[CANCER RESEARCH
45, 3878-3890,
August 1985]
Intermediate-sized Filaments and Specific Markers in a Human Salivary Gland
Adenocarcinoma Cell Line and Its Nude Mouse Tumors1
Mitsunobu Sato,2 Yoshio Hayashi, Tetsuo Yanagawa, Hideo Yoshida, Yoshiaki Yura, Masayuki Azuma, and
Akemichi Ueno
Second Department of Oral and Maxillofacial Surgery [M. S., T. Y., H. Y., Y. Y., M. A.], Central Laboratory for Clinical Investigation [Y. H.], and Department of Biochemistry
[A. U.], Tokushima University School of Dentistry, 3 Kuramoto-cho, Tokushima 770, Japan
ABSTRACT
The adenocarcinoma cell line HSG from human salivary gland,
which proliferates in vitro or in nude mice, was examined by the
immunoperoxidase method for the expression of three different
types of intermediate-sized filaments (IPs) and of specific anti
gens such as carcinoembryonic antigen, S-100 protein, secretory
component, lactoferrin, myosin, tropomyosin, and actin. The
cultured HSG cells were found to express three different types
of IPs defined by antibodies to keratin, vimentin, and desmin. In
HSG cells proliferating in vitro at 34°C and 37°C but not at
39°C, the expression of tropomyosin and carcinoembryonic
antigen was observed, although myosin and S-100 protein were
not detected. The expressions of actin, lactoferrin, and secretory
component were restricted to cultured HSG cells at 39°Cand
37°C,respectively. Transplantation of HSG cells into nude mice
resulted in the establishment of a nude mouse system with
malignant characteristics such as invasion and metastasis. The
expression of IPs in the primary tumors was restricted to keratin
and desmin IPs, whereas coexpression of keratin, vimentin, and
desmin IPs was observed in some neoplastic cells present in the
metastatic tumors in regional lymph nodes and lung. In addition,
expression of actin, myosin, tropomyosin, and S-100 protein was
found in the metastatic tumors, whereas myosin and S-100
protein were not detected in the primary tumors. Moreover, the
metastatic tumors were almost occupied by the neoplastic cells
with oncocytic changes, although oncocytic change was not
found in the cultured HSG cells and their primary tumors.
INTRODUCTION
It is well known that the epithelial tumors arising in human
salivary glands show a variety of histopathological features and
that their neoplastic cells are mostly composed of an admixture
of transformed epithelial duct and myoepithelial cells. Since the
cells that are pluripotential in differentiation such as the interca
lated duct and myoepithelial cells are present in the salivary
glands and can differentiate into the other types of cells such as
squamous and chondroid cells (10, 15, 21, 28, 37), tumor for
mation in human salivary glands can be regarded as a conse
quence of neoplastic transformation and its proliferation in the
process of differentiation of one or both of the 2 types of cells.
However, on the basis of their morphological findings in the
surgical specimens of salivary gland neoplasms, many investi
gators have suggested that the epithelial tumors of salivary gland
1This investigation was supported in part by Grants-in-Aidfor Cancer Research
58015080 and 59015074 from the Ministry of Education, Science and Culture of
Japan.
2To whom requests for reprints should be addressed.
Received5/15/84; revised 5/3/85; accepted 5/3/85.
CANCER
RESEARCH
arise from the intercalated ducts and that the myoepithelial cells
including duct-like epithelial cells function as supporters of the
formation of certain salivary gland neoplasms (9, 11, 20, 22).
This may indicate that the salivary gland neoplasms stem from
the intercalated duct cells and that the neoplastic cells can
represent both phenotypes of epithelial and mesenchymal cells.
Recently, neoplastic epithelial duct cells of human salivary gland
origin have been isolated and characterized (38). This cell line,
HSG, has an ultrastructure similar to that of intercalated duct
cells, and transplantation of HSG cells into the backs of athymic
BALB/c nude mice results in a formation of tumors which are
histopathologically interpreted as adenocarcinoma. Thus, by uti
lizing this cell line, we have conducted investigations of differ
entiation of salivary gland duct cells. In particular, we have taken
much interest in the mode of expression of IPs3 and specific
markers in this HSG cell line.
Recent studies on IPs have revealed that epithelial cells are
characterized by keratin IPs, whereas mesenchymal cells contain
mainly IPs of the vimentin type and IF constituents of muscle
cells are the desmin type (3, 4, 13, 27, 36). Moreover, several
investigators have recently reported coexpression of keratinand vimentin-type IPs in human metastatic carcinoma cells or in
mouse keratinocytes in culture (12, 32). On the other hand, the
neoplastic cells present in various types of salivary gland neo
plasms have recently been found to express specific markers
such as CEA and S-100 protein (5, 6, 8, 18, 25, 29, 30, 33). We
have also detected CEA, SC, and LF in the HSG cells (35).
In the present communication, we report the coexpression of
3 different types of IPs and temperature-dependent expression
of several specific markers in neoplastic intercalated duct cells
(HSG) proliferating in vitro. In addition, we establish the nude
mouse system with biological characteristics such as invasion
and metastasis similar to those of human salivary gland adeno
carcinoma, and we describe its expression and its characteriza
tion of IPs and specific markers in the nude mice tumors.
MATERIALS AND METHODS
Culture and Media. A neoplastic epithelial duct cell line, HSG, estab
lished from a human salivary gland was grown in Eagle's minimal
essential medium supplemented with 10% calf serum and 2 mw Lglutamine in the presence of 5% CO2 in a temperature gradient incubator
at various temperatures (34°C, 37°C, and 39°C). This incubator (Toyo
Ind., Co., Tokyo, Japan) is composed of 6 chambers, and the tempera3The abbreviationsused are: IPs,intermediate-sizedfilaments;CEA, carcinoem
bryonic antigen; SC, secretory component; LF, lactoferrin; PBS, phosphate-buff
ered saline (100 mw sodium phosphate:120 HIMNaCI, pH 7.2); PAP, peroxidaseantiperoxidase; SDS, sodium dodecyl sulfate; SDSiPAGE, sodium dodecyl sul
fate:polyacrylamidegel electrophoresis.
