STAT3 Activation Abrogates Growth Factor Dependence ... - CiteSeerX

Vol. 13, 355–362, August 2002

Cell Growth & Differentiation

STAT3 Activation Abrogates Growth Factor Dependence and Contributes to Head and Neck Squamous Cell Carcinoma Tumor Growth in Vivo1 Taro Kijima, Hideo Niwa, Richard A. Steinman, Stephanie D. Drenning, William E. Gooding, Abbey L. Wentzel, Sichuan Xi, and Jennifer Rubin Grandis2 Departments of Otolaryngology [T. K., H. N., S. D. D., A. L. W., S. X., J. R. G.], Pharmacology [J. R. G., R. A. S.], Medicine [R. A. S.], and Biostatistics [W. E. G.], University of Pittsburgh School of Medicine and University of Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania 15213

Abstract Epidermal growth factor receptor (EGFR) is upregulated and contributes to the loss of growth control in squamous cell carcinoma of the head and neck (SCCHN). Previously, we reported an association between autocrine stimulation of EGFR and constitutive signal transducers and activators of transcription (STAT) 3 activation in SCCHN cells in vitro and in vivo. Here, we evaluated the role of activated STAT3 in tumor progression and EGFR-independent mitogenic signaling. We found that SCCHN cells stably transfected with a dominant active STAT3 construct expressed elevated levels of STAT3 target genes, including Bcl-XL and cyclin D1, and demonstrated increased proliferation in vitro and more rapid tumor growth rates in vivo. Cell cycle analysis demonstrated an increased proportion of STAT3 construct transfectants in G2-M. These findings provide evidence that constitutive STAT3 activation contributes to tumor growth in SCCHN, independent of the EGFR autocrine axis.

Introduction The EGFR3 is overexpressed and has been implicated in the growth of a wide variety of carcinomas, including SCCHN. Up-regulation of EGFR and its ligand, TGF-␣, has

Received 11/15/01; revised 6/3/02; accepted 6/17/02. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by NIH Grants CA77308 and DE13059 (to J. R. G.). 2 To whom requests for reprints should be addressed, at Departments of Otolaryngology and Pharmacology, University of Pittsburgh School of Medicine and University of Pittsburgh Cancer Institute, 200 Lothrop Street, Suite 500, Pittsburgh, PA 15213. E-mail: [email protected]. 3 The abbreviations used are: EGFR, epidermal growth factor receptor; SCCHN, squamous cell carcinoma of the head and neck; STAT3-C, STAT3 construct; TGF, transforming growth factor; STAT, signal transducers and activators of transcription; EMSA, electrophoretic mobility shift assay; BrdUrd, bromodeoxyuridine; HRP, horseradish peroxidase; cdk, cyclin-dependent kinase; SFM, serum-free medium; EGF, epidermal growth factor; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.

been identified as an early event in head and neck carcinogenesis (1, 2). Stimulation of the EGFR axis by ligand (derived from autocrine or paracrine sources) increased the proliferation of cells that overexpressed EGFR, including SCCHN (3, 4). Further investigation demonstrated that expression levels of TGF-␣ or EGFR in the primary SCCHN tumor predicted patient survival, independent of other clinical and pathological parameters, including nodal staging (5). However, the precise signaling pathways that regulate EGFR-mediated SCCHN proliferation are incompletely understood. We have demonstrated previously that down-modulation of EGFR or TGF-␣ expression in SCCHN inhibits SCCHN cell proliferation in vitro (6). TGF-␣ initiates cell signaling through stimulation of EGFR, which contains a cytoplasmic domain with intrinsic protein tyrosine kinase activity. In response to EGFR ligands, including TGF-␣, EGFRs dimerize and become phosphorylated on multiple tyrosine residues. These phosphotyrosines, in turn, facilitate the recruitment of STATs to specific tyrosine residues (Y1068 and Y1086) in the cytoplasmic domain of the receptor (7). Direct interaction between STAT protein Src homology 2 domains and the activated EGFR results in STAT phosphorylation followed by dimerization and translocation to the nucleus. In the nucleus, STATs bind to DNA response elements in promoters and regulate growth factor- and cytokine-directed gene expression (8, 9). STAT3, one of seven STATs that have been identified to date, is constitutively activated in a variety of human malignancies, including breast cancers and breast cancer-derived cell lines compared with benign lesions and normal breast epithelium (10 –12), prostate cancers (13, 14), and myelomas (15). We reported previously that both TGF-␣ expression and EGFR expression in SCCHN cells are linked to constitutive STAT3 activation in vitro (16). Further analysis demonstrated that EGFR was associated with STAT3 activation in vivo (17). Abrogation of STAT3 using antisense oligonucleotides or transfection of dominant negative mutants of STAT3 resulted in the growth inhibition of SCCHN cells. These studies established a requirement for STAT3 in SCCHN growth but did not establish whether activated STAT3 was sufficient to sustain cell proliferation. It is notable that EGFR binding activates several signaling pathways including ras/mitogenactivated protein kinase; phospholipase C ␥, and STATs. A requirement for coincident activation of several of these pathways in cell growth and survival could not be ruled out by the studies to date. The present study was undertaken to gain further insight into the role of activated STAT3 in both EGFR-dependent and EGFR-independent mitogenic signaling in SCCHN. STAT3 activation levels in eight SCCHN cell lines were determined and correlated with TGF-␣ and EGFR mRNA and

