2,3,7,8-Tetrachlorodibenzo-p-dioxin-dependent Regulation of Transforming Growth Factors-a and -B2 Expression in a Human Keratinocyte Cell Line Involves Both Transcriptional and Post-transcriptional Control*

modulator of cell growth and differentiation, has been shown to induce transforming growth factor (TGF)-a in cultures of human keratinocytes and in the human keratinocyte cell line, SCC-12F. In this report, we investigated the mechanisms by which TCDD alters TGF-a expression. In addition, we studied the actions of TCDD on TGF-& and TGF& expression. unchanged. Changes in TGF-a and TGF-B2 expression were h. No change in the rate of transcription of TGF-a was detected following treatment with TCDD as determined by nuclear run-off analysis. resulted in a stabilization of TGF-a mRNA as judged by an approximately 2-fold higher level of TGF-a mRNA in uersus control cells in the presence of actinomycin In contrast to TGF-a, the rate of transcription of TGF-B2 was significantly reduced following TCDD treatment. These findings dem- onstrate that the induction of TGF-a expression in SCC-12F cells by TCDD occurs post-transcriptionally, primarily by mRNA stabilization, while They suggested that of TGF-a important We report that the increase in TGF-a mRNA in cells TCDD a stabilization of the TGF-a mRNA. In addition, we show that TCDD decreases the level of TGF-P2 mRNA by a mechanism involv-ing transcriptional control of TGF-P2 gene expression.

gramming of cell growth and differentiation is essential for the pathogenesis of chloracne and can be modeled in uitro using cultures of human epidermal keratinocytes and cell lines derived from squamous cell carcinomas (SCC) (4)(5)(6)(7). Treatment of cultured human epidermal cells with TCDD results in an increase in the number of cells undergoing terminal differentiation (4,6,8). The SCC-12F cell line used in these studies, a nontumorigenic keratinocyte cell line derived from an SCC of the face (9), also responds to TCDD with an increase in the number of cells undergoing terminal differentiation (6).
Many of the actions of TCDD have been shown to be mediated through an intracellular binding protein, designated the Ah receptor (10). In most cells the Ah receptor mediates the TCDD-dependent expression of a battery of genes that includes cytochrome P1-450 (CYPlAl) (10,11). Transcriptional activation of CYPlAl is a primary response to TCDD resulting from the binding of the TCDD-Ah receptor complex to specific enhancer regions upstream of the CYPlAl transcription start site (11,12). In certain cells responsive to TCDD, including human keratinocytes, the TCDD-Ah receptor complex is postulated to alter the expression of genes involved in growth control (6,13).
Transforming growth factors, originally identified by their ability to stimulate growth of rodent fibroblasts in soft agar (14, E ) , have since been found to affect the growth and differentiation of a number of different cell types (16)(17)(18), including keratinocytes (19,20). Transforming growth factora (TGF-a) shares structural homology with epidermal growth factor (EGF), utilizes the same receptor as EGF, and possesses a nearly identical spectrum of activity (16,21). The TGF-ps are a family of related proteins which are structurally and biologically distinct from TGF-a (17,22,23). Both TGF-a and the TGF-/3s are thought to act as autocrine regulators of keratinocyte proliferation and differentiation (20, [24][25][26]. TGF-a is mitogenic for keratinocytes, while TGF& and TGF-P2 are reported to inhibit keratinocyte proliferation (17)(18)(19).
Recently Choi et al. (27) reported that treatment of cultured human keratinocytes or SCC-12F cells with TCDD resulted in an increase in the levels of TGF-a mRNA and secreted TGF-a peptide. They suggested that overexpression of TGFa is an important mechanism by which TCDD produces its effects. We report that the increase in TGF-a mRNA in SCC-12F cells following TCDD treatment results from a stabilization of the TGF-a mRNA. In addition, we show that TCDD decreases the level of TGF-P2 mRNA by a mechanism involving transcriptional control of TGF-P2 gene expression.

