PII: 80964-1955 (97)00047-X
Oral Oncology, Vol. 33, No. 6, pp. 426-430, 1997
3'/ 1997 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
Expression of E1AF, an ets-Family Transcription Factor, is
Correlated with the Invasive Phenotype of Oral Squamous
K. Hida, 1 M. Shindoh, 2,3 K. Yoshida, 3 A. Kudoh, 1 K. Furaoka, 1 T. Kohgo, 2 K. Fujinaga 3
and Y. Totsuka I
1Department of Oral Surgery; 2Department of Oral Pathology, Hokkaido University School of Dentistry; and
3Department of Molecular Biology, Cancer Research Institute, Sapporo Medical University School
of Medicine, Sapporo, Japan
E1AF is a newly identified ets-oncogene family transcription factor. Previous reports have noted that
EIAF can upregulate promoter activities of several matrix metalloproteinase (MMP) genes and
showed that invasive potentials of oral squamous cell carcinoma-derived cell lines are correlated with
expression of EIAF and MMPs. The invasive phenotype is restrained by transfection with an antisense
EIAF expression vector. Thus, EIAF is thought to be highly correlated with malignant potentials of
cancer cells. However, little is known about EIAF expression and cancer cell malignancies in in vivo
tumours. In the present study, 27 oral squamous cell carcinoma (SCC) specimens were examined
using RT-PCR, Southern blot hybridisation and in situ hybridisation (ISH) and compared to the
clinicopathological parameters. Among the 27 patients, EIAF was detected in 15 cases. EIAF mRNA
was detected in 13 of 17 invasive SCCs, whereas the majority of SCCs not expressing EIAF showed an
expansive growth pattern. Increased prevalence of EIAF-positive oral SCC was observed in cases with
nodal metastasis. These results indicate that EIAF may be involved in cancer cell malignancies
through its ability to promote invasive potential. () 1997 Elsevier Science Ltd. All rights reserved.
Key words: E1AF, invasion and metastasis, oral squamous cell carcinoma
Oral Oncology, Vol. 33, No. 6, pp. 426-430, 1997
The prognosis of patients with oral squamous cell carcinoma
(SCC) is correlated with invasive and metastatic potentials of
cancer cells, and proteolytic degradation of the extracellular
matrix is required for tumour cells to invade basement mem-
brane, stromal matrix and vascular tissue . Matrix metal°
loproteinases (MMPs) are thought to have important roles in
degradation of extracellular matrices . Expression of
MMPs as well as other extracellular matrix-degrading pro-
teinases is elevated in a variety of invasive carcinomas [3-5],
including oral squamous cell carcinoma [5, 6]. So MMPs are
thought to be closely related to invasive potential and lymph
node metastasis. The transactivating mechanism of MMP
genes is still not fully understood. However, it has been
shown that expression of some MMPs is regulated by ets-
related transcription factors [7, 8].
E1AF is a newly identified ets-oncogene family transcrip-
tion factor . Chloramphenicol acetyltransferase (CAT)
assay revealed that E1AF can upregulate promoter activities
of MMP-1, -3 and -9 . We have already reported the
Correspondence to M. Shindoh.
Received 11 Feb. 1997; accepted 4 July 1997.
possible correlation of MMP-1, -9 and EIAF expression in
highly invasive oral squamous cell carcinoma-derived cell
lines , and that the ectopic expression of the E1AF gene in
non-invasive human breast cancer cells leads to acquisition of
the invasive phenotype . Moreover, we have also found
that cells transfected with an antisense-E1AF expression
vector had reduced invasive ability accompanied by reduction
of MMP-1, -3 and -9 expression . Therefore, E1AF is
thought to be closely associated with invasion and metastasis
of cancer cells by activating MMP genes. However, the rela-
tionship between E1AF expression and cancer cell malig-
nancies has not yet been completely understood in in vivo
tumours. The present study is intended to investigate the
ratio of E1AF mRNA expression in oral squamous cell carci-
nomas and to compare the clinical parameters.
MATERIALS AND METHODS
27 patients diagnosed as having oral squamous cell carci-
noma were investigated. Fresh biopsy specimens in oral
SCCs were collected at the Department of Oral Surgery,
Hokkaido University Dental Hospital, Sapporo, Japan from
January 1994 to June 1996. Clinical and pathological features
E1AF Expression in Oral SCC
are shown in Table 1. Clinical staging of the disease was
determined according to the UICC criteria (1987) .
