A New

Immunocompetent

Murine

Model for Oral Cancer

Bert W.

O'Malley, Jr, MD; Keary A. Cope;

Candace S. Johnson, PhD;

Objective: To develop and characterize a new immu-

nocompetent murine model that attempts to parallel the

clinical and biological nature of head and neck cancer.

Mary

Schwartz,

MD

of the mouth of C3H/HeJ mice after 5

to 7 days. Local
mylohyoid musculature and mandible
was present. Cervical lymph node and pulmonary me-

invasion into the

tastases were

Design: The growth rate and histologic characteristics

R.

identified between 2 and 3 weeks.

of the SCC VII/SF cell line were initially determined in

tissue culture experiments. Animal experiments were subsequently performed on C3H/HeJ mice. Using direct injection, 5 \m=x\105 SCC VII/SF cells were delivered to the floor
of the mouth of each animal. Animals were killed after
1, 2, and 3 weeks, and tumor growth, invasion, and regional and distant metastases were evaluated.

Conclusions: This study introduces a new oral cancer animal model that shows initial locoregional tumor invasion,
direct extension into the neck, early cervical metastases,
and pulmonary metastases. These clinical and histopathologic attributes reflect the biological behavior and tumor
progression seen in human oral cancer and therefore provide a model for clinically applicable research for primary
and metastatic head and neck cancer.

Results: Squamous cell carcinomas that could be palpated and measured externally were identified in the floor

Arch

The

use

Otolaryngol Head Neck Surg. 1997;123:20-24

of animal models to

study human head and neck

cancer is not new;

to our

however,

knowledge, the devel

opment of an immunocom

petent squamous cell cancer model with tu

growing in a natural head and neck site

has not been described. The ability to study
human head and neck cancer centers on the
use of an athymic, nude mouse first described
in 1968by Pantelouris.1 This model was used
to study human head and neck carcinomas
implanted subcutaneously in the flank by
Braakhuis et al,2 who observed tumor growth
in only 26% of the mice. In this study of 130
transplanted human head and neck cancers,
no evidence of mtastases was found, sug
gesting that the true biological behavior of
the disease was not evident. In other reports
of various human tumor implantations in
nude mice, mtastases have not been ob
served34 or have occurred in fewer than 2%
of the cases.3 More recent reports have shown
mtastases in 30% of animals after human
tumor implantations, but most were adenocarcinomas.6 In this latter study, only 2 of
17 animals that underwent implantation with
squamous cell cancer had mtastases. Over
all, most studies with tumors implanted into
the subcutaneous tissue of the abdomen,
nape of the neck, or flank show a benign
growth pattern. These subcutaneous immors

Departments of
Otolaryngology-Head and

From the

Surgery, The Johns

Hopkins University, Baltimore,
Md (Dr O'Malley) and the
University of Pittsburgh,
Pittsburgh, Pa (Ms Cope and
Dr Johnson), and the
Department of Pathology,
Baylor College of Medicine,
Neck

Houston,

Tex

(Dr Schwartz).

plants grow in a confined, rounded pattern

and lack notable invasion of surrounding tis

and aggressive metastatic activity.7 Be

human patients have aggressive car
cinomas that locally invade and metastasize,
this lack of invasive behavior in established
animal tumor models may limit the appli
cability of new cancer therapies or treatment
sues

cause

strategies.

More consistent animal models for

studying tumor mtastases have been de

scribed, but they entail intravenous injection
into a tail vein or intrasplenic or intraperitoneal injections.89 These models do not truly
represent the spread of tumor seen in patients
with head and neck cancer. They are not
spontaneously occurring and they result from
large inocula of tumor cells into sites distant
to

and not inclusive of head and neck can

cer

regions.

