Histone gene (H3) expression in chemically transformed oral keratinocytes

EXPEUIMENTAL

AND

Histone

MOLECULAR

PATHOLOGY

206-214 (1988)

Gene (H3) Expression in Chemically Oral Keratinocytes DAVID

Department

49,

of Oral

T. W. WONG

Medicine and Oral Pathology, Longwood Avenue, Boston,

Received

October

13, 1987,

Transformed

Harvard School of Dental Massachusetts 02115

and in revised

form

March

II,

Medicine,

188

1988

The hamster cheek pouch is an excellent target tissue for the experimental study of oral carcinogenesis. In the course of searching for molecular alterations during the malignant transformation process, the necessity for a molecular marker for cellular proliferation became apparent. In this report, we show that the cellular level of the histone H3 mRNA is valid as a molecular index of proliferation for cycling cell populations. H3 is known to be proliferation dependent for its expression in cultured animal cells. This study shows that H3 retains its cell-cycle-dependent expression in chemically transformed oral keratinocytes. The onset of H3 mRNA synthesis couples to the onset of DNA synthesis (S-phase). The cellular level of H3 mRNA therefore is proportional to the fraction of cells in the S-phase of the cell cycle. This conveniently allows us to correlate, in asynchronized cell populations, the expression of cellular genes to their proliferation rates. We demonstrate the usefulness of this proliferation marker by presenting data that different chemically induced oral carcinomas, but not normal cheek pouch tissues, contain readily detectable levels of c-Ki-ras proto-oncogene mRNA. Probing the same RNA blot to quantitate H3 mRNA levels allowed us to conclude that the high levels of c-Ki-ras mRNA in tumor tissues was likely due to the increased growth rate of the tumor tissues and not due to the deregulated expression of this Cehkir

prOtO-OnCOgene.

0 1988 Academic

Press. Inc.

INTRODUCTION Deregulated growth is a hallmark of neoplasia. In experimental carcinogenesis the ability to monitor the proliferation of both normal and transformed cells is essential in the overall experimental design. [3H]Thymidine incorporation, flow cytometry, and other methods have been used successfully in different experimental systems. We have been studying molecular changes during the chemical induction of epidermoid carcinomas in the cheek pouch of Syrian hamsters (Wong and Biswas, 1987; Wong, 1987; Wong et al., 1988). Since many transformationassociated altered cellular gene expressions can simply be a consequence of the altered proliferation rate of transformed cells, a genetic marker to monitor cellular proliferation would be helpful in evaluating the importance of these “altered molecular events.” Recently Calabretta et al. (1986) have demonstrated that in asynchronous cellular systems, there is a correlation in the expression of specific cell-cycle genes. They have demonstrated that in asynchronous systems, G,- and S-phase genes would be correlated in their expression if precise temporality of their expression is required for an ordered cell-cycle progression. Thus comparing the mRNA ratio of a gene of interest (e.g., an oncogene whose expression seems to be altered) to a cell-cycle-specific gene (e.g., histone) in normal and transformed tissues should allow us to determine whether there is true deregulation. The need to search for such a cell-cycle-specific gene expressing in both normal and transformed tissues then becomes an important task in the molecular study of experimental carcinogenesis. 206 0014-4800188 $3.00 Copyright 0 1988 by Academic Press, Inc. AU rights of reproduction in any form reserved.

