Visualization of CD146 dimerization and its regulation in living cells
Pengcheng Bua,b, Jie Zhuanga,b, Jing Fenga, Dongling Yanga, Xun Shena, Xiyun Yana,⁎
aNational Laboratory of Biomacromolecules, Institute of Biophysics, Chinese Academy of Sciences, 15 Datun Road, Beijing 100101, China
bGraduate School, Chinese Academy of Sciences, Beijing 100039, China
Received 25 September 2006; received in revised form 30 December 2006; accepted 19 January 2007
Available online 27 January 2007
Our previous study showed that the adhesion molecule CD146 as a biomarker is over-expressed on activated endothelium during angiogenesis,
which was induced by tumor conditional medium and inhibited by anti-CD146 monoclonal antibody (mAb AA98). However, the CD146
molecular organization on the cells is unknown. Here, using immunoprecipitation, we found that the dimerization of CD146 occurs in both normal
and tumor cells. However, the dimer/monomer ratio was higher in tumor cells than in normal cells. Moreover, we found that CD146 dimerization
was up-regulated by tumor conditional medium through the NF-kappa B pathway and down-regulated by mAb AA98. To further confirm that
CD146 dimerization occurs in living cells, we used fluorescence resonance energy transfer (FRET) with melanoma Mel888 cells co-expressing
CFP/YFP-tagged CD146 fusion proteins. By acceptor photobleaching, we observed a strong FRET signal produced by these two fluorescence-
tagged proteins. The FRET efficiency reached 20.1%. Our data provide the first evidence that CD146 dimerization occurs in living cells and is
regulated within the tumor microenvironment, implying that dimerization of CD146 may be associated with malignancy.
© 2007 Elsevier B.V. All rights reserved.
Keywords: CD146; Dimerization; FRET; NF-kappa B
Adhesion molecule CD146 (100–130 kDa) belongs to the
immunoglobulin superfamily. The sequence of CD146 is highly
homologous to a number of cell adhesion molecules such as
NCAM  and contains a V-V-C2-C2-C2 Ig-like extracellular
domain, a single membrane-spanning, and a short cytoplasmic
domain. The CD146 molecule was originally identified as a
biomarker for melanoma [2–4] and is involved in tumor
progression and metastasis. It has been shown that the enforced
expression of CD146 in melanoma cells increases melanoma
growth and metastasis , whereas the decrease of CD146
expression results in reduced tumorigenicity . Mills et al.
have reported that anti-CD146 antibody significantly inhibits
tumor growth and metastasis of human melanoma . These
data indicate that CD146 plays an important role in promoting
Another important set of data has identified CD146 as a
marker on angiogenic vascular endothelium [8,9] and a
structural component of interendothelial junctions . A
number of studies have addressed the biological function of
CD146 in blood vessels. For instance, Chan et al. reported that
the suppression of CD146 protein expression by antisense
oligonucleotides results in poor vascular development in
zebrafish . Our results showed that the CD146 molecule
is over-expressed on the activated endothelium of many
different tumors, and this expression is induced by tumor
conditional medium, resulting in increased cell adhesion,
migration and angiogenesis. Conversely, the induced endothe-
lial cell activities and angiogenesis could be inhibited by anti-
CD146 monoclonal antibody (mAb AA98) . In addition, we
and others found that CD146 plays an important role in cell
migration and invasion based on the observation that CD146 is
predominantly expressed on invasive intermediate trophoblasts
rather than non-invasive cytotrophoblasts [12–14]. Our finding
that CD146 expression was absence on the placenta of pre-
eclampsia  also supports the idea that the CD146 is a critical
molecule in embryo implantation and pregnancy disorder.
Biochimica et Biophysica Acta 1773 (2007) 513–520
⁎Corresponding author. Tel.: +86 10 64888583; fax: +86 10 64888584.
E-mail address: email@example.com (X. Yan).
0167-4889/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
Although the CD146 ligand has not been identified, the
engagement of CD146 with anti-CD146 antibodies has
profound effects on cellular functions. It has been shown
that CD146 mediates Ca2+-independent homotypic and
heterotypic cell–cell interaction on cultured endothelial cells
and is involved in control of intercellular permeability .
Additional studies indicated that CD146 induces the associa-
tion of tyrosine kinase p59fynwith the cytoplasmic tail of
CD146 and the phosphorylation of p125FAKand paxillin in
human umbilical vein endothelial cells . These findings
indicate that CD146 does not act as merely “molecule glue”,
but is actively involved in outside-in signaling and in the
control of cell–cell contact. However, the underlying me-
chanism remains unclear.