EXPERIMENTAL PROCEDURES
Cells and Culture Conditions-SCC-12F cells (provided by Dr. J. Rheinwald, Biosurface Technology, Inc., Cambridge, MA) were plated in plastic culture dishes (Falcon Plastics, Oxnard, CA) at 2.5 X 10' cells/cm2 in Dulbecco's modified Eagle's medium (Mediatech, Herndon, VA) supplemented with 5% fetal calf serum (Hyclone, Logan, UT), plus a final concentration of 40 mM L-glutamine and 1 mM sodium pyruvate and incubated at 37 "C in a humidified atmosphere of 5% COn/air. We have found that addition of fresh media plus serum altered the expression of many of the genes which we were investigating. To avoid this effect we do not add fresh media at the time of TCDD treatment. After incubation for 48 h, culture media was removed and pooled. To the pooled media, TCDD (KOR Isotopes Inc., Cambridge, MA) in MezSO (Sigma) or 0.1% Me2SO as control was added and the media was then aliquoted back into the cultures.
Determination of Cell Number and Colony Forming Efficiency-Cell number was determined in triplicate for control and TCDDtreated cultures in 12-well plates using a Coulter counter (Coulter Electronics, Inc., Hialeah, FL). Colony forming efficiency was determined by plating cells in 5 ml of Dulbecco's modified Eagle's medium + 5% fetal calf serum in triplicate in 60-mm dishes (5000 cells/dish).
After incubation for 1 week, cells were washed once in phosphatebuffered saline, fixed in methanol, and stained with 10% Giemsa (J. T. Baker, Inc., Phillipsburg, NJ). Colony numbers were determined and results expressed as percent of colonies formed per total number of cells plated. Colonies were counted if they contained 10 or more cells.
Cross-linked Envelope Assay-Cells were collected by trypsinization, resuspended in 5 ml of Leibovitz L-15 medium (GIBCO-BRL); split into 2 tubes with lo6 cells/tube, spun down, and resuspended in 0.5 ml of L-15 medium. Spontaneous envelopes were determined by adding 0.5 ml of lysis buffer (50 mM dithiothreitol, 0.5% SDS) to the first tube and counting the number of envelopes/ml with a hemacytometer. Competent envelope formation was determined by adding 2.5 pl of Can+ ionophore A23187 (Sigma, 1 mg/ml in MezSO) and incubating with mixing for 3-5 h at 37 "C. After incubation the procedure followed was the same as with spontaneous envelope formation.
Northern Analysis-Cells were collected by scraping and mRNA was isolated using the procedure supplied with the FastTrack mRNA isolation kit (Invitrogen, San Diego, CA). mRNA was precipitated in 2.5 volumes of ethanol, resuspended in water, and quantified. For Northern analysis, mRNA (5 pg/sample) was denatured with glyoxal, size separated by agarose gel electrophoresis, and transferred to a nylon membrane (Bio-Rad). Hybridization of RNA with the indicated DNA probes was carried out overnight at 42 "C in a solution of 50% formamide, 5 X SSPE (1 X SSPE, 0.15 M sodium chloride, 0.01 M sodium phosphate, 1 mM EDTA), 5 X Denhardt's, 0.5% SDS, and 100 pg/ml salmon sperm DNA. The presence of specific RNA was detected by autoradiography and quantified by densitometry. After quantification, filters were stripped and reprobed with each of the indicated DNA probes. A glyceraldehyde phosphate dehydrogenase DNA probe was used as a loading and transfer control to ensure equal quantities of RNA in control and TCDD-treated samples. To confirm results obtained with glyceraldehyde phosphate dehydrogenase, an actin DNA probe was used as a second loading and transfer control. Results obtained with actin were similar to those obtained with glyceraldehyde phosphate dehydrogenase. Experiments were run 2-5 times with different RNA preparations.
TGF-a mRNA Stabilization Assay-Cells were plated and treated as described above. After incubation for 12 h, media was removed from control or TCDD-treated cultures and pooled. Actinomycin D (5 pg/ml, Sigma) was added to the pooled media and the media was then added back to the cultures. This concentration of actinomycin D inhibited [3H]uridine incorporation by 90-95% (data not shown).