H3 EXPRESSION

IN

ORAL

CANCER

CELLS

207

In cultured animal cells, histone synthesis is coordinated with DNA replication and therefore takes place during S-phase (Stein et al., 1984). Recently, the hamster histone H3 gene has been molecularly cloned by Lee and her co-workers (Artishevsky et al., 1985). They have elegantly shown that the expression of the H3 gene is indeed cell-cycle controlled, tightly coupled to DNA synthesis. In this report we demonstrate that the hamster H3 gene retains its cellcycle-dependent expression in dimethylbenzanthracene (DMBA)-transformed hamster oral keratinocytes. Data are presented to show that we can use this genetic marker to demonstrate proliferative heterogeneity of different DMBAinduced oral carcinomas. Furthermore, we use the tumor H3 mRNA levels to illustrate that the apparent overexpression of the c-J&-us proto-oncogene in DMBA-transformed hamster oral tissues is not due to deregulated expression but rather is related to the increased growth rate of the tumor tissues. MATERIALS

AND METHODS

Cell Culture

HCPC-1 cells are DMBA-transformed Syrian hamster oral epidermoid carcinoma cells isolated and established by Odukoya et al. (1983). They are routinely maintained at 37°C 5% CO* in FlO medium supplemented with 10% fetal bovine serum and antibiotics (penicillin, 100 units/ml; streptomycin, 100 kg/ml; amphotericin B, 0.25 Kg/ml) (Whittaker MA Bioproducts, Walkersville, MD). Total and Cytoplasmic

RNA

Isolation

Total RNA from DMBA-induced hamster cheek pouch tumors was prepared by the guanidine isothiocyanate method according to Davis et al. (1986). The preparation of cytoplasmic RNA from cultured HCPC-1 was carried out according to a method communicated to us by Dr. M. C. Hung (Department of Tumor Biology, M. D. Anderson Hospital and Tumor Institute, Houston, TX). Briefly, about 5 x lo6 cells were pelleted from a lOO-mm-diameter culture dish and then lysed with 0.6% NP-40 in 250 p,l of 10 mM Tris * HCl (pH 7.6), 0.5 M NaCl, 1 n& EDTA, and 10 m&f vanadyl-ribonucleoside complexes (Bethesda Research Laboratory, Gaithersburg, MD). Total cytoplasmic RNA was extracted from the NP-40 supernatant by two extractions with phenol-chloroform (1: 1) and one extraction with chloroform. The aqueous phase was then adjusted to 0.3 M NaOAc (pH 6.0) and the RNA was precipitated with 2.5 vol of absolute ethanol at -20°C. RNA

Blotting

The cytoplasmic RNA isolated from the HCPC-1 cells and the total RNA from tumor tissues was fractionated by 1% agarose denaturing formaldehyde gel electrophoresis and blotted onto AMF Zetabind nylon membrane as previously described (Wong et al., 1984; Wong and Biswas, 1985). Molecular

Probes

The hamster H3 probe (pAAD3.7) was provided for us by Dr. A. S. Lee (University of Southern California, Los Angeles, CA) (Artishevsky et al., 1985). The mouse B-actin probe (pAI) was provided by Dr. D. W. Cleveland (National Institute of Health) (Cleveland et al., 1980). The c-Ki-rus probe was the human cKi-ras-2 cDNA (41027) purchased from the American Type Culture Collection

208

DAVID

T.

W.

WONG

(Rockville, MD). These probes were nick-translated to a specific activity of 1 x IO8 cpm/Fg DNA, and 0.25 x lo6 cpm/ml of hybridization solution was typically used. DNA

Synthesis:

Tritiated

Thymidine

Incorporation

The level of DNA synthesis during the experimental periods was quantitated by pulse-labeling (30 min) with 1 .O &i of [‘Hlthymidine (Amersham, 16.0 Ci/mmole) per ml of media. The amount of incorporated thymidine was measured by counting the radioactivity recovered in the trichloroacetic acid-precipitated material (Melero and Fincham, 1978). DMBA-Znduced

Hamster

Cheek Pouch Tumors

Epidermoid carcinomas were induced in the cheek pouch of male Syrian hamsters according to the protocol of Shklar (1972). Briefly, 0.5% DMBA in mineral oil was applied thrice weekly to the left cheek pouch of each experimental animal for 14 weeks. Exophytic tumors greater than 10 mm in diameter were excised and a sample was saved for histopathological examination. The remaining tissue was immediately processed for RNA isolation by the guanidine isothiocyanate method (Davis et al., 1986). All tissues used for this study were histologically malignant (squamous cell carcinomas). RESULTS Cell-Cycle S-Phase-Specific Chemically Transformed

