| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Tumor Biology |
Institut National de la Santé et de la Recherche Médicale U331, Faculty of Medicine René Laënnec, 69372 Lyon Cedex 08 [H. B., P. B.]; Department of Pathology, Anti-Cancer Center Léon Bérard, 69373 Lyon Cedex 08 [E. T.]; Department of Dermatology, Hôpital de l’Antiquaille, 69321 Lyon Cedex 08 [F. B., B. B., F. W., H. P., L. T.]; and Institut National de la Santé et de la Recherche Médicale U80, Faculty of Medicine René Laënnec, 69372 Lyon Cedex 08 [K. S.], France
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
vß3, a subtractive immunization
approach was used to elicit host antibody response against cell surface
proteins expressed on metastatic variants. In this study, a new
monoclonal antibody (MoAb), LY1, that is highly specific for the 7GP
and T1P26 variants, was isolated. LY1 identifies a membrane protein of
Mr 55,000 on melanoma variants with epitopes
that were resistant to sugar-cleaving enzymes. Immunostaining cells
from variants by LY1 showed that staining is distributed to the cell
periphery with high labeling intensity at the cell-to-cell contact
points. This MoAb significantly inhibited invasion of metastatic
variants through a reconstituted basement membrane (Matrigel) in
vitro. Moreover, tumor growth of melanoma variants was
dramatically affected in vivo with this MoAb. In
vitro studies indicate that the LY1 MoAb does not inhibit
chemotactic migration of the metastatic variants, the adhesion of tumor
cells to vitronectin, collagen IV, fibronectin, and laminin, or cell
proliferation. Expression of this antigen is high in human striated
muscle, heart, spleen, brain, and lung and absent in kidney, liver, and
pancreas. Using 59 fixed, paraffin-embedded archival tissues of human
melanomas and nevi, LY1-reactive cells were not observed in
melanocytes, nevi, or radial growth phase primary melanomas. In sharp
contrast, LY1 selectively stained melanocytes derived from the vertical
growth phase of many primary melanomas and metastatic melanomas. These
results provide evidence that the Mr 55,000
protein expressed by selected variants with increased metastatic
properties in vivo plays a functionally important role
in determining metastasis. This molecule may represent a new metastatic
risk marker in human melanoma and may be of biological importance in
the identification of fatal metastatic subpopulations that have
acquired competence for metastasis production. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Benign to malignant progression results from the sequential emergence
of a series of aggressive subpopulations of cells that preexisted in
the parent tumor (3
, 4)
. Such metastatic variants differ
from the parental tumor in a number of properties such as transcription
factors, cell surface enzymes, motility, signaling, angiogenic factors,
and cell adhesion molecules (5, 6, 7, 8)
. Recently, receptors
that mediate cell-to-cell and cell-to-substratum adhesion were shown to
be key components in the metastatic cascade (9, 10, 11, 12)
. An
integrin such as vß3
has been shown to be expressed by an increasing proportion of primary
melanomas in the transition from
RGP4
to VGP, suggesting that this integrin plays an active role in melanoma
progression (13
, 14)
. Likewise, the intercellular adhesion
molecule 1, MUC18, HLA-DR, and gangliosides
GD2 and
GD3 exhibit enhanced
expression on melanoma cells as they progress to a more metastatic
phenotype (15, 16, 17, 18, 19)
. However, the absence of
expression of many of these antigens was frequently observed on tissue
sections (13
, 14
, 20)
, suggesting that additional tumor
cell markers that are indicative for metastatic capacity are needed.
Using a metastasis model, which mimics the early events of metastasis
in humans, we recently selected from a poorly
vß3-positive
metastatic human melanoma cell line, M4Beu, a
whole series of
vß3-negative variants
that generate a higher frequency of spontaneous metastases into
immune-suppressed neonatal rats and aggressively proliferate in
vivo after s.c. implantation into athymic nude mice
(21)
. These results suggest that expression of such a
phenotype can confer metastatic ability to variants that have emerged
from the in vivo selection pressure. Such dramatic changes
in the expression of a major integrin on the surface of biologically
aggressive subpopulations of tumor cells prompted us to search for the
nature of the cell membrane molecules that mediated the metastatic
ability of tumor cells.
