
tion as those observed in the human gene, indicating
that it is exactly the same ortholog as that of human
WISP1v.
Regulated expression of WISP1v along with the
chondrocytic differentiation of RGC cells
In order to assess whether WISP1v is associated with
the development of growth cartilage, primary growth
cartilage cells collected from rabbits were cultured until
they became confluent. Afterwards, a long-term culture
for the induction of terminal differentiation was car-
ried out for over one month. Fluctuation in type II
collagen as a marker gene of chondrocyte differenti-
ation showed that RGC cells were properly differenti-
ated in vitro (Fig. 5C). Alizarin red staining of RGC
cells revealed increased mineralization by the cells from

(B) Agarose gel electrophoresis analysis confirming the quality of the total RNA at each stage. (C) Expression of the type II collagen gene as
evaluated by quantitative real-time PCR. (D) RT-PCR analysis with the primers recognizing WISP1v. (E) A quantitative real-time PCR analysis
performed with rabbit-specific WISP1v primers. The sampling time points are represented as the numbers of weeks after the cell
reached confluence. Representative results of four independent series of experiments are displayed with error bars (SD of real-time PCR
evaluation).
T. Yanagita et al. CCN4 splicing variants in chondrocytes
FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS 1659
Therefore following mRNA analysis, we examined the
distribution of CCN4 ⁄ WISP1 and WISP1v proteins in
developing mouse growth cartilage by immunohisto-
chemistry (Fig. 6). Because no anti-(CCN4 ⁄ WISP1)
serum was able to detect the loss of VWC, which is
the structural determinant of WISP1v that distingui-
shes it from full-length WISP1 mRNA, we used an
antibody that recognized a common module between
the full-length CCN4 ⁄ WISP1 mRNA and WISP1v
proteins. As a result, positive signals representing full-
length CCN4 ⁄ WISP1 mRNA and WISP1v were
observed, and were particularly strong in the hyper-
trophic zone of the developing tibial sections of the
mouse embryo. Although these signals also include
those from the full-length CCN4 ⁄ WISP1 mRNA, these
in vivo findings are consistent with the total outcome
of the constitutive expression of the full-length and the
differentiation stage-dependent expression of WISP1v
in vitro.
Effect of overexpression of CCN4/WISP1 and its
variants on the phenotypes of human
chondrocytic HCS-2 ⁄ 8 cells
Finally, the biological function of WISP1v and

ticular, expression of WISP1v among normal chondro-
cytes and osteoblastic cells was confirmed, which has
heretofore not been elucidated. Expression of WISP1v
increased during the course of chondrocytic terminal
differentiation, thus suggesting a role for WISP1v
therein. This hypothesis was supported by the fact that
overexpression of WISP1v enhanced the gene expres-
sion of one of the mineralization markers in chondro-
cytic cells. Moreover, the presence of a novel splicing
variant not described previously was discovered in
HCS-2 ⁄ 8 cells. These findings are summarized in
Fig. 8.
To date, WISP1v has been investigated only from
the point of view of tumorigenesis and malignant phe-
notypes of tumors [16]. Indeed, expression of WISP1v
has been reported only in malignant tumor cells [7,17].
Scirrhous carcinoma of the stomach is known to be
associated with WISP1v and it is characterized by
rapid growth with a vast fibrous stroma, high invasive-
ness and a substantially poor prognosis [20]. Because
little is known about the molecular pathogenesis of the
disease, a pathological role for WISP1v should be
investigated further. It should be noted that most
CCN family members are associated with certain types
of malignancy.
Regarding physiological functions, some CCN fam-
ily members play an important role in skeletal growth,
as typically represented by CCN2 ⁄ CTGF [1,4,5].
CCN2 ⁄ CTGF was found to be highly expressed in
hypertrophic chondrocytes, as shown by in situ hybrid-