VOL. 45 AUGUST 1985
3878
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC MARKERS
OF SALIVARY
ture in each chamber is controlled gradiently. The temperature regulation
is within the limits of ±0.2°C(SD). The temperature conditions, regulated
by a thermister
PID (proportional-plus-integral-plus-derivative-control)
membrane (Millipore Co., San Francisco, CA) and used as an absorbed
antiserum for the present study. The acetone spleen powder was made
as follows. Washed, diced spleen was ground with acetone in a mortar
and filtered or squeezed onto coarse grade filter paper. After the deposit
was washed 5 times with acetone, this was dried overnight at room
temperature and used as the acetone spleen powder in the present
study. A blocking test in the use of anti-CEA serum was provided by
incubating parallel samples with the anti-CEA serum absorbed with a
viability as judged by trypan blue exclusion. The HSG cells (38) were
isolated from the submandibular salivary gland which was taken by
surgical operation from a patient who had received therapeutic irradiation
for carcinoma of the floor of the mouth. Transplantation of these cells
into the backs of BALB/c nude mice results in the production of adenocarcinoma with a solid and trabecular pattern. This cell line HSG has
been maintained in monolayer culture, and subculture has been made
on 5- or 6-day intervals.
When these HSG cells accomplished 400 cell generations from the
start of the culture, the colony formation of HSG cells in semisolid agar
was performed, as described previously (38). Briefly, suspension of 102
or 103 cells in the growth medium containing 0.3% Special Agar Noble
in 3 ml of growth medium in 30-mm plastic Petri dishes until confluent
cell monolayers were formed. The growth medium was changed twice a
week during that period. Of 10 colonies isolated, one clone which showed
most stable growth was used for this study.
Heterotransplantation of HSG Cells into Nude Mice. For the purpose
of the present study, 3-week-old female BALb/c nude mice were pur
chased from Clea Japan, Inc., Tokyo, Japan. All mice were maintained
in a pathogen-free environment in a vinyl isolator (Clea Japan).
Mice received a total number of 107 cells s.c. through an injection
CELLS
incubated at room temperature for 2 h. Thereafter, this mixture was
added with 0.8 ml of PBS, and the supernatant obtained by centrifuging
at 500 x g for 15 min at 4°Cwas filtered through a type HA Millipore
thermocontroller,
are monitored by the automatic recording system
throughout the experiments. The HSG cell population which were cul
tured in this incubator for 3 days at 34°C, 37°C, and 39°C had >90%
(Difco Laboratories, Detroit, Ml) were poured into the agar medium which
had been hardened by the addition of 0.6% agar into the growth medium.
The colonies formed after 20 days of incubation at 37°C in a 5% CO2
incubator were isolated with Pasteur's pipets, and the cells were cultured
ADENOCARCINOMA
CEA extract of human colon cancer tissues. The negative control was
provided by incubations of parallel samples with a relevant dilution of
normal rabbit serum instead of the rabbit specific antiserum of the initial
incubation. To test the specificities of antisera used for detection of actin,
myosin, tropomyosin, S-100 protein, and LF, each specific antiserum
was replaced by either PBS, normal rabbit serum, or the specific antiserum absorbed with each corresponding antigen. Antibody solutions
diluted 5-fold in PBS (1 ml) were mixed with 200 ^g of specific antigen
and incubated at room temperature for 1 h. Thereafter, they were
centrifuged at 500 x g for 15 min at 4°Cand then filtered through a type
HA Millipore membrane to remove microaggregates. The supernatants
were used as absorbed antisera for the present study. These antibody
samples gave negative results. Actin and myosin from chicken muscle,
tropomyosin from rabbit muscle, and LF from human milk were pur
chased from Sigma Chemical Co., Saint Louis, MO. S-100 protein was
purchased from Calbiochem-Behring, La Jolla, CA. For controls in stain
ing with anti-SC serum, the primary antiserum was replaced with normal
rabbit serum. Also, parallel samples were treated with anti-human IgA,
IgG, and IgM (Dakopatts, Copenhagen, Denmark) to rule out any positive
reaction due to these substances.
Rabbit antisera to CEA, SC, and LF were purchased from Dakopatts,
and the rabbit antiserum to S-100 protein was purchased from Wako
needle in the anterior aspect of the lateral thoracic region. Tumor growth
was observed weekly, and animals were sacrificed by cervical dislocation
at the time ranging from 3 to 10 months after the inoculation of HSG
cells. Tumors, regional and distant lymph nodes, and representative
sections from various organs were fixed in phosphate-buffered
10%
Pure Chemicals Co., Osaka, Japan. Rabbit antisera to human keratin,
human vimentin, chicken desmin, chicken actin, chicken myosin, and
chicken tropomyosin were purchased from Transformation Research,
Inc., Framingham, MA. In the experiments, anti-CEA and anti-S-100
protein sera were used at a dilution of 1:1000, anti-SC and anti-LF sera
were used at a dilution of 1:500, anti-vimentin, anti-desmin, anti-actin,
anti-myosin, and anti-tropomyosin sera were used at a dilution of 1:100;
and anti-keratin serum was used at a dilution of 1:60.
formalin.
Immunoperoxidase Staining. The IPs and specific markers of tumor
cells were essentially investigated by the PAP method described by
Stemberger ef al. (39). The specific antigens chosen were keratin,
vimentin, desmin, actin, myosin, tropomyosin, CEA, S-100 protein, LF,
lute methanol for 10 min to block endogenous peroxidase. After being
washed 3 times with PBS, they were incubated with an appropriate
dilution of rabbit antiserum to each of the above-described antigens for
Light and Electron Microscopic Observation. The paraffin sections
of nude mice tumors were stained with hematoxylin and eosin.
For transmission electron microscopy, the tumor tissues were fixed in
4% glutaraldehyde buffered with 0.1 M phosphate (pH 7.3) containing
0.005 M calcium chloride for 1 h and then dissected into small pieces of
about 2 cu mm. After these materials were washed in PBS, they were
postfixed with 2% osmium tetroxide, dehydrated step by step with
ethanol, and embedded in Epon 812. The ultrathin sections were cut on
a LKB ultramicrotome and stained with uranyl acetate and lead hydrox
ide. The finished preparations were observed under a Hitachi Model H-
30 min at room temperature followed by a rinse with PBS for 30 min.
Thereafter, goat anti-rabbit IgG (Medical and Biological Laboratory, Ltd.,
500 electron microscope.