In this study, a chemical immunosuppression approach (22) was used to raise monoclonal antibodies to functional cell surface antigens that showed minimal expression on the parental melanoma cell line M4Beu. but that significantly enhanced on metastatic melanoma variants. Immunosuppression was achieved by the use of the tolerizing cytotoxic drug cyclophosphamide, which reduces the recognition of the common immunodominant cell surface antigens present on the two phenotypically distinct melanoma cell types by selectively killing proliferating clones of B cells (23) . This approach has proved to be a powerful technique for the development of function-blocking mouse monoclonal antibodies directed against less immunodominant proteins such as maturation-specific sperm surface molecules, cerebellar neuron, nasal retinal axons, and axonal proteins expressed by premigratory, migrating retinal ganglion cells (22) and to specific-tumor associated antigens (24) .
We first immunized mice with a poorly metastatic M4Beu. melanoma cell line. The immunosuppressive drug cyclophosphamide was then used to tolerize mice to M4Beu. followed by multiple immunizations with a highly metastatic melanoma variant, T1P26. These immunization procedures enabled us to produce several potential MoAbs that have a selective ability to bind to highly metastatic tumor variants. Here, we report the characterization of one MoAb (LY1) that recognizes a Mr 55,000 surface antigen (p55), which plays a functional role in the process of tumor progression and metastasis. Expression of p55 was preferentially detected in primary melanomas in the VGP as well as metastatic lesions but not on melanocytes, nevi, or primary melanomas in the RGP. This cell surface protein of Mr 55,000 may represent a new tumor progression and metastatic risk marker in human melanoma.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 (clone HP2B6), 2
(Gi9), 3 (clone M-KID 2),
4 (HP2/1), 5 (SAM1),
6 (GoH3), v (clone
AMF7), and ß3 (clone SZ21) integrins were
purchased from Immunotech (Marseille, France). Other reagents were
obtained as follows: NuSerum, Matrigel basement membrane matrix, and
Biocoat cell culture inserts were from Becton Dickinson Labware
(Bedford, MA); cyclophosphamide was from Sigma (St. Louis, MO),
prestained high molecular weight protein marker and horseradish
peroxidase color development reagent were from Bio-Rad (Hercules, CA);
protease inhibitor mixture was from ICN (Costa Mesa, CA); Hybond-C and
[3H]thymidine were from Amersham Pharmacia
Biotech (Little Chalfont, United Kingdom); a cell-proliferating kit,
N-glycosidase F, recombinant O-glycosidase (from
Diplococcus pneumoniae), and neuraminidase (from
Arthrobacter ureafaciens) were from Boehringer Mannheim
(Indianapolis, IN); and a RAL 555 staining kit was from Reactif RAL
(Paris, France).
Cell Culture and Selection of
vß3-deficient Cells
The human M4Beu. cell line was derived
from a lymph node metastasis of a patient with malignant melanoma
(25)
. M4Beu. cells expressing
vß3
(vß3-positive) are
unable to give lung metastases after orthotopic injection of the
cultured cells into immunosuppressed newborn rats (low incidence of
spontaneous lung colonization; Ref. 21
). The
vß3-negative
metastatic variant, designated T1P26 (high incidence of spontaneous
lung colonization), was established in culture after two direct
successive orthotopic transplantations of M4Beu.
tumors into immunosuppressed newborn rats (21)
. The
vß3-negative 7GP (high
incidence of spontaneous lung colonization) metastatic variant was
selected in vivo through eight serial transplantations of
lymph node metastases of the M4Beu. cell line
into immunosuppressed newborn rats (21)
. Human melanoma
cell lines T1P26L and T1P26R (high incidence of spontaneous lung
colonization) are clones that were sorted from T1P26 cells on the basis
of their ability to bind the peanut agglutinin lectin using a
fluorescence-activated cell sorter (26)
. These melanoma
cell line variants have a doubling time (30 ± 0.5 h) similar to that of the parent M4Beu.
Cell Line
The cells were cultured as monolayers in RPMI 1640 or McCoy’s
5A medium supplemented with 10% fetal bovine serum as described
previously (21)
. Cultures were routinely checked and found
free of Mycoplasma contamination using the Hoechst 33258
fluorescence staining procedure. To minimize the possibility of
phenotypic drift, variants were maintained in culture for 4–6-week
intervals after which they were replaced with new frozen stock.
Passages 26–40, counted from the first passage of the original
variant, were used in these studies. SK-MEL-2 and SK-MEL-24 human
melanomas, MG-63 human osteosarcoma, and MCF-7 human breast carcinoma
cell lines were obtained from the American Type Culture Collection
(Manassas, VA). The DEV human medulloblastoma cell line was a gift from
Dr. B. Jacquemont (INSERM U433, Lyon, France; Ref.
