Uncoupling between Epidermal Growth Factor Receptor and Downstream Signals Defines Resistance to the Antiproliferative Effect of Gefitinib in Bladder Cancer Cells

Activation of the epidermal growth factor receptor (EGFR) and downstream signaling pathways, such as phosphatidylinositol-3 kinase/Akt and Ras/mitogen-activated protein kinase (MAPK), have been implicated in causing resistance to EGFR-targeted therapy in solid tumors, including the urogenital tumors. To investigate the mechanism of resistance to EGFR inhibition in bladder cancer, we compared EGFR tyrosine kinase inhibitor (Gefitinib, Iressa, ZD1839) with respect to its inhibitory effects on three kinases situated downstream of EGFR: MAPK, Akt, and glycogen synthase kinase-3B (GSK-3B). We found that the resistance to the antiproliferative effects of gefitinib, in vitro as well as in vivo in nude mice models, was associated with uncoupling between EGFR and MAPK inhibition, and that GSK-3B activation and degradation of its target cyclin D1 were indicators of a high cell sensitivity to gefitinib. Further analysis of one phenotypic sensitive (253J B-V) and resistant (UM-UC13) cell lines revealed that platelet-derived growth factor receptor-B (PDGFRB) activation was responsible for short circuiting the EGFR/MAPK pathway for mitogenic stimuli. However, invasion as well as actin dynamics were efficiently reduced by EGFR inhibition in UM-UC13. Chemical disruption of signaling pathways or of PDGFR kinase activity significantly reduced the inactive pool of cellular GSK-3B in UM-UC13 cells. In conclusion, our data show that the uncoupling of EGFR with mitogenic pathways can cause resistance to EGFR inhibition in bladder cancer. Although this uncoupling may arise through different mechanisms, we suggest that the resistance of bladder cancer cells to EGFR blockade can be predicted early in the course of treatment by measuring the activation of GSK-3B and of nuclear cyclin D1. (Cancer Res 2005; 65(22): 10524-35)


Introduction
Transitional cell carcinoma (TCC) of the bladder is the fourth most common solid tumor malignancy in the United States.The American Cancer Society has estimated that there will be about 13,180 deaths from TCC in the United States in 2005, most of which will occur as the result of metastatic disease (1).Virtually, all patients with distant metastatic bladder cancer die of the disease, with a median survival duration of no better than 18 months even with the best available treatment regimens.
The epidermal growth factor receptor (EGFR) is expressed or overexpressed in a variety of human solid tumors, including nonsmall cell lung carcinoma (NSCLC), breast, colorectal, gastric, ovarian, prostate, and bladder cancer (2).Numerous studies have suggested that expression of high levels of EGFR in tumors is associated with advanced disease, development of a metastatic phenotype, and poor prognosis (3).For these reasons, inhibition of EGFR function is clearly an attractive target for solid tumor therapy.Although the mechanism by which EGFR regulates tumor biology in bladder cancer is not clearly defined, it has been shown that EGFR signaling regulates cellular proliferation, differentiation, survival, and invasion, and that it is implicated in the induction of tumor-induced angiogenesis and metastasis in bladder cancer (4).Over the past decade, drug discovery efforts have produced a variety of chemical structures that inhibit the EGFR tyrosine kinase, and several of these agents are currently under clinical development (5,6).Tyrosine kinase inhibitors (TKI) share some of their mechanisms of action with anti-EGFR monoclonal antibodies, suggesting that blocking ligand binding with antibodies or preventing kinase activation with specific inhibitors results in a similar shutdown of EGFR-dependent processes (7)(8)(9).
Gefitinib (Iressa, ZD1839; AstraZeneca, London, United Kingdom) is an orally active, selective EGFR/TKI that causes complete inhibition of EGF-stimulated EGFR autophosphorylation in cell lines at submicromolar concentrations (IC 50 = 0.02-0.08Amol/L; ref. 10).In preclinical studies, gefitinib showed antitumor activity in a variety of human cancer cell lines expressing EGFR, including ovarian, breast, and colon, and it was active in a range of xenograft models, including colon, NSCLC, and prostate (10,11).In human xenograft models, gefitinib, like other EGFR inhibitors, in combination with standard cytotoxic agents caused both delayed tumor growth and tumor regression irrespective of the level of EGFR (12).To date, two phase II trials of gefitinib for lung cancer, one done in the United States and one done in Japan, have shown that EGFR inhibition effectively induced tumor shrinkage only in patients bearing EGFR gain-of-function mutations (13,14).Other studies have shown that tumor cells may acquire resistance to anti-EGFR therapies without altering EGFR expression but rather through up-regulation and activation of other proliferative and/or antiapoptotic activities: G-coupled protein receptors (15)(16)(17), insulin-like growth factor receptor-I (IGFR-I), and downstream signal transduction through the phosphatidylinositol 3-kinase/Akt pathway and extracellular signal-regulated kinase 1/2 (16,18,19), or PDGFR (20,21).
Glycogen synthase kinase-3h (GSK-3h) is a serine/threonine kinase that plays a crucial role in mammalian development by regulating the Wnt signaling pathway (22).GSK-3h is also a critical component in several receptor-coupled signaling pathways (23), including those emanating from growth factor-stimulated receptors that activate the intermediary protein kinase Akt or ribosomal S6 kinase (RSK), which in turn phosphorylates and inhibits GSK-3h and other signaling pathways (23)(24)(25)(26)(27). Recent discoveries suggest that receptor activation of Ras promotes the accumulation of active cyclin D1/cyclin-dependent kinase 4 (cdk4) complexes via at least two pathways.First, activated Ras promotes transcription of cyclin D1 through a kinase cascade involving Raf1/MAPK kinase/ RSK (24,28,29).Second, the rate of cyclin D1 proteasomal degradation is mediated by GSK-3h-dependent phosphorylation of a single threonine residue (Thr 286 ) near the COOH terminus of cyclin D1 (30).Mitogens, such as EGF, inactivate GSK-3h via a pathway involving Ras/phosphatidylinositol-3 kinase (PI3K)/protein kinase B/Akt or Ras/MAPK/RSK (31,32).Furthermore, alterations in the subcellular distribution of cyclin D1 during the cell cycle may also regulate cyclin D1/cdk4 function.Thus, cyclin D1 accumulates in the nucleus throughout G 1 phase, but it relocalizes to the cytoplasm during the remainder of interphase (33).Cyclin D1 redistribution and its degradation are correlated with its phosphorylation on Thr 286 by GSK-3h (34).