Given that many receptors and adhesion molecules are
functional as dimers or oligomers when interacting with
extracellular factors [17,18], which is a key event in signaling
transduction , we hypothesized that the oligomerization of
CD146 might occur in cells and be associated with cell activity.
To test this, we performed FRET and immunoprecipitation to
study the molecular organization of CD146 in living cells. Our
results provide the first evidence that CD146 dimerization
occurs in living cells, and is down-regulated by mAb AA98 and
up-regulated by tumor conditional medium through the NF-
kappa B pathway.
2. Materials and methods
2.1. Cell lines and antibodies
Human melanoma cell A375 and normal vascular endothelial cell
ECV304, both of which are natively CD146-expressing cells, were obtained
from American Type Culture Collection (Rockville). Mel888 is a human
melanoma cell line, which does not express CD146. CD146-positive Mel888
is the Mel888 cell transfected with CD146, which was kindly provided by Dr.
Judith P. Johnson (University of Munich, Germany). All the cells were
cultured in Dulbecco modified Eagle medium (DMEM) containing 10% fetal-
calf serum (FCS).
Antibodies used in this study were anti-NF-kappa B (p65), anti-IκBα, anti-
β-actin purchased from Santa Cruz, horseradish peroxidase (HRP)-conjugated
anti-mouse or anti-rabbit IgG purchased from Pierce, and rabbit anti-CD146
antibody and mAb AA98 produced in our lab.
2.2. Preparation of conditional medium
Melanoma cells A375 were cultured with DMEM in a 10 cm culture plate.
After 24-h culture, the supernatant from the culture was collected and used as
conditional medium (named as A375-CM).
2.3. Construction of tagged CD146
The cDNA encoding full length of human CD146 with its signal peptide was
amplified by PCR using pUC18-CD146 plasmid as template (kindly provided
by Dr. Judith P. Johnson, University of Munich, Germany). The primers used
were 5′-cgggaattcatggggcttcccaggctg-3′ (EcoRI underlined) and 5′-cccggatc-
catgcctcagatcgatgtatttc-3′ (BamHI underlined). The PCR products of CD146
gene were then inserted into the vector of pEyFP-N1 and pEcFP-N1 (Clontech),
producing CD146YFPand CD146CFPconstructs.
2.4. Fluorescence microscopy
Mel888 cells were co-transfected with CD146YFPand CD146CFPconstructs.
After 48 h culture, these transfected cells were fixed in 4% cold formaldehyde/
PBS for 4 min at room temperature, followed by observation under a confocal
laser scanning microscope (FV500, Olympus). For mAb AA98 staining, these
cells were continuously incubated overnight at 4 °C with mAb AA98, then with
biotin labeled anti-mouse IgG for 1 h, and finally with avidin-conjugated cy-
chrome (BD Pharmingen) for 1 h. After carefully washing with PBS, the cells
were examined under a fluorescence microscope with excitation at 510 nm and
emission at 630 nm.
For detection of the regulation of NF-kappa B activity, the endothelial
cells ECV304 were treated for 30 min with an inhibitor of NF-kappa B,
BAY 11-7082 (10 nM, Calbiochem) and then cultured with A375-CM for 2
days. The treated cells were fixed in 4% cold formaldehyde/PBS for 4 min,
permeabilized with 0.2% Triton X-100, and incubated for 30 min in 5%
goat serum/PBS. After reaction with anti-NF-kappa B (p65) overnight at
4 °C, and then incubation with FITC-conjugated anti-mouse IgG antibody
for another 1 h, the cells were examined under a confocal laser scanning
2.5. FRET measurements
To generate cells co-expressing CD146YFP/CD146CFPcells, the Mel888
cells were co-transfected with 2 μg of the CD146YFP/CD146CFPconstructs
mixed with Lipofectamine 2000 according to the instructions of the
manufacturer (Invitrogen). After 48 h of culture, these cells were washed with
cold PBS and fixed in cold 4% formaldehyde/PBS for 4 min at room temp-
erature. Image processing was performed under the confocal laser scanning
microscope. Filters used for observing CD146CFPallowed excitation at 458 nm
and emission at 465–495 nm; for CD146YFPexcitation at 515 nm and emission
Fig. 1. CD146 dimerization is detected in various cells. Cell lysates from melanoma A375, vascular endothelial ECV304 and CD146-forced expressing Mel888
CD146-Mel888 were immunoprecipitated with anti-CD146 mAb AA98, followed by Western blotting with rabbit anti-CD146 antibody. The CD146 dimer
(200 kDa) and monomer (100 kDa) were specifically recognized by the two antibodies (A, C, and E). No cross-reaction with other proteins was detected by
secondary anti-rabbit IgG antibodies (B, D, and F). The ratio of the CD146 dimer and monomer were qualified by densitometry using the data from three
independent experiments (G).