27
). IAP-negative, vector-transfected OVO10 and
IAP-transfected OVO10+ ovarian carcinoma cell
lines were kindly supplied by Dr. F. P. Lindberg (Washington
University School of Medicine, St. Louis, MO; Ref. 28
).
Subtractive Immunization
The technique and schedule used to immunize mice were as
described previously (22)
. Briefly, subconfluent monolayer
cultures of poorly metastatic melanoma cells were detached by the
addition of nonenzymatic cell dissociation solution and washed twice in
sterile PBS. The washed cells were resuspended in the same buffer and
injected i.p. (2 x 106 cells per
animal) into 8-week-old female BALB/c mice. Twenty-four and 48 h
later, these mice were injected i.p. with 200 mg/kg cyclophosphamide
diluted in sterile PBS. Two weeks later, blood was collected, and the
serum was screened for the presence of anti-poorly metastatic melanoma
antibodies using an ELISA technique (29)
. Three days after
the last injection, cells from the metastatic variant were detached
from the culture plate as described above, and 2 x 106 cells were inoculated i.p. into mice.
Boosting was performed 3 weeks later by injecting these animals i.p.
with 3 x 106 metastatic variant
cells. Three days after the last injection, sera were collected and
screened for differential binding toward the poorly metastatic and
metastatic cells by FACScan (Becton Dickinson). The next day,
the mouse with a selective differential immune response to highly
metastatic versus poorly metastatic melanoma cells was
killed by cervical dislocation, and the spleens from the hosts were
taken for fusion with the BALB/c myeloma SP2/0Ag14. A control study was
performed with animals inoculated with the poorly metastatic parental
cell line and metastatic variants in the absence of cyclophosphamide.
Production of Monoclonal Antibodies
Details of the fusion procedure were as described previously
(30)
using standard procedures (31)
.
Supernatants from the wells containing growing hybrids were screened
for their reactivity to highly metastatic variants and poorly
metastatic parental cell line using ELISA and FACScan techniques. The
positive clones were cloned by limiting dilution, and monoclonal
antibodies were purified from ascitic fluid by gel filtration on
Sephacryl S-300. Classes and subclasses of the monoclonal antibodies
were determined by the Amersham isotyping kit (Amersham International,
les Ullis, France) following the instructions given by the
manufacturer.
Screening of Monoclonal Antibodies
In the first stage of screening, an ELISA was used to evaluate
antibody binding to poorly metastatic and highly metastatic melanoma
cells (29)
. Briefly, melanoma cells were harvested with
nonenzymatic cell dissociation buffer and added (2 x 104
/0.1 ml) to each microtiter well containing 50
µg/ml poly-L-lysine. Fifty to 100 µl of hybridoma
culture supernatant were added to wells and incubated for 1 h at
37°C. Bound antibodies were detected after sequential addition of
peroxidase-conjugated rabbit anti-mouse IgG and IgM (heavy and light
chain specific; Dakopatts, Copenhagen, Denmark) and the substrate
1,2-phenylendiamin. Nonspecific binding was performed with preimmune
sera and subtracted from the total binding. In the second stage of
screening, positive hybridoma culture supernatants were analyzed by
flow cytometry as described previously (21)
using either
culture medium or preimmune sera as a control. Of 66 culture
supernatants tested, 8 (LY1–LY8) hybridomas were found by FACScan
containing antibodies directed against the metastatic variant and not
to the poorly metastatic parental cell line. One of these monoclonal
antibodies, designated LY1, was chosen for its strong reactivity and
was used for further studies. For immunocytochemical staining, cells
grown on 12-well multitest slides for 48 h were fixed for 1 h
in cold methanol before the addition of the first MoAb or normal mouse
serum as described previously (10)
. Thereafter, the cells
were washed extensively with Ca2+- and
Mg2+-free PBS followed by sequential incubation
with the biotinylated antimouse immunoglobulin and streptavidin
conjugated to alkaline phosphatase using the DAKO (Glostrup, Denmark)
LSAB+ kit. Staining
was completed after incubation with the naphtol-fuchin solution.