Through studies done on a panel of 10 bladder cancer cells, we found that EGFR blockade suppresses cell proliferation in a subgroup of cell lines in a manner that was not tightly linked to the receptor's expression levels.In fact, the gefitinib-specific target EGFR was effectively inhibited in all of the cell lines tested, but the antiproliferative effect was obtained only in some cell lines, suggesting that activation of specific downstream pathways was crucial to maintaining the proliferation and survival of neoplastic bladder cells.Based on this novel observation, we began to explore the activity of other kinase pathways that may be relevant to the biological effect exerted by gefitinib.We hypothesized that specific signal coupling between the EGFR and downstream kinases would dictate the biological response obtained when the cells are treated with an EGFR inhibitor.In this study, we provide evidence that in vitro gefitinib inhibition of the MAPK pathway is associated with GSK-3h activation, cyclin D1 degradation, and cell cycle inhibition in G 0 -G 1 .We also show that EGFR-dependent GSK-3h regulation may predict the cytostatic effect of gefitinib on bladder tumor cells xenografted in nude mice.
Immunoprecipitation and Western blot analysis.EGF-stimulated and nonstimulated cells were treated with gefitinib for 1 hour.Cells were harvested at f75% to 80% and lysed, and protein concentration was assayed by the Bio-Rad protein assay reagent (Bio-Rad Laboratories, Hercules, CA).For immunoprecipitation analysis, anti-phosphotyrosine antibody was used according to a standard protocol, as previously reported (36).For Western blot analysis, protein samples were boiled in sample buffer [62.5 mmol/L Tris-HCl (pH 6.8), 10% (w/v) glycerol, 100 mmol/L DTT, 2.3% SDS, 0.002% bromophenol blue] for 5 minutes and cooled on ice for 5 minutes.Samples were loaded, and separated on 10% SDS-PAGE at 120 V in electrophoresis buffer [25 mmol/L Tris-HCl (pH 8.3), 192 mmol/L glycine, 0.1% SDS].Proteins in the gels were electrophoretically transferred onto polyvinylidene difluoride membrane in transfer buffer (25 mmol/L Tris-HCl, 192 mmol/L glycine, 20% methanol) at 30 mV overnight at 4jC.The membranes were washed in blocking buffer [TBS: 10 mmol/L Tris-HCl (pH 8.0), 150 mmol/L NaCl], with 4% bovine serum albumin or with 5% nonfat milk for 2 hours at room temperature with shaking and then rinsed once briefly with TBS (99.9% TBS, 0.1% Tween 20).The membranes were incubated with 1:500 diluted primary antibodies overnight then with diluted (1:3,000) second antibodies [anti-mouse or anti-rabbit immunoglobulin, horseradish peroxidase-linked F(ab)2 fragment from mouse] for 1 hour at room temperature with shaking.The probed proteins were detected using the enhanced chemiluminescence system (Amersham Biosciences, Piscataway, NJ) according to the manufacturer's instructions.
Cell proliferation assay.Cells (5 Â 10 3 ) were plated in 96-well plates for 24 to 48 hours and then treated with or without gefitinib at increasing concentrations in EGF-stimulated and nonstimulated environments for 48 to 72 hours.A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was done to determine the relative cell numbers based on the conversion of MTT to formazan in viable cells.MTT (40 Ag/mL) was added to each well, and the cells were incubated for 2 hours.The media were removed, and 100 AL of DMSO were added to lyse the cells and solubilize the formazan.A standard microplate reader was used to determine the absorbance.Each experimental data point represented average values obtained from four replicates, and each experiment was done at least twice.We also measured the effects of gefitinib on DNA synthesis by pulse labeling cells with [ 3 H]thymidine.Cells were plated in 96-well plates as described above and exposed to various treatment conditions for 24 hours.Medium was removed and replaced with fresh medium containing 5 ACi/mL [ 3 H]thymidine (MP Biomedicals, Irvine, CA).Cells were pulsed with [ 3 H]thymidine for 1 hour and lysed by addition of 100 AL of 0.1 mol/L NaKOH.Cells were harvested onto fiberglass filters, and incorporated tritium was quantified in a h-counter.For rescue experiments, cells were cultured in 96-well plates and transfected with SA-GSK-3h and wt-GSK-3h (37) or T286T-CyD1 and wt-CyD1 (34) using FUGENE 6/DNA (Roche Applied Sciences, Indianapolis, IN) in a 3:1 ratio for 6 hours before overnight treatment with gefitinib.
Boyden chamber invasion assays.Polyvinylprolidone-free polycarbonate filters (8-Am pore size; Nucleopore, Becton Dickinson Labware, Franklin Lakes, NJ) were coated with a mixture of basement membrane components (Matrigel, 25 Ag/filter) and placed in modified Boyden chambers.The cells (50 Â 10 5 ) were released from their tissue culture flask by EDTA (1 mmol/L), centrifuged, resuspended in 0.1% DMEM, and placed in the upper compartment of the Boyden chamber.Fibroblast-conditioned medium was placed in the lower compartment as a source of chemoattractants.After incubation for 24 hours at 37jC, noninvading cells on the upper surface of the filter were removed.The cells on the lower surface of the filter were stained with Diff-Quick (American Scientific Products, McGaw Park, IL).
Invasive activity was measured by counting the cells that had migrated to the lower side of the filter.
In vivo assays.Male nude mice, 4 to 6 weeks old, were purchased from Charles Rivers Laboratories, Inc. (Wilmington, MA).The research protocol was approved, and mice were maintained according to the M.D. Anderson Cancer Center's animal care guidelines.Mice were acclimatized at our institution for 2 weeks before being injected with cancer cells.Mice were injected in the subcutis (10 6 cancer cells per injection) with four bladder carcinoma cell lines (253J B-V, UM-UC3, KU7, and UM-UC13), which were suspended with 200 AL of Matrigel (BD Biosciences Co., Franklin, NJ).With each cell line, two groups of animals were assigned: a treatment arm and a placebo arm.After 1 to 2 weeks (when tumors reached 4-5 mm in diameter), six mice (12 tumors) per group were treated i.p. on days 1 to 5 each week with gefitinib 2 mg/dose/mouse in the treatment arm, and an equal number of mice were treated with placebo (Hank's solution) in the control arm for a total of 3 weeks.Tumors were harvested at the end of treatment, and the weights were measured.