355

356

STAT3-mediated EGFR-independent Growth in SCCHN

Fig. 1. EGFR protein levels correlate with constitutive STAT3 activation levels in eight representative SCCHN cell lines. Immunoblotting (A) and EMSA (B) showing EGFR protein expression and corresponding constitutive STAT3 activation in SCCHN cell lines. The SIF-A band represents STAT3 homodimers. C, scatterplot showing correlation of EGFR protein expression levels and STAT3 activation levels in eight SCCHN cell lines (r ⫽ 0.905, p ⫽ 0.02).

protein expression levels. A representative SCCHN cell line was stably transfected with a constitutively active form of STAT3 (STAT3-C; Ref. 18). Clones were enumerated and characterized in vitro and in vivo for growth determinations, STAT3 activation, STAT3-mediated gene expression, and responsiveness to the EGFR autocrine pathway.

Results EGFR Expression Levels Correlate with STAT3 Activation Levels in SCCHN Cell Lines. We previously reported a strong correlation between TGF-␣ or EGFR mRNA and protein levels in SCCHN cell lines and tissues (2). Northern blotting, immunoblotting, and EMSAs were used to determine the levels of TGF-␣, EGFR, and STAT3 activation in each of eight SCCHN cell lines. Although the levels varied among the cell lines tested, there was good correlation between EGFR mRNA and EGFR protein expression, as described previously (data not shown). Constitutive STAT3 activation levels varied among the eight SCCHN cell lines examined. Strong correlation was observed between STAT3 activation levels and EGFR mRNA levels or EGFR protein levels in all eight SCCHN cell lines (r ⫽ 0.976, P ⫽ 0.01; r ⫽ 0.905, P ⫽ 0.02, respectively; Fig. 1). In contrast, there was no significant correlation between TGF-␣ mRNA expression levels and STAT3 activation levels in the cell lines tested, suggesting that receptor rather than ligand availability is the rate-limiting step in EGFR-mediated STAT3 activation in SCCHN cells (data not shown).

Increased STAT3-mediated Gene Expression in SCCHN Cells Expressing Constitutively Activated STAT3. To determine whether constitutively activated STAT3 contributes to cell growth in SCCHN, a representative SCCHN cell line, UM-22B, which expresses relatively low levels of activated STAT3, was stably transfected with a constitutively active STAT3 vector (FLAG-tagged STAT3-C) or empty vector (RcCMV-Neo). Representative clones were isolated and expanded under selection pressure. Coimmunoprecipitation using a FLAG monoclonal antibody was performed to verify that each clone expressed the vector (Fig. 2). Constitutive STAT3 activation levels in STAT3-C clones and RcCMV-Neo clones were determined by EMSA as well as by immunoblotting utilizing a phosphotyrosine-specific STAT3 antibody. STAT3-C-transfected cells demonstrated higher levels of tyrosine phosphorylated STAT3 and constitutive STAT3 activation on EMSA compared with vector-transfected control cells or the parental cell line (Fig. 2). STAT3 homodimers (or heterodimers), when phosphorylated, translocate to the nucleus and bind to STAT3 response elements in promoter regions of target genes, thus stimulating gene transcription. STAT3-mediated tumor progression presumably relies on the transcription of genes stimulated by STAT3 binding to specific regulatory regions. STAT3responsive regions have been identified in several genes that influence cell cycle progression or abrogate apoptosis, including Bcl-XL and cyclin D1 (15, 19). To determine the consequences of STAT3 activation of target gene expression, Bcl-XL and cyclin D1 expression levels were analyzed in STAT3-C-transfected cells and vector-transfected control


Fig. 3. STAT3 target gene expression level in STAT3-C-transfected SCCHN cells (UM-22B). Immunoblotting of representative STAT3-Ctransfected clones and vector control-transfected clones was performed to determine expression levels of (A) Bcl-xL and (B) cyclin D1 and (C) a ␤-actin loading control. Fig. 2. Increased STAT3 activation levels in STAT3-C-transfected clones. A representative SCCHN cell line (UM-22B) was stably transfected with STAT3-C. A, coimmunoprecipitation of representative clones showing FLAG expression in STAT3-C clones but not in controls. Cells were immunoprecipitated with anti-FLAG M2 monoclonal antibody followed by immunoblotting with the same antibody. B, immunoblotting of representative clones showing higher phospho-STAT3 expression in STAT3-C clones compared with vector-transfected control clones. C, EMSA showing higher constitutive STAT3 activation levels (SIF-A represents STAT3 homodimers) in STAT3-C clones compared with vector-transfected control clones.