Indirect Immunofluorescence
Nagoya, Japan) was applied in a dilution of 1:40 in PBS for 30 min at
room temperature. After being washed with PBS for 30 min, these
samples were incubated with rabbit horseradish PAP complex (Miles
Laboratories, Inc., Elkhart, IN), diluted 1:50 in PBS for 30 min, and rinsed
with PBS for 20 min. Finally, the peroxidase was localized by treatment
of the samples with a fresh mixture of 0.05% 3,3-diaminobenzidine and
0.01% H2O2 in 10 mw Tris-HCI buffer (pH 7.6) for 5 min; after being
beling technique was performed essentially as described by Ramaekers
ef al. (32). Cells grown in glass coverslides were fixed in acetone at
—20°Cfor 10 min and then subjected to indirect immunofluorescence
and SC.
The cultured cells were washed 3 times with PBS and fixed with 95%
ethanohacetone (4:6) for 10 min at 4°C. The paraffin-embedded speci
mens were cut in serial sections approximately 4 /¿mthick. The paraffin
sections were deparaffinized with xylene and rehydrated step by step
with descending concentrations of ethanol.
These preparations were incubated at 37°Cwith 0.3% H2O2 in abso
washed with distilled water, these samples were counterstained with
hematoxylin.
The following controls were used. The rabbit antiserum to CEA was
absorbed with normal human spleen extract to remove nonspecific crossreacting antigen as follows. Antibody sample (0.2 ml) was mixed with
100 nQ of acetone powder prepared from normal human spleen and was
CANCER
RESEARCH
Technique. The double-antibody la
microscopy. The fixed cells were incubated with the first antibodies, with
a subsequent washing step between the incubations. The first antibodies
(Transformation Research) used in this double label experiment were:
rabbit antisera to human vimentin and chicken desmin (diluted 1:20
each); a mouse monoclonal IgG antibody to human keratin (not diluted);
or a mouse monoclonal IgM antibody to human vimentin (not diluted).
After an incubation step of 30 min with mouse monoclonal anti-keratin
and rabbit antiserum to human vimentin, mouse monoclonal anti-vimen
tin, and rabbit antiserum to chicken desmin, or mouse monoclonal antikeratin and rabbit antiserum to chicken desmin in a humidified box at
VOL. 45 AUGUST 1985
3879
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC MARKERS
OF SALIVARY
room temperature, the sampleswere washed with PBScontaining0.25%
Triton X-100 in 2 subsequent washing steps of 10 min each and
thereafter were washed with the PBS alone. These preparations were
then incubated for another 30 min with the second antibodies, again with
a subsequent washing step between the incubations.The labeledsecond
antibodies (Cappel Laboratories, Cochranville, PA) were: fluoresceinconjugated goat anti-mouse IgG and IgM (diluted 1:25 each); and rhodamine-conjugated goat anti-rabbit IgG (diluted 1:20). After a second
series of washes, the samples were mounted with 50% glycerol in PBS.
Cells were examined under a Nikon microscope, equipped with epifluorescent illumination(HBO 100-watt bulb), by using the appropriate filler
systems for fluorescein and rhodamine fluorescence. Overlap between
the 2 channelswas checked by using cells labeledwith a single antibody.
Tri-X film from Kodak was used for photographs.
Immunoblotting.The preparationsenrichedin IFs were performed
according to the method of Franke ef al. (14). Briefly, HSG monolayers
formed by culturing for 3 days at 37°Cwere rinsed twice in 10 mw TrisHCI:140 mM NaCI:5 mw MgCI2buffer (pH 7.6) and then incubated for
about 3 min in 10 HDMTris-HCI:140 mM NaCI buffer (pH 7.6) containing
1% Triton X-100 at room temperature. The cell residues remaining
attached to the substratum were then incubated for 30 min at room
temperature in 10 mM Tris-HCI:140 mM NaCI buffer (pH 7.6) containing
0.5% Triton X-100 and 1.5 M KCI (pH 7.6). The pellet obtained by
centrifuging the dispersed cell material at 3500 x g for 10 min was
washed twice by brief dispersion in 10 mMTris-HCI(pH 7.6) followed by
centrifugation. This material was further incubated for 15 to 25 min in 10
HIMTris-HCI buffer (pH 7.6) containing 0.2 mM MgCI2 and pancreatic
DNase I (100 i¡glm\;Worthington Biochemicals Corp., Freehold, NJ) at
room temperature in order to digest the DNA present in these cytoskele
ton-enriched preparations. For examination of the protein by
SDS:PAGE,precipitates obtained with 5% cold trichloroacetic acid were
washed with acetone and taken up in 10 mM Tris-HCI buffer (pH 7.6)
with 2% SDS and 10 mM dithiothreitol. These preparations were soni
cated and boiled for 2 min to solubilize IF proteins.
The procedure for SDS:PAGE used in the present study essentially
followed the method of Laemmli(26). The slab gels measured 8.0 x 12.0
x 0.2 cm. The stacking gel contained 5% acrylamide(Wako Pure Chem
icals Co.) and 0.13% A/,A/'-methylenebisacrylamide(Wako Pure Chemi
cals Co.). The separation gel contained 12.5% acrylamide and 0.1% N,
W'-methylenebisacrylamide. Electrophoresis was performed at room
temperature for 4 h at 4 ma/cm width of gel. The migration front was
followed by adding 0.002% bromophenol blue to the electrode buffer.
Gels were stained for 12 h using 0.25% Coomassie Brilliant Blue in 35%
ethanol:! 0% acetic acid in water and were destained in the same solution
lacking the stain. Proteins from gels were transferred to nitrocellulose
paper as reported by Towbin ef al. (42). Detection of IFs on strips was
performed by the PAP method as described above. Briefly, the nitrocel
lulose sheets were immersed with PBS containing 3% (w/v) bovine
serum albumin (Sigma) and then were incubated with a 10-fold dilution
of rabbit antiserumto humankeratin, human vimentin,or chicken desmin,
followed by reaction with goat anti-rabbit IgG (1:20) and PAP complex
(1:40).The peroxidase was localized by treatment of the sheets with the
fresh mixture of 0.05% 3,3-diaminobenzidineand 0.01% H2O2in 10 mM
Tris-HCI buffer (pH 7.6).