In comparison with CCN1 ⁄ Cyr61 and CCN2 ⁄
CTGF, the role of CCN4 ⁄ WISP1 in chondrocyte dif-
ferentiation appears more specific to the mineralizing
stage of endochondral ossification. This fundamental
difference may be interpreted from an evolutionary
point of view. Genome-wide prediction of CCN ortho-
logs in several species indicated that CCN4 ⁄ WISP1 or-
thologs were found only in animals with a calcified
skeleton, whereas CCN1 ⁄ Cyr61 and CCN2 ⁄ CTGF
could be identified in invertebrates such as Ciona intes-
tinalis. Thus, CCN4 ⁄ WISP1 is thought to be required
for the establishment of a calcified endoskeleton. This
hypothesis is consistent with our findings on WISP1v.
However, it may not apply exactly to full-
length CCN4 ⁄ WISP1, because no clear change was
observed according to the chondrocytic differentiation.
Fig. 8. Formation and function of
CCN4 ⁄ WISP1 variants. Structure, expres-
sion and function of the CCN4 ⁄ WISP1 vari-
ants are summarized. All of the findings
were obtained in this study except for that
of gastrointestinal cancers reported by Tan-
aka et al. [17,20]. ++, enhanced expression
evaluated by real-time RT-PCR; ±, modest
expression evaluated by real-time RT-PCR;
–, expression undetectable.
T. Yanagita et al. CCN4 splicing variants in chondrocytes
FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS 1661
In addition, no substantial induction of ALPase activ-
ity was seen in the overexpression experiment (data

tion of minimum essential medium (Sigma, St Louis, MO)
containing 10% fetal bovine serum. Cell cultures were
maintained in a humidified 5% CO
2
atmosphere at 37 °C
without any passage until analysis.
Alizarin red staining
RGC cells in six-well plates were washed with NaCl ⁄ P
i
and
fixed using 4% paraformaldehyde ⁄ NaCl ⁄ P
i
. Following fix-
ation cells were washed again in NaCl ⁄ P
i
and stained with
a 1% alizarin red solution, as described previously [34].
Macroscopic images were captured after 15 extensive
washes with NaCl ⁄ P
i
.
DNA transfection
HCS-2 ⁄ 8 cells were seeded onto six-well plates at 500 000
cells per well. Cells were transfected with 2 lg of plasmid
DNA using 8 lL of FuGENE6 transfection reagent (Roche
Applied Science, Penzberg, Germany) according to the
manufacturer’s instructions. Cells were lysed in 50 lLof1·
passive lysis buffer (Promega, Madison, WI), and collected
for western blotting at 48 h after transfection. RNA was
also extracted from cells for RT-PCR at 48 h after transfec-

PCR products were electrophoresed in a 1% agarose gel
containing ethidium bromide and visualized under UV
light. Photographs of the stained gels were taken and
analyzed quantitatively with transilluminator fasiii
(Toyobo).
Antibodies and molecular clones
An anti-(CCN4 ⁄ WISP1 H-55) serum recognizing the CT
module (Santa Cruz Biotechnologies, Santa Cruz, CA) and
an anti-(FLAG M2) mAb (Sigma) were utilized in western
blotting or immunohistochemistry.
For molecular cloning of wisp1-related cDNA fragments,
PCR products were fractionated by agarose gel electrophor-
esis and were extracted from the gels using the Gel Extrac-
tion kit (Qiagen). Next, purified PCR products were
inserted into pGEM T-easy vector according to the manu-
facturer’s instructions (Promega).
CCN4 splicing variants in chondrocytes T. Yanagita et al.
1662 FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS
The original human CCN4 ⁄ WISP1 and WISP1v mam-
malian expression vectors were given by Shinji Tanaka
(Tokyo Medical and Dental University), in which human
cDNAs were inserted in pCR3.1 (Invitrogen, San Diego,
CA) [17]. Three FLAG epitope-tagged expression vectors
were newly constructed.
ccn4 ⁄ wisp1 and wisp1v cDNAs were excised from the
pCR3.1-based plasmids by digestion with HindIII and
XbaI, and then inserted between the corresponding restric-
tion enzymatic sites of the Mammalian Amino-Terminal
FLAG
Ò