SCCHN cells. As shown in Fig. 3, both Bcl-XL and cyclin D1 were up-regulated in SCCHN cells stably transfected with dominant active STAT3 compared with vector control transfectants. STAT3 Activation Contributes to SCCHN Cell Growth in Vitro and in Vivo. To examine the growth rate of STAT3-Ctransfected cells, cell counts were performed using vital dye exclusion. As shown in Fig. 4, STAT3-C clones proliferated more rapidly compared with vector-transfected control cells. Observed counts of STAT3- and Neo-transfected cells were successfully fit to distinct polynomial regression curves. The models showed significant increased cell count on days 4 (P ⫽ 0.0012), 6 (P ⬍ 0.0001), and 8 (P ⬍ 0.0001), but not on days 1 and 2 (P ⬎ 0.10). By day 6, Neo-transfected cells were decreasing, whereas STAT3-transfected cells were continuing to proliferate. To determine the effect of activated STAT3 on tumor growth in vivo, STAT3-C-transfected SCCHN cells were inoculated s.c. in nude mice, and tumor growth rates were monitored. Tumors that developed from cells transfected with dominant active STAT3 grew more rapidly than tumors that developed from vector-transfected control cells or the parental (untransfected) cell line (P ⫽ 0.05; Fig. 5). STAT3 activation promotes G1-S transition and leads to accumulation of cells in G2-M phase. To better understand

Fig. 4. STAT3 contributes to SCCHN growth in vitro. Growth rate of a representative STAT3-C-transfected clone ({) compared with the growth rate of a representative vector control-transfected clone (‚). Viable cells were counted via vital dye exclusion over the course of 8 days.

the accelerated growth of STAT3-C-expressing SCCHN cells, cell cycle analysis of STAT3-C and control transfectants was undertaken using BrdUrd pulsing and DNA content analysis. Interestingly, in log phase cells, the proportion of cells in G0-G1 phase of the cell cycle was 3-fold higher in vector transfectants than STAT3-C transfectants (39.4% versus 12.6%). In the STAT3-C-expressing cells, a much higher percentage of the cells were in G2-M phase (40.1% versus 15.9%). The cell cycle profile of STAT3-C-transfected cells and vector-transfected control cells was then determined in the presence of serum starvation. STAT3-C-transfected cells or vector control-transfected cells were serum-starved, and cell cycling was analyzed by flow cytometry 1, 2, 3, or 4 days later. At each time point, a higher percentage of STAT3-Ctransfected cells were cycling (in S-G2-M) compared with

357

358


Fig. 5. Differential growth rate of tumors in vivo derived from STAT3-C-transfected and vector control-transfected SCCHN cells. Tumor volumes of 10 mice randomized to receive STAT3-C-transfected or vector controltransfected SCCHN cells. Tumor dimensions were measured 11–31 days after tumor inoculation. A mixed quadratic regression model was fit to the log-transformed tumor volumes. Despite a greater tumor volume in the controltransfected SCCHN xenografts on day 11, the curves eventually cross, and the growth rate was significantly greater in the STAT3-Ctransfected tumors (likelihood ratio test, p ⫽ 0.0306). Symbols represent the actual tumor volumes by day. Lines are the predicted tumor volumes based on the fitted model.

Table 1

Cell cycle analysis of dominant active STAT3-transfected cells

Dominant active STAT3 transfectants and control transfectants were serum-starved for 1, 2, 3, or 4 days followed by BrdUrd labeling and flow cytometry. % cells in G0-G1 Time point

1 2 3 4

day days days days

% cells in S

STAT3-C

Vector control

39.0 41.6 46.7 52.6

81.8 88.3 90.5 84.3

vector-transfected control cells (Table 1). It is notable that serum starvation was ineffective in inducing G0-G1-phase growth arrest in the STAT3-C transfected cells. Because of the ineffectiveness of low serum in maintaining a G1-S-phase checkpoint in STAT3-C transfectants, the levels of several cell cycle regulators were measured in serum-starved Neo and STAT3-C transfectants (Fig. 6). A decrease in p21 expression in the STAT3-C transfected cells was evident, but overall, there was little change in G1-S regulators (p27, cdk2, and cyclin D3) or G2-M checkpoint regulators (cyclin B and cdk1) at the steady-state protein level (Fig. 6). Increased STAT3 Activation Reduces Proliferative Response to EGFR Ligand. EGFR stimulation by ligand in many cancer cells, including SCCHN, results in cell proliferation. We reported previously that TGF-␣/EGFR activation is linked to STAT3 signaling in SCCHN cells in vitro or in vivo (16, 17). Cell growth assays were then performed in the presence of EGFR ligand or an EGFR-specific inhibitor to further investigation of the ability of constitutive STAT3 activation to overcome the EGFR autocrine pathway. These studies demonstrated that SCCHN cells expressing dominant active STAT3 did not proliferate in response to EGFR ligand and were not inhibited by blockade of EGFR tyrosine kinase activity (Fig. 7).