RESULTS
IFs and Specific Markers in Human Salivary Gland Adenocarcinoma Cell Line HSG Proliferating in Vitro. HSG cells were
cultured in a temperature gradient incubator for 3 days at 34°C,
37°C,or 39°C,and then they were examined by the PAP method
for the expression of IFs and specific markers. Figs. 1 and 2
show representative samples of the immunohistochemical ap
pearance of HSG cells which were cultured at various tempera
tures. Table 1 summarizes all of the results obtained by the
ADENOCARCINOMA
CELLS
Table 1
Summary of the immunohistochemical characteristics ol the duct-type HSG cells
HSG cells were cultured for 3 days at 34°C, 37°C,and 39°C.Thereafter, cells
were examined by the PAP method for the presence of the specific antigens. The
percentage of positive cells was counted microscopically. Values are averages of
2 experiments. A blocking test in the use of rabbit antiserum raised against human
keratin, human vimentin, or chicken desmin was provided by incubating parallel
samples with 5-fold diluted antibody samples (1 ml) absorbed with 600 ^g of the
lyophilized preparation enriched in IFs from HSG cells. These antibody samples
gave negative results.
cellsSpecific
% of positive
antigensKeratinVimentinDesminActinMyosinTropomyosinCEAS-1
proteinLFSC34°C95959000958000037°C95959500959502.52.539°C9575075000000
00
immunohistochemical study on HSG cells.
As can be seen in Fig. 1, a to f, and Table 1, cultivation of
HSG cells at 34°C or 37°C resulted in the coexpression of 3
different types of IFs defined by the antisera to keratin, vimentin,
and desmin. Although only IFs of the desmin type disappeared
in the cultured HSG cells at 39°C, both keratin and vimentin
were detected in the cultivation of HSG cells at 39°C. Fig. 1, g
to /', illustrates the mutual exclusiveness of the reactions of 3
antisera to keratin, vimentin, or desmin on frozen sections from
normal human oral tissues. It is noted that keratin-specific antiserum gives very strong positive staining for epithelium and that
each antiserum to vimentin or desmin reacts exclusively with
submucosal mesenchymal tissue or with muscle tissues includ
ing smooth muscle cells present in the vascular region, respec
tively.
On the other hand, we have already demonstrated the pres
ence of CEA, LF, and SC in the cultured HSG cells at 37°C(35).
Thus, HSG cells were examined by the PAP method for expres
sion and its temperature dependency with these specific markers
including cytoskeletal proteins such as actin, myosin, and tropomyosin. As shown in Fig. 2 and Table 1, the expression of
actin was clearly observed in the cultured HSG cells at 39°C,
whereas actin was never detected in HSG cells which were
cultured at either 34°Cor 37°C. The presence of tropomyosin
was demonstrated in the HSG cells which were cultured at either
34°Cor 37°C,whereas tropomyosin did not occur in cultivation
of HSG cells at 39°C.The expression of CEA was demonstrated
in the cytoplasms of HSG cells which were cultured at either
34°Cor 37°C but not at 39°C. The expression of LF and SC
was restricted to the cultured HSG cells at 37°C. However,
myosin and S-100 protein were not observed in HSG cells under
any experimental conditions used for the present study.
Since it could be considered from the results obtained by the
PAP method that cultured HSG cells express simultaneously 3
different types of IFs, cultured HSG cells at 37°Cwere further
examined by double immunofluorescence labeling by using vi
mentin and desmin antibodies raised in a rabbit and a mouse
monoclonal antibody directed against human keratin or human
vimentin. Fig. 3 shows some typical examples of cultured HSG
cells stained by the double label technique, illustrating the CO-
CANCER RESEARCH VOL. 45 AUGUST 1985
3880
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC MARKERS
OF SALIVARY
existence of keratin and vimentin IPs, of keratin and desmin IPs,
or of vimentin and desmin IPs. In the cultured HSG cells, keratinpositive as well as keratin-negative cells could be demonstrated.
The same was true for vimentin and desmin. It also was obvious
that keratin, vimentin, and desmin are distributed differently
throughout the cytoplasm of the cell. Surprisingly, many HSG
cells gave the impression that 3 different IF systems co-localize.
Moreover, to confirm that intermediate-size filament proteins
reactive to antibodies directed against keratin, vimentin, and
desmin exist in the cultured HSG cells, these antisera were
applied to nitrocellulose transfers of SDS:PAGE containing the
preparation enriched in IPs from cultured HSG cells. After reac
tion of the nitrocellulose sheets with rabbit anti-keratin antibod
ies, followed by goat anti-rabbit IgG and PAP complex, a distinct
reaction is seen with M, 46,000,56,000 and 67,000 polypeptides
of cultured HSG cells (Fig. 4, Lane a). Rabbit anti-vimentin
antibodies bound to a M, 59,000 polypeptide present in cultured
HSG cells, and rabbit anti-desmin antibodies bound to a M,
53,000 polypeptide (Fig. 4, Lanes b and c).
IPs and Specific Markers in the HSG Nude Mice Tumors.
Inoculation of 107 HSG cells s.c. into nude mice resulted in
production of tabulated tumor masses 8 to 13 mm in diameter in
the inoculation site about 30 days after transplantation. These
tumors became aggressively enlarged and, about 3 months after
the inoculation, a majority of the tumor-bearing mice had formed
ulcers in the outer skin adjacent to the tumor. However, these
ulcers had a tendency to form a crest and were encapsulated by
fibrous tissue. In addition, the formation of new tumor masses
in the areas corresponding to axillary lymph nodes in the homolateral or less frequently bilateral sides was observed in almost
all of the tumor-bearing mice. Since the tumor masses which
could be considered the axillary lymph node metastasis became
expansively enlarged, the animals were sequentially killed by
cervical dislocation, and autopsy was performed.
The primary tumors which occurred in the sites inoculating
ADENOCARCINOMA
CELLS
HSG cells were histopathologically interpreted as adenocarcinoma with solid and trabecular pattern as described previously
(38) (Fig. 5a). The tumor was composed of almost cuboidal cells
with occasional anisocytism and mitosis. In addition, a group of
larger cells which had pale-staining eosinophilic cytoplasm ap
peared in some areas of the tumor tissue. Although the tumor
was surrounded by the pseudocapsule, some groups of neo
plastia cells frequently invaded the pseudocapsule or occasion
ally infiltrated the surrounding fibrous and muscular tissues (Fig.