Specimens were fixed in 10% neutral-buffered formalin,
embedded in paraffin and examined histopathologically. His-
topathological estimation of turnout differentiation was
undertaken using the WHO classification, and patterns of
cancer invasion were determined according to Totsuka et al.
RT-PCR and Southern blot hybridisation
Tumour tissues were snap-frozen in liquid nitrogen and
stored at -80°C until use. Approximately 50 mg of each tis-
sue sample was homogenised, and RNA was isolated using
Trizol reagent (GibcoBRL, Gaithersburg, Maryland, U.S.A.)
as described by the manufacturer, cDNA was synthesised
using oligo(dT) primer (GibcoBRL) and Superscript II
reverse transcriptase (GibcoBRL). E1AF in cDNA templates
was amplified by PCR. We used a sense primer, 5 t-
140-169), and an antisense primer,
TAGGGGCGACTG-3 ~ (n.n. 378-397), corresponding to
the sense sequences of the E1AF acidic domain and the
antisense sequences of the Q-rich domain, respectively. The
sequences of these primers are highly specific for E1AF.
cDNA was placed in 50 ~tl of 1 ×PCR buffer (10mM Tris-
HC1, pH 8.3, 50mM KC1, 1.5mM MgC12, 0.01% (w/v)
gelatin), dATP, dGTP, dCTP, dTTP (0.2 mM each), primers
(0.2 pM each) and 2.5 units of Taq-polymerase. Amplifica-
tion with 30 cycles was carried out on a DNA thermal cycler
(Perkin-Elmer Cetus, Norwalk, Connecticut, U.S.A.) with
1 min denaturation at 94°C, 2min annealing at 55°C and
2 min extension at 72°C. After amplification, 15 ktl aliquots of
the amplified reaction mixture were subjected to electro-
phoretic analysis on 1% agarose and stained with ethidium
bromide. They were denatured twice in 0.5 M NaOH, 1.5 M
NaCl for 20min, neutralised twice in 0.5M Tris-HC1 (pH
7.5), 3 M NaC1 for 20 min then transferred to nitrocellulose
membranes (Schleichner & Schell, Dassel, Germany) by the
capillary method. They were hybridised at 42°C for 12 h with
a 32p-labelled E1AF cDNA probe containing the n.n. 196-
706 fragment by using the BcaBest labelling kit (Takara,
Tokyo, Japan). Filters were washed once in 2xSSC, 0.1%
SDS for 15min at room temperature (20°C), once in
1 ×SSC, 0.1% SDS for 15 min at room temperature and then
twice in 0.2xSSC, 0.1% SDS for 15min at 50°C. The filters
were then exposed to KODAK XOmat X-ray film for 3-12 h
In situ hybridisation
Paraffin embedded specimens were cut at 5pm and
mounted on 3-amino-propyl-triethoxysilane-coated
slides. They were deparaffinised, rehydrated and digested as
described previously . After acetylation (0.1 M triethano-
lamine and 0.25% acetic anhydride), tissue specimens were
covered with an in situ hybridisation mixture containing 50%
formamide, 0.6M NaC1, 10mM Tris-HC1 (pH 7.5), 1 mM
EDTA, 0.Smg/ml ssDNA, 0.5mg/ml tRNA, 0.1% SDS,
1 xDenhardt's solution and a digoxigenin-labelled riboprobe.
A 472 bp E1AF fragment was subcloned into Bluescript II
KS(+) vector for in vitro transcription. Hybridisation was
Table 1. Clinical and genetic findings in 27 patients with oral SCC
expression No. Age Sex Primary site T N
*FOM, floor of mouth.
K. Hida et al.
carried out at 50°C for 36h. After hybridisation, RNAse
digestion and a high stringent wash were carried out. Tissue
specimens were incubated with an anti-digoxigenin polyclo-
nal antibody conjugated with immunogold and visualised by
silver enhancement reagents (Boehringer Manheim, Yama-
nouchi, Japan), followed by counterstaining with methyl-
Independence of clinical parameters and histological fea-
tures concerning E1AF expression was examined using the
1 2 3 4 5 6 7 8 9 10
Fig. 1. Results of RT-PCR (A) and Southern blotting (B). A
257bp amplified fragment of E1AF mRNA expression was
detected in RT-PCR (A) and confirmed by Southern blotting
Fig. 2. (A) Histological findings of E1AF-positive infiltrative
cancer tissue (case no. 10) (H & E: original magnification
250x). (B) E1AF mRNA was detected in the cytoplasm of
cancer cells invading into connective tissue by ISH (ISH: ori-
ginal magnification 400×).