Expression of the Histone Oral Epithelial Cells

H3 Gene in the

We determined whether the expression of the H3 gene is cell-cycle regulated in the HCPC-1 cells by simultaneously measuring the cellular levels of H3 mRNA and the rate of [3H]thymidine incorporation by pulse-labeling synchronized cultures (serum deprivation for 24 hr). Figure 1A shows that the level of H3 mRNA was at a minimum immediately after serum deprivation (0 hr). Its level steadily rose to a maximum at 5 hr after replenishing with 10% fetal bovine serum. Thereafter it decreased gradually. The same blot is rehybridized with a mouse B-actin probe to demonstrate the integrity of the isolated RNA. The ethidium bromidestained gel before transfer is also shown to show that a similar amount of cytoplasmic RNA has been loaded onto each lane. Figure 1B illustrates the results of [3H]thymidine pulse-labeling carried out in HCPC-I cultures in parallel with the plates destined for cytoplasmic RNA isolation. The H3 signals on the autoradiogram were scanned with a densitometer and superimposed onto the data of [3H]thymidine incorporation in Fig. 1B. Figure 2 shows a similar experiment except that the HCPC-1 cultures were serum deprived for a total of 36 hr (12 hr more than the previous experiment). This produced better synchrony of these cultures. Here the cells took 12.5 hr (compare to 5 hr in the first experiment) to reach a level for maximal H3 expression. Thereafter another peak appears at 22.5 hr. The difference between the two peaks is 10 hr, which is in good agreement with the finding of Odukoya et al. (1983) that the doubling time of these cells is about 12 hr. Quantitating H3 mRNA Pouch Carcinomas

Figure 3 is a Northern

Levels in DMBA-Znduced

Hamster

Cheek

blot analysis of total RNA isolated from eight DMBA-

H3 EXPRESSION Hous YolYS2

A

ao(VV)lhrr

0 kb

28s

-5.1

188-

-2.0

IN

ORAL

CANCER

209

CELLS -

23 2221-

- 1.5 - 1.4 - 1.3 -1.2 - 1.1

28s 18s. “3.

I)*

1.8 - 1.7 - 1.6

v-09

-0.3 -0.8

’

-0.7 - 0.6 -0.6 - 0.4 A0

g 0 ; E m * L s if

- 0.3

.2.l A

0.2 0.0

2.5

Q 3-Hmymldlm

6.0 7.6 Time bmr)

10.0 +

12.3

H3 mRNA

1. Cell cycle S-phase specific-expression of the H3 gene in HCPC-1 cells. (A) Semicontluence cultures of HCPC-1 cells were grown in FlO medium supplemented with 10% FBS. These cultures were then serum starved for 24 hr (FlO alone without any FBS). Cells appeared very healthy at the end of the starvation period with an obvious increase in cell number. The cultures were then fed with fresh medium containing 10% FBS. (A) Asynchronized culture before serum starvation, O-, 2.5, 5.0-, 7.5, lO.O-, and 12.5hr HCPC-1 cultures after serum starvation. Top: Ethidium bromide-stained gel of electrophoresed cytoplasmic RNA isolated before transfer. Twenty micrograms of cytoplasmic RNA was used in each lane. 28 S (5.1 kb) and 18 S (2.0 kb) are positions and approximate sizes of ribosomal RNAs (Davis er al., 1986). Middle: Autoradiogram from hybridization with 32P-hamster H3 insert DNA (Artishevsky et al., 1985). Size of H3 signal is approximately 0.9 kb. Bottom: Rehybridization of the H3 blot with 32P-mouse p-actin insert DNA (Cleveland et al., 1980). Size of @actin signal is approximately 2.1 kb. (B) DNA synthesis in HCPC-1 cultures assayed by [3H]thymidine incorporation carried out in parallel with those described in (A). Results from quadriplicate cultures are determined and error bars indicating one standard deviation are also included. The relative H3 mRNA concentration, obtained by determining the relative optical density of the H3 signals in (A), is superimposed onto this chart at the same time points (Phototronix densitometer, Model 73). FIG.