514P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520
at 565–595 nm was used; for FRETof CD146YFPand CD146CFPexcitation was
at 458 nm and emission at 565–595 nm.
Acceptor photobleaching was performed to test the specificity of FRET. The
CD146YFPas acceptor fluorophore was selectively photobleached for 2 min at
515 nm (meanwhile the 458 nm scanning window was set to zero for prevent
any CFP photobleaching), and the increase in fluorescence intensity of the donor
CD146CFPwas measured. The FRET efficiency E is given by: E=1−(Fda/Fd),
where Fdais the intensity of donor fluorescence when both donor and acceptor
are present, and Fdis the donor fluorescence intensity after the acceptor was
The cells of A375, ECV304 and CD146-positive Mel888 were respectively
lysed in a culture dish by adding 0.6 mL ice-cold lysis buffer A (150 mM NaCl,
1 mM EDTA, 50 mM Tris, pH 8.0, 10% glycerol, 1% Triton X-100, 1 mM
phenylmethylsulfonyl fluoride, and 25 μg/mL aprotinin). The supernatants were
collected by centrifugation at 12,000×g at 4 °C for 10 min and then pre-cleaned
with protein A-Sepharose (Sigma) to remove the protein A-bound proteins. The
total amount of proteins in the pre-cleaned supernatants was measured by
Bradford kit (Bio-Rad). Each sample containing total protein 3.2 mg was
immunoprecipitated with either mAb AA98 or control mIgG at 4 °C for 2 h,
followed by incubation with protein A-Sepharose for 1 h. Immunoprecipitates
were washed twice with the lysis buffer and then boiled for 5 min in loading
2.7. Preparation of cytoplasmic and nuclear extracts
For analysis of NF-kappa B activation in the cells, ECV304 cells were
treated with an inhibitor of NF-kappa B, BAY 11-7082 (10 nM) for 30 min, and
then co-cultured with A375-CM for 2 days. After that, the cytoplasmic and
nuclear extracts of the ECV304 cells were prepared. Briefly, the cells were
washed twice and scraped into 1.0 mL of ice-cold PBS. After centrifugation at
3000×g at 4 °C for 5 min, the cell pellet was lysed in 60 μL lysis buffer B
(10 mM Tris, pH 8.0, 1.5 mM MgCl2, 1 mM 1, 4-Dithiothreitol (DTT), 0.1%
NP-40, 1 mM phenylmethylsulfonyl fluoride and 25 μg/mL aprotinin) and then
incubated on ice for 15 min followed by centrifugation at 12,000×g at 4 °C for
15 min. The supernatants (cytoplasmic fraction) were stored at 4 °C and the
pelletscontainingnucleiwere suspended in 60μLbuffer C (10 mM Tris, pH8.0,
50 mM KCl, 100 mM NaCl, 1 mM phenylmethylsulfonyl fluoride and 25 μg/
mL aprotinin) at 4 °C for 30 min. After centrifuged at 12,000×g at 4 °C for 15
min, the supernatants were saved as nuclear extracts and boiled for 5 min in
2.8. Western blotting
The sample containing 25 μg proteins from the cytoplasmic fraction and
containing 20 μg proteins from the nuclear extract were loaded into 8% SDS-
PAGE and transferred onto a Hybond membrane (Amersham). The membranes
were blocked 1 h with 5% milk/PBS, incubated for 2 h with the primary
antibodiesanti-NF-kappa B (p65), anti-IκBα, anti-β-actin (Santa Cruz) or rabbit
anti-CD146. Then the membranes were probed with HRP-conjugated anti-
mouse or anti-rabbitIgG (Pierce), and detected by enhancedchemiluminescence
3.1. Detection of CD146 dimerization in various cell lysates
Using immunoprecipitating with mAb AA98 and Western
blotting with anti-CD146 rabbit polyclonal antibodies, we
observed two bands at 100-kDa and 200-kDa in the
immunoprecipitates of various cell lyses, including melanoma
A375, vascular endothelial ECV304, and CD146-forced
expressing Mel888 cells (Fig. 1A, C, E). The 100-kDa band
Fig. 2. Schematic representation of fluorescence tagged CD146 constructs. The
CD146CFPand CD146YFPwere constructed by fusing CFP or YFP at the
C-terminus of CD146. The N-terminus of CD146 contains a signal peptide.