Table 1 shows clinicopathological and genetic findings in
27 patients. The patients were 28-88 years old, including 13
cases of tongue SCC, 4 of the cheek, 2 of the floor of the
mouth, 5 of the gingiva, 2 of the lip, and 1 of maxillary SCC.
Clinical staging of patients was undertaken utilising the cri-
teria of the UICC (1987). Among the 27 patients, 6 cases
were classified as T1, 16 cases as T2 and 5 cases as T3/T4.
20 cases were classified as NO, 4 as N1, and 3 were N2 in
primary clinical examination. Subsequent metastasis was
found in 4 of the 20 NO cases in the follow-up period of 6-30
months (Table 1).
The WHO classification was applied to assess histological
malignancy. There were 11 cases of grade I and 16 cases of
grade II. The mode of tumour invasion was estimated by the
criteria of Totsuka et al. . 17 cases showed an infiltrative
pattern and 10 cases were expansive.
27 oral squamous cell carcinoma specimens were investi-
gated for the expression of EIAF mRNA using RT-PCR,
Southern blot hybridisation and in situ hybridisation. A
257bp amplified fragment of E1AF mRNA expression was
detected in 15 of the 27 oral squamous cell carcinomas
(55%) and it was confirmed by Southern blotting (Fig. 1).
Moreover, we carried out in situ hybridisation to investigate
the localisation of E1AF mRNA expression, using an EIAF
antisense RNA probe. Signals of EIAF were seen in the
cytoplasm of cancer cells invading into the connective tissue
(Fig. 2). E1AF expression was detected in 3 of 6 TI, 9 of 16
T2, 1 of I T3 and 2 of 4 T4 stage cases, and there seemed to
be no significant difference among the turnouts expressing
E1AF mRNA and the T stage (Table 1, Fig. 3).
The incidence of cases expressing EIAF mRNA was more
frequent in cases in the N1/N2 stage and in the cases with
subsequent metastases than in the cases with no nodal
 N1/N2 (N+)
(NO -~- N+)
• NO (N-)
+ - + - + - + -
T1 T2 T3 T4
Fig. 3. Clinical classification and E1AF expression. E1AF
expression was more frequently detected in the cases of N+ or
with subsequent metastasis.
E1AF Expression in Oral SCC
metastasis (Fig. 3), especially in the cases with early lesions of
T1/T2. All advanced T3/T4 cases were N1/N2 stage (N+)
whether they were EIAF-positive or -negative (Table 1,
Fig. 3). However, primary metastasis or subsequent metasta-
sis was observed in 5 of 12 E1AF-positive T1/T2 cases
whereas only 1 of 10 E1AF-negative T1/T2 cases had subse-
quent metastasis (Fig. 3).
Histologically, E1AF mRNA was detected in 7 of 11 grade
I SCCs, and 8 of 16 grade II SCCs (Fig. 4). E1AF mRNA
was identified in 13 (76%) of 17 invasive carcinomas, but
only in 2 (20%) of 10 expansive carcinomas (Fig. 4). E1AF
expression was significantly frequent in invasive tumours
compared to expansive tumours (P< 0.04).
Invasion and metastasis are closely associated with poor
patient prognosis , and MMPs have been shown to be
involved in this step by degrading various extracellular
matrices . It has been reported that there is a close rela-
tionship between MMP expression and malignancy of cancer
cells [17, 18]. Expression levels of mRNA of MMPs have
been reported to be elevated in human invasive squamous cell
Fig. 4. Pathologic findings and E1AF expression. No correla-
tion could be found between WHO classification and E1AF
expression (A). However, E1AF-positive cases are more
numerous in infiltrative tumours
than expansive ones
carcinomas and to be strongly related to cancer metastasis
[5,6, 19,20]. It has been shown that there are Ets binding
motifs in the transcription regulatory regions of many MMP
genes, and suggested that the Ets binding site plays an
important role in activating several MMP genes [7, 8, 21,22].
E1AF is an adenovirus 5 E1A enhancer-binding protein
. Its cDNA was isolated from a HeLa cell X gt 11
expression library, and contains the Ets-domain in the C-
terminal of its amino acid sequence, which is identical to the
ets-oncogene family . E1AF was shown to be located on
chromosome 17q21  and its translocation was found in
Ewing's sarcoma and an undifferentiated sarcoma of infancy
[25, 26]. Thus, EIAF is thought to have oncogenic proper-
Higashino et al. , using a transient expression assay,
reported that E1AF can upregulate promoter activities of
MMP-1, -3, and -9, and our previous report showed that
ectopic expression of E1AF induces an invasive phenotype of
non-invasive breast carcinoma . We have reported that
E1AF expression is highly correlated with malignant pheno-
type in oral cancer cell lines by activating MMP protein pro-
duction , and cells transfected with an antisense E1AF
expression vector caused downregulation of MMP-1, MMP-
3 and MMP-9, and consequently markedly reduced the
invasive ability of cancer cells .