% cells in G2-M

STAT3-C

Vector control

STAT3-C

Vector control

10.3 6.26 3.47 3.15

10.4 4.72 1.75 1.77

45.7 46.8 41.9 35.4

6.83 6.09 6.51 10.4

Discussion STAT proteins serve the dual function of transmitting a signal from the cell surface to the nucleus and directly participating in gene regulation (9). We demonstrated previously that TGF␣/EGFR signaling in SCCHN was linked to STAT3 activation in vitro and in vivo. These prior investigations used strategies that selectively abrogated TGF-␣ or EGFR and examined the consequences of such a blockade on constitutive STAT3 activation (20). In the present study, we investigated the role of activated STAT3 independent of EGFR. We transfected SCCHN cells with a constitutively active STAT3 vector generated by the substitution of two cysteine residues within the COOH-terminal loop of the Src homology 2 domain, producing a molecule that dimerizes spontaneously, binds to DNA, and activates transcription (18). STAT3-C has been previously shown to cause transformation of fibroblasts as demonstrated by growth in soft agar and tumor formation in nude mice, thus defining activated STAT3 as an oncogene (18). The present study provides the first evidence for a critical role of STAT3 activation in xenograft tumor growth. STAT3 is widely expressed and activated in response to a large number of cytokines and growth factors, as well as by oncogenic receptor and nonreceptor (Src-like) tyrosine kinases (21). Many cancers are characterized by constitutive


Fig. 6. Expression of cell cycle regulators in control and STAT3-Cexpressing cells. Cells were serum-starved for 24 h, and then they were exposed to medium alone (⫺), medium containing EGF (30 ng/ml), or medium containing TGF-␣ (30 ng/ml) for 2 h. Extracts were then prepared and immunoblotted for cell cycle regulators as indicated.

STAT3 activation including lymphomas, leukemias, multiple myeloma, and gliomas as well as carcinomas of the breast, prostate, and head and neck (12, 13, 17, 22). Several strategies have been used to block STAT3 and examine the consequences on tumor cell growth including introduction of antisense oligonucleotides, antisense gene therapy, forced overexpression of dominant negative forms of STAT3, or inhibition of upstream kinases. STAT3-specific blockade in tumor-derived cell lines has resulted in growth inhibition and/or apoptosis (14 –17). In several studies, down-modulation of STAT3 was accompanied by decreased expression of STAT3-mediated genes. Constitutive activation of STAT3 has been shown to be accompanied by up-regulation of cyclin-D1, C-Myc, and Bcl-x (23). To examine the oncogenic effects of STAT3, a constitutively active molecule, STAT3-C, was generated. Murine fibroblasts transformed by stable introduction of this oncogene were capable of forming tumors in nude mice and were more resistant to proapoptotic stimuli compared with nontransformed fibroblasts (18, 24). In the present study, forced overexpression of STAT3-C was accompanied by increased expression levels of STAT3 target genes, including Bcl-XL and cyclin D1. STAT proteins, including STAT3, are critical mediators of many external signals initiated by growth factors and cytokines. The specific role of STAT3 in cell cycle regulation is incompletely understood. The constitutive expression of activated STAT3 led to an altered cell cycle profile including a

Fig. 7. Resistance of dominant active STAT3 transfectants to the EGFR autocrine axis. A, representative MTT assay demonstrating growth stimulation of vector control transfectants relative to untreated control cells, compared with STAT3-C-transfected cells treated with exogenous EGFR ligand (EGF, 30 ng/ml). The experiment was repeated three times with 3 wells/condition for each experiment. B, representative MTT assay demonstrating relative growth inhibition of control-transfected cells compared with STAT3-C transfectants treated with an EGFR-specific tyrosine kinase inhibitor at either of two doses (䊐, 100 nM PD153035; u, 1000 nM PD153035).

decreased proportion of cells in G1-G0 phase, decreased restriction point control, and an increased percentage of cells in G2-M phase. Steady-state levels of a number of cell cycle regulators did not vary significantly between STAT3-C and vector transfectants or between unstimulated cells and vector-transfected cells stimulated with ligand to cycle. This suggested that constitutive undermining of the G1-S checkpoint in STAT3 transfectants may be related to alterations in complex formation by cell cycle regulators, or by c-myc induction by STAT3-C. The cell cycle regulator p21 was expressed at a lower level in serum-starved STAT3-C cells, but this may not account for the loss of restriction point control in these cells, which has generally been related to p27 rather than p21 (25). Whereas the transition through G1 phase was accelerated in STAT3-C-transfected cells, passage through G2-M phase was relatively retarded. We speculate that constitutive activation of STAT3-C may unmask a G2-M checkpoint function of STAT3 that is not apparent after the transient activation of STAT3 by ligand. Alternatively, constitutive expression of STAT3-C may cause a more profound acceleration of passage through G1 and S cell cycle