Immunofluorescence, confocal analysis, and ELISA.Cellular localization of EGFR phosphorylation and filamentous-actin (F-actin) was determined using indirect immunofluorescence as described (36).Briefly, cells grown on glass coverslips were fixed (without permeabilization) in 3.7% paraformaldehyde at room temperature for 10 minutes and then extracted with ice-cold acetone.Cells were treated with or without anti-EGFR rabbit polyclonal antibody and then treated with Alexa-488-labeled goat anti-rabbit antibody and Alexa 546-labeled phalloidin for actin fibers (Molecular Probes, Inc., Eugene, OR).For Y216 GSK-3h staining, or Flagtagged cyclin D1 (34), we used a similar protocol.Confocal analysis was carried out using a Zeiss laser-scanning confocal microscope and Figure 1.Protein expression or activation of EGFR does not alter sensitivity to gefitinib in TCC cells.A, protein expression of three EGFR family members, EGFR, HER2, and HER4 in 10 TCC lines.Total protein levels were determined by Western blot analysis using specific antibodies.Cell lysates (150 Ag) derived from cells cultured in 5% FCS supplemented DMEM were analyzed for each lane.Baseline activation of EGFR was determined by Western blot analysis using specific anti-Y1068-phosphorylated residue of EGFR.B, comparison of EGFR protein expression levels and its autophosphorylation levels with or without EGF in the absence or presence of various doses of gefitinib.Protein extracts were resolved by 7.5% SDS-PAGE and probed with either antibody.Immunoreactive proteins were visualized by enhanced chemiluminescence.C, cytostasis and invasion are modulated differently by gefitinib.Fluorescence-based cell cycle sort analysis using propidium iodide staining of two bladder cancer cell lines treated with gefitinib for 24 hours.Representative results of three different experiments done on duplicate dishes.D, numeric values for percentage of cells in each cell cycle phase.Note the absence of sub-G 0 -G 1 cells in the presence of gefitinib.Immunofluorescence staining of (E ) 253J B-V and (F ) of UM-UC13 bladder cancer cell lines for phosphorylated EGFR and cytoskeleton protein F-actin.Active EGFR staining with anti-Y1068 phosphorylated antibody followed by FITC-labeled goat anti-rabbit IgG (green ).F-actin staining with Alexa 546-labeled phalloidin (red).pEGFR appeared on highly dynamic F-actin-containing structures and was partially internalized upon EGF addition.Arrowhead, Y1068/EGFR-negative cell lacking F-actin dynamic structures after gefitinib treatment in UM-UC13.
established methods, involving processing of the same section for each detector (two excitations corresponding to 546 and 488, or 633 for the Flag staining and 488 for F-actin staining) and comparing images pixel by pixel.Colocalization of the two proteins (F-actin and pEGFR) is indicated by the presence of yellow color as a result of overlapping red and green pixels.Quantification of the inactive pools of GSK-3h was done using an ELISA kit (Abcam, Cambridge, MA) according to the manufacturer's instructions.
Cell cycle analysis.Cells were grown in six-well plates in the presence of 10% DMEM.After reaching 70% confluence (within 24 hours), the cells were exposed to various concentrations of gefitinib for 48 to 72 hours.Cells were harvested by trypsinization and pelleted by centrifugation.The pellets were then resuspended in PBS containing 50 Ag/mL propidium iodide, 0.1% Triton X-100, and 0.1% sodium citrate.Propidium iodide fluorescence was measured by fluorescence-activated cell sorting analysis (FL-3 channel, Becton Dickinson, Mountain View, CA).Cells displaying a hypodiploid content of DNA indicative of DNA fragmentation were scored as apoptotic.
Small interfering RNA transfection, cell extracts, and immunoblotting analysis.Transfections of the small interfering RNA (siRNA)-targeting endogenous EGFR and unrelated target were done using oligofectamine (Invitrogen Corp., Carlsbad, CA).A cocktail of four custom-designed siRNA duplex oligomers were purchased from Dharmacon Research.Transient transfection using DNA vectors (Flag/pFlex/CyD1 and Flag/pFlex/T286) were done with FUGENE 6 (Roche Applied Sciences) before biochemical assays.Total protein lysate was collected 48 hours after transfection (unless indicated otherwise) and analyzed by Western blotting.To prepare cell extracts, cells were washed thrice with PBS and then lysed with radioimmunoprecipitation assay buffer [50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 0.5% NP40, 0.1% SDS, 0.1% sodium deoxycholate, protease inhibitor cocktail (Roche Applied Science), and 1 mmol/L sodium orthovanadate] for 20 minutes on ice.

Results
Cytostatic effects of gefitinib do not correlate with epidermal growth factor receptor expression or activation.We characterized a panel of 10 human urothelial carcinoma cell lines for EGFR, HER2, and HER4 expression, as well as pEGFR, and found that all the cell lines tested expressed various levels of HER2 and EGFR protein and that some cell lines displayed higher levels of activated EGFR (Fig. 1A).
It has been reported that the sensitivity to EGFR inhibitors correlates with the level of EGFR expression or activation in certain human cancer cell lines (38).To determine whether this was true of human urothelial carcinoma cells, we tested the antiproliferative effect of the inhibitor by thymidine incorporation (Table 1).Cells were treated with gefitinib at concentrations ranging from 0.01 to 5 Amol/L for 24 hours.Sensitivity to gefitinib was defined as a z50% inhibition of cell growth at a concentration of V1 Amol/L.Of the 10 cell lines tested, four lines (UM-UC5, 253J P, 253J B-V, and UM-UC1) were most sensitive to the antiproliferative effect of gefitinib, with an IC 50 < 0.5 Amol/L; whereas five lines were relatively resistant to gefitinib (KU7, UM-UC14, UM-UC13, UM-UC3, and RT4) based on the criteria specified before.Because KU7, 253J P, RT4, UM-UC13, and UM-UC14 express similar amounts of EGFR as tested by Western blot but displayed differences in gefitinib sensitivity, we concluded that expression of EGFR does not predict sensitivity to the antiproliferative effects of gefitinib.A high baseline activation of the EGFR did not seem to be a predictor of gefitinib sensitivity, because UM-UC5 and UM-UC6, which displayed the highest activation, have also different response patterns.Lower levels of EGFR autophosphorylation were also detectable in 253J B-V cells, as well as in UM-UC13 cells, which differ significantly in their antiproliferative response to gefitinib.Thus, we further characterized potential mechanisms of resistance to EGFR blockade in UM-UC13.The achievable levels of EGFR autophosphorylation upon ligand addition was very similar in both cell lines (Fig. 1B) when normalized per unit of total EGFR, suggesting that receptor sensitivity to EGF does not cause gefitinib resistance.Furthermore, the addition of 1 Amol/L gefitinib effectively blocked EGF-induced phosphorylation in both cell lines (Fig. 1B).The filters were further reprobed with an anti-EGFR for gel loading control (Fig. 1B, bottom).
Cytostatic effect of gefitinib does not parallel inhibition of cell invasion.To see if gefitinib exerts a uniquely cytostatic effect on 253J B-V cells, we examined cell cycle changes in 253J B-V and UM-UC13 following treatment with gefitinib at concentrations up to 1 Amol/L (Fig. 1C).The percentage of 253J B-V cells in the S phase decreased from 35.6% to 23.7% after treatment with 1 Amol/L gefitinib, whereas the percentage of cells in the sub-G 0 -G 1 population was near to zero, indicating that gefitinib has a cytostatic effect on these cells (Fig. 1C and D, top).As predicted by the results of the thymidine incorporation assay, gefitinib did not reduce S phase or change other cell cycle phases in UM-UC13 cells (Fig. 1C and D, bottom).