induced hamster oral carcinomas, one mineral oil-treated control cheek pouch, and the HCPC-1 cell line. Histological examination of representative sections of these tumors revealed that tumors 4 and 7 were well differentiated, tumor 6 was most anaplastic, and the others exhibit intermediate degree of differentiation. All of the eight tumors contained readily detectable H3 mRNA levels. The relative abundance of the H3 mRNA, however, varies from tumor to tumor. Tumors 4 and 7 had the lowest H3 mRNA concentration, tumor 6 had the highest, and the others had intermediate levels. Note that the normal control cheek pouch had a low but detectable signal. Notice also the intense signal generated by the RNA from the HCPC-1 cells, suggesting the highest fraction of the cells going through S-phase of the cell cycle and thus implying the highest rate of proliferation. Levels of c-Ki-ras Proto-oncogene mRNA in DMBA-Induced Pouch Carcinomas

Hamster Cheek

The RNA blot shown in Fig. 3 was rehybridized with the human Ki-ras cDNA probe (Fig. 4). Ki-ras hybridizing mRNA sequences can be readily detectable in all of the DMBA-induced hamster cheek pouch tumors and the HCPC-1 cell line. The intensity of the c-Ki-ras signals varies from tumor to tumor. Tumor 7 contained the least c-Ki-ras mRNA, tumor 6 contained the most, and all other tumors contained intermediate levels. The normal hamster cheek pouch tissues (Fig. 4,

210

DAVID

T. W. WONG

285, 185,

FIG. 2. Prolonged starvation produced better synchronization of HCPC-1 cells. HCPC-1 cultures were serum starved (FlO alone without FBS) for 36 hr instead of 24 hr as in Fig. 1. At Time 0, fresh medium with 10% FBS was added to these cultures to synchronize progression through the cell cycle. Top: Ethidium bromide-stain of electrophoresed cytoplasmic RNA before transfer. Twenty micrograms of cytoplasmic RNA was used in each lane. Middle: Autoradiogram hybridized with 32P-hamster H3 insert DNA. H3 ., S-hr exposure at room temperature; H3*,, overnight exposure at -70°C to Kodak XAR-5 films. Arrows within the autoradiogram indicate time points with maximum concentration of H3 mRNA at 12.5 and 22.5 hr. Bottom: Rehybridization of H3 blot to 32P-mouse B-actin insert DNA. Notice the cycle-dependent nature of its expression.

lane 9) contained no detectable Ki-ras mRNA (the apparent faint band was due to the over-exposed signal generate from the HCPC-1 cells in lane IO). The size of the c-Ki-ras transcript detected in these hamster oral tumor tissues is about 6 kb. DISCUSSION Deregulated growth is a hallmark of tumor cells. The ability to monitor the growth of tumor cells is essential in experimental cancer research. This is particularly important when determining whether the altered expression of a cellular gene (e.g., oncogene) during malignant transformation is true deregulation or simply related to increased proliferation of the transformed tissue. 12

3

4

56

78

910

kb 28S-

-5.1

18%

-2.0

H3,

-0.9

FIG. 3. Elevated levels of H3 mRNA in DMBA-induced hamster oral carcinomas. Lanes l-8, different DMBA-induced squamous cell carcinomas; lane 9, normal cheek pouches; lane 10, HCPC-I cells. Twenty micrograms of total RNA were used.

H3 EXPRESSION

IN

ORAL

CANCER

CELLS

211

12345678910

28s.