359

360


phases, leading to accumulation of cells in a normal-length G2-M phase. These possibilities are under investigation. The detection of EGFR overexpression and STAT3 activation in the same cancer cell does not necessarily imply that EGFR stimulation directly mediates STAT3 activation. In A431 vulvar squamous cell carcinoma cells, STAT3 was constitutively complexed with EGFR and rapidly phosphorylated in response to EGF (26). However, EGFR kinase activity was not required for constitutive STAT3 activation in EGFR-expressing breast cancer cell lines (12). Studies to date suggest that constitutive STAT3 activation in SCCHN results from chronic stimulation of EGFR, most likely by production of autocrine or paracrine growth factors including TGF-␣ (27). In SCCHN cells, constitutive STAT3 activation by EGFR autocrine stimulation results in proliferative and antiapoptotic effects on tumor cells. In the present study, we explored the effects of increasing constitutive STAT3 activation by forced overexpression of a dominant active STAT3 construct. In addition to increased proliferation in vitro and tumor growth in vivo, we determined that STAT3 activation could facilitate the resistance of SCCHN cells to the EGFR autocrine axis. Cumulative evidence supports the role of EGFR as a potential therapeutic target in cancers that express elevated levels of this growth factor receptor. Clinical trials are under way to investigate the anti-tumor efficacy of EGFR targeting strategies, including monoclonal antibodies and EGFRspecific tyrosine kinase inhibitors (28). In a Phase I study of head and neck cancer patients refractory to the therapeutic effects of cisplatin, additional treatment with a monoclonal antibody against the EGFR plus cisplatin resulted in a high response rate (29). The results of the present investigation suggest that STAT3 activation can serve as a growthpromoting signal, independent of the EGFR axis, in human tumors that express high levels of this growth factor receptor. Therefore, strategies that selectively abrogate STAT3 may demonstrate antitumor efficacy and should be considered when designing molecular targeting cancer strategies.

Materials and Methods Cell Lines. Eight cell lines were previously derived from patients with SCCHN. Several of these SCCHN cell lines are part of a large collection established in the Department of Otolaryngology at the University of Pittsburgh and Pittsburgh Cancer Institute (30). The cell line 1483 is a well-characterized SCCHN cell line derived from a tumor of the retromolar trigone region of the oropharynx (31). Several cell lines were generously provided by Dr. Susanne M. Gollin from the University of Pittsburgh (SCC-104, SCC-207, SCC-203, and SCC-200). UM-22B is a well-characterized SCCHN cell line developed at the University of Michigan (32). All cells were grown in DMEM (Cellgro, Washington, D.C.) supplemented with 15% fetal bovine serum, 100 units/ml penicillin, and 100 units/ml streptomycin (all from Life Technologies, Inc., Grand Island, NY). All of the SCCHN cell lines tested negative for Mycoplasma infection by the Gen-Probe Mycoplasma T.C. rapid detection system before use in experiments (GenProbe, Inc., San Diego, CA).

Transfection of Dominant Active STAT3 (STAT3-C) into SCCHN Cells. The cell line UM-22B was chosen for transfection with constitutively active STAT3 (STAT-3C). The UM22B cell line demonstrated relatively low constitutive activation levels of STAT3 and grows well as a xenograft in nude mice. The UM-22B cells were stably transfected with FLAGtagged expression vectors encoding STAT3-C or empty vector RcCMV-Neo as a negative control using Gene Porter (Gene Therapy Systems, San Diego, CA) according to the manufacturer’s instructions. Both the STAT3-C vector and the empty vector were a kind gift from Dr. Jacqueline F. Bromberg of The Rockefeller University (18). Briefly, cells were grown in 6-well plates containing DMEM until they reached 80% confluence. Cells were incubated overnight with fresh serum-free media containing Gene Porter reagent and vector plasmid. The medium was substituted with DMEM containing G418 (2000 ␮g/ml) 20 h after transfection. Several clones (three to five clones) from each vector were isolated using a standard cloning ring technique and expanded. Each clone was passaged at least 10 times and examined for Mycoplasma infection by a Mycoplasma Detection Kit (Roche Diagnostics GmbH, Mannheim, Germany). In Vivo and in Vitro Growth Determinations. Cell counts using vital dye exclusion were performed in duplicate to determine the growth rate of STAT3-C-transfected clones. SCCHN cells were plated in 24-well plates in serum-containing medium and counted in duplicate every 2 days for a total of 8 days. For in vivo studies, female athymic nude mice (4 – 6 weeks old; weighing approximately 20 –25 g; Harlan Sprague-Daley, Indianapolis, IN) were assigned to each tumor group. In the first experiment, mice in each group (5–7 animals/group) received s.c. implant of 1.5 ⫻ 106 cells (UM-22B STAT-3-C clones or UM-22B RcCMV-Neo clones) in both the right and left flank with a 26-gauge needle/1-ml tuberculin syringe (Becton Dickinson, Rutherford, NJ). Once established tumor nodules formed (approximately 2 weeks after tumor cell implantation), tumor volumes were calculated (tumor volumes ⫽ length ⫻ width2/2). Tumor measurements were taken three times a week for a total of 3 weeks. Tumors were harvested immediately after sacrifice and snap-frozen in liquid nitrogen for subsequent studies. In a second experiment, 10 animals received s.c. implant in the right of STAT3C-transfected cells in the right flank and Neo-transfected UM-22B cells in the left flank so that each animal could serve as its own control. Immunoblotting and Immunoprecipitation. Cell lines were lysed in detergent containing 1% NP40, 0.1 mM phenylmethylsulfonyl fluoride, 1 mg/ml leupeptin, and 1 mg/ml aprotinin, and protein levels were determined using the BioRad protein assay method (Bio-Rad Laboratories, Hercules, CA). Fifty ␮g of total protein were separated on 10% SDSPAGE gels and transferred to nitrocellulose membranes using semi-wet blotting. Filters were blocked with 5% bovine serum albumin/Tris-buffered saline with Tween 20 solution overnight, rinsed three times in Tris-buffered saline with Tween 20, and incubated for 90 min with a mouse antihuman EGFR monoclonal antibody (Transduction Laboratories, Lexington, KY) to determine EGFR protein expression levels of SCCHN cell lines or for 60 min with mouse polyclonal anti-