Because recent reports have indicated that internalization of EGFR following treatment with an inhibitor predicts for the sensitivity of NSCLC to gefitinib (39), we determined the subcellular localization of the activated EGFR in 253J B-V and UM-UC13 cells by immunofluorescence microscopy following gefitinib treatment.Confocal analysis of tumor cells double-stained for autophosphorylated EGFR (green) and F-actin (red) confirmed the existence of baseline and EGF-inducible EGFR autophosphorylation (Fig. 1E and F, top, respectively).These studies also revealed that an important pool of the activated EGFR was localized on the dynamic F-actin containing structures, such as ruffles, as they appeared in yellow, especially in UM-UC13 (Fig. 1E and F, top).Finally, EGF-induced EGFR internalization in a pattern resembling endosomal localization was documented in both cell lines (Fig. 1E  and F).We observed complete inhibition of baseline and EGFinduced EGFR phosphorylation coupled with the loss of EGFinduced spindled cell morphology in 253J B-V cells following treatment with gefitinib (Fig. 1E).A similar analysis of UM-UC13 cells revealed greater baseline EGFR autophosphorylation together with the existence of highly dynamic F-actin-containing ruffles and fillopodias.Gefitinib inhibited significant EGFR activation and cellular ruffling in these cells as well (Fig. 1F).Interestingly, when UM-UC13 cells were pretreated with gefitinib before EGF stimulation, we could still detect autophosphorylated EGFR in f70% of the cells; in this case, however, the active EGFR was strictly localized to the cellular focal adhesion points, and cell ruffling was significantly diminished (Fig. 1F, bottom right, arrowhead).
To determine whether these phenotypic changes reflected changes in cellular motility, we measured the migration of 253J B-V and UM-UC13 cells through Matrigel membranes in Boyden chambers following gefitinib treatment (Fig. 2A).We observed that baseline invasion by UM-UC13 was 5-fold higher than that of 253J B-V (Fig. 2A, white columns).Interestingly, invasion of both cell lines was inhibited by gefitinib, although the efficiency was cell type dependent (Fig. 2A, black columns).These results emphasize the complexity of predicting the response to gefitinib and suggest that different downstream signaling pathways are preferentially affected in different tumor cells following exposure to the drug.This point is illustrated by our observation that although gefitinib efficiently inhibited EGFR phosphorylation, EGF-induced EGFR internalization, and invasion through Matrigel in both 253J B-V and UM-UC13, it failed to inhibit the proliferation of UM-UC13.
Uncoupling between epidermal growth factor receptor and downstream pathways.We next investigated the pattern of EGFinduced activation of PI3K/Akt and of Ras/MAPK pathways in both cell lines following gefitinib treatment.The addition of 1 Amol/L of Exponentially growing cancer cells in 10% FCS-containing DMEM were serum starved for 24 hours and treated with EGF for 30 minutes with or without pretreatment of various gefitinib concentrations for another 30 minutes.Protein extracts were resolved by 10% SDS-PAGE and probed with either antibody.PDK1, Akt, and extracellular signal-regulated kinase 1/2 activity were determined using corresponding anti-phosphorylation antibodies.B, 253J-BV cells; C, UM-UC13 cells.Bottom, MAPK and Akt phosphorylation was quantified by densitometry and expressed as a ratio of the total protein, respectively.D, siRNA-targeted EGFR does not impair extracellular signal-regulated kinase 1/2 baseline activation in UM-UC13 cells.UM-UC13 cells untransfected or transfected with oligofectamine and a pool of four RNA oligos (SMART pool) designed against human EGFR mRNA or unrelated/ scrambled RNA oligos (SCR ).Left, cellular extracts prepared 24, 48, and 72 hours after transfection were immunoblotted and probed with anti-EGFR and anti-HER2 antibodies, with the latter serving as a loading control.Right, cell extracts from 72 hours after transfection were Western blotted and probed against pEGFR, phosphorylated extracellular signal-regulated kinase 1/2, or the total EGFR and total extracellular signal-regulated kinase 1/2 proteins.
gefitinib effectively blocked PDK1 phosphorylation in 253J B-V but not in UM-UC13 cells and was associated with the inhibition of its substrate, Akt phosphorylation (Fig. 2B, top and bottom, black columns).However, EGF had a moderate stimulatory effect on Akt/ Ser 473 phosphorylation, and 1 Amol/L gefitinib could block its increase in UM-UC13 cells (Fig. 2C, top and bottom, black columns).Interestingly, the addition of as little as 0.1 Amol/L gefitinib almost completely blocked the MAPK pathway in 253J B-V cells (Fig. 2B, top and bottom, white columns), although as much as 1 Amol/L of the inhibitor had no effect on UM-UC13 cells (Fig. 2C, top and bottom, white columns).
Because the uncoupling between EGFR and MAPK activation in UM-UC13 cells could be the result of the unique presence of EGFR and thus serve as scaffold protein for other signaling molecules, we next decided to down-regulate EGFR expression in UM-UC13.Down-regulation of EGFR protein via siRNA-mediated knockdown (Fig. 2D, left) had no effect on MAPK phosphorylation (Fig. 2D, right), confirming that EGFR activation was uncoupled with the MAPK activity in UM-UC13 cells and suggesting that other signals or growth factor receptor(s) may be responsible for MAPK activation and cell proliferation in these cells.