Western Blot Analysis
Cells were lysed in Lubrol lysis buffer [20 mM Tris
(pH 7.4), 0.15 M NaCl, and 1% Lubrol] containing a
protease mixture cocktail according to the instructions of the
manufacturer (ICN). The lysates were centrifuged at 12,000 x g at 4°C for 15 min and separated by
electrophoresis in a 10% polyacrylamide gel containing SDS according
to the method of Laemmli (32)
. The proteins were then
electrotransferred to nitrocellulose paper (Hybond-C). The nonspecific
binding was blocked by soaking the electrophoretic blots for 1 h
in Tris-buffered saline [0.02 M Tris-HCl (pH
7.6), and 0.15 M NaCl] containing 3% BSA. The
membranes were incubated with purified MoAb for 1 h. Bound
antibodies were detected after sequential addition of a goat antimouse
IgM antibody conjugated to peroxidase (1:2000 dilution, Dakopatts) and
a horseradish peroxidase color development reagent obtained from
Bio-Rad. Control experiments were performed with a nonimmune mouse
serum or an irrelevant mouse IgM isotype antibody.
Enzyme Treatment
The cell lysates were subjected to enzyme digestion according to
the instructions of the manufacturer (Boehringer Mannheim). Briefly, 20
µg of the lysates were heat denatured in the presence of 1% SDS and
diluted to 0.1% SDS. Denatured samples were then treated with 4
milliunits of neuraminidase, 5 milliunits of O-glycosidase F
alone, 4 units of N-glycosidase alone, or
N-glycosidase in combination with O-glycosidase F
for 17 h at 37°C before the Western blot analysis. Control
samples were incubated without enzymes.
Adhesion to Extracellular Matrix Proteins
The cell adhesion assay was performed as described previously
(33)
. Briefly, 1–2 µg of fibronectin, laminin, collagen
IV, or vitronectin diluted in PBS (pH 7.2) were adsorbed on each well
of the microtiter plate for 1 h at 37°C. Cells
(104
) resuspended in serum-free RPMI 1640
containing 0.35% BSA were then added to coated wells after 20 min of
incubation with purified MoAb (10 µg/106
cells) or with an irrelevant mouse IgM isotype antibody. After a
further incubation at 37°C for 60–90 min, nonadherent cells were
removed by gently washing the wells. One hundred µl of medium and 10
µl of MTT labeling reagent (0.5 mg/ml) were added to each well and
incubated for 4 h at 37°C following the manufacturer’s
instructions (Boehringer Mannheim). Then 100 µl of 10% SDS in 0.01
M HCl were added for 12 h. The solubilized formazan
product in the wells was spectrophotometrically quantified using an
ELISA reader.
Migration (Chemotaxis) Assay
Chemotaxis was assayed using uncoated 24-well BioCoat cell
culture inserts (Becton Dickinson Labware, MA) with an
8-µm-porosity polyethyleneterepthylate membrane. Briefly, cells were
removed with nonenzymatic cell dissociation buffer and resuspended at
5 x 105
/ml in serum-free RPMI
1640 containing 0.35% BSA. Cells (5 x 104
) pretreated for 20 min with 5–10 µg/ml
MoAb or an irrelevant mouse IgM isotype antibody were then added to the
upper compartments of the BioCoat chambers (6.25-mm membrane size). The
wells of the lower chamber were filled with RPMI 1640 containing 10%
NuSerum, and the chambers were each assembled by placing the uncoated
membrane between the lower and upper compartments according to the
manufacturer’s instructions (Becton Dickinson). The migration assay
was carried out at 37°C for 15 h, after which the filters were
removed, fixed, and stained with the RAL 555 staining kit. Cells on the
upper surface of the filters were removed by wiping with a cotton swab,
and migration was determined by counting of the cells that had migrated
to the lower side of the filter with a microscope at x100
magnification. Experiments were assayed in triplicate, and at least
eight fields were counted in each experiment.
Invasion Assay
Invasive ability of the melanoma cells in vitro was
carried out as described previously (34)
with
modifications. Invasion was measured by using 24-well BioCoat cell
culture inserts with an 8-µm-porosity polyethyleneterepthylate
membrane coated with Matrigel basement membrane matrix (100
µg/cm2). Briefly, the Matrigel was allowed to
rehydrate for 2 h at room temperature by adding warm, serum-free
RPMI 1640. The wells of the lower chamber were filled with RPMI 1640
containing 10% NuSerum, and the chambers were each assembled similarly
to the method described above for migration assay. Cells
(105
) were seeded in the upper compartment
(6.25-mm membrane size) in serum-free RPMI 1640. Cells were treated
with 0.5–1 µg of either MoAb or an irrelevant mouse IgM isotype
antibody for 20 min before the invasiveness assay. The invasion assay
was carried out at 37°C in a 5% CO2 humidified
incubator for 48–72 h. Fresh antibody was added each day. At the end
of the invasion assay, filters were processed and quantitated as for
the migration assay. Experiments were assayed in triplicate, and at
last 10 fields were counted in each experiment.