In this study, we investigated the E1AF expression in 27
cases of oral SCC using RT-PCR, Southern blotting and in
situ hybridisation. E1AF mRNA was detected in 15 patients
(55%) and there were no significant differences between
E1AF mRNA expression by gender, age or primary site
(Table 1). E1AF-positive tumours were found in all T stages,
and no correlation could be found between tumour size and
E1AF expression (Fig. 3, Table 1). On the other hand, E1AF
expression was more frequently found in N1/N2 cases
(Fig. 3). The advanced T3/T4 cases had nodal metastasis
whether they were E1AF-positive or -negative. It is interest-
ing to note that 5 of 12 E1AF-positive T1/T2 cases were
identified with nodal metastases, although only 1 of 10
E1AF-negative cases showed subsequent metastasis from the
early lesion of SCC. Five of 6 cases in the early stage tumour
(T1/T2) with nodal metastasis were E1AF-positive. Histolo-
gically, no correlation between WHO classification and E 1AF
expression could be found. However, the incidence of E1AF
expression was more frequent in invasive turnouts than in
expansive tumours (P< 0.04) (Fig. 4). We carried out in situ
hybridisation to identify the localisation of the E IAF mRNA-
expressing cells. E1AF mRNA was detected in cancer cells
invading into stromal connective tissue.
These results indicate that E 1AF is actually associated with
invasive status in in vivo oral SCC, presumably activating
MMP gene transcription as shown in our previous in vitro
experiments [6, 10, 11]. The present study confirmed our
previous in vitro reports [6, 11] that E1AF is highly correlated
with invasion and metastasis of cancer cells through tran-
scriptional activation of MMP genes, and it is suggested that
E1AF expression might be an important prognostic factor for
in vivo tumours, especially in the early stages of squamous
cell carcinomas in the oral cavity.
1. Liotta, L. A., Thorgeirsson, U. P. and Gatbisa, S., Cancer
metastasis and angiogenesis: an imbalance of positive and nega-
tive regulation. Cell, 1991, 64, 327-336.
K. Hida et al.
2. Liotta, L. A., Stetler-Stevenson, W. G. and Steeg, P. S., Cancer
invasion and metastasis: positive and negative regulatory ele-
ments. Cancer Investigation, 1991, 9, 543-551.
3. Grigioni, W. F., D'Errico, A., Fortunato, C., et al., Prognosis of
gastric carcinoma revealed by interactions between tumor cells
and basement membrane. Modern Pathology, 1994, 7, 220-225.
4. Shima, I., Sasaguri, Y., Kusukawa, J., et al., Production of matrix
metalloproteinase 9 (92-kDa gelatinase) by human oesophageal
squamous cell carcinoma in response to epidermal growth factor.
British Journal of Cancer, 1993, 67, 721-727.
5. Juarez, J., Clayman, G., Nakajima, M., et al., Role and regula-
tion of expression of 92-kDa type-IV collagenase (MMP-9) in 2
invasive squamous-cell-carcinoma cell lines of the oral cavity.
International Journal of Cancer, 1993, 55, 10-18.
6. Shindoh, M., Higashino, F., Kaya, M., et al., Correlated
expression of matrix metalloproteinases and ets family transcrip-
tion factor E1A-F in invasive oral squamous-cell-carcinoma-
derived cell lines. American Journal of Pathology, 1996, 148, 693-
7. Gaire, M., Magbanua, Z., McDonnell, S., McNeil, L., Lovett,
D. H. and Matrisian, L. M., Structure and expression of the
human gene for the matrix metalloproteinase matrilysin. Journal
of Biological Chemistry, 1994, 269, 2032-2040.
8. Wasylyk, B., Hahn, S. I. and Giovane, A., The ets family of
transcription factors. European Journal of Biochemistry, 1993, 211,
9. Higashino, F., Yoshida, K., Fujinaga, Y., Kamio, K. and Fuji-
naga, K., Isolation of a cDNA encoding the adenovirus E1A
enhancer binding protein: a new human member of the ets
oncogene family. Nucleic Acids Research, 1993, 21, 547-553.