I c-ki-ras

i as.

FIG. 4. Levels of c-Ki-ras mRNA in DMBA-induced hamster oral carcinomas. The RNA blot in Fig. 3 was rehybridized with a 32P-human c-Ki-ras cDNA. The size of the hamster c-Ki-ras transcript is about 6 kb.

The working system in this study is an epidermoid carcinoma cell line, HCPC-1, derived from a DMBA-induced squamous cell carcinoma from the cheek pouch of a Syrian hamster (Odukoya et al., 1983). Odukoya et al. (1983) have demonstrated that this cell line shares many common features with similarly derived in vivo tumors. Recently we reported that the c-erb BZ oncogene is overexpressed and amplified in both HCPC-1 cells and oral carcinomas (Wong and B&was, 1987; Wong, 1987). In this report we present experimental data that the hamster histone H3 gene retains cell-cycle-dependent expression in chemically transformed oral epithelial cells. The results in Fig. 1 demonstrate that in synchronized HCPC-1 cultures the extent of i3H]thymidine incorporation closely approximates the level of H3 mRNA, suggesting a tight linkage of DNA synthesis and onset of H3 mRNA synthesis. Thus the cellular level of H3 mRNA can indirectly determine the proliferative rates of these tissues. More important is the fact that we can now have a convenient genetic marker to correlate the activity/expression of cellular genes to the proliferative rate of normal and tumor tissues. Figure 2 shows that the generation time of the HCPC-1 cells is about 10 hr. It further shows that the expression of the p-actin gene is also cell-cycle regulated in the HCPC-1 cells. This observation has been made by McCairns et al. (1984) in lectin-activated lymphocytes. These investigators noted that in lectin-stimulated lymphocytes, the maximal level of p-actin was detected during G,-phase of the cell-cycle. This seems to be our observation as well, since the peak levels of p-actin mRNA occur about 5 hr before the peak levels for the H3 mRNA. We demonstrated that this genetic marker can be used in experimental molecular carcinogenesis to quantitate cellular proliferation by determining the levels of H3 mRNA in different DMBA-induced oral carcinomas (Fig. 3). Nagamine (1978) demonstrated that in DMBA-induced hyperplastic and neoplastic cheek pouch epithelium, the cell-cycle time was decreased to approximately two-thirds and one-third that of normal control, respectively. Furthermore, Thilagaratnam and Main (1972) reported that DMBA-induced cheek pouch carcinomas proliferate about 10 times faster than those of control cheek pouch epithelium. The results presented in Fig. 3 show that we can quantitate the proliferative rates of normal and transformed oral tissues by quantitating the levels of H3 mRNA. All tumors