The development of an oral tumor

model was first described by Fitch et al,10 but
they did not quantify and describe the growth

patterns. In a more complete study, Dinesman et al7 described in detail a floor-of-themouth (FOM) human tumor model in nude

See Materials and Methods

Downloaded From: http://archotol.jamanetwork.com/ by a University of British Columbia User on 05/31/2016

on

next

page

mice. In this report, 80% of the animals had bone invasion,

MATERIALS AND METHODS

IN VITRO

The SCC VII/SF cell line is a spontaneously arising

squamous cell carcinoma syngeneic to C3H/HeJ
mice.13 SCC VII/SF cells were cultured in tissue cul
ture flasks (T-75, Corning Ine, Corning, NY) con
taining 30 mL of RPM1-1640 media (Sigma Chemi
cal Co, St Louis, Mo), 12% bovine calf serum, 1% dual
penicillin and streptomycin, and 1% L-glutamine. Cells
were maintained in 10% carbon dioxide incubators.
Two days after the initial plating, the cells were har
vested for the first time by trypsinization and then
expanded using standard techniques. After 2 days,
the cells were harvested, concentrated using centrifugation, and then counted using a hemocytometer. Viable cells were assessed using trypan blue ex
clusion and resuspended at a final concentration of
5X106 viable cells per milliliter in Hanks buffered
saline solution for use in the in vivo experiments.
An SCC VII/SF growth rate experiment was per
formed in duplicate 6-well tissue culture plates us
ing 1X103 cells per well harvested after the first pas
sage. At days 1, 2, and 3, cells were harvested from 4
individual wells and counted as described herein.

and 44% of the animals experienced pulmonary mtastases.

This head and neck tumor model was later adapted to study
the effects of "suicide" gene therapy for the treatment of hu
'!
man head and neck cancer. The T-cell deficient nude mouse
model, however, has a major disadvantage, because stud
ies with immunothrapies using direct cy tokine gene trans
fer and cancer vaccine strategies are impossible.
To study antitumor immune responses involving
cells, multiple immunocompetent murine models for
studying nonhead and neck tumors have been devel
oped. The use of an immunocompetent murine model
to study the possible effects of immunothrapies for solid
squamous cell tumors established in the flanks of C3H/
HeJ mice also has been described.12 These models have
the same disadvantage of other subcutaneous flank mod
els, because they produce localized tumors lacking natu
ral mtastases or showing patterns dissimilar to those of
patients with head and neck cancer.
The first description of the C3H/HeJ mouse dates to
1968, when Sultzer13 discovered its unique resistance to li
popolysaccharide endotoxins. This resistance is highly spe
cific, and C3H/HeJ mice can respond normally to other
bacterial B- and T-cell activators with complete im-

munocompetence.14

study was to use the C3H/HeJ

develop an immunocompetent murine model
that parallels more closely the clinical and biological na
The purpose of this

mouse

IN VIVO

Animal experiments, including designations for sur

vival outcomes, were approved by the Johns Hop
kins University Animal Care and Use Committee. All

experiments were performed on C3H/HeJ mice (Jack

son

ImmunoResearch Labs Ine, West Grove, Pa)

us

ing sterile techniques under a laminar flow hood. Mice

7 to 8 weeks old were anesthetized by inhalation of

metofane (Pitman Moore Co, Mundelein, 111) in a
closed container system. Using a 1-mL tuberculin
syringe (Hamilton Co, Reno, Nev) and 23-gauge
needle, a 100- volume of 5X103 SCC VII/SF cells
was injected into the FOM of the mouse. The cell sus
pension was slowly injected at the depth of the mylohyoid muscle. The animals were then maintained
in standard housing conditions until they were killed.
After they were killed, a standard apron-flap in
cision was made in the neck of the experimental ani
mals. Sharp dissection with surgical scissors pro
vided access to the tumors growing in the FOM
musculature. The tumors could be identified origi
nating in the mylohyoid musculature, and most
growth extended into the submandibular area at 7
days. Because of this growth pattern, 3-dimensional
access for caliper measurements was possible. All tu
mors were measured in ventral-to-dorsal, left-toright, and cephalad-to-caudad directions.

HISTOLOGIC ANALYSIS
At the time of necropsy, harvested tumor, surround
ing tissues, and distant organs were fixed in 10% buff
ered formaldehyde. The specimens were then em
bedded in paraffin, sectioned, and stained with

hematoxylin-eosin.

to

ture of head and neck cancer. We characterize the inva

sive growth pattern of squamous cell cancer in this model

and show regional lymph node and eventual pulmonary

mtastases. This new model should provide a founda
tion for studying novel immunothrapies, including
schemes for cytokine delivery for head and neck cancer.
RESULTS