10. Higashino, F., Yoshida, K., Noumi, T., Seiki, M. and Fujinaga, K.,
Ets-related protein E1A-F can activate three different matrix
metaUoproteinase gene promotor. Oncogene, 1995, 10, 1461-1463.
11. Kaya, M., Yoshida, K., Higashino, F., Mitaka, T., Ishii, S. and
Fujinaga, K., A single ets-related transcription factor, E1AF,
confers invasive phenotype on human cancer cells. Oncogene,
1996, 12, 221-227.
12. Hida, K., Shindoh, M., Yasuda, M., et al., Antisense E1AF
transfection restrains oral cancer invasion by reducing matrix
metalloproteinase activities. American Journal of Pathology, 1997,
13. Spissel, B., Beahrs, O., Hermanek, P., et al., Head and Neck
Tumors. Springer-Verlag, Berlin, 1989.
14. Totsuka, Y., Usui, Y., Tel, K., et al., Mandibular involvement by
squamous cell carcinoma of the lower alveolus: analysis and
comparative study of histologic and radiologic features. Head &
Neck, 1991, 13, 40-50.
15. Garbisa, S., Scagliotti, G., Masiero, L., et al., Correlation of
serum metalloproteinase levels with lung cancer metastasis and
response to therapy. Cancer Research, 1992, 52, 4548 4549.
16. Vogelstein, B. and Kinzler, K. W., The multistep nature of can-
cer. Trends in Genetics, 1993, 9, 138-141.
17. Matrisian, L. M., McDonnell, S., Miller, D. B., Navre, M.,
Seftor, E. A. and Hendrix, M. J., The role of the matrix metal-
loproteinase stromelysin in the progression of squamous cell
carcinomas. American Journal of Medical Science, 1991,302, 157-
18. Nakagawa, H. and Yagihashi, S., Expression of type IV collagen
and its degrading enzymes in squamous cell carcinoma of lung.
Japanese Journal of Cancer Research, 1994, 85, 934-938.
19. Alessandro, R., Minafra, S., Pucci, M. I., et al., Metallo-
proteinase and TIMP expression by the human breast carcinoma
cell line 8701-BC. International Journal of Cancer, 1993, 55, 250-
20. Bernhard, E. J., Gruber, S. B. and Muschel, R. J., Direct evi-
dence linking expression of matrix metalloproteinase 9 (92-kDa
gelatinase/collagenase) to the metastatic phenotype in trans-
formed rat embryo cells. Proceedings of the National Academy of
Sciences USA, 1994, 91, 4293-4297.
21. Wasylyk, C., Gutman, A., Nicholson, R. and Wasylyk, B., The
c-Ets oncoprotein activates the stromelysin promoter through the
same elements as several non-nuclear oncoproteins. EMBO
Journal, 1991, 10, 1127-1134.
22. Wernert, N., Gilles, F., Fafeur, V., et al., Stromal expression of
c-Ets 1 transcription factor correlates with tumor invasion. Cancer
Research, 1994, 54, 5683 5688.
23. Yoshida, K., Narita, M. and Fujinaga, K., Binding sites of HeLa
cell nuclear proteins on the upstream region of adenovirus type 5
E1A gene. Nucleic Acids Research, 1989, 17, 10015-10034.
24. Isobe, M., Yamagishi, F., Yoshida, K., Higashino, F. and Fuji-
naga, K., Assignment of the ets-related transcription factor E 1A-F
gene (ETV4) to human chromosome region 17q21. Genomics,
1995, 28, 357-359.
25. Kaneko, Y., Yoshida, K., Handa, M., et al., Fusion of an ets-
family gene, E1AF, to EWS by t(17-22) (q12-q12) chromosome
translocation in an undifferentiated sarcoma of infancy. Genes,
Chromosomes and Cancer, 1996, 15, 115-121.
26. Urano, F., Umezawa, A., Hong, W., Kikuchi, H. and Hata, J., A
novel chimera gene between EWS and E1A-F, encoding the
adenovirus E1A enhancer-binding protein, in extraosseous
Ewing's sarcoma. Biochemical and Biophysical Research Commu-
nications, 1996, 219, 608-612.
Acknowledgements--We thank Professor M. Mori for his helpful
discussion and kind advice on our manuscript. We are also grateful to
Mr M. K. Barrymore for his editorial assistance. This study was
supported in part by a Grant-in-Aid for Scientific Research from the
Ministry of Education, Science and Culture of Japan, and by a
Grant-in-Aid from the Kurozumi Foundation.