Fig. 3. Cell localization of CFP/YFP tagged CD146 in living cells. Mel888 cells were co-transfected with CD146YFPand CD146CFPconstructs. After 48 h culture,
the cells were observed under a confocal microscope at the CFP (A) or YFP (B) channel. The CD146 fusion proteins on the cell surface were also recognized by mAb
AA98 with cy-chrome (C). (D) and (E) show the merging of AA98/cy-chrome with either CFP or YFP. (F) shows the corresponding phase-contrast images of the
515P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520
corresponds to the size of CD146 monomer and the 200-kDa
corresponds to the size of CD146 dimer. No cross-reaction was
detected with a secondary HRP-conjugated anti-rabbit IgG (Fig.
1B, D, F), and also no bands were seen in the immunoprecipita-
tions using mIgG as negative control (data not shown),
indicating that the CD146 dimer and monomer were specifically
recognized by the two different anti-CD146 antibodies. These
results demonstrate that the CD146 dimer and its monomer
were both present in all CD146-expressing cells tested, whether
they were tumor cell (A375), normal cell (ECV304) or the cells
transfected with a CD146 expression plasmid (CD146-positive
Mel888). Interestingly, we also observed that the dimer/
monomer ratio in melanoma cell line A375 and CD146-positive
Mel888 was significantly higher than that in the normal cell line
ECV304 (Fig. 1G), implying that dimerization of CD146 may
be associated with malignancy.
3.2. Localization of fluorescence-fused CD146 in living cells
In order to study the localization and dimerization of CD146
in living cells, we prepared YFP and CFP-tagged CD146
constructs (CD146YFPand CD146CFP) (Fig. 2). The CD146
negative Mel888 cells were pairwise co-transfected with these
constructs, and the clones co-expressing CD146YFP and
CD146CFPwere selected. Expression of CD146 fusion proteins
in these doubly transfected cells was detected under a confocal
laser scanning microscope. Fig. 3 shows two representative
double transfected clones. A high degree of peripheral
fluorescence was seen on the cell surface and no fluorescence
was detected in either cytosol or nucleus, indicating that the
fluorescence-tagged CD146 presented the same cellular loca-
lization as its native status. In addition, these images indicated
that CD146CFP(Fig. 3A) and CD146YFP(Fig. 3B) were co-
expressed within the same population of cells. To further
confirm that the yellow and cyan fluorescence on the cell
surface represent the YFP- and CFP-tagged CD146, but not
YFP and CFP alone, we stained these cells using mAb AA98
combined with cy-chrome dye giving a red fluorescence (Fig.
3C). The confocal images of either CFP or YFP fluorescence
merged with AA98 cy-chrome fluorescence (Fig. 3D and E)
indicate that the CD146-tagged fusion proteins are co-expressed
on cell membrane.
Fig. 4. FRET measurement. The Mel888 cells were transfected with CD146YFPand CD146CFPindividually and 48 h later, the cells were visualized under a confocal
microscope with optimal filter sets. Cells containing CD146CFPwere observed with excitation of CFP and collection CFP image (A), with no crossover with the YFP
channel (B). Conversely, in the cells containing CD146YFP, YFP was excited and visualized in the YFP image only (E) without crossover with the CFP channel. The
cells containing both CD146YFPand CD146CFPwere observed at the excitation of the donor (CFP) and collection of both donor (CFP) (G) and acceptor (YFP)
emission (H) using the optimal FRET settings. The corresponding phase-contrast images of these cells are presented in (C), (F) and (I).