5o). Moreover, in all of 15 mice exhibiting macroinvasion into the
outer skin, tumors invaded the local lymphatic channels which
appeared dilated and filled with the clusters of neoplastia cells
(Fig. 5c). Lymph node involvement was found in 13 of the
examined 15 mice, which was restricted to the axillary lymph
nodes. This was characterized by the presence of neoplastic
cells in the peripheral sinus of lymph nodes with extension into
the paracortical and medullary areas. Nodal capsular invasion
and tumor extension into the surrounding tissue was also ob
served. The metastatic tumors were histopathologically inter
preted as adenocarcinoma with oncocytic change (Fig. 5d). The
tumor was composed of sheets of predominantly polygonal or
spindle-shaped large cells with occasional mitoses, the cyto
plasm of which contained many eosinophilic granules. Ultrastructurally, numerous mitochondria with crescent-shaped crista or
concentric tubular crista were observed in the cytoplasm of
neoplastic cells (Fig. 5e). Two neighboring tumor cells were often
attached with continuous tight junction, and distinct desmosomes were scant. Intercellular digitations formed by infoldings
of cytoplasmic processes were sometimes observed. The nuclei
were oval, and the nuclear chromatin was evenly distributed. In
the lungs of 2 of the examined 15 mice, métastaseswere found
in areas adjacent to small bronchi and bronchioli (Fig. 5f). A
majority of the metastatic tumors present in the lung showed
morphological characteristics identical to those in the lymph
nodes.
The HSG cells used for this study were found to express actin,
tropomyosin, LF, and SC, including 3 different types of IPs. Thus,
a search for these antigens including myosin and S-100 protein
on tissue sections of the primary and metastatic tumors was
performed immunohistochemically. As shown in Fig. 6 and Table
2, keratin and desmin were observed in the neoplastic cells
present in the primary tumors, although vimentin was not de-
94K-
t7K-
Tab^
Summaryof the immunohistochemicalcharacteristics of the primary and
metastatic tumors, which were produced by inoculation of HSG cells into
nude mice
To localize the above antigens in tumor transplants, formalin-fixed paraffinembedded materials were examined by the PAP method.
43K-ÃŽOK-
tumorsLymph
>OK~
»4K-
Specific
antigensKeratin
RESEARCH
nodes
+
Vimentin
Desmin
Actin
Myosin
Tropomyosin
CEA
S-100 protein
LF
SCMetastatic
* +, positive; -, negative.
Fig.4. Immunoblotting of HSG IPs proteins with antibodies to human keratin
and vimentin and to chicken desmin. The preparationenriched in protein of IPs was
made from the cultured HSG cells at 37°Caccording to the method of Franke et
al. (14)and was run on a slab gel. After the protein was transferred to nitrocellulose,
the blot was cut into strips and incubated with rabbit antiserum specific for human
keratin (Lane a), human vimentin (Lane b), or chicken desmin (Lane c) or stained
with Amido Black 108 (Lane d). The immunoreaction was detected by the PAP
method. Abscissa, molecular weight markers x 103 (K). Phosphorylase b, M,
94,000; bovine serum albumin, M, 67,000; ovalbumin, M, 43,000; carbonic anhydrase, M, 30,000; soybean trypsm inhibitor, M, 20,000; a-lactalbumin,M, 14,000.
CANCER
Primary
Lung+"
tumors
VOL. 45 AUGUST 1985
3881
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
+
IPs AND SPECIFIC
MARKERS
OF SALIVARY
tected. The tissue sections from primary tumors showed positive
stainings of actin, tropomyosin, and CEA (Fig. 7). On the other
hand, in addition to the cytoskeletal proteins such as actin,
myosin, and tropomyosin, the expression of 3 different types of
IPs including S-100 protein was detected in some neoplastic
cells present in the metastatic tumors in the lymph nodes and
lung, although CEA was not detected in the metastatic tumors
(Figs. 6 and 7).
This paper reports expression and its characterization of IFs
and specific markers in a neoplastic epithelial duct cell line (HSG)
of human salivary gland in origin, which propagate in vitro and in
nude mice. Since Hanna (17) has reported that tumors which are
nonmetastatic in nude mice older than 10 weeks of age will
metastasize at a high frequency if inoculated into 3-week-old
animals, we attempted to establish a nude mouse system with
biological actions similar to those of the adenocarcinoma arising
from human salivary glands by inoculating HSG cells into the
younger animals because transplantation of HSG cells into the
backs of nude mice results in a formation of adenocarcinoma.
Consequently, the nude mouse system established in the present
study represented growth patterns such as local invasion and
metastasis into the regional lymph nodes or lung.
The data presented here provide the evidence for coexpression of keratin, vimentin, and desmin in the HSG cells propagating
in vitro. Although coexpression of keratin- and vimentin-type IFs
has been demonstrated in human metastatic carcinoma cells
(32) or in mouse keratinocytes in culture (12), eukaryotic cells in
which 3 different types of IFs are coexpressed have not yet been
reported excluding HSG cells used for the present study, as far
as we know. However, we must stress that the 3 IF systems
were not observed simultaneously in the same cell in HSG cells
tested. The HSG cells which were cultured at either 34°C or
37°C were found to express 3 different types of IFs together,
although desmin-type IFs alone disappeared in the cultivation of
HSG cells at 39°C. This may indicate that the expression of
desmin in the HSG cells proliferating in vitro is temperature
dependent, although the possibility that desmin is degraded in
the cultured HSG cells at 39°Cand cannot be detected under
the present experimental conditions is not excluded.