Tumorigenicity in Nude Mice
Swiss nude mice (nu/nu) were purchased from
IFFA-CREDO (Arbresle, France). Subconfluent cultures of melanoma
cells were harvested with nonenzymatic cell dissociation buffer, washed
three times with serum-containing medium, and then resuspended in PBS
(pH 7.2). Two hundred µl of cells (2 x 106) were inoculated s.c. on the belly of nude
mice as described previously (10)
, after preincubation for
20 min with either MoAb (16 µg/106
cells) or an isotype control mouse IgM. Tumor growth rate was
determined by weekly measure of two perpendicular diameters of the
tumor. All experimental groups contained five mice. The effect of the
MoAb on cell growth, viability, and cytotoxicity was measured by using
[3H]thymidine uptake or the MTT cell
proliferation kit following the manufacturer’s instructions
(Boehringer Mannheim, IN) as described previously
(33)
.
Immunohistochemical Analysis of Normal and Melanoma Tissues
Tissues Studied.
Histopathological sections of the following normal human adult tissues
were investigated: heart, brain, kidney, liver, lung, pancreas, spleen,
and muscle. Serial 5-µm sections of paraffin-embedded tissue were
obtained from Novagen (Madison, WI). Surgical biopsy material of
malignant lesions of melanocyte origin was obtained from the
Departments of Pathology and Dermatology at Hôpital de
l’Antiquaille. The lesions were classified by histopathological
examination of paraffin sections. Tumor thickness, Clark’s level of
invasion, and the growth phase of each lesion were determined on
hematoxylin-stained sections as described previously (35)
.
The depth of invasion of lesions used in this study ranged from levels
I to V, and the vertical tumor thickness varied between 0 (in
situ) and 10 mm. Primary cutaneous melanomas were considered to be
in RGP if the epidermal component was larger than the dermal component
and if atypical melanocytic cells lay beyond the basement membrane and
extended into the papillary dermis. Lesions that did not fulfill these
criteria were considered to be in VGP wherein tumor cells grow downward
into the underlying mesenchymally derived dermis. The tumor lesions
were classified as belonging to one of the following groups: benign
skin lesions including 16 nevi (8 dermal, 4 compound, 3 congenital, and
1 junctional; n = 16) and 2 dermatofibromas
(n = 2); 10 cutaneous melanomas in RGP (tumor
thickness 
0.75 mm; n = 10);
23 cutaneous melanomas that had entered VGP (tumor thickness 
1.0 mm; n = 23); and 8 metastatic
melanomas to lymph nodes (n = 8). Samples
that show both the RGP and VGP compartments were analyzed separately.
Immunohistological Labeling.
Immunochemistry on paraffin-embedded tissue sections (5–6 µm) was
performed using a catalyzed amplification system (Dako) following the
manufacturer’s instructions. Briefly, sections were incubated for
2 h with purified MoAb (1:500) or an isotype control mouse IgM.
Sections were then incubated serially with link antibody (biotinylated
rabbit antimouse immunoglobulin), streptavidin-biotin complex,
amplification reagent (biotinyl-tyramide and hydrogen peroxide), and
streptavidin-peroxidase. Bound MoAb was detected with AEC, which gives
rise to a red chromogen. These sections were counterstained with
Harris’s hematoxylin and finally mounted in aqueous mounting medium.