516 P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520
3.3. FRET imaging CD146 dimerization in living cells
In order to visually monitor the dimerization of CD146 in
living cells, FRET measurements were performed using a
fluorescence microscope system. We first optimized the filter
settings to make sure no crossover excitation of YFP occurred
when CFP was excited. For this purpose, we transfected either
CD146YFPor CD146CFPinto Mel888 cells individually. The
results from microscope images are shown in Fig. 4. The
CD146CFPalone transfectants showed a cyan fluorescence with
excitation at 458 nm and emission at 465–495 nm, and no
fluorescence signal was detected in the bandwidth of 565–
595 nm (CD146YFPchannel) (Fig. 4, panel A and B). Similarly,
the cell transfected CD146YFPalone showed a yellow color with
excitation at 515 nm and emission at 565–595 nm without
crossover to the CFP channel (emission at 465–495 nm) (Fig. 4,
panel D and E). At the optimal FRET setting, excitation of
CD146CFP(donor) at 458 nm produced a diffuse image of
CD146YFP(acceptor) at 565–595 nm,indicating that a transfer
of energy was occurred from CD146CFPto CD146YFP.
Another measure to ensure the specificity of FRET was
acceptor photobleaching. The CD146YFPas acceptor fluoro-
phore was selectively photobleached for 2 min at 515 nm and
then the fluorescence intensity of the donor CD146CFPwas
measured. As shown in Fig. 5, the sharp drop in the
fluorescence intensity of the acceptor YFP (Fig. 5A and B)
was compensated for by the increase in the intensity of the
donor CFP (Fig. 5C and D). This increase was due to the lack of
acceptor absorbing the energy liberated by the donor, which in
turn leads to a higher emission from the donor. Moreover, we
calculated the FRET efficiency as 20.1%. Based on the above
observations, we concluded that the FRETsignal observed upon
co-expression of CD146YFP and CD146CFP resulted from
specific dimerization of CD146 in living cells.
3.4. Regulation of CD146 dimerization by A375-CM and mAb
We have previously observed tumor conditional medium
induced endothelial cell adhesion and migration. However,
these cell activities were inhibited by mAb AA98. To test the
molecular organization of CD146 on the cell surface under the
influence of extracellular stimulators, we used tumor condi-
tional medium (A375-CM) as stimulator to culture ECV304
cells, and then visualized the CD146 dimer/monomer ratio in
the treated cells using immunoprecipitation and Western
blotting analysis. The results showed that CD146 dimerization
was greatly increased in the endothelial cells after treated with
A375-CM. Interestingly, the higher level of CD146 dimeriza-
tion was dramatically reduced by mAb AA98, regardless of
whether CD146 dimerization induced by A375-CM in
endothelial cells or naturally existed in melanoma A375 cells
(Fig. 6). These results demonstrate that the CD146 dimerization
can be up-regulated by A375-CM and down-regulated by mAb
3.5. CD146 dimerization is associated with the NF-kappa B
It has been shown that many tumors secret several
cytokines which trigger signaling and induce NF-kappa B
activation, resulting in changes in cell behavior. NF-kappa B is
Fig. 5. Acceptor photobleaching. In CD146YFPand CD146CFPconstructs transfected Mel888 cells, the acceptor (YFP) was photobleached by repeated scanning with
the 515-nm laser. (A) and (B) present pre- and post-bleached images of the acceptor. (C) and (D) are corresponding donor (CFP) images in the same cell. (E) shows the
corresponding phase-contrast images of the cell. (F) Quantification of FRET efficiency (E, %) using the acceptor photobleaching method. The FRET efficiency was
517 P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520
an inducible transcription factor that belongs to the Rel family.
In its resting state, NF-kappa B complexed with IκBα exists in
cytoplasm. When a cell receives certain extracellular signals,
IκBα is rapidly phosphorylated and degraded via the
ubiquitin-proteasome pathway, and NF-kappa B is redistrib-
uted from the cytosol into the nucleus, regulating the
expression of many genes. Based on previous studies, we
wondered whether the A375-CM enhanced CD146 dimeriza-
tion might be associated with NF-kappa B activation. To test
this, we analyzed the NF-kappa B nuclear translocation in
ECV304 cells after A375-CM treatment. The fluorescence
microscope images using anti-NF-kappa B antibody combined
with FITC are shown in Fig. 7A. We found that NF-kappa B
was presented in the cytosol of ECV304 cells when cultured
in normal DMEM medium. After stimulation with A375-CM,
the NF-kappa B was redistributed from the cytosol into the
nuclei. However, an NF-kappa B inhibitor, BAY11-7082,
suppressed the nuclear translocation of NF-kappa B which
was induced by A375-CM.