At the indicated times after addition of actinomycin D, RNA was collected and processed for Northern analysis as described above.
Nuclear Run-off Assay-Nuclear run-off transcription assay was performed as described by Nevins (32) with slight modifications. Briefly, nuclei collected by Nonidet P-40 lysis were added directly to 2 X reaction buffer (20 mM Tris (pH 8), 20% glycerol, 140 mM KCl, 10 mM MgC12, 1 mM MnCIZ, 14 mM 2-mercaptoethanol, 1 mM ATP, CTP, and GTP, 60 mM phosphocreatine, 200 units of creatine kinase, 0.8 mM phenylmethylsulfonyl fluoride, 8 mM dithiothreitol, 0.8 units/ pl RNasin (Promega, Madison, WI), and 250 pCi of [32P]UTP (800 Ci/mmol, Amersham Corp.)) and incubated for 30 min at 25 "C. The nascent transcripts were processed for hybridization by the hot phenol method as described by Nevins (32). An equal amount of radiolabeled RNA was added to DNA immobilized on nylon filters and hybridized for 3 days at 42 "C in 50% formamide hybridization buffer prepared as described above. Filters were prepared by adding 5 pg of alkalidenatured plasmid per slot and UV cross-linked with a UV Stratalinker (Stratagene, La Jolla, CA). After incubation, the filters were washed as described above, exposed to Kodak XAR-2 film, and the activity quantified by densitometry. Glyceraldehyde phosphate dehydrogenase was used as a control in all experiments. In some experiments glyceraldehyde phosphate dehydrogenase was compared with actin and similar results were obtained.
Statistical Analysis-Unless otherwise noted, values presented in this study represented the mean * S. E. from at least 3 separate experiments. Statistical comparisons between treatment and control values were analyzed using a paired t test. Significance for all experiments was determined at the 5% level.

RESULTS
To characterize the response of SCC-12F cells to TCDD under the culture conditions used for this study, we examined several parameters of growth and differentiation. Treatment of SCC-12F cultures with TCDD resulted in an increase in the state of differentiation as judged by: 1) a reduction, relative to control, in total cell number (Fig. lA); 2) an increase in the number of cells competent to form crosslinked envelopes (Fig. 1B); and 3) an increase in involucrin mRNA levels (data not shown). Colony-forming efficiency was not significantly altered by TCDD (Fig. IC). These results are in agreement with the reported actions of TCDD on normal and transformed human keratinocytes (4, 6, 33) and on human bronchial epithelial cells (34) and confirm that under the conditions used in this study, SCC-12F cultures respond to TCDD with a reduction in cell growth and an increase in differentiation.
The level of TGF-a mRNA is reported to be increased 96 h after treatment of confluent serum-free cultures of human keratinocytes and SCC-12F cells with TCDD (27). We have characterized the time course for changes in mRNA expression of TGF-a and two other transforming growth factors, TGF-PI and TGF-P, (Fig. 2). Expression of CYPlAl, a gene under direct transcriptional control by TCDD (35) Treatment of cells with the protein synthesis inhibitor, cycloheximide (10 pg/ml), for 12 h either in the presence or absence of TCDD resulted in substantial increases in the mRNAs for TGF-a and TGF-P,, relative to both control and TCDD-treated cultures (Fig. 2C). No changes in either TGFa (0.99 & a range of 0.28; based on the results of two separate experiments) or TGF-6, mRNA levels (0.88 5 a range of 0.12; based on the results of two separate experiments) were observed when cells treated with TCDD plus cycloheximide were compared with cultures treated with cycloheximide alone. Unlike CYPlAl mRNA, which is superinduced in the presence of cycloheximide (12,44), TCDD-specific alterations in TGF-a and TGF-P, mRNA levels are not observed in the presence of cycloheximide.