Glycogen synthase kinase-3B activation by gefitinib associates with cytostasis.Our observation that the EGFR/Ras/MAPK and EGFR/PDK1/Akt axes are intact in 253J B-V cells, which also respond to gefitinib, is in agreement with recent findings by Dominguez-Escrig et al. (40).The uncoupling between Ras/MAPKs and EGFR activation in the resistant UM-UC13 cell line suggests that the ability of EGFR inhibitors to exert their antiproliferative effect was dependent upon the balance between the activation status of MAPK and Akt, which could be reflected by the activation Figure 3. Gefitinib-induced cytostasis is mediated by GSK-3h activation.A, dose-dependent inhibition of EGF-induced GSK-3h phosphorylation on Ser 9 residue (inactivation site).Serum-starved 253J B-V cells (top ) and UM-UC13 cells (bottom ) were treated for 30 minutes with the indicated concentrations of gefitinib followed by the addition of EGF (50 ng/mL) for 4 hours.Protein extracts were resolved by 10% SDS-PAGE and probed with anti-phosphorylated GSK-3h or with anti-h actin antibody, with the latter serving as the loading control.B, immunofluorescence staining of 253J B-V and UM-UC13 bladder cancer cells for phosphorylated GSK-3h on Y216 residue (activation site).A-B, active GSK-3h staining followed by FITC-labeled goat anti-rabbit IgG (green ).Arrowheads, nuclear translocation of active GSK-3h in gefitinib-treated 235J B-V cells.C, signaling pathways regulating GSK-3h activity.Serum-starved 253J B-V cells (white columns ), or UM-UC13 cells (black columns ) were treated with EGF for 30 minutes with or without pretreatment of gefitinib, MAPK inhibitor PD098059, PI3K inhibitor LY294002, or p38MAPK SB239063.The amount of inactive GSK-3h was determined from total protein extracts using the ELISA method.Normalized by the total protein content using the Bradford method.D, GSK-3h-dependent cytostasis of UM-UC13 cells.Cell cycle sensitivity to exogenous active GSK-3h was determined by changes in thymidine incorporation of cells after 24 hours after transfection in the absence or presence of various doses of gefitinib.Cell cycle inhibition was most effective when dominant active GSK-3h construct was expressed compared with vector control.Columns, averages of triplicate values; bars, SD.E, GSK-3h-dependent cytostasis of two TCC cells lines.Cell cycle sensitivity to GSK-3h activation induced by gefitinib was determined by changes in thymidine incorporation of cells after 24 hours of EGF treatment in the absence or presence of various doses of gefitinib.Cell cycle inhibitory effect was rescued by GSK-3h inhibition using ODZT pretreatment in 253J B-V (white columns ) compared with UM-UC13 (black columns).Columns, averages of triplicate values; bars, SD.
of a common downstream kinase, such as GSK-3h.EGFR signaling through MAPK/RSK or PI3K/Akt inactivates this kinase by phosphorylating its Ser 9 residue, the predominant pathway that modulates GSK-3h phosphorylation being dictated by the specific cellular environment (24,30).
To determine whether the inhibition of EGFR phosphorylation altered the balance between active and inactive GSK-3h, we quantified by Western blot analysis the level of its inactive pool in 253J B-V and UM-UC13 cells in the presence or absence of gefitinib (Fig. 3A).We observed a significant reduction in the inactive form of GSK-3h in 253J B-V cells following treatment with 0.5 Amol/L gefitinib (Fig. 3A, top), whereas the addition of the inhibitor failed to reduce the size of inactive pool of GSK-3h in UM-UC13 (Fig. 3A, bottom).Because phosphorylation on the Y216 residue of GSK-3h is responsible for the increased activity of the kinase and because activation is associated with specific target recognition in welldefined cellular locations (25), we next determined the subcellular localization of the active pool of Y216-GSK-3h using laser-scanning microscopy (Fig. 3B).Confocal analysis showed intranuclear accumulation Y216-phosphorylated GSK-3h in 253J B-V cells after gefitinib treatment (Fig. 3B, top right, arrowhead), whereas no active kinase was detected in UM-UC13 cells (Fig. 3B, bottom right).To determine whether the amount of inactive GSK-3h could be reduced by blocking the EGFR or components of its downstream signaling pathways (26), we treated 253J B-V and UM-UC13 with gefitinib or specific chemical inhibitors to the MAPK, PI3K, or p38MAPK signaling pathways in the presence of EGF stimulation and quantified the level of inactive GSK-3h by ELISA (Fig. 3C).Gefitinib and inhibitors of the MAPK (PD098059) and PI3K (LY294002) pathways effectively reduced the inactive pools of GSK-3h in 253J B-V cells (Fig. 3C, left).Although gefitinib did not reduce the levels of inactive GSK-3h in UM-UC13 cells, inhibitors of the MAPK, PI3K, or p38MAPK signaling pathways resulted in the efficient reduction of inactive GSK-3h in this cell line (Fig. 3C, right).These results suggest that modulation of GSK-3h activity reflects the uncoupling between the EGFR and its downstream signaling pathways.To further show that DNA synthesis can be reduced in the presence of active GSK-3h in cells that were resistant to gefitinib, we did tritium-labeled thymidine incorporation assays using 1 Amol/L gefitinib in UM-UC13 cells transfected with constitutively active GSK-3h construct or with control vector (Fig. 3D).Increasing the active pool of GSK-3h in UM-UC13 resulted in 40% reduction of DNA synthesis, and addition of gefitinib further increased this inhibition to only another 10%, suggesting that the balance between active and inactive pool of this kinase is an important factor for cell cycle progression in these cells.
To show that the cytostatic effect of gefitinib relies on a GSK-3hdependent mechanism in 253J B-V cells, we did tritium-labeled thymidine incorporation assays using different concentrations of gefitinib in the presence or absence of a specific GSK-3h inhibitor (OTDZT; ref. 24).We found that 1 Amol/L gefitinib inhibited 65% of EGF-stimulated thymidine incorporation, which was rescued in a dose-dependent manner by the kinase inhibitor (Fig. 3E, white columns), suggesting that cell cycle inhibition may be a GSK-related mechanism in 235J B-V cells.As expected, no effect of the EGFR inhibitor on thymidine incorporation could be observed in UM-UC13, although a slight rescue by the addition of the inhibitor could be detected, suggesting that low levels of active GSK-3h might be present in these cells and that the release from GSK-3h activation may further enhance cellular proliferation (Fig. 3E, black columns).
Cyclin D1 degradation by glycogen synthase kinase-3B regulation is responsible for the antiproliferative effect of gefitinib.Because active GSK-3h can directly bind, phosphorylate, and degrade cyclin D1 (34), we investigated whether increased GSK-3h activation was associated with decreased cyclin D1.Significant reduction of cyclin D1 could be detected after 1 Amol/L gefitinib treatment in 253J B-V cells (Fig. 4A, left), whereas no change in its levels was observed in UM-UC13 cells (Fig. 4A, right).To further show that cyclin D1 regulation represents a GSK-3h-dependent event, we measured DNA synthesis of transfected 253J B-V cells with wild-type (WT) and mutant (T286) cyclin D1 constructs (34) in the presence or absence of gefitinib (Fig. 4B).Addition of a GSK-3h nondegradable cyclin D1 resulted in an efficient rescue of the DNA synthesis from 50% inhibition to a 25% inhibition in the presence of 0.5 Amol/L gefitinib.Because the addition of WT cyclin D1 was not as efficient as the mutant T286cyclin D1, we did confocal microscopy analysis of transfected cells stained for Flag-tag to visualize the subcellular localization of the exogenous cyclin D1 protein (blue).These experiments revealed that only the mutant form was efficiently translocated to the nucleus in the presence of gefitinib (Fig. 4C, arrowheads), whereas the WT cyclin D1 was frequently localized into the cytosol (Fig. 4C, arrows).Similar results were obtained using other gefitinibsensitive cell lines, such as 253J P or UM-UC5 (data not shown).