IN VITRO

Growth curve experiments show that SCC VII/SF cells

grow logarithmically in a monolayer in vitro. The cell dou
bling time is estimated at 18 to 24 hours based on these
data (Figure 1).
IN VIVO

A pilot study to perfect the method of establishing consis

tent FOM tumors was performed. After appropriate anes
thesia, 20 C3H/HeJ mice received FOM injections of 5X105
SCC VII/SF cells. Five of the initial animals injected with

cells died shortly after the implantation procedure

completed. These animals underwent necropsy, and
the findings were consistent with edema of the base of the
tongue and hypopharynx. The cause of death was attrib
uted to a technical error in delivering the tumor volume
too deep and posteriorly in the FOM, resulting in airway
compromise. The technique was modified to deliver the
tumor cells more anteriorly in the FOM near the genial tu
bercle of the mandible, and no further deaths occurred.
Animals (5 each) were then killed at 4, 7, or 10 days
after tumor implantation to evaluate tumor growth and
tumor

was

develop a consistent measurement technique. At 4 days,

Downloaded From: http://archotol.jamanetwork.com/ by a University of British Columbia User on 05/31/2016

Figure 1. Growth curve for SCC VII/SF cells in vitro. The cells grow
logarithmically in a monolayer with a doubling time estimated at 18 to 24
hours. Error bars indicate SD.

of the tumors were smaller than 50 mm3 and were

difficult to recognize accurately and measure with the cali
pers without substantial dissection in the FOM. At 7 days,
tumor sizes ranged from 70 mm3 to 140 mm3. At 10 days,
the tumor sizes ranged from 160 to 200 mm3.
From previous studies in the nude mouse,11 it was
determined that the desirable tumor size should be at least
50 mm3. Tumors smaller than 50 mm3 are not detected
easily in the FOM musculature, which precludes accu
rate identification and caliper measurement. Moreover,
a critical tumor volume existed in the gene therapy strat
egy: animals with tumors larger than 150 mm3 showed
lower treatment and survival outcomes. Based on these
data, the time of 7 days after tumor implantation was cho
sen for subsequent experimentation.
To further define the growth pattern of the SCC VII/
SF cell line in vivo, 30 C3H/HeJ mice received FOM injec
tions of 5 X103 SCC VII/SF cells. Animals were observed daily,
and a clinically apparent fullness in the FOM was detected
between days 5 and 7, although no appreciable measurement
of tumor size could be obtained until after the animals were
killed. At 1 week, 10 animals were killed using the inhalational agent metofane. A surgical incision was then made
in the midneck that resembles a classic apron-flap incision.
Dissection to the mandible disclosed tumors growing from
the FOM that ranged from 70 to 160 mm3 as determined by
caliper measurement. Microscopic examination of tumor
specimens disclosed a poorly differentiated squamous cell
cancer (Figure 2, top). All tumors clinically invaded the
FOM musculature and the larger tumors abutted the man
dible but were easily dissected in a tissue plane from the peri
osteum. There was no evidence of lymphadenopathy. The
FOM musculature invasion was confirmed by microscopic
examination (Figure 2, bottom). The liver and lungs from
3 of the animals were harvested and sent for evaluation of
mtastases. Microscopic examination disclosed normal tis
sue architecture and no evidence of mtastases.
After 2 weeks, 10 animals were killed and the FOM
was dissected in the same fashion. Tumors ranged from 1500
to 2250 mm3 by caliper measurement. All tumors invaded
the FOM and tongue musculature, but the predominance
most

Figure 2. Microscopic examination of tumor specimens growing in vivo in

the floor of the mouth of C3H/HeJ mice. Top, Poorly differentiated squamous
cell carcinoma with many mitotic figures. Subsequent locally invasive tumor
and mtastases had an identical histologie appearance. Bottom, Invasion of
floor of mouth musculature (right) by squamous carcinoma (left)
(hematoxylin-eosin. original magnification 500 top] 250 bottom]).

of growth was into the subcutaneous space of the neck. This

growth produces a palapble mass that can be measured ex
ternally with calipers. Eight of 10 mice had frank mandible
invasion, 7 of 10 had clinically enlarged regional lymph nodes,
and 5 of 10 had gross tumor invasion into the salivary gland.