reaction mixture (Toyobo) containing 2 lL cDNA, 5 lm
each primer and 1· SYBR Green Master Mix. The primer
sets used for the evaluation of ccn4 ⁄ wisp1 and its variants
were the same as the regular RT-PCR. The nucleotide
sequences of the primer sets for the quantitative evaluation
of human alkaline phosphatase (ALP), type II collagen and
aggrecan cDNAs were: human alp, sense 5¢-TGG
AGCTTCAGAAGCTCAACACCA-3¢ and antisense 5¢-AT
CTCGTTGTCTGAGTACCAGTCC-3¢ (443 bp); human
type II collagen, sense 5¢-GAGGGCAATAGCAGGTTCA
CGTA-3¢ and antisense 5¢-TGGGTGCAATGTCAATGA
TGG-3¢ (133 bp); human aggrecan, sense 5¢-TCTTCA
GTCCCGTTCTCCAC-3¢ and antisense 5¢-AACATCACT
GAGGGCGAAGC-3¢ (93 bp), The amplification condi-
tions were as follows: for human ccn4 ⁄ wisp1 and human
alp cDNA, 95 °C (1 min) for 1 cycle, followed by 95 °C
(1 s), 59 °C (1 s), 72 °C (30 s) for 50 cycles, and a melting
process from 50 to 95 °C for 5 min, for rabbit ccn4 ⁄ wisp1
and rabbit ccn4 ⁄ wisp1v cDNAs, 94 °C (5 min) for 1 cycle,
followed by 94 °C (5 s), 60 °C (5 s), 72 °C (30 s) for 45
cycles, and melting process from 50 to 95 °C for 5 min,
human aggrecan,94°C (30min) for 1 cycle, followed by
94 °C (5 s), 60 °C (0 s), 72 °C (30 s) for 45 cycles, and
melting process from 50 to 95 °C for 5 min. Similarly, rab-
bit gapdh cDNA fragment (283 bp) was analyzed with sense
(5¢-TCACCATCTTCCAGGAGCGA-3¢) and antisense
(5¢-CACAATGCCGAAGTGGTCGT-3¢) primers under the
following conditions: 94 °C (30 s) for 1 cycle, followed by
94 °C (5 s), 65 °C (10 s), 72 °C (15 s) for 45 cycles, and
melting process from 50 °Cto95°C for 5 min. Rabbit

to block endogenous peroxidase activity. To reduce nonspe-
cific binding, 10% goat serum (Vector Laboratories, Burlin-
game, CA) in NaCl ⁄ P
i
(GS-NaCl ⁄ P
i
) was applied to the
specimens for 30 min. Sections were then incubated with a
specific antibody against CCN4 ⁄ WISP1 (1 : 50) diluted in
GS-NaCl ⁄ P
i
at 4 °C overnight.
Specific signals were probed by a horseradish peroxy-
dase-conjugated secondary antibody (Histofine: Nichirei),
and visualized using 3,3-diaminobenzidine tetrachloride
(Sigma). Finally, the samples were counterstained with
methyl green solution (Wako Pure Chemical Industries,
Osaka, Japan).
For negative controls, we skipped the primary antibody
reaction step.
T. Yanagita et al. CCN4 splicing variants in chondrocytes
FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS 1663
Western blotting
The cell lysate was prepared in a SDS sample buffer con-
taining 2.5% b-mercaptoethanol. Each sample (25 lL) was
heated for 5 min at 95 °C and then was separated by 15%
PAGE, then the separated proteins were transferred to a
poly(vinylidene difluoride) membrane (Immobilon, Milli-
pore, Bedford, MA). Following the established protocol for
western blotting, FLAG-tagged proteins were detected

1 Perbal B & Takigawa M (2005) CCN Proteins: A New
Family of Cell Growth and Differentiation Regulators.
Imperial College Press, London.
2 Brigstock DR (1999) The connective tissue growth fac-
tor ⁄ cysteine-rich 61 ⁄ nephroblastoma overexpressed
(CCN) family. Endocrin Rev 20, 189–206.
3 Perbal B (2001) NOV (nephroblastoma overexpressed)
and the CCN family of genes: structural and functional
issues. Mol Pathol 54, 57–79.
4 Takigawa M (2003) CTGF ⁄ Hcs24 as a multifunctional
growth factor for fibroblasts, chondrocytes and vascular
endothelial cells. Drug News Perspect 16, 11–21.
5 Takigawa M, Nakanishi T, Kubota S & Nishida T
(2003) Role of CTGF ⁄ HCS24 ⁄ ecogenin in skeletal
growth control. J Cell Physiol 194, 256–266.
6 Blacketer MJ, Koehler CM, Coats SG, Myers AM &
Madaule P (1993) Regulation of dimorphism in Sac-
charomyces cerevisiae: involvement of the novel protein
kinase homolog Elm1p and protein phosphatase 2A.
Mol Cell Biol 13, 5567–5581.
7 Brigstock DR, Goldschmeding R, Katsube KI, Lam
SC, Lau LF, Lyons K, Naus C, Perbal B, Riser B,
Takigawa M et al. (2003) Proposal for a unified CCN
nomenclature. Mol Pathol 56, 127–128.
8 Pennica D, Swanson TA, Welsh JW, Roy MA,
Lawrence DA, Lee J, Brush J, Taneyhill LA, Deuel B,
Lew M et al. (1998) WISP genes are members of the
connective tissue growth factor family that are up-
regulated in wnt-1-transformed cells and aberrantly
expressed in human colon tumors. Proc Natl Acad Sci