These observations were further confirmed using Western
blotting analysis as shown in Fig. 7B. In ECV304 cells grown
in normal medium, most IκB was found in the cytosol
fraction and very little NF-kappa B was found in the nuclear
fraction. However, after stimulation with A375-CM, the NF-
kappa B in the nuclear fraction was dramatically increased
and IκB in the cytosol fraction was decreased. Similarly, as
seen in Fig. 7A, the amount of NF-kappa B in the nuclear
fraction was reduced by an NF-kappa B inhibitor. These
observations indicate that NF-kappa B activation in endothe-
lial cells is regulated by tumor conditional medium from
Interestingly, we found that the NF-kappa B activation was
associated with CD146 dimerization. As shown in Fig. 7C, the
majority of CD146 is in the monomeric form in normal
ECV304 cells. After treatment with A375-CM, the dimer/
monomer ratio in ECV304 cells was significantly increased.
However, this increased CD146 dimerization could be sup-
pressed by the NF-kappa B inhibitor, BAY 11-7802. Whereas
the BAY 11-7802 did not dissociate the CD146 dimer existed in
the ECV304 without A375-CM treatment (data no shown). This
suggests that CD146 dimerization is triggered by A375-CM and
regulated through the NF-kappa B pathway.
In this study, we provided the first evidence that CD146
dimerization occurs in living cells by demonstrating that (1) the
CD146 dimer (200 kDa) was specifically pulled-down with
mAb AA98, and recognized by anti-CD146 rabbit antibody, (2)
A specific FRETsignal was observed in living cells which were
co-expressing CFP/YFP-tagged CD146, indicating CD146
homophilic interaction. This observation confirms the occur-
rence of CD146 dimerization in living cells.
Another important finding in this study is that CD146
dimerization was regulated by anti-CD146 mAb AA98 and
A375-CM through the NF-kappa B pathway. However, the
molecular organization of CD146 in the presence of stimulator
or inhibitor remained unclear. Here, we provided several lines
of evidence to support the hypothesis that CD146 dimerization
is regulated by mAb AA98 and A375-CM through the NF-
kappa B pathway. First, we found that the dimer and monomer
of CD146 both existed in the cells tested and the dimer/
monomer ratio in tumor cells (A375) was significantly higher
than that in normal endothelium. Second, CD146 dimerization
in endothelial cells was up-regulated by A375-CM. After
culture of the endothelial cells in A375-CM for 2 days, the
amount of the dimeric form of CD146 was greatly enhanced.
In contrast, we found that mAb AA98 inhibited the di-
merization of CD146 in melanoma A375 cells. These results
strongly suggest that CD146 dimerization is a dynamic process
in living cells and responds to the stimuli in the tumor micro-
environment and to inhibitors like mAb AA98. Finally, A375-
CM-induced CD146 dimerization could be reduced by an NF-
kappa B inhibitor BAY 11-7082, indicating that CD146
dimerization up-regulation by A375-CM is through the NF-
kappa B pathway.
Many factors in tumor conditional medium, such as TNF-α,
can activate the NF-kappa B pathway, inducing the expression
of many genes involved in tumor progression. Our previous
studies have demonstrated that tumor conditional medium up-
regulated the CD146 expression and induced the migration and
Fig. 6. CD146 dimerization is enhanced by A375-CM but suppressed by mAb
AA98. (A) ECV304 cells were treated with or without AA98 and A375-CM,
and with only one of each. (B) A375 cells were treated with or without mAb
AA98. The cell lysates were immunoprecipitated with monoclonal anti-CD146
antibody, AA98, followed by Western blotting with rabbit anti-CD146 antibody.
The ratio of the CD146 dimer and monomer were qualified by densitometry
using the data from three independent experiments (C and D).