The percentage of stained cells was visually estimated by three
independent observers (B. B., E. T., and H. B.). Tumors were scored
positive if 10% or more reactive tumor cells showed staining with
MoAb, and nevi were scored positive if any reactive areas were seen.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. This MoAb was observed by immunochemistry to bind to the
surface of melanoma cell variants (Fig. 1B)
. This staining
appeared localized to the cell periphery with high labeling intensity
at the cell-cell contact sites (Fig. 1B)
.
|
, LY1 binds selectively to highly metastatic melanoma variants but
failed to bind to the poorly metastatic parental cell line
M4Beu. Moreover, this MoAb also stained SK-MEL-2
and SK-MEL-24, two human melanoma cell lines derived, respectively,
from lymph node or cutaneous metastases (Table 1)
. In contrast, no
staining was observed with this MoAb on human breast carcinoma,
osteosarcoma, and medulloblastoma cell lines and IAP-negative (OVO10)
and IAP-positive (OVO10+) ovarian carcinoma cell
lines. In addition, no significant difference in expression of
1, 2,
3, 4,
5 6, and
v was observed between the melanoma variants
(T1P26 and 7GP) and the parental cell line M4Beu. (data not
shown), except for the integrin subunit ß3,
which was previously shown to be expressed to a high level on parental
M4Beu. cells but was dramatically reduced to a
background level on the metastatic T1P26 and 7GP melanoma variants
(21)
.
|
, LY1 immunoblotted a Mr
55,000 major protein from lysates of metastatic variant T1P26 under
nonreducing or reducing conditions compared with an irrelevant mouse
IgM isotype antibody of the same class as LY1 (Lanes 1 and
2). Other faint immunoreactive protein bands migrating at
Mr 85,000,
Mr 110,000, and
Mr 150,000 were variably detected in
the sample of cell lysates (results not shown). In contrast, no
identical set of bands could be detected from lysates of the poorly
metastatic parental cell line M4Beu, providing further
evidence that LY1 is selectively directed against a
metastasis-associated antigen (Lane 3). Treatment of the
cell lysate with neuraminidase, with N-glycosidase alone, or
with N-glycosidase in combination with
O-glycosidase did not affect antibody binding (Fig. 2B)
.
|
, the invasive ability of the metastatic variant TIP26 cell line was
significantly affected (75% decrease in the invasive potential of the
cells) by the LY1 MoAb compared with an irrelevant mouse IgM isotype
antibody. In contrast, the invasive ability of the poorly metastatic
parental cell line M4Beu. was not significantly
affected in the presence of the LY1 MoAb (Fig. 3)
.
|
, melanoma variant T1P26 yielded readily
visible tumors as early as 15 days that gave rise to large tumors at 30
days. In contrast, with LY1 added, melanoma variant T1P26 gave rise to
small tumors growing at a lower rate than in the controls. Controls
included an irrelevant mouse IgM isotype antibody used at the same
concentration as that of LY1. To exclude the possibility that LY1 might
have a direct cytotoxic effect, cell variants were preincubated with
excess LY1 (10 µg/ml) and monitored for cell viability and cell
growth. No loss of melanoma cell viability was shown by trypan blue
exclusion. Moreover, control and LY1-treated cells showed a similar
level of MTT formazan production and cell proliferation in
vitro (results not shown). Furthermore, in the presence of fresh
rabbit or nude mice sera containing complement, LY1 did not support
lysis of melanoma cells, suggesting that the observed inhibition by LY1
in vivo was not due to complement-dependent cytotoxicity or
opsonization (data not shown).
|
, LY1 had no effect on the attachment of the
metastatic variant T1P26 to various purified matrix proteins compared
with an irrelevant mouse IgM isotype antibody. To examine further the
mechanism allowing LY1 to affect tumor growth or invasion, we
looked at the ability of LY1 to affect migration of T1P26 cells in the
chemotaxis assay. Fig. 5B
shows that migration of T1P26
cells toward lower chambers was not affected in the presence of LY1.
|
. A typical example of the staining pattern of skin containing a benign
melanocytic nevus or RGP of a superficial spreading malignant melanoma
is shown in Fig. 6
. Cells from nevi (Fig. 6A)
and RGP (Fig. 6B)
lesions were not stained with LY1 compared with a control section
stained with a IgM isotype antibody. In all lesions, staining of
keratinocytes was observed. In marked contrast, VGP cells and
metastatic cells stained strongly with this antibody (Fig. 6, C and D)
. There appears to be significant
heterogeneity on the staining pattern of VGP and metastatic lesions.