212

DAVID

T. W.

WONG

examined contained concentrations of H3 mRNA higher than those of control cheek pouch (Fig. 3, lane 9). However each tumor contains a different amount of H3 mRNA. This suggests different rates of proliferation of these tumors and implies proliferative heterogeneity in these tumors. However it is possible that in these in vivo tumors, the differences in H3 mRNA levels were due to dilution of tumor cells by connective tissues and nontransformed epithelial cells. None of the in vivo tumors contains H3 mRNA as high as that of the HCPC-I cells. This is likely due to the fact that HCPC-1 is a clonal malignant cell line, while each of the in vivo tumors might contain a mixture of other tissues (e.g., normal epithelial cells and connective tissue cells), thus diluting the effective concentration of H3 mRNA in the tumor cells. These results of determining tumor proliferation heterogeneity by quantitating tumor H3 mRNA levels agree quite well with the histological grading. The welldifferentiated tumors (Fig. 3, tumors 4 and 7) contained the least amount of H3 mRNA, the most anaplastic tumor (Fig. 3, tumor 6) contained the highest abundance of H3 mRNA, and all the other tumors that exhibited an intermediate degree of differentiation contained intermediate levels of H3 mRNA. These data further assure us that the tumor level of H3 mRNA can be correlated to histopathological grading. We illustrated a major application of this molecular marker of cellular/tumor proliferation by demonstrating the cellular/tumor mRNA level of the cellular proto-oncogene c-Ki-ras. Figure 4 showed that we can readily detect c-Ki-rus mRNA in all DMBA-induced hamster cheek pouch tumors and the HCPC-1 cell line. No c-Ki-ras mRNA can be detected in the normal cheek pouch tissues. From these results alone, we might conclude that the c-Ki-rus proto-oncogene is overexpressed in DMBA-transformed hamster oral tissues and may even further suggest that c-Ki-rus might play a role in the malignant transformation of the oral mucosal tissues. However comparing individual tumor c-Ki-rus mRNA levels to the H3 mRNA level (Fig. 3) will demonstrate to us that a positive correlation exists between the rate of proliferation of these oral tumors and the levels of c-Ki-rus mRNA. The higher the rate of proliferation of these oral tumors, the higher the level of c-Ki-rus mRNA. Tumor 6 was histologically most anaplastic. Compared with the other tumors, it had the highest level of H3 mRNA and the highest level of c-Ki-rus mRNA. Tumor 7 was histologically a well-differentiated carcinoma. It had one of the lowest levels of H3 mRNA and correspondingly one of the lowest levels of c-Ki-rus mRNA. HCPC-1 cells, being composed of homogenous tumor cells, contained the highest levels of H3 and c-Ki-rus mRNAs. A careful densitometric scanning of these radiographic signals will likely reveal to us that a similar ratio of c-Ki-rus to H3 mRNA exists in these tumor tissues examined. Thus instead of concluding overexpression or deregulated expression of the c-Ki-rus proto-oncogene in these chemically transformed hamster oral tumors, the readily detectable levels of c-Ki-rus mRNA in these tumor tissues was more likely related to the higher rate of proliferation relative to the normal tissues (Fig. 3, lane 9). Indeed, the expression of the c-Ki-ras proto-oncogene has been reported to be proliferation dependent in normal and chemically transformed tissues (Campisi et al., 1984; Goyette et al., 1984). In the study of molecular carcinogenesis, it is very important to have a molecular marker for proliferation. This is because upon malignant transformation, the

H3 EXPRESSION

IN ORAL

CANCER

CELLS

213

expression of various cellular genes could have been altered. Altered cellular gene expression could be due to truly deregulated expression or simply be related to the altered growth rate of the transformed tissues. In this report, we have demonstrated that the cellular level of the H3 mRNA can be used as a molecular marker of proliferation in chemically transformed hamster oral tissues. Since RNA blotting is such a common technique for studying altered gene expression in molecular carcinogenesis, quantitating the cellular/tumor levels of H3 mRNA (on the same or duplicate blot) will allow us to determine if the altered expression of a cellular gene is indeed deregulated or simply related to the altered growth of the tumor tissues. ACKNOWLEDGMENTS The plasmid pAAD3.7 was provided by Dr. A. S. Lee (University of Southern California). Plasmid pA1 was provided by Dr. D. W. Cleveland (National Institute of Health). This work was supported by a grant from the Smokeless Tobacco Research Council, Inc., No. 0128. Also aided by a Cancer Research Scholar Award from the American Cancer Society, Massachusetts Division, Inc.