518 P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520
angiogenesis of endothelial cells, whereas these cell activities
were suppressed by mAb AA98 via inhibition of NF-kappa B
[8,20]. Here we found that A375 conditional medium induced
the CD146 dimerization though the NF-kappa B pathway. And
the mAb AA98 and NF-kappa B inhibitor both dissociated
CD146 dimerization. Based on these observations, we
hypothesize that mAb AA98 inhibits NF-kappa B translocation
via its suppression of CD146 dimerization. However, it is still
not clear which is the first event, CD146 dimerization or NF-
kappa B activation, inhibited by mAb AA98. Another
explanation might be that certain factors in A375 conditional
medium activate NF-kappa B, resulting in the expression of
many genes, possibly including the ligand of CD146. As a
result, the increased amount of ligand may engage with the
CD146 molecule and further induce its dimerization. Like
many other membrane receptors , CD146 dimerization
may initiate signaling transduction since the CD146 molecule
possesses potential recognition sites for protein kinases in its
cytoplasmic domain. The inhibitory effect of mAb AA98 on
CD146 dimerization might be due to blocking interaction
between CD146 and its ligand. Although the ligand of CD146
has not been identified, there is evidence that many melanoma
cells express the CD146 ligand , and CD146 mediated
cell–cell adhesion through its unidentified ligand may
represent an important role in cell biological function, and
perhaps in tumor growth and metastasis. It will be of interest to
identify the CD146 ligand and better understand the role of
CD146 in signaling.
In summary, this study provides the first evidence that
CD146 dimerization occurs in living cells and is regulated by
the tumor microenvironment through the NF-kappa B pathway.
The dynamic nature of CD146 dimerization may be associated
with tumor malignancy.
We thank Dr. J. Johnson for supplying the CD146-
transfected Mel 888 cells, the parental Mel 888 cells and
CD146 gene contained in pUC-CD146 plasmids. We also thank
Dr. Sarah Perrett for helpful suggestions and critical reading of
the manuscript. This work was supported by the National 863
grant, the National Natural Sciences Foundation, a Chinese
Academy of Sciences grant and the Program of Founding
Research Centers for Emerging and Reemerging Infectious
Diseases by the Ministry of Education, Culture, Sports, Science
and Technology (MEXT) of Japan.
 C.L. Holness, D.L. Simmons, Structural motifs for recognition and
adhesion in members of the immunoglobulin superfamily, J. Cell Sci. 107
 J.P. Johnson, U. Rothbacher, C. Sers, The progression associated antigen
MUC18: a unique member of the immunoglobulin supergene family,
Melanoma Res. 3 (1993) 337–340.
 J.M. Lehmann, G. Riethmuller, J.P. Johnson, MUC18, a marker of tumor
progression in human melanoma, shows sequence similarity to the neural
cell adhesion molecules of the immunoglobulin superfamily, Proc. Natl.
Acad. Sci. U. S. A. 86 (1989) 9891–9895.
 J.M. Lehmann, B. Holzmann, E.W. Breitbart, P. Schmiegelow, G.
Riethmuller, J.P. Johnson, Discrimination between benign and malignant
cells of melanocytic lineage by two novel antigens, a glycoprotein with a
Fig. 7. CD146 dimerization is associated with NF-kappa B pathway. (A) Immunofluorescence of the ECV304 cells stained with anti-p65 antibody and observed under
a confocal laser scanning microscope. NF-kappa B translocation was induced by A375-CM and suppressed by the NF-kappa B inhibitor, BAY 11-7082. No NF-kappa
B nuclear translocation was detected in the ECV304 cultured in DMEM medium. This observation was confirmed by Western blotting (B). The NF-kappa B in the
nuclear extracts was detected by anti-p65 antibody, whereas the IκB and β-actin as control in the cytosol extracts were detected by anti-IκB and anti-β-actin antibody.
(C) The immunoprecipitations with monoclonal anti-CD146 antibody, AA98 were stained using rabbit anti-CD146 antibody. Western blotting showed that the CD146
dimerization in the ECV304 cells was enhanced by A375-CM and suppressed by the NF-kappa B inhibitor, BAY 11-7082. (D) The ratio of the CD146 dimer and
monomer were qualified by densitometry using the data from three independent experiments.
519 P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520
molecular weight of 113000 and a protein with a molecular weight of
76000, Cancer Res. 47 (1987) 841–845.
 S. Xie, M. Luca, S. Huang, M. Gutman, R. Reich, J.P. Johnson,M. Bar-Eli,
Expression of MCAM/MUC18 by human melanoma cells leads to
increased tumor growth and metastasis, Cancer Res. 57 (1997)
 K. Satyamoorthy, J. Muyrers, F. Meier, D. Patel, M. Herlyn, Mel-CAM-
specific genetic suppressor elements inhibit melanoma growth and
invasion through loss of gap junctional communication, Oncogene 20
 L. Mills, C. Tellez, S. Huang, C. Baker, M. McCarty, L. Green, J.M.
Gudas, X. Feng, M. Bar-Eli, Fully human antibodies to MCAM/MUC18
inhibit tumor growth and metastasis of human melanoma, Cancer Res. 62
 X. Yan, Y. Lin, D. Yang, Y. Shen, M. Yuan, Z. Zhang, P. Li, H. Xia, L. Li,
D. Luo, Q. Liu, K. Mann, B.L. Bader, A novel anti-CD146 monoclonal
antibody, AA98, inhibits angiogenesis and tumor growth, Blood 102
 C.B. St, C. Rago, V. Velculescu, G. Traverso, K.E. Romans, E.