In HSG cells proliferating in vitro at certain conditioned tem
peratures, the expression of actin, tropomyosin, CEA, LF, and
SC was observed, although the occurrence of myosin and S100 protein in the cultured HSG cells was never detected under
the present experimental conditions. The presence or absence
of the cytoskeletal proteins such as actin and tropomyosin in the
cultured HSG cells at different temperatures strongly suggests
that the change of cellular morphology accompanying alterations
of the microfilamentary system within the HSG cells occurs
according to various cultivation temperatures. It has been desribed that the presence of CEA in human salivary gland neo
plasms is related mainly to the well-differentiated ductal tissue in
the neoplasms (29). Moreover, Caselitz ef al. (7) have recently
identified the occurrence of CEA in the acinar and intercalated
duct cells of the normal human parotid gland. Since HSG cells
have a fine structure similar to that of intercalated duct cells (38),
the expression of CEA in the HSG cells strongly suggests that
the HSG cells are of the intercalated duct type. Recently, the
RESEARCH
CELLS
presence of LF has been demonstrated in the normal human
salivary gland (33) and in certain human salivary gland neoplasms
including the benign and malignant tumors (8). This LF can be
considered to be related to the glandular function of the normal
tissue, and an elevation of LF in saliva frequently occurs during
the course of chronic recurrent parotitis (40). On the other hand,
it is well known that human salivary glands contain IgA-producing
DISCUSSION
CANCER
ADENOCARCINOMA
cells and shed naturally IgA or SC into saliva (5). In addition,
several investigators have reported the synthesis of SC in human
tumors such as mammary, colonie, and pulmonary neoplasms
(16,19, 23, 31). However, Rognum ef al. (34) have reported that
the malignant alterations in large bowel adenomas impair the
capacity of the columnar epithelium to produce SC and to
express it as a functional receptor mediating uptake of dimeric
IgA. The above findings about CEA, LF, and SC strongly suggest
that the neoplastic intercalated duct cell line HSG with functional
similarity to that of the normal duct cells is well differentiated.
On the other hand, the expression of IFs in the primary tumors
arising in the inoculation site of HSG cells was essentially re
stricted to keratin- and desmin-type IFs, whereas coexpression
of keratin-, vimentin-, and desmin-type IFs was observed in some
neoplastic cells present in metastatic tumors in the regional
lymph nodes and lung. In addition, the expressions of the cyto
skeletal proteins such as actin, myosin, and tropomyosin and of
S-100 protein were found in some neoplastic cells present in the
metastatic tumors, although myosin and S-100 protein were not
detected in the primary tumors. Recently, we have demonstrated
the presence of myosin and S-100 protein in a neoplastic myoepithelial cell line of human salivary gland origin, but not in the
duct-type HSG cells (35). In addition, Archer and Kao (1) also
have identified immunohistochemically the presence of actomyosin in myoepithelium of human tissue. Moreover, a cell
transitional between duct-type and myoepithelial cells has been
demonstrated in the salivary gland tumor (24). This suggests a
possibility of mutual conversion between duct-type and myo
epithelial cells. Thus, the neoplastic cells containing myosin, S100 protein, or desmin can be considered to represent exclu
sively the phenotype of myoepithelial cell lineage. This may
indicate that some steps of differentiation towards the myoepi
thelial cells of the duct-type HSG cells can be represented in the
process of metastasis in nude mice. Therefore, we now examine
whether the neoplastic cells with an ultrastructure like myoepi
thelial cells or a cell type intermediate between the duct-type and
myoepithelial cells are present in the metastatic tumors. Sheets
of the neoplastic cells in the tissue sections from metastatic
tumors of the regional lymph nodes and lung were almost
occupied by the neoplastic cells showing oncocytic change,
although these cells were not choked with many mitochondria
as observed in the oncocytomas reported by Balogh and Roth
(2), and Tandler ef al. (41). Oncocytic change was not found in
the cultured HSG cells at the 3 different temperatures for 3 days
and in the primary HSG tumors. This morphological observation
probably indicates that oncocytic cells in the salivary gland arise
from the intercalated duct cells. Moreover, based on the above
findings about the specific markers of neoplastic cells present in
the primary and metastatic tumors, it can be considered that
other differentiation stages different from the original HSG cells
are induced during the course of metastasis in nude mice.
VOL. 45 AUGUST 1985
3882
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC
MARKERS
OF SALIVARY
ACKNOWLEDGMENTS
The authors wish to express their appreciation to Nanayo Furumoto for secre
tarial assistance.
REFERENCES
1. Archer, F. L., and Kao, V. C. Y. Immunohistochemicalidentification of actomyosin in myoepithelium of human tisssue. Lab. Invest., 78: 669-674, 1968.
2. Balogh, K., Jr., and Roth, S. I. Histochemicaland electron microscopic studies
of eosinophilic granular cells (oncocytes) in tumors of the parotid gland. Lab.
Invest., 74:310-320,1965.
3. Bennet, G. S., Fellini.S. A., Croop, J. M., Otto, J. J., Bryan, J., and Holzer, H.
Differences among 100-Äfilament subunits from different cell types. Proc.
Nati. Acad. Sci. USA, 75: 4364-4368,1978.
4. Bennetj G. S., Fellini, S. A., and Holtzer, H. Immunofluorescent visualization
of 100A filaments in different cultured chick embryo cell types. Differentiation,
72:71-82,1978.
5. Brandzaeg, P. The secretory immune system of lactating human mammary
glands compared with other exocrine organs. Ann. NY Acad. Sci., 409: 353382, 1983.
6. Caselitz, J., Jaup, T., and Seifert, G. Immunohistochemicaldetection of carcinoembryonic antigen (CEA)in parotid gland carcinomas.Analysis of 52 cases.
Virchows Arch. Abt. A Pathol. Anat., 394: 49-60,1981.
7. Caselitz, J., Seifert, G., and Jaup, T. Presence of carcinoembryonic antigen
(CEA) in the normal and inflamed human parotid gland. J. Cancer Res. Clin.
Oncol., 700:205-211, 1981.
8. Caselitz, J., Seifert, J., and Jaup, T. Tumor antigens in neoplasms of the
human parotid gland. J. Oral Pathol., 77: 374-386,1982.
9. Chisholm, D. M., Waterhouse, J. P., Kraucunas, E., and Sciubba, J. J. A
quantitative ultrastructural study of the pleomorphic adenoma of human minor
salivary gland. Cancer (Phila.),34: 1631-1641,1974.
10. Doyle, L. E., Lynn, J. A., Panopio, I. T., and Crass, G. Ultrastructure of
chondroid regions of benign mixed tumor of salivary gland. Cancer(Phila.),22:
225-233,1968.
11. Evans, P. W., and Cruickshank, A. H. Epithelialtumors of the salivary glands.
In: P. W. Evans and A. H. Cruickshank (eds.), The Series Major Problems in
Pathology, Vol. 1, pp. 209-226. Philadelphia:W. B. Saunders, 1970.
12. Franke, W. W., Schmid, E., Breitkerutz, D., Lüder,
M., Boukamp, P., Fusenig,
N. E., Osbom, M., and Weber, K. Simultaneous expression of two different
types of intermediate sized filaments in mouse keratinocytes proliferating in
vitro. Differentiation, 74: 35-50,1979.