Platelet-derived growth factor receptor short circuits epidermal growth factor receptor to induce mitogen-activated protein kinase and cell proliferation in gefitinib-resistant cells.Because EGFR apparently does not have a major role in maintaining MAPK activation in UM-UC13 (EGF could not stimulate it further, siEGFR had no effect of basal MAPK phosphorylation, and gefitinib could not block its activation), we next hypothesized that other tyrosine kinase receptors, such as IGFR, PDGFR, or c-kit, might be responsible for the EGFR/MAPK uncoupling in these cells.Western blot analysis revealed that PDGFRh was undetectable or expressed at very low levels in all four gefitinib-sensitive cell lines (253J P, 253J B-V, UM-UC5, and UM-UC1) and expressed at higher levels in all resistant cell lines (Fig. 5A).None of the resistant cell lines tested expressed c-kit (negative data, not shown), and IGFR-I expression pattern did not unveil any correlation to gefitinib-sensitivity in the cells tested (Fig. 5A).
To further show that MAPK phosphorylation reflects PDGFRh activation in UM-UC13 cells, we treated both cell lines with EGF or PDGF, with or without gefitinib and PDGFR inhibitor, and did Western blot analysis for PDGFR and MAPK phosphorylation (Fig. 5B).Interestingly, both EGF and PDGF addition only slightly increased the baseline PDGFR phosphorylation in UM-UC13 (Fig. 5B, left), whereas addition of PDGFRh inhibitor significantly reduced the phosphorylation levels of both the PDGFR and MAPK below the baseline levels (Fig. 5B, right).As expected, neither PDGFRh phosphorylation nor MAPK regulation by PDGFR ligand or its inhibitor was detected in gefitinib-sensitive cell lines (Fig. 5B, left).In addition, PDGFRh inhibition was associated with dephosphorylation of GSK-3h in UM-UC13, whereas gefitinib efficiently reduced the Akt phosphorylation in both cells (Fig. 5C), suggesting that MAPK represents an important upstream regulator for GSK-3h in UM-UC13 cells.Overall, our results suggest that PDGFRh-triggered signals activate MAPK, which inactivates GSK-3h and consequently maintains an increased cyclin D1 level in UM-UC13.The EGFR or PDGFR coupling with the GSK-3h modulation could be shown in other gefitinib-sensitive cells, such as 253J P and UM-UC5 (Fig. 6A), or gefitinib-resistant cells, such as KU7 and UM-UC3 (Fig. 6B).Thus, the quantified data presented in Fig. 6A and B (bottom) shows that although gefitinib reduced significantly the GSK-3h phosphorylation in 253J P and UM-UC5, it had no effect in KU7 and UM-UC3 cells.
To further investigate whether the effects of the receptor TKIs on the downstream GSK-3h are translated into changes of DNA synthesis in cell lines other than 253J B-V and UM-UC13, we measured cell proliferation changes in six more urothelial carcinoma cells (Fig. 6C).Interestingly, a total dependency of the DNA synthesis on the PDGFRh activity was observed only in UM-UC13 and KU7, whereas in UM-UC3 and UM-UC14, simultaneous inhibition of both receptors had growth effects on cell proliferation that were absent following addition of EGFR inhibitor alone.This suggests that the biological functions of EGFR and PDGFR are far from being redundant in bladder cancer cells.
Glycogen synthase kinase-3B regulation as surrogate marker for the cytostatic effect of gefitinib.The association among GSK-3h activation, loss of cyclin D1, and the cytostatic effects of gefitinib suggests that loss of phosphorylated GSK-3h after the first treatment might predict the antiproliferative effects of the inhibitor in bladder tumors.To test this hypothesis, and based on the previous in vitro observations, we s.c.injected four bladder cell lines, one sensitive (253J B-V) and three resistant (UM-UC3, UM-UC13, and KU7), into nude mice.Ten days after injection, we initiated treatment of 0.4 mg/dose/mice gefitinib administered i.p. for 3 weeks, 5 days per week.After 1 week of treatment, we sampled the tumors from each group to measure the levels of Ser 9phosphorylated GSK-3h by Western blot analysis (Fig. 6D, top) and of the phosphorylated (active) MAPK (Fig. 6D, middle).Tumor weights were measured at the end of treatment and expressed as percentiles from control, vehicle-treated tumors (Fig. 6E).Interestingly, the reduction in the inactive pool of GSK-3h between the gefitinib-treated and untreated tumors very closely reflected the effect observed at the end of the treatment (Fig. 6D and E).Thus, the most significant reduction in phosphorylated GSK-3h and phosphorylated MAPK was obtained in the 253J B-V tumors (Fig. 6D, left and E, *), in which gefitinib exerted its most potent .DNA synthesis inhibition is mediated through cyclin D1 regulation by GSK-3h.A, dose-dependent regulation of cyclin D1 protein expression.Serum-starved 253J B-V (left) or UM-UC13 (right ) cells were treated for 30 minutes with the indicated concentrations of gefitinib followed by the addition of EGF (20 ng/mL) for 30 minutes.Protein extracts were resolved by 10% SDS-PAGE and probed with anti-cyclin D1 or anti h-actin, with the latter serving as a loading control.B, impaired cyclin D1 degradation rescues DNA synthesis in gefitinib-treated cells, 253J B-V were transiently transfected with wild-type (wt) or T286 cyclin D1 constructs and treated with gefitinib.DNA synthesis was measured using tritiated thymidine incorporation 24 hours after transfection.The efficiency of transfection was measured through immunohistochemistry using anti-Flag (M2) antibody and estimated between 35% and 50% in 253J B-V cells.C, confocal microscopy showing the subcellular localization of Flag-tagged cyclin D1 constructs transiently transfected into gefitinib-sensitive cells in the presence or absence of 0.5 Amol/L gefitinib.Arrows, cytosolic localization of Flag-tagged protein; arrowheads, nuclear localization of mutant T286-cyclin D1 (T286-cyD1 ) protein.
cytostatic effect (60% difference).The fact that reduction of GSK-3h phosphorylation was reflected in the cytostatic effect of the EGFR inhibitor, even in such a setting, strongly suggests that it represents a potential surrogate marker for gefitinib sensitivity.

Discussion
In this study, we characterized sensitivity of a panel of urothelial carcinoma cell lines to an EGFR TKI with respect to baseline EGFR expression or activation and regulation of downstream pathways.We found that although the target was efficiently blocked in all cell lines, the cell growth inhibitory effects were dependent on regulation of specific downstream kinases.Thus, we found that cell cycle inhibitory effects of gefitinib were associated with GSK-3h activation and cyclin D1 degradation.Using two bladder cancer cell lines, the sensitive 253J B-V cell line and the resistant UM-UC13 cell line, we further investigated the mechanism of gefitinib resistance.We found that PDFGR-initiated signals were responsible for maintaining high levels of active MAPK and an inactive GSK-3h in the resistant cell line, although other biological functions, such as migration and actin remodeling, were affected by the EGFR inhibition.