TCDD-induced changes in TGF-a and TGF-P, mRNA levels were concentration dependent (Fig. 3). The concentration dependence for these growth factors was similar to that for the induction of CYPlAl and was comparable to the induction of the TGF-a peptide in cultures of human epidermal cells and the SCC-12F cell line (27). The active TCDD congener, 2,3,7,8-tetrachlorodibenzofuran, altered TGF-a and TGF-0, to a similar degree as TCDD, while the inactive congener, 2,7-dichlorodibenzo-p-dioxin, did not alter TGF-a nor TGF-P, gene expression (data not shown). The observed concentration dependence (Fig. 3) and stereospecificity support the involvement of the Ah receptor, previously shown to be present in these cells (36).
Nuclear run-off analysis a t 90 min, 12 h and 24 h following TCDD treatment was performed to determine whether changes in TGF-a and TGF-P, expression occurred at the level of mRNA transcription. In addition, cultures were treated with TCDD plus cycloheximide (10 pg/ml) or cycloheximide alone to evaluate the role of protein synthesis in the actions of TCDD on gene transcription. Results were com-   Table I). By 24 h, the relative rate of CYPlAl transcription was approximately 3.5-fold above control levels. In agreement with earlier reports (12,44), the rate of CYPlAl transcription was superinduced in the presence of TCDD plus cycloheximide. The rate of transcription of TGF-(Y was not significantly different in treated versus control cultures a t any of the time points examined, either in the presence or absence of cyclo- Nuclei were collected at 24 h and nuclear run-off analysis was performed as described under "Experimental Procedures." Actin and glyceraldehyde phosphate dehydrogenase (GAPDH) were used as loading controls. CYPlAl was used as a positive control for the induction of transcription by TCDD. Plasmid DNA was used as a control for nonspecific binding.

TABLE I
Effect of TCDD on rate ofgene transcription Cells were plated as described in the legend to Fig. 1 and treated with 0.1% Me2S0 or 10 nM TCDD in the presence or absence of 10 pg/ml cycloheximide (CH). Nuclei were isolated a t 90 min, 12 h, and 24 h after treatment and processed for Nuclear Runoff analysis as described under "Experimental Procedures." Densitometric tracings were normalized to GAPDH. Normalization to actin gave similar results. Results are expressed as the relative level of transcription in TCDD-treated cultures uersus control cultures and are the mean f S.E. from 3 seDarate exDeriments. heximide. The rate of transcription of TGF-P2 was not changed a t 90 min, but was significantly reduced in the TCDD-treated cultures a t 12 and 24 h. This 70% decrease in TGF-P2 transcription rate corresponds with the 70% decrease in TGF-P2 mRNA levels (Fig. 2), indicating that the decrease in TGF-P, mRNA levels following treatment with TCDD could be directly accounted for by the decrease in TGF-P, gene transcription. The rate of TGF-P2 transcription was not altered in cells treated with TCDD plus cycloheximide compared to cultures treated with cycloheximide alone, indicating that protein synthesis is probably required for this effect on Since no changes were detected in rate of transcription, increases in TGF-a mRNA following treatment with TCDD must occur post-transcriptionally. Studies performed to analyze the stability of TGF-a mRNA in the presence of actinomycin D showed a two-component decay curve in control and treated cultures relative to glyceraldehyde phosphate dehydrogenase mRNA (Fig. 5). Approximately 50% of the TGF-a mRNA in the absence of TCDD decayed in 3 h revealing a subpopulation of TGF-a mRNA relatively resistant to degradation. Treatment with TCDD caused an approximately 2fold increase in the subpopulation of TGF-a mRNA resistant to degradation. The two-component decay curve was reproducible as judged by the results of two separate experiments within the same time period shown in Fig. 5. A two-component decay curve has been described previously for human growth hormone and CYPlAl mRNA (45,46). We must interpret the actinomycin D data with caution, since actinomycin D may modify mRNA half-lives indirectly, presumably by their effects on labile proteins or RNAs (47). Control experiments, however, showed no effects of TCDD plus actinomycin D on the decay of either actin (Fig. 5) or TGF-P2 (data not shown) mRNAs relative to SCC-12F cells treated with actinomycin D alone. These results suggest that the 2-fold increase in the resistant subpopulation of TGF-a mRNA that occurs following treatment with TCDD may be due in part to a stabilization of this TGF-a mRNA. The mechanisms involved in this stabilization of the TGF-a mRNA are not clear at this time.