phosphotyrosine STAT3 to determine expression levels of activated STAT3 or 60 min with polyclonal anti-FLAG antibody (Sigma Chemicals, St. Louis, MO) monoclonal Bcl-xL antibody (Santa Cruz Biotechnology, Santa Cruz, CA), polyclonal cyclin D1 antibody (Santa Cruz Biotechnology), or STAT3 phosphotyrosine-specific monoclonal antibody (Santa Cruz Biotechnology). The other antibodies used included monoclonal cdk1, cyclin D3, cyclin B1, and HRP-conjugated p27 (Transduction Laboratories). In addition, polyclonal HRP-conjugated p21 (Santa Cruz Biotechnology) and polyclonal anti-cdk2 antibodies were used (Neomarkers, Inc., Freemont, CA). FLAG expression was determined by immunoprecipitation with an anti-FLAG monoclonal antibody (Sigma Chemicals), followed by immunoblotting with the same anti-FLAG antibody. Membranes were then incubated for 45 min with HRP-conjugated secondary antibody (Bio-Rad Laboratories). Enhanced chemiluminescence (New Life Science Inc., Boston, MA) technology was used to detect protein expression levels. Membranes were exposed to Kodak XOMAR film (Eastman Kodak Co., Rochester, NY) for 10 –30 s. Quantification of protein expression levels was performed by scanning the autoradiogram using a Molecular Dynamics Personal Densitometer SI and Imagequant software (Molecular Dynamics, Sunnyvale, CA). Only comparisons made between samples run on the same gel were considered valid. When quantitative determinations were made on several gels, a positive control/standard was run on each gel to allow for normalization of designated values. EMSA. Cells were harvested when 50 – 80% confluent, and whole cell extracts were prepared. EMSA was performed on 4% native polyacrylamide gels as described previously (33, 34). STAT3 activation was evaluated by using binding reaction with 20 ␮g of extracted protein and radiolabeled high-affinity serum inducible element duplex oligonucleotide, used to clone and characterize STAT3 (35). Quantification of STAT3 activation levels was performed utilizing a Molecular Dynamics Personal Densitometer SI and Imagequant software (Molecular Dynamics). Background density was calculated and subtracted from all bands. Activation levels of STAT3 were assessed by determining the density of the SIF-A band (representing STAT3 homodimer) only as described previously (17). Only comparisons made between samples run on the same gel were considered valid. Cell Cycle Analysis. UM-22B-Neo and UM-22B-STAT3C-transfected cells were grown in DMEM with 12% fetal bovine serum in 10-cm plastic tissue culture dishes. Cells were either maintained in log phase or rested for 24, 48, 72, or 96 h in SFM, and then they were incubated for 30 min in SFM with 30 ng/ml recombinant TGF-␣ or EGF. In a separate experiment, cells were rested for 24 h in SFM and then incubated for 1 h at 37°C in 10 ␮M BrdUrd in SFM and harvested as per the manufacturer’s instructions (BrdUrd flow kit; BD PharMingen, San Diego, CA). Flow cytometric analysis (CoulterXL flow cytometer) was used to measure cellular incorporation of BrdUrd (stained with FITC anti-BrdUrd) and the total DNA content (stained with 7-aminoactinomycin D). A dot plot of 7-AAD levels versus BrdUrd incorporation was analyzed as described in the BrdUrd flow

kit manual to determine the percentage of the cell populations in the G0-G1, S, and G2-M stages of the cell cycle. Cell Growth Assays. Cells were plated at 5000 cells/well in a 48-well plate. After 1 day of growth in 12% serum DMEM, cells were starved for 24 h in serum-free DMEM. Cells were treated for 60 min with 30 ng/ml EGF and then rested for 30 min in serum-free DMEM. The percentage of stimulation versus untreated control was determined by metabolic activity assay (MTT; Sigma). Determinations in 6 wells were averaged for each data point. For the EGFR inhibition studies, cells were plated at 5000 cells/well in a 48-well plate. After 1 day of growth in 12% serum DMEM, cells were treated with an EGFR-specific tyrosine kinase inhibitor (100 or 1000 nM PD153035; Parke-Davis). Control wells were treated with DMSO in a volume equivalent to that used in the 1000 nM dose. The percentage of survival versus DMSO vehicle control was determined by metabolic activity assay (MTT; Sigma). Results from triplicate wells were averaged for each data point. Statistical Analysis. Correlations between EGFR protein levels, TGF-␣ mRNA levels, EGFR mRNA levels, and STAT3 activation levels were examined using the Spearman rank correlation coefficient. Cell counts from nine grids on a slide from each of 3 wells were obtained on days 1, 2, 4, 6, and 8. Cell counts were assumed to follow a Poisson distribution and were square root-transformed to stabilize variances. A mixed linear model was fit to the transformed cell counts for each day. A within-slide covariance matrix was estimated to capture the correlation between adjacent grids. Regression slopes were tested for homogeneity. In vivo tumor growth rates were compared by testing differences in sample means at each time and by comparing growth curves. Briefly, quadratic growth curves were fit by regressing log-transformed tumor volumes on days of tumor volume measurement. A mixed model approach (36) was used that represents individual mice as random effects and accounts for between and within-mouse variation. A likelihood ratio test was constructed to determine the appropriate choice of model parameters. Time by group interaction terms were also tested.