Initially, EGFR was found to form specific ligand-dependent homodimers or heterodimers with HER2 to HER4, each pair being activated by one of multiple ligands in the EGF and neuregulin families.HER2, the one family member without a true ligand, is the preferred heterodimerization partner of every other family member.By heterodimerizing with them, HER2 can alter their signaling output and duration (41,42).However, blocking HER2 phosphorylation with Herceptin did not have any effect on downstream pathways, such as MAPK in UM-UC13 (negative data, not shown), which ruled out its participation in the gefitinib resistance mechanism.Recent reports have shown that EGFR transactivation can be detected using a variety of G proteincoupled receptor agonists, phorbol esters, cytokines, chemokines, estrogen, and cell stress signals, making this receptor a central player in cellular responses (43,44).When these alternative pathways phosphorylate EGFR on tyrosine residues, EGFR behaves in a manner indistinguishable from that of activation by the exogenous addition of EGF family ligands.Moreover, EGFR transactivation has also been shown through other receptor tyrosine kinases, including IGFR-I and the PDGFRh.The fact that the EGF addition induced a slight PDGFR transactivation in UM-UC13 suggests that PDGFR/EGFR crosstalk may be a possible mechanism for diverse stimuli that feed into associated mitogenic or migration/invasion pathways, although additional data is required.Interestingly, we were able to identify the presence of an autocrine loop for PDGF-CC in UM-UC13 (data not shown), which may explain the basal activity of PDGFRh and the insensitivity to PDGF-BB addition, as shown in Fig. 5C.We were able to document heterodimer formation between EGFR and PDGFRh in UM-UC13 (data not shown); however, the presence of heterodimers in the other cell lines and their biological roles remain to be further established.This effect, termed ''transmodulation, '' was described a decade ago and seems to serve in decreasing the binding affinity of EGF to EGFR (45)(46)(47).
The first demonstration of functional EGFR/PDGFR cooperation was observed using murine B82L fibroblasts (48).The authors showed that enhanced motility correlated with PDGF-stimulated EGFR tyrosine phosphorylation and was prevented by expression of a catalytically inactive or truncated EGFR.In our system, the catalytic activity of EGFR was significantly blocked by the gefitinib addition, and this was translated into significant inhibition of the migratory phenotype of these cells in association with cytoskeleton remodeling.Interestingly, recent studies have shown that selective inhibition of the EGFR with a chemical inhibitor (AG1478) or the use of EGFR-deficient cells abolished the activation of p21activated kinase (PAK), indicating that PAK may be responsible for PDGF-dependent, EGFR-induced cell motility (48).Moreover, because the EGFR contains a specific actin-binding motif not found in the PDGFR (49), it has been proposed that coactivation of EGFR-containing heterodimeric receptors may enhance the ability of PDGF to mediate cell motility through modulation of actin dynamics that are necessary for cell migration (50).Thus, it is possible that EGFR transactivation by such a mechanism short circuits the need for exogenous EGF ligands.Similarly, in our system, the EGFR in UM-UC13 has been short circuited for the mitogenic function by the PDGFR, although the actin dynamics and migration still remained dependent on a kinase-active EGFR.It would be interesting to find out whether other biological functions, such as survival, which are relevant for sensitization to conventional chemotherapy in bladder cancer, are still dependent on the EGFR kinase activity or are bypassed by other stimuli.
Activated or overexpressed GSK-3h has been shown to promote apoptotic signaling in a number of conditions through a p53dependent mechanism (51), and several apoptotic stimuli were recently found to cause the accumulation of GSK-3h in the nucleus, colocalizing it with p53 (51) and contributing to p53-mediated p21/ Waf1/Cip1 induction and caspase-3 activation (52).In our studies, although the nuclear localization of active GSK-3h did not induce apoptosis in 253J B-V bladder cancer cells, it was associated with degradation of cyclin D1, a driving force for cell cycle progression.During G 0 -G 1 phase, the D-type cyclins (D1, D2, and D3) accumulate and assemble with either cdk4 or cdk6 in response to mitogenic growth factors.The active cyclin D1 holoenzyme promotes G 1 progression by inactivating the growth-suppressive properties of the retinoblastoma protein (Rb) through site-specific phosphorylation and by virtue of its ability to titrate cdk inhibitors, such as p27 Kip-1 and p21 Cip1 (32,53).Titration of 27 Kip1 and p21 Cip1 , in turn, facilitates activation of the cyclin E/cdk2 complex and subsequent entry and progression through the DNA synthetic (S) phase of the cell cycle.Although p27 Kip1 and p21 Cip1 are effective inhibitors of cyclin E/cdk2 and cyclin A/cdk2 complexes, recent evidence shows that they promote the assembly of cyclin D/cdk complexes (32) and are found in catalytically active cyclin D/cdk complexes in vivo (53).Extracellular mitogens promote cellular proliferation via receptor-mediated signaling, in our case, EGFR, PDGFR, or both.These signals ultimately converge on the cell cycle machinery, in which cyclin D may function as a critical sensor of this information.Based on recent observations, increased expression of cyclin D1 is not sufficient for cellular transformation (54).However, increased expression of GSK-3h nondegradable cyclin D1 has been shown to permanently reside within the nucleus and to have transforming properties (54).Our studies showed that GSK-3h activation using the EGFR or PDGFR inhibitor is a necessary event for cell cycle inhibition in vitro and suggest a mechanism that involves cyclin D1 regulation.Initiation of DNA replication at the G 1 -S phase boundary is a meticulously regulated process that ensures that the cell has assembled the appropriate machinery needed for high-fidelity genome duplication.Before S-phase entry, pre-replication complexes form at replication origins during G 1 phase.Furthermore, it has been shown that cyclin D1/cdk4 kinase is incorporated into chromatin-bound protein complexes with the same kinetics as minichromosome maintenance (MCM) proteins and dissociates Rb-MCM7 complexes, thereby facilitating establishment of the pre-replication complexes composed of cyclin D1/MCM7 (55).Because cyclin D1 nuclear export has been shown to correlate with GSK-3h regulation (56), it seems that deregulation of GSK-3h kinase activity through receptor TKI may be closely associated to DNA replication in tumor cells.Using a cyclin D1 in which the Thr 286 residue has been mutated (34), we were able to show that GSK-3h-dependent degradation of cyclin D1 by gefitinib is important for blocking DNA synthesis.The diagram presented in Fig. 6F summarizes the signal transduction events found in sensitive (EGFR dependent) and resistant (PDGFR dependent) bladder cancer cells (Fig. 6F).Thus, our results link the deregulation of EGFR signals to gefitinib resistance and cell cycle arrest through modulation of GSK-3h and cyclin D1 (Fig. 6F).In gefitinib-resistant bladder cell lines, measurement of GSK-3h activation together with nuclear cyclin D1 detection may represent potential surrogate prognostic markers for estimating the cytostatic effect of receptor tyrosine kinase inhibition in bladder cancer.