Microscopic examination disclosed local tumor invasion

of the mandible and salivary gland and regional lymph node

(Figure 3). The livers and lungs of each ani
mal were microscopically evaluated, and no mtastases were
found.
Six animals died on days 18 and 19, and necropsy dis
closed a large FOM tumor in each animal that extended into
the neck, but with no involvement of the base of the tongue,
hypopharynx, or larynx. There were visible and palpable nod
ules entirely infiltrating the lungs, with the remaining ab
dominal organs containing no palpable masses. However,
because of rigor mortis and the presence of tissue autolysis, no definite pathologic evaluation was possible. It could
only be presumed that the animals died of pulmonary m
tastases. The remaining 4 animals were killed immediately
because of the onset of anorexia and cachexia (failure to eat
and/or loss of more than 10% body weighta designated
end point by the Johns Hopkins University Animal Care and
Use Committee) 3 weeks after the original tumor implan
tation. External clinical measurements, with accuracy con
firmed by open surgical dissection, disclosed massive tumors
mtastases

Downloaded From: http://archotol.jamanetwork.com/ by a University of British Columbia User on 05/31/2016

Figure 3. Regional metastasis to cervical lymph node (normal follicle seen

at lower left) by poorly differentiated squamous cell carcinoma (lower
right) (hematoxylin-eosin, original magnification 125).

Figure 4. Metastatic squamous cell carcinoma in lung (hematoxylin-eosin,

original magnification 250).

ranging from 2500 to 4500 mm3 with regions of focal cen

in human patients, but this model is designed to

closely approximate human disease, thereby reducing the
need for subjective extrapolation.
An even greater strength is the actual anatomical site
and the initial locoregional aggressiveness of this tumor,
with early direct extension into the neck and later cervical
lymph node and pulmonary mtastases. These clinical and
histopathologic attributes reflect the biological behavior and

tral necrosis. Cervical lymphadenopathy, frank bony inva

sion, and direct tumor extension to the deep structures and
vasculature of the neck were present in all animals. The base
of tongue, hypopharynx, and larynx showed no evidence
of tumor invasion, so there was no evidence of upper air
way compromise. The lungs, livers, kidneys, esophagus, stom
ach, and small intestine were harvested from each of the 4
animals and sent for histologie evaluation. Metastatic car
cinoma was found in all 4 lung specimens and consumed
more than 70% to 80% of alveolar tissue as estimated by his
tologie examination. The remaining organs were free of tu
mor. The metastatic tumors were composed of poorly dif
ferentiated squamous cell cancer and there was a high level
of mitotic activity (Figure 4). No necrosis was identified
in the metastatic lesions.
COMMENT

Syngeneic tumor implants into the nape or flank of mice

often have been used to study tumor biological character

istics and cancer therapies. The development of a realistic
clinical correlate of true human disease in immunocom

petent animals, however, is uncommon. We are unaware

of a tumor model using immunocompetent animals that
parallels the natural progression of human head and neck

closely as in our C3H/HeJ mouse model. Al

though adapted from reports using nude mice for head and
neck cancer studies,711 this model has the advantage of imcancer as

munocompetence, which allows the use of immune modi

fiers such as cytokines in the investigation of novel cancer

therapies, including direct cytokine gene delivery, a strat

egy now under investigation in our laboratory.
This new model has 2 major strengths. The first ma
jor strength is the confirmed tumor histologie character
istic of squamous cell carcinoma, which is the most com
mon tumor in the head and neck region. Because it is well
known that different tissue types vary widely in their clini
cal behavior and response to select therapies, the focus on
squamous cell cancer may allow a more effective human
application of the proposed novel therapies. We under
stand that there is subjectivity in translating animal tumor
pathologic characteristics and animal study results to out-

comes

tumor progression seen in human head and neck cancer.