A & Montalto G (2004) Expression of WISPs and of their
novel alternative variants in human hepatocellular carci-
noma cells. Ann NY Acad Sci 1028, 432–439.
17 Tanaka S, Sugimachi K, Kameyama T, Maehara S,
Shirabe K, Shimada M, Wands JR & Maehara Y
(2003) Human WISP1v, a member of the CCN family,
is associated with invasive cholangiocarcinoma.
Hepatology 37, 1122–1129.
18 YuC, Le AT, Yeger H, Perbal B & Alman BA (2003)
NOV (CCN3) regulation in the growth plate and CCN
CCN4 splicing variants in chondrocytes T. Yanagita et al.
1664 FEBS Journal 274 (2007) 1655–1665 ª 2007 The Authors Journal compilation ª 2007 FEBS
family member expression in cartilage neoplasia.
J Pathol 201, 609–615.
19 Soon LL, Yie TA, Shvarts A, Levine AJ, Su F &
Tchou-Wong KM (2003) Overexpression of WISP-1
down-regulated motility and invasion of lung cancer
cells through inhibition of Rac activation. J Biol Chem
278, 11465–11470.
20 Tanaka S, Sugimachi K, Saeki H, Kinoshita J, Ohga T,
Shimada M & Maehara Y (2001) A novel variant of
WISP1 lacking a von Willebrand type C module overex-
pressed in scirrhous gastric carcinoma. Oncogene 20,
5525–5532.
21 Nakanishi T, Yamaai T, Asano M, Nawachi K, Suzuki
M, Sugimoto T & Takigawa M (2001) Overexpression
of connective tissue growth factor ⁄ hypertrophic
chondrocyte-specific gene product 24 decreases bone
density in adult mice and induces dwarfism. Biochem
Biophys Res Commun 281, 678–681.

28 Kireeva ML, Mo FE, Yang GP & Lau LF (1996)
Cyr61, a product of a growth factor-inducible immedi-
ate-early gene, promotes cell proliferation, migration,
and adhesion. Mol Cell Biol 16, 1326–1334.
29 Leu SJ, Liu Y, Chen N, Chen CC, Lam SC & Lau LF
(2003) Identification of a novel integrin alpha 6 beta 1
binding site in the angiogenic inducer CCN1 (CYR61).
J Biol Chem 278, 33801–33808.
30 Leu SJ, Chen N, Chen CC, Todorovic V, Bai T, Juric
V, Liu Y, Yan G, Lam SC & Lau LF (2004) Targeted
mutagenesis of the angiogenic protein CCN1 (CYR61).
Selective inactivation of integrin alpha6beta1–heparan
sulfate proteoglycan coreceptor-mediated cellular func-
tions. J Biol Chem 279, 44177–44187.
31 Babic AM, Kireeva ML, Kolesnikova TV & Lau LF
(1998) CYR61, a product of a growth factor-inducible
immediate early gene, promotes angiogenesis and tumor
growth. Proc Natl Acad Sci USA 95, 6355–6360.
32 Takigawa M, Tajima K, Pan HO, Enomoto M,
Kinoshita A, Suzuki F, Takano Y & Mori Y (1989)
Establishment of a clonal human chondrosarcoma cell
line with cartilage phenotypes. Cancer Res 14,
3996–4002.
33 Simmons NL (1990) A cultured human renal epithelioid
cell line responsive to vasoactive intestinal peptide. Exp
Physiol 75, 309–319.
34 Kawata K, Eguchi T, Kubota S, Kawaki H, Oka M,
Minagi S & Takigawa M (2006) Possible role of LRP1,
a CCN2 receptor, in chondrocytes. Biochem Biophys
Res Commun 345, 552–559.