DISCUSSION
TCDD is postulated to elicit its effects on keratinocyte growth and differentiation by altering expression of critical growth regulatory genes (1,2). Recently, it was reported that treatment of cultured human keratinocytes with TCDD resulted in an increase in TGF-a expression (27). In this study  Fig. 1 with 0.1% Me2S0 as a control or TCDD. After incubation for 12 h, media from control or treated cultures was removed, pooled, and 5 pg/ml actinomycin D was added. Media was then added back to culture and RNA was collected at the indicated time points for Northern analysis as described under "Experimental Procedures." 0, control TGF-a; 0, TCDD TGF-a; 0, control actin; ., TCDD actin. Quantification is based on densitometric analysis of a single experiment with glyceraldehyde phosphate dehydrogenase used as a loading control. A repeat of this experiment gave similar results with TGF-a mRNA levels in TCDD-treated cultures being approximately 2-fold above control 10 h after treatment.
we confirm the induction of TGF-a mRNA by TCDD in cultured human keratinocytes and show that the increase in TGF-a mRNA is accompanied by a reduction in the expression of TGF-P2. Accompanying this increase in TGF-a mRNA is an increase in the levels of TGF-a protein measurable in the medium from TCDD-treated cultures (27).2 TGF-a is similar to EGF in its ability to bind, activate, and downregulate the EGF receptor (16,21). It has been proposed that the decrease in EGF receptor binding that occurs following TCDD treatment is due to the overexpression of TGF-a protein induced by TCDD (27).
TCDD causes an increase in TGF-a mRNA levels as a result of a post-transcriptional stabilization of the TGF-a message (Fig. 5). TCDD has also been shown to regulate the tissue-specific expression of CYPlAl and CYPlA2, in part, through a post-transcriptional process (37,38). Thus, posttranscriptional control may represent an important pathway by which TCDD regulates gene expression.
In contrast to TGF-a, the decrease in the steady state level of TGF-fi, mRNA following TCDD treatment is due to a decrease in the rate of TGF-P2 gene transcription. Transcriptional down-regulation is specific for TGF-P2, since transcription of TGF-P1 is not affected by TCDD. Although the observed changes in TGF-P2 mRNA levels involve direct transcriptional regulation of the TGF-P2 gene, this may represent a secondary response, not necessarily regulated directly by the TCDD-Ah receptor complex, since changes in the rate of transcription of TGF-P, occur later (after 90 min) than with CYPlAl and do not occur in the presence of cycloheximide (Table I).TGF-P1 and TGF-P2 are closely related polypeptides encoded by separate genes and regulated independently (17,23). For the most part, the biological activities of TGF-P1 and TGF-b2 appear to be similar. However, TGF-P1 and TGF-P, show some differences in their expression and activity. TGF-,B1 mRNA is detected in most tissues and tumor cells investigated, while the expression of TGF-P, is more confined to specific cell types (39,40). In addition, TGF-P1 and TGF-P2 differentially regulate hemopoietic cell differentiation, inhibition of endothelial cell proliferation, and mesoderm induction (41-43). Furthermore, TGF-& and TGF-p, are differentially expressed during the induction of terminal differentiation of cultured murine keratinocytes by calcium (24). Thus, the regulated expression of TGF-P1 and TGF-P, is important in the normal growth and differentiation process and the differential regulation of TGF-P1 and TGF-P2 expression may be of significance in the cellular response to TCDD.
In this study we have demonstrated that TCDD induces TGF-a expression via a post-transcriptional mechanism and reduces expression of TGF-P2 at the level of gene transcription. However, regulation of TGF-P2 expression does not appear to be under direct transcriptional control of the Ah receptor. Thus, regulation of gene expression by TCDD results in differential changes in mRNA levels for specific genes and can occur by multiple mechanisms. These mechanisms include, direct transcriptional control (11,12,44), transcriptional control mediated as a secondary response, and posttranscriptional stabilization of mRNA.