13. Franke, W. W., Schmid, E., Osbom, M., and Weber, K. Different intermediatesized filaments distinguished by immunofluorescencemicroscopy. Proc. Nati.
Acad. Sci. USA, 75. 5034-5038, 1978.
14. Franke, W. W., Schmid, E., Weber, K., and Osbom, M. HeLa cells contain
intermediate-sized filaments of the prekeratin type. Exp. Cell Res., 778: 95109,1979.
15. Friedman, M., and Hall, J. W. Radiation-induced squamous metaplasia and
hyperplasia of the normal mucous glands of the oral cavity. Radiology, 55:
848-856,1950.
16. Gotoh, T., Takishita, Y., Doi, H., and Tsubura, H. Secretory componentproducing lung cancer with hypergammaglobulinemiaof secretory IgA. Cancer
(Phila.),48: 1776-1782, 1981.
17. Hanna, N. Expression of metastatic potential of tumor cells in young nude
mice is correlated with low levels of natural killer cell-mediated cytotoxicity.
Int. J. Cancer, 26: 675-680, 1980.
18. Hará,K., Ito, M., Takeuchi, J., lijima, S., Endo, T., and Hidaka, H. Distribution
of S-1OObprotein in normalsalivaryglands and salivarygland tumors. Virchows
Arch. [Pathol. Anat.], 407: 237-249, 1983.
19. Harris, J. P., Caleb, M. H., and South, M. A. Secretory component in human
mammary carcinoma. Cancer Res., 35:1861-1864,1975.
20. Hoshino, M., and Yamamoto, I. Ultrastructure of adenoid cystic carcinoma.
CANCER
RESEARCH
ADENOCARCINOMA
CELLS
Cancer (Phila.),25: 186-198, 1970.
21. Hübner,
G., Klein,H. J., Kleinasser,0., and Schiefer, H. G. Roleof myoepithelial
cells in the development of salivary gland tumors. Cancer (Phila.),27: 12551261,1971.
22. Hübner,
H., Kleinasser,O., and Klein, H. J. Zur Finestrukter und Genese der
Cylindsomeder Speicheldrüsen
Weitere Untersuchungerzur Rolle myoepithe
lialdifferenzierterZellen in Speicheldrüsen-Geschwülsten.
Virchows Arch. Abt.
A Pathol. Anat., 347: 296-315, 1969.
23. Hung, S. W., Fogh, J., and Hong, R. Synthesis of secretory component by
human colon cancer cells. Scand. J. Immunol., 5: 263-268, 1976.
24. Kierzenbaum,A. L. The ultrastructure of mixed salivary tumors. Lab. Invest.,
78:391-396,1968.
25. Korsrud, F. R., and Brandtzaeg, P. Immunohistochemical studies on the
epithelial and lymphoid components of Warthin's tumor. Acta Otolaryngol.
Suppl., 360: 221-224, 1979.
26. Laemmli.U. K. Cleavageof structural proteins during the assembly of the head
of bacteriophageT4. Nature (Lond.), 227: 680-685, 1970.
27. Lazarides,E. Intermediatefilamentsas mechanicalintegrators of cellularspace.
Nature (Lond.), 283: 249-256,1980.
28. Luna, M. A., Mackay, B., and Gamez-Araujo,J. Myoepitheliomaof the palate.
Report of a case with histochemical and electron microscopic observations.
Cancer (Phila.),32:1429-1435,1973.
29. McDicken, W., and Scott, J. The presenceof distribution of carcinoembryonic
antigen in tumors of human minor salivary gland. J. Oral Pathol., 70: 296-303,
1981.
30. Nakajima,T., Watanabe, S., Sato, Y., Kameya, T., Hirota, T., and Shimosato,
Y. An immunoperoxidase study of S-100 protein distribution in normal and
neoplastic tissues. Am. J. Surg. Pathol., 6: 715-727,1982.
31. Poger, M. E., Hirsch, B. R., and Lamm, M. E. Synthesis of secretory component
by colonie neoplasms. Am. J. Pathol., 82: 327-338,1976.
32. Ramaekers, F. C. S., Haag, D., Kant, A., Moesker, O., Jap, P. H. K., and
Vooijis,G. P. Coexpressionof keratin-and vimentin-typeintermediatefilaments
in human metastatic carcinoma cells. Proc. Nati. Acad. Sci. USA, 80: 26182622, 1983.
33. Reitamo, S., Kontinen, Y. T., and Segergerg-Kontinen, M. Distribution of
lactoferrin in human salivary gland. Histochemistry, 66: 285-291,1980.
34. Rognum, T. O., Pausa, 0., and Brandtzaeg, P. Immunohistochemicalevalua
tion of carcinoembryonic antigen, secretory component, and epithelial IgA in
tubular and villous large-bowel adenomas with different grades of dysplasia.
Scand. J. Gastroenterol., 77: 341-348, 1982.
35. Sato, M., Hayashi, Y., Yoshida, H., Yanagawa, T., Yura, Y., and Nitta, T.
Search for specific markers of neoplastic epithelial duct and myoepithelialcell
lines established from human salivary gland and characterization of their
growth in vitro. Cancer (Phila.),54: 2959-2967, 1984.
36. Schmid, E., Tapscott, S., Bennet, G. S., Croop, J., Fellini, S. A., Holzer, H.,
and Franke, W. W. Differential location of different types of intermediate-sized
filaments in various tissues of the chicken embryo. Differentialen, 75: 27-40,
1979.
37. Shafer, W. G., Hine, M. K., and Levy, B. M. Tumors of the Salivary Glands.In:
W. G. Schafer,M. K. Hine,and B. M. Levy (eds.),A Textbook of Oral Pathology,
pp. 232-233. Philadelphia:W. B. Saunders Co., 1974.
38. Shirasuna, K., Sato, M., and Miyazaki, T. A neoplastic epithelial duct cell line
established from an irradiated human salivary gland. Cancer (Phila.),48: 745752,1981.
39. Stemberger, L. A., Hardy, P. H., Cuculins,J. J., and Meyer, H. G. The unlabelled
antibodyenzyme methodof immunohistochemistry:Preparationand properties
of soluble antigen-antibody complex (horseradish peroxidase-anti-horseradish
peroxidase) and its use in identification of spirochetes. J. Histochem. Cytochem., 78:315-333,1970.