The importance of this point is obvious to the head and
neck surgeon, who is well aware of the critical vascular and
nerve structures present in the neck. No other model pro
vides these critical structures in proximity to squamous cell
carcinoma. This factor becomes important in assessing safely
issues of cytokine delivery, which can have profound toxic
effects of capillary leak and blood vessel injury.16 The more
anatomical model described in this work may help pre
dict such outcomes and allow strategy changes before en
tering clinical trials and encountering these potential com
plications. With respect to the metastatic aspect of this
model, studies by Smith17 based on autopsies of patients
with head and neck cancer disclosed a 17% to 57% inci
dence of distant mtastases, but at the time of diagnosis of
the primary cancer, only 1% to 14% of these patients had
distant disease. This result correlates with our findings of
mtastases at the level of end-stage disease or after pro
longed presence of the primary tumor.
Although we report percentages of invasion and meta
static patterns, these numbers were not calculated to be com
pared with human values. They were included to help read
ers assess the usefulness of this model in their investigations
of oral cancer. The rate of metastasis and extent of local
tissue invasion can be operator-dependent when consid
ering the technique of creating the FOM tumors. We de
livered the tumor cells anteriorly in the FOM muscula
ture and in a very slow manner to avoid direct vascular
injection. Extensive previous experience in a nude mouse
model by one of us (B.W.O.) provides consistency with this
tumor implantation technique, thus allowing for credibil
ity in interpretation of these data. Furthermore, because
the tumors show early extension into the neck, external
measurements are possible and can be used to provide an
accurate measurement of tumor growth or inhibition in

Downloaded From: http://archotol.jamanetwork.com/ by a University of British Columbia User on 05/31/2016

Suggested Preliminary Experiments to Develop

Immunocompetent Murine Model for Oral Cancer
Point of Interest
Established tumor and
floor-of-mouth invasion
Mandible and salivary
gland invasion
Lymph node mtastases
Pulmonary mestastases

No. of
Animals

Time to Occurrence
After Tumor
Implantation, d

20

5-7

10
15
15

10-14
14
18-21

studies with new cancer therapies. The Table gives our sug
gestion for the number of animal procedures that should
be performed to develop the basic technical proficiency re
quired to achieve consistent tumor implantation and con
sistent identification of nodal and pulmonary mtastases.
The total number of 60 animals was derived from efforts
to teach young investigators who have had no previous ani
mal research experience. The Table is intended to be a gen
eral guide to those who may consider application of this
model to their experimental systems.
For long-term studies, animal care regulations rightly
require designation of the experimental end point and im
mediate killing ofthe animals when tumors reach sizes more
than 2 cm, when animals have cachexia or anorexia, or when
tumor erosion through skin occurs. These defined end points
will allow standardization and comparison between each in
vestigator who uses this model to study cancer therapies.
A potentially important drawback of this model is the
rapid tumor growth and identification of pulmonary m
tastases at or near the time of designated experimental end
points. Distant mtastases therefore occur late in the course
of disease in these mice. This finding may limit experi
mentation directed at developing new therapies for the pre
vention or treatment of distant mtastases. Although not
performed at present in our laboratory, one possible solu
tion to this drawback would be to excise the primary tu
mor before massive tumor growth (ie, about week 2), which
may provide survival beyond 3 weeks. Although techni
cally feasible, pilot experimentation would be required to
assess whether survival can be increased and whether m
tastases still develop at the same time interval.
The overall benefit of studying tumors in this system
is apparent, because strategies for novel therapies such as
cancer vaccines and gene therapies may be altered by muscle
and bone invasion, tumor proximity to critical vascular and
nerve structures, or the presence of regional or distant m
tastases. Using this model to investigate gene delivery strat
egies may allow tailoring the choice and method of deliv
ery of therapeutic genes to the level and aggressiveness of
the tumor and the proximity to critical structures. This may
enhance the outcomes and safety of novel therapies when
applied to human clinical situations. The presence of regional
lymphadenopathy with histopathologic confirmation of m
tastases is also important in the planning of new cancer thera
pies. The treatment of head and neck cancer beyond stage
I should include some consideration of cervical nodal m
tastases. The incidence of clinically apparent or occult m
tastases is estimated at more than 50% for stage II and higher

head and neck cancers.18 This model allows for treatment

of nodal disease surgically or with novel therapies.
We do not propose that experimentation and treat
ment outcomes in animal models can be correlated di
rectly with human results. Animal models, however, are
a necessity to establish safety and consistency and to maxi
mize the therapeutic strategy before embarking on hu
man clinical trials. The ability to parallel the biological
behavior of local invasion and regional and distant m
tastases may allow investigators to develop more effec
tive therapies for human head and neck cancer.

Accepted for publication October 11, 1996.