References 1. Grandis, J. R., and Tweardy, D. J. Elevated levels of transforming growth factor ␣ and epidermal growth factor receptor messenger RNA are early markers of carcinogenesis in head and neck cancer. Cancer Res., 53: 3579 –3584, 1993. 2. Grandis, J. R., Melhem, M. F., Barnes, E. L., and Tweardy, D. J. Quantitative immunohistochemical analysis of transforming growth factor-␣ and epidermal growth factor receptor in patients with squamous cell carcinoma of the head and neck. Cancer (Phila.), 78: 1284 –1292, 1996. 3. Yoneda, T., Alsina, M. M., Watatani, K., Bellot, F., Schlessinger, J., and Mundy, G. R. Dependence of a human squamous carcinoma and associated paraneoplastic syndromes on the epidermal growth factor receptor pathway in nude mice. Cancer Res., 51: 2438 –2443, 1991. 4. Ozawa, S., Ueda, M., Ando, N., Hirai, M., and Shimizu, N. Stimulation by EGF of the growth of EGF receptor-hyperproducing tumor cells in athymic mice. Int. J. Cancer, 40: 706 –710, 1987. 5. Grandis, J. R., Melhem, M. F., Gooding, W. E., Day, R., Holst, V. A., Wagener, M. M., Drenning, S. D., and Tweardy, D. J. Levels of TGF-␣ and EGFR protein in head and neck squamous cell carcinoma and patient survival. J. Natl. Cancer Inst. (Bethesda), 90: 824 – 832, 1998. 6. Grandis, J. R., Chakraborty, A., Zeng, Q., Melhem, M. F., and Tweardy, D. J. Downmodulation of TGF-␣ protein expression with antisense oligo-

361

362


nucleotides inhibits proliferation of head and neck squamous carcinoma but not normal mucosal epithelial cells. J. Cell Biochem., 69: 55– 62, 1998. 7. Coffer, P., and Kruijer, W. EGF receptor deletions define a region specifically mediating STAT transcription factor activation. Biochem. Biophys. Res. Commun., 210: 74 – 81, 1995. 8. Ihle, J. N. STATs: signal transducers and activators of transcription. Cell, 84: 331–334, 1997. 9. Darnell, J. E., Jr. STATs and gene regulation. Science (Wash. DC), 277: 1630 –1635, 1997. 10. Garcia, R., Yu, C. L., Hudnall, A., Catlett, R., Nelson, K. L., Smithgall, T., Fujita, D. J., Ethier, S. P., and Jove, R. Constitutive activation of Stat3 in fibroblasts transformed by diverse oncoproteins and in breast carcinoma cells. Cell Growth Differ., 8: 1267–1276, 1997. 11. Watson, C. J., and Miller, W. R. Elevated levels of members of the STAT family of transcription factors in breast carcinoma nuclear extracts. Br. J. Cancer, 71: 840 – 844, 1995. 12. Garcia, R., Bowman, T. L., Niu, G., Yu, H., Minton, S., Muro-Cacho, C. A., Cox, C. E., Falcone, R., Fairclough, R., Parsons, S., Laudano, A., Gazit, A., Levitzki, A., Kraker, A., and Jove, R. Constitutive activation of Stat3 by the Src and JAK tyrosine kinases participates in growth regulation of human breast carcinoma cells. Oncogene, 20: 2499 –2513, 2001. 13. Lou, W., Ni, Z., Dyer, K., Tweardy, D. J., and Gao, A. C. Interleukin-6 induces prostate cancer cell growth accompanied by activation of Stat3 signaling pathway. Prostate, 42: 239 –242, 2000. 14. Ni, Z., Lou, W., Leman, E. S., and Gao, A. C. Inhibition of constitutively activated Stat3 signaling pathway suppresses growth of prostate cancer cells. Cancer Res., 60: 1225–1228, 2000. 15. Catlett-Falcone, R., Landowski, T. H., Oshiro, M. M., Turkson, J., Levitzki, A., Savino, R., Ciliberto, G., Moscinski, L., Fernandez-Luna, J. L., Nunez, G., Dalton, W. S., and Jove, R. Constitutive activation of Stat3 signaling confers resistance to apoptosis in human U266 myeloma cells. Immunity, 10: 105–115, 1999. 16. Rubin Grandis, J., Drenning, S. D., Chakraborty, A., Zhou, M., Zeng, Q., Pitt, A., and Tweardy, D. Requirement of Stat3 but not Stat1 for EGFR-mediated cell growth in vitro. J. Clin. Investig., 102: 1385–1392, 1998. 17. Rubin Grandis, J., Drenning, S. D., Zeng, Q., Watkins, S., Melhem, M., Endo, S., Johnson, D., Huang, L., and Kim, J. Constitutive activation of Stat3 signaling abrogates apoptosis in squamous cell carcinogenesis in vivo. Proc. Natl. Acad. Sci. USA, 97: 4227– 4232, 2000. 18. Bromberg, J. F., Wrzeszczynska, M. H., Devgan, G., Zhao, Y., Pestell, R. G., Albanese, C., and Darnell, J. E., Jr. Stat3 as an oncogene. Cell, 98: 295–303, 1999. 19. Oritani, K., Tomiyama, Y., Kincade, P. W., Aoyama, K., Yokota, T., Matsumura, I., Kanakura, Y., Nakajima, K., Hirano, T., and Matsuzawa, Y. Both Stat3-activation and Stat3-independent BCL2 downregulation are important for interleukin-6-induced apoptosis of 1A9-M cells. Blood, 93: 1346 –1354, 1999. 20. Grandis, J. R., Zeng, Q., and Drenning, S. D. Epidermal growth factor receptor-mediated Stat3 signaling blocks apoptosis in head and neck cancer. Laryngoscope, 110: 868 – 874, 2000.