40. Tabak, L., Mandel, I. D., Kariamd, D., and Baurmash, H. Alterations in lacto
ferrin in salivary gland disease. J. Dent. Res., 57: 43-47, 1978.
41. Tandler, B., Hutter, R. V. P., and Eriandson,R. T. Ultrastructureof oncocytoma
of the parotid gland. Lab. Invest., 23: 567-580, 1970.
42. Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of proteins
from polyacrylamidegels to nitrocellulose sheets: procedure and some appli
cations. Proc. Nati. Acad. Sci. USA 76: 4350-4354, 1979.
VOL. 45 AUGUST 1985
3883
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC
MARKERS
OF SALIVARY
ADENOCARCINOMA
CELLS
*
.-•''*
Id
*
Ib
le
If
Fig. 1. Detection of 3 different types of IPs in the cultured HSG cells at 37°C by the immunoperoxidase method, using anti-keratin (a), anti-vimentin (t>), and antidesmin (c), and anti-keratin (d), anti-vimentin (e) and anti-desmin (f), which were absorbed with the preparation enriched in IPs from cultured HSG cells. Hematoxylin
counterstain. a to I, x 480. Note the mutual exclusiveness of the reactions with these sera on frozen tissue sections from the normal human oral tissues. Anti-keratin
stained only the epithelium (g), whereas anti-vimentin (h) and anti-desmin (/) stained exclusively submucosal mesenchymal tissues and muscles, respectively. Hematoxylin
counterstain. g, h, x 120; /', x 360.
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3884
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC
MARKERS
OF SALIVARY
ADENOCARCINOMA
CELLS
19
v»
r
\
*vi
'
h l
i
^
v
* J
-.>"\-\ç \
-
s.
1tr
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3885
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC
.
MARKERS
OF SALIVARY
ADENOCARCINOMA
CELLS
i-,n
*
.-
m
^
i
Jfc
";.
b
Fig. 2. Immunohistochemical detection of the cytoskeletal proteins and specific markers in the cultured HSG cells at different temperatures, a, actin (39°C);b,
tropomyosin (34°C);c, CEA (37°C);d, LF (37°C);e, SC (37°C).Arrows in d and e, positive staining. Hematoxylincounterstain, x 480.
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3886
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC MARKERS
OF SALIVARY
ADENOCARCINOMA
CELLS
t «3b
Fig. 3. Indirect ¡mmunofluorescence
microscopy of cultured HSG cells at 37°C.Double label immunofluorographsshowing the simultaneous presence of keratin (a)
and vimentin (£>),
of keratin (c) and desmin (d), or of vimentin (e) and desmin (r). First antibodies used: a, c, mouse monoclonal anti-keratin IgG; b, rabbit anti-vimentin;
d, f, rabbit anti-desmin;e, mouse monoclonal anti-vimentinIgM. Second antibodies used: a, c, fluorescein-conjugatedanti-mouse IgG; b, d, I, rhodamine-conjugatedantirabbit IgG; e, fluorescein-conjugatedanti-mouse IgM. x 640.
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3887
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IFs AND SPECIFIC MARKERS
OF SALIVARY
ADENOCARCINOMA
CELLS
5d
Fig. 5. Histological appearance of the primary and metastatic tumors, which were produced by inoculation of HSG cells into nude mice, a, primary tumor. This is
interpreted as adenocarcinomawith a solid and trabecular pattern. Anisocytism and mitosis are occasionallyobserved. H & E, x 240. b, photomicrograph demonstrating
invasion of the skin covering the primary tumors. H & E, x 120. c, photomicrograph demonstrating the presence of neoplastic cells in vessels adjacent to the primary
tumor bed (arrows). H & E, x 120. d, representative histological appearanceof metastasis in the axillary lymph nodes. Oncocytic change is observed in many neoplastic
cells. H & E, x 360. e, ultrastructure of the metastatic tumor formed in the axillary lymph nodes. Numerous mitochondria with various sizes and shapes are observed in
the cytoplasm of neoplastic cells, x 15.000. f. representative histological appearance of metastasis in the lung. H & E, x 120. Black arrow, neoplastic cell nests; white
arrow, bronchiolus; arrowhead, pulmonary alveolus.
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3888
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
IPs AND SPECIFIC
MARKERS
OF SALIVARY
ADENOCARCINOMA
CELLS
" ''$kv*
|
6d
6e
6b
«
v.e,
Fig. 6. Immunohistochemicaldetection of IFs in the primary and metastatic nude mice tumors. Primary tumors: a, anti-keratin (positive);b, anti-vimentin(negative);c,
anti-desmin(positive). Hematoxylin countersign, a to c, x 360. Metastatic tumors in the lung: d. anti-keratin (positive);e, anti-vimentin(positive);f, anti-desmin(positive).
Hematoxylin countersign, d to f, x 360. The tissue sections from metastatic tumors present in the regional lymph nodes showed the positive stainings for keratin,
vimentin, and desmin in a similar fashion as observed in the metastatic tumors in the lung (data not shown).
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3889
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
*
" -. :/
'
.
J
Fig. 7. Immunohistochemical detection of specific markers in the primary and metastatic nude mice tumors. Primary tumors: a, actin (positive); 6, myosin (negative);
c. tropomyosin (positive); d. CEA (positive); e. S-100 protein (negative) Hematoxylin counterstain. a to e, x 360. Metastatic tumors in the lung: /, actin (positive); g.
myosin (positive, arrows); ft, tropomyosin (positive); /'. CEA (negative); /', S-100 protein (positive). Hematoxylin counterstain. f to/, x 360. The tissue sections from
metastatic tumors present in the regional lymph nodes showed the positive stainings for the other specific antigens excluding CEA, in a fashion similar to that observed
in the metastatic tumors in the lung (data not shown).
CANCER
RESEARCH
VOL. 45 AUGUST 1985
3890
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.
Intermediate-sized Filaments and Specific Markers in a Human
Salivary Gland Adenocarcinoma Cell Line and Its Nude Mouse
Tumors
Mitsunobu Sato, Yoshio Hayashi, Tetsuo Yanagawa, et al.
Cancer Res 1985;45:3878-3890.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/45/8/3878
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1985 American Association for Cancer Research.