This project was supported in part by a Lucille Markey
Trust grant, Pittsburgh Cancer Institute, Pittsburgh, Pa.
We thank Annetta Walker, Dermatopathology Labo
ratory, Baylor College of Medicine, Houston, Tex, for prepa
ration of histologie slides, and Ming-Jei Chang, MD, Ruth
A. Modzelewski, PhD, and Derick M. Rssel for assis
tance and expertise that enabled completion of this project.
Reprints: Bert W. O'Malley, Jr, MD, Department of
Otolaryngology-Head and Neck Surgery, The Johns Hop
kins University, PO Box 41402, Baltimore, MD 21287.
REFERENCES
1. Pantelouris EM. Absence of
370-371.

thymus

in a mouse mutant. Nature.

1968;217:

2. Braakhuis, BJM, Sneeuwloper G, Snow GB. The potential of the nude mouse
xenograft model for the study of head and neck cancer. Arch Otorhinolaryn-

gol. 1984;239:69-79.

3. Povlsen CO, Rygarrd J. Heterotransplantation of human epidermoid carcinomas to the mouse mutant 'nude.' Acta Pathol Microbiol Scand A. 1972;80:
713-717.
4. Sordat M, Merenda C, Carrel S. Invasive growth and dissemination of human
solid tumors and malignant cell lines grafted subcutaneously to newborn nude
mice. In: Proceedings of Second Workshop on Nude Mice; 1977; Tokyo, Japan. Pages 313-325.
5. Sharkey FE, Fogh J. Metastasis of human tumors in athymic nude mice. Int J
Cancer. 1979;24:733-738.
6. Neulat-Duga I, Sheppel A, Marty C, et al. Metastases of human tumor xenografts in nude mice. Invasion Metastasis. 1984;4:209-224.
7. Dinesman A, Haughey B, Gates GA, Aufdemorte T, Von Hoff DD. Development
of a new in vivo model for head and neck cancer. Otolaryngol Head Neck Surg.

1990;103:766-774.

8. Naito S, Von Eschenbach AC, Giavazzi R, Fidler IJ. Growth and metastasis of
tumor cells isolated from a human renal cell carcinoma implanted into different organs of nude mice. Cancer Res. 1986;46:4109-4115.
9. Naomoto Y, Kondo H, Tanaka N, Orita K. Novel experimental models of human cancer metastasis in nude mice. J Cancer Res Clin Oncol. 1987;113:544-549.
10. Fitch KA, Somers KD, Schecter GL. The development of a head and neck tumor
model in the nude mouse. In: Wolf GT, Carey TE, eds. Head and Neck Oncology
Research. Amsterdam, the Netherlands: Kuglere & Ghedini; 1988:187-190.
11. O'Malley BW Jr, Chen S-H, Schwartz MR, Woo SLC. Adenovirus-mediated gene
therapy for human head and neck squamous cell cancer in a nude mouse model.
Cancer Res. 1995;55:1080-1085.
12. Grandis JR, Chang M-J, Yu W-D, Johnson CS. Antitumor activity of interleukin-1\g=a\and cisplatin in a murine model system. Arch Otolaryngol Head Neck

Surg. 1995;121:197-200.

13. Sultzer BM. Genetic control of leucocyte responses to endotoxin. Nature. 1968;
219:1253-1254.
14. Sultzer BM, Goodman GW. Endotoxin protein: a B-cell mitogen and polyclonal
activator of C3H/HeJ lymphocytes. J Exp Med. 1976;144:821-827.
15. Fu KK, Rayner PA, Lam KN. Modification of the effects of continuous lowdose radiation by concurrent chemotherapy infusion. Int J Radiant Oncol Biol

Phys. 1984;10:1473-1478.

16. West WH, Tauer KW, Yannelli JR. Constant infusion of recombinant interleukin-2 in adoptive immunotherapy of advanced cancer. N Engl J Med. 1987;
316:898-905.
17. Smith JD. Distant metastasis in carcinoma of the head and neck. In: English GM,
ed. Otolaryngology. New York, NY: HarperCollins Publications Inc; 1984;5:1-13.
18. Mendenhall WM, Million RR, Cassisi NJ. Elective neck irradiation in squamous
cell carcinoma of the head and neck. Head Neck. 1980;3:15-20.

Downloaded From: http://archotol.jamanetwork.com/ by a University of British Columbia User on 05/31/2016