Approximately 30–90% of the melanoma cells were stained with LY1.
|
|
.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In addition to the above-mentioned results, primary melanoma
cells of the RGP consisting of cells with no demonstrable competence
for metastasis (usually those with a thickness <0.76 mm) failed to
express p55 (35)
. In striking contrast, the expression of
this antigen is dramatically increased in the primary melanoma cells of
the VGP. These results have clinical relevance, because previous
observations have shown that melanomas in the VGP lesions of >0.76 mm,
as used in this study, often carry a much worse prognosis and behave
genotypically and phenotypically similar to the lesions of distant
metastasis (35)
. These data, together with the fact that
the cell surface expression of p55 is associated with the metastatic
potential of a panel of in vivo-selected melanoma variants,
lend further support for a critical role of this melanoma-associated
antigen in the aggressive phenotype of melanoma cells. It was
remarkable to note that our experimental selection procedure led to the
appearance of a phenotypic marker of primary melanomas undergoing the
radial to VGP transition. It must be noted, however, that the
ß3 integrin subunit with a level of expression
that was shown to correlate with the aggressiveness of the tumor
(10
, 11
, 38)
is not expressed at any detectable level in
melanoma variants used during the course of this study. Whether such
vß3-negative
metastatic variants expressing p55 recapitulated their behavior in the
human patients remains to be determined.
Further characterization of the tissue distribution of the p55 cell surface protein in several normal human tissues indicated that this melanoma-associated antigen is expressed on striated muscle, heart, spleen, brain, and lung but not kidney, liver, and pancreas. These results suggest that the expression of p55 is not melanocytic specific and is not dependent on the histogenetic origin of the tissue. Indeed, epithelial as well as nonepithelial cells were stained with the LY1 MoAb. Furthermore, a number of epithelial cells (particularly glandular cells) as well as nonepithelial cells were also stained with LY1.
The LY1 MoAb did not inhibit adhesion of the metastatic variants to
vitronectin, collagen IV, fibronectin, or laminin (Fig. 5)
. The
observations that LY1 strongly stained the cell-cell contacts (Fig. 1B)
raise the possibility that p55 is not functioning as an
extracellular matrix protein receptor but rather as a cell-cell
adhesion molecule. Indeed, the human organs that need strong
intercellular cohesion (e.g., muscle, heart, spleen, brain,
and lung) were those showing high expression of p55 (Table 3)
. Its
function as an adhesive molecule necessary for cell-cell interaction is
further supported by our results showing that consistent staining of
keratinocytes was noted at the cell membranes but not at the basement
membrane-associated basal pole of epithelial cells. Such a role of the
p55 cell surface protein in mediating cell-cell interactions is very
similar to that of previously described adhesion molecules such as
CD146 (also known as melanoma cell adhesion molecule,
MUC18, A32 antigen, or S-Endo-1; Ref. 39
) or
activated leukocyte cell adhesion molecule (40)
. In these
studies, it was shown that transfected melanoma cells with these
adhesion molecules exhibited increased homotypic adhesion and increased
invasiveness through Matrigel-coated filters. These observations,
together with previous work showing a direct correlation between the
tendency of B16 melanoma variants to undergo homotypic aggregation
in vitro and their metastatic potential in vivo
(41)
, strongly suggest a role of p55 in allowing melanoma
cells to adhere to each other in various steps of the metastatic
process.
Although it is generally believed that loss of or a reduction in homotypic interactions may be required to facilitate cell detachment from the primary tumor and subsequent migration (9 , 19) , the increased cell-cell adhesion by p55 seems to contradict the paradigm of tumor progression predicated on the need for tumor cells to reduce cell surface expression of adhesion molecules such as E-cadherin to enable invasion. However, metastasis is a dynamic process characterized by subclonal evolution (3) . Up-regulation of p55 in melanomas may provide melanocytic cells with a new phenotype that enhances the ability of tumor cells to reestablish intercellular contact, thereby increasing their ability to evade immune destruction and establish metastatic foci (42) .
This metastasis-associated antigen is unrelated to IAP (known also as
CD47) in both molecular weight (Mr,
50,000) and specificity of association with
vß3 integrin in
nucleated cells (43)
. Moreover, the MoAb LY1 does not bind
to the IAP-negative, vector-transfected ovarian carcinoma cell line
OVO10 and to IAP-transfected OVO10+ cells,
suggesting that it recognizes a different antigen than CD47
(28)
. Attempts to obtain the
NH2-terminal sequence of the
Mr 55,000 melanoma cell surface
protein have been unsuccessful due to blocked
NH2-terminal residue. Further analysis of the p55
antigen sequence by gene cloning would provide pertinent information
about its biological importance in the malignant progression of human
cutaneous melanoma.