REFERENCES ARTISHEVSKY, A., GRAFSKY, A., and LEE, A. S. (1985). Isolation of a mammalian sequence capable of conferring cell cycle regulation to a heterologous gene. Science 230, 1061-1063. CALABRETTA, B., VENTURELLI, D., KACZMAREK, L., NARNI, F., TALPAZ, M., ANDERSON, B., BERAN, M., and BASERGA, R. (1986). Altered expression of Gl-specific genes in human malignant myeloid cells. Proc. Natl. Acad. Sci. USA 83, 1495-1498. CAMPISI, J., GRAY, H. E., PARDEE, A. B., DEAN, M., and SONENSHEIN, G. E. (1984). Cell-cycle control of c-myc but not c-ras expression is lost following chemical transformation. Cell 36, 241-247. CLEVELAND, D. W., LOPATA, M. A., MACDONALD, R. J., COWAN, N. J., RUTTER, W. J., and KIRSCHNER,M. W. (1980). Number and evolutionary conservation of a-and B-tubulin and cytoplasmic B-and o-actin genes using specific cloned cDNA probes. Cell 20, 95-105. DAVIS, L. G., DIBNER, M. D., and BATTEY, J. F. (1986). “Basic Methods in Molecular Biology.” p. 150. Elsevier, New York. GOYETTE, M., PETROPOULOS, C. J., SHANK, P. R., and FAUSTO, N. (1984). Regulated transcription of c-Ki-ras and c-myc during compensatory growth of rat liver. Mol. Cell. Biol. 4, 1493-1498. MCCARINS, E., FAHEY, D., MUSCAT, G. E. O., MURRAY, M., and ROWE, P. B. (1984). Changes in levels of actin and tubulin mRNAs upon the lectin activation of lymphocytes. Mol. Cell. Biol. 4, 1754-1760. MELERO, J. A., and FINCHAM, V. (1978). Enhancement of the synthesis of specific cellular polypeptides in a temperature-sensitive Chinese hamster cell line (K12) defective for entry into S phase. J. Cell.

Physiol.

95, 295-306.

NAGAMINE, Y. (9178). Changes of the cell cycle during carcinogenesis of hamster cheek pouch epithehum by 7,12-dimethylbenzanthracene. Gann 69, 317-322. ODUKOYA, O., SCHWARTZ, J., WEICHSELBAUM, R., and SHKLAR, G. (1983). An epidermoid carcinoma cell line derived from hamster 7,12-dimethlybenz[a]-anthracene-induced buccal pouch tumors. J. Natl. Cancer Inst. 71, 1253-1264. STEIN, G. S., STEIN, J. L., and MARZLUFF. W. F. (1984). “Histone Genes. Structure, Organization and Regulation.” Wiley, New York. SHKLAR, G. (1972). Experimental oral pathology in the Syrian hamster. Progr. Exp. Tumor Res. 16, 518-538. THILAGARATNAM,

C. N., and MAIN, J. H. P. (1972). Changes in cell cycle characteristics in hamster cheek pouch epithelium during treatment with DMBA. J. Oral Pathol. 1, 89-102 WONG, D. T. W. (1987). Amplification of the c-erb BI oncogene in DMBA-induced oral carcinomas. Carcirzogenesis

8, 1963-1967.

WONG, D. T. W., and BISWAS, D. K. (1985). Mechanism of benzo(a)pyrene induction of human chorionic gonadotropin gene expression in human lung tumor cells. J. Cell Biol. 101, 2245-2252.

214

DAVID

T. W. WONG

WONG, D. T. W., and BISWAS, D. K. (1987). Activation of c-e& B oncogene during DMBA-induced hamster cheek pouch carcinogenesis. Oncogene 2, 67-72. WONG, D. T. W., and GALLAGHER, G. T. (1987). Effect of sialadenectomy on DMBA-induced hamster cheek pouch carcinogenesis. J. Dent. Res. 66, 333. [Abstract No. 18131 WONG, D. T. W., GALLAGHER, G. T., GERTZ, R., CHANG, L. C., and SHKLAR, G. (1988). Transforming growth factor (Yin chemically transformed hamster oral keratinocytes. Cancer Res., 48, 313s 3134. WONG, D. T. W., HARTIGAN, J. A., and BISWAS, D. K. (1984). Mechanism of induction of human chorionic gonadotropin in lung tumor cells in culture. J. Bid/. Chem. 259, 10,738-10,744.