Montgomery, A. Lal, G.J. Riggins, C. Lengauer, B. Vogelstein, K.W.
Kinzler, Genes expressed in human tumor endothelium, Science 289
 N. Bardin, F. Anfosso, J.M. Masse, E. Cramer, F. Sabatier, B.A. Le, J.
Sampol, F. gnat-George, Identification of CD146 as a component of the
endothelialjunction involvedin the control of cell–cell cohesion,Blood 98
 B. Chan, S. Sinha, D. Cho, R. Ramchandran, V.P. Sukhatme, Critical roles
of CD146 in zebrafish vascular development, Dev. Dyn. 232 (2005)
 Q. Liu, X. Yan, Y. Li, Y. Zhang, X. Zhao, Y. Shen, Pre-eclampsia is
associated with the failure of melanoma cell adhesion molecule (MCAM/
CD146) expression by intermediate trophoblast, Lab. Invest. 84 (2004)
 I.M. Shih, R.J. Kurman, Expression of melanoma cell adhesion molecule
in intermediate trophoblast, Lab. Invest. 75 (1996) 377–388.
 Q. Liu, X. Zhao, Y. Zhang, Y. Shen, Y. Liu, X. Yan, Melanoma cell
adhesion molecule (MCAM/CD146) is a critical molecule in trophoblast
invasion, Prog. Biochem. Biophys. 34 (2004) 309–312.
 A.N. Solovey, L. Gui, L. Chang, J. Enenstein, P.V. Browne, R.P. Hebbel,
Identification and functional assessment of endothelial P1H12, J. Lab.
Clin. Med. 138 (2001) 322–331.
 F. Anfosso, N. Bardin, V. Frances, E. Vivier, L. Camoin-Jau, J. Sampol, F.
gnat-George, Activation of human endothelial cells via S-endo-1 antigen
(CD146) stimulates the tyrosine phosphorylation of focal adhesion kinase
p125 (FAK), J. Biol. Chem. 273 (1998) 26852–26856.
 I. Hunter, H. Sawa, M. Edlund, B. Obrink, Evidence for regulated
dimerization of cell–cell adhesion molecule (C-CAM) in epithelial cells,
Biochem. J. 320 (1996) 847–853.
 T. Zhao, P.J. Newman, Integrin activation by regulated dimerization and
oligomerization of platelet endothelial cell adhesion molecule (PECAM)-1
from within the cell, J. Cell Biol. 152 (2001) 65–73.
 M. Vicente-Manzanares,M. Rey, D.R. Jones,D. Sancho,M. Mellado,J.M.
Rodriguez-Frade, M.A. del Pozo, M. Yanez-Mo, A.M. de Ana, A.
Martinez, I. Merida, F. Sanchez-Madrid, Involvement of phosphatidyli-
nositol 3-kinase in stromal cell-derived factor-1 alpha-induced lymphocyte
polarization and chemotaxis, J. Immunol. 163 (1999) 4001–4012.
 P. Bu, L. Gao, J. Zhuang, J. Feng, D. Yang, X. Yan, Anti-CD146
monoclonal antibody AA98 inhibits angiogenesis via suppression of
nuclear factor-kappaB activation, Mol. Cancer Ther. 5 (2006) 2872–2878.
 S. Bishayee, S. Majumdar, J. Khire, M. Das, Ligand-induced dimerization
of the platelet-derived growth factor receptor. Monomer-dimer intercon-
version occurs independent of receptor phosphorylation, J. Biol. Chem.
264 (1989) 11699–11705.
 I.M. Shih, D. Speicher, M.Y. Hsu, E. Levine, M. Herlyn, Melanoma cell–
cell interactions are mediated through heterophilic Mel-CAM/ligand
adhesion, Cancer Res. 57 (1997) 3835–3840.
520 P. Bu et al. / Biochimica et Biophysica Acta 1773 (2007) 513–520