The major characteristic of malignant tumor cells is their ability to invade foreign tissues and form metastatic foci at distant locations in the body (44) . Such a process requires tumor cell attachment to various matrix proteins followed by migration of the surrounding stroma by tumor cells. As shown in the present report, adhesion of the metastatic variants to major matrix proteins as well as their chemotactic migration were not affected by the LY1 MoAb.
These results suggest that the mechanism by which this antibody affects tumor cell invasion is apparently not related to alterations in the interaction of invasive tumor cells with the basement membrane extracellular matrix or motility of tumor cells. Preliminary results indicate that such variant cell lines produced more of the Mr 92,000 collagenase type IV compared with the parental cell line.5 One possibility is that melanoma-melanoma cell interactions mediated by p55 may regulate (through intracellular signals) the expression of matrix metalloproteinases and that the binding of LY1 to p55 may affect the production of enzymes required for degrading the basement membrane and invasion (39) . Indeed, several studies have shown that the invasive potential of tumor cells could be altered via the signal transduction by the cell adhesion molecules (39 , 45) .
Tumor cell growth is a complex interaction between tumor cells and their surrounding connective tissue components (46) . Because the LY1 MoAb significantly inhibited the growth of melanoma cell variants, p55 must play an important role in the proliferation of melanoma in vivo. The finding that the LY1 MoAb did not completely inhibit tumor growth suggests that other adhesion proteins contribute to the full expression of the tumorigenic phenotype of the cells. Maintenance of intercellular contacts by p55 may provide tumor cells with a better chance of survival, as shown previously for E-cadherin-mediated cell-cell adhesion (47 , 48) . Blocking p55 with LY1 may lead to apoptosis of tumor cells. Alternatively, inhibition of tumor growth may be due to interference of the LY1 MoAb with tumor-host interactions.
In summary, we have identified in this study a
Mr 55,000 cell surface protein on
vß3-negative
metastatic melanoma variants that plays a critical role in the
malignant phenotype of human melanoma. Evidence is provided that
expression of this cell surface protein in situ is increased
dramatically in the melanocytic tumors known to have a relatively high
risk of metastasizing. This molecule may represent a new progression
marker of biological importance to identify fatal metastatic
subpopulations that have acquired competence for metastasis production.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported by Association Pour la
Recherche sur le Cancer Grant 1124, Fédération Nationale
des Groupement des Entreprises Françaises de Lutte contre le
Cancer, and specific grants from Comité
Départemental du Rhône and Saône-et-Loire de la Ligue
Nationale Française Contre le Cancer. 
2 These authors contributed equally to this
work.
3 To whom requests for reprints should be
addressed, at Institut National de la Santé et de la Recherche
Médicale Unit 331, Faculté de Médecine René
Laënnec, 8 Rue Guillaume Paradin, 69372 Lyon Cedex 08, France.
Phone: (33) 4-78-78-57-11; Fax: (33) 4-78-77-86-12; E-mail: Habib.Boukerch{at}laennec.univ-lyon1.fr
4 The abbreviations used are: RGP, radial growth
phase; VGP, vertical growth phase; MoAb, monoclonal antibody; IAP,
integrin-associated protein; MTT,
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; AEC,
3-amino-9-ethylcarbazole.
5 K. Pellier and H. Boukerche, unpublished data.
Received 4/24/00. Accepted 8/16/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
v gene expression in human melanoma tumorigenicity. J. Clin. Invest., 89: 2018-2022, 1992.
This article has been cited by other articles:
|
|
 |
|
  H. Boukerche, Z.-z. Su, C. Prevot, D. Sarkar, and P. B. Fisher mda-9/Syntenin promotes metastasis in human melanoma cells by activating c-Src PNAS, October 14, 2008; 105(41): 15914 - 15919. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Boukerche, Z.-z. Su, L. Emdad, D. Sarkar, and P. B. Fisher mda-9/Syntenin Regulates the Metastatic Phenotype in Human Melanoma Cells by Activating Nuclear Factor-{kappa}B Cancer Res., February 15, 2007; 67(4): 1812 - 1822. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Boukerche, Z.-z. Su, L. Emdad, P. Baril, B. Balme, L. Thomas, A. Randolph, K. Valerie, D. Sarkar, and P. B. Fisher mda-9/Syntenin: A Positive Regulator of Melanoma Metastasis Cancer Res., December 1, 2005; 65(23): 10901 - 10911. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