21. Schindler, C., and Darnell, J. E., Jr. Transcriptional responses to polypeptide ligands: the JAK-STAT pathway. Annu. Rev. Biochem., 64: 621– 651, 1995. 22. Bowman, T., Garcia, R., Turkson, J., and Jove, R. STATs in oncogenesis. Oncogene, 19: 2474 –2488, 2000. 23. Turkson, J., and Jove, R. STAT proteins: novel molecular targets for cancer drug discovery. Oncogene, 19: 6613– 6626, 2000. 24. Shen, Y., Devgan, G., Darnell, J. E., Jr., and Bromberg, J. F. Constitutively activated Stat3 protects fibroblasts from serum withdrawal and UV-induced apoptosis and antagonizes the proapoptotic effects of activated Stat1. Proc. Natl. Acad. Sci. USA, 98: 1543–1548, 2001. 25. Coats, S., Flanagan, W. M., Nourse, J., and Roberts, J. M. Requirement of p27Kip1 for restriction point control of the fibroblast cell cycle. Science (Wash. DC), 272: 877– 880, 1996. 26. Olayioye, M. A., Beuvink, I., Horsch, K., Daly, J. M., and Hynes, N. E. ErbB receptor-induced activation of stat transcription factors is mediated by Src tyrosine kinases. J. Biol. Chem., 274: 17209 –17218, 1999. 27. Song, J. I., and Grandis, J. R. STAT signaling in head and neck cancer. Oncogene, 19: 2489 –2495, 2000. 28. Lango, M. N., Shin, D. M., and Grandis, J. R. Targeting growth factor receptors: integration of novel therapeutics in the management of head and neck cancer. Curr. Opin. Oncol., 13: 168 –175, 2001. 29. Shin, D. M., Donato, N. J., Perez-Soler, R., Shin, H. J., Wu, J. Y., Zhang, P., Lawhorn, K., Khuri, F. R., Glisson, B. S., Myers, J., Clayman, G., Pfister, D., Falcey, J., Waksal, H., Mendelsohn, J., and Hong, W. K. Epidermal growth factor receptor-targeted therapy with C225 and cisplatin in patients with head and neck cancer. Clin. Cancer Res., 7: 1204 – 1213, 2001. 30. Heo, D. S., Snyderman, C., Gollin, S. M., Pan, S., Walker, E., Deka, R., Barnes, E. L., Johnson, J. T., Herberman, R. B., and Whiteside, T. L. Biology, cytogenetics, and sensitivity to immunological effector cells of new head and neck squamous cell carcinoma lines. Cancer Res., 49: 5167–5175, 1989. 31. Sacks, P. G., Oke, V., Amos, B., Vasey, T., and Lotan, R. Modulation of growth, differentiation and glycoprotein synthesis by b-all-trans retinoic acid in a multicellular tumor spheroid model for squamous cell carcinoma of the head and neck. Int. J. Cancer, 44: 926 –933, 1989. 32. Jetten, A. M., Kim, J. S., Sacks, P. G., Rearick, J. I., Lotan, R., Hong, W. K., and Lotan, R. Inhibition of growth and squamous cell differentiation markers in cultured human head and neck squamous carcinoma cells by b-all-trans retinoic acid. Int. J. Cancer, 45: 195–202, 1990. 33. Sadowski, H. B., Shuai, K., Darnell, J. E., Jr., and Gilman, M. Z. A common nuclear signal transduction pathway activated by growth factor and cytokine receptors. Science (Wash. DC), 261: 1739 –1744, 1993. 34. Wong, P., Severns, C. W., Guyer, N. B., and Wright, T. M. A unique palindromic element mediates gamma interferon induction of mig gene expression. Mol. Cell. Biol., 14: 914 –922, 1994. 35. Wagner, R. L., Hayward, N. K., Houghton, P. J., and Cochran, B. H. The SIF binding element confers sis/pdgf inducibility onto the c-fos promoter. EMBO J., 9: 4477– 4484, 1990. 36. Zeger, S. L., and Diggle, P. J. Semiparametric models for longitudinal data with application to CD4 cell numbers in HIV seroconverters. Biometrics, 50: 689 – 699, 1994.