Because the majority of our cell lines proliferate in an EGFRindependent manner in vitro, by focusing on EGFR alone, we may be overlooking an important therapeutic target in this disease.We identified a subset of EGFR inhibitor-insensitive cell lines that are sensitive to chemical PDGFR inhibitors in vitro.In addition, several of the cell lines tested responded to a combined blockade of the EGFR and PDGFR.Based on our observations and the published literature, it is likely that both EGFR and PDGFR cooperate in maintaining the proliferative state in many tumors.This suggests that greater therapeutic activity might be obtained by using combinations of EGFR and PDGFR antagonists (and possibly antagonists of other receptors); alternatively, it may be necessary to phenotype each tumor with respect to its dependence on EGFR or PDGFR (or other growth factor receptor) signaling and tailor therapy on a patient-by-patient basis, to exploit the biological effects of EGFR (or PDGFR) inhibition in human urothelial carcinoma.A challenge that remains is the development of the methodology to query human tissue to define a set of pharmacodynamic markers that identify growth factor receptor-dependent proliferation.
In conclusion, we propose that as long as the upstream stimulatory signals (i.e., PDGFR and other mitogenic signals) act independently of the EGFR kinase activity and keep GSK-3h in a preponderant inactive form with an active cyclin D1 inside the nucleus, the cytostatic effect of the EGFR inhibitor will be minor or null.Alternatively, inhibition of the MAPK pathway and/or GSK-3h activation will translate into cell arrest with cyclin D1 degradation and impaired DNA replication.

Figure 2 .
Figure 2. Inhibition of cell invasion by gefitinib.A, cells plated in Matrigel on top of migration chambers were allowed to migrate through the filter towards the lower chamber containing 3T3-conditioned medium as chemoattractant in the absence or presence of 1 Amol/L gefitinib.Representative results of three different experiments.Columns, average of triplicate filters; bars, SD.Dose-dependent inhibition by gefitinib of Akt and extracellular signal-regulated kinase 1/2 phosphorylation under EGF-stimulated conditions in two human TCC cell lines.Exponentially growing cancer cells in 10% FCS-containing DMEM were serum starved for 24 hours and treated with EGF for 30 minutes with or without pretreatment of various gefitinib concentrations for another 30 minutes.Protein extracts were resolved by 10% SDS-PAGE and probed with either antibody.PDK1, Akt, and extracellular signal-regulated kinase 1/2 activity were determined using corresponding anti-phosphorylation antibodies.B, 253J-BV cells; C, UM-UC13 cells.Bottom, MAPK and Akt phosphorylation was quantified by densitometry and expressed as a ratio of the total protein, respectively.D, siRNA-targeted EGFR does not impair extracellular signal-regulated kinase 1/2 baseline activation in UM-UC13 cells.UM-UC13 cells untransfected or transfected with oligofectamine and a pool of four RNA oligos (SMART pool) designed against human EGFR mRNA or unrelated/ scrambled RNA oligos (SCR ).Left, cellular extracts prepared 24, 48, and 72 hours after transfection were immunoblotted and probed with anti-EGFR and anti-HER2 antibodies, with the latter serving as a loading control.Right, cell extracts from 72 hours after transfection were Western blotted and probed against pEGFR, phosphorylated extracellular signal-regulated kinase 1/2, or the total EGFR and total extracellular signal-regulated kinase 1/2 proteins.

Figure 4
Figure 4. DNA synthesis inhibition is mediated through cyclin D1 regulation by GSK-3h.A, dose-dependent regulation of cyclin D1 protein expression.Serum-starved 253J B-V (left) or UM-UC13 (right ) cells were treated for 30 minutes with the indicated concentrations of gefitinib followed by the addition of EGF (20 ng/mL) for 30 minutes.Protein extracts were resolved by 10% SDS-PAGE and probed with anti-cyclin D1 or anti h-actin, with the latter serving as a loading control.B, impaired cyclin D1 degradation rescues DNA synthesis in gefitinib-treated cells, 253J B-V were transiently transfected with wild-type (wt) or T286 cyclin D1 constructs and treated with gefitinib.DNA synthesis was measured using tritiated thymidine incorporation 24 hours after transfection.The efficiency of transfection was measured through immunohistochemistry using anti-Flag (M2) antibody and estimated between 35% and 50% in 253J B-V cells.C, confocal microscopy showing the subcellular localization of Flag-tagged cyclin D1 constructs transiently transfected into gefitinib-sensitive cells in the presence or absence of 0.5 Amol/L gefitinib.Arrows, cytosolic localization of Flag-tagged protein; arrowheads, nuclear localization of mutant T286-cyclin D1 (T286-cyD1 ) protein.

Figure 5 .
Figure 5. Presence of active PDGFRh alters gefitinib sensitivity by mocking the EGFR/MAPK for mitogenic signals.A, protein expression of PDGFRh, IGFRI, and c-kit.Total protein levels were determined by Western blot analysis using specific antibodies.B, PDGFRh phosphorylation modulation and associated MAPK phosphorylation in cells treated with either ligands for EGFR or PDGFR or their inhibitors: top left, 253J B-V; top right, UM-UC13; bottom, graphic representation of optical densities for the phosphorylated proteins after normalization using vinculin as loading control.C, EGFR phosphorylation modulation and associated Akt and GSk-3h phosphorylation in cells treated with either ligands for EGFR or PDGFR or their inhibitors: top left, 253J B-V; top right, UM-UC13; bottom, graphic representation of optical densities for the phosphorylated proteins after normalization using vinculin as loading control.

Figure 6 .
Figure 6.Validation of GSK-3h regulation importance for cell proliferation in vitro and in vivo .Western blot analysis for (A ) phosphorylated GSK-3h in gefitinib-sensitive cells, 253J B-V (left) and UM-UC5 (right ) or (B ) gefitinib-resistant cells, UM-UC3 (left) and KU7 (right ) cultured in vitro.C, in vitro EGFR and PDGFR inhibition modulation of DNA synthesis in TCC cells.Columns, averages of three values; bars, SD.D-E, gefitinib-induced tumor growth inhibition of human bladder cancer cell lines injected in nude mice.Four bladder cancer cell lines were injected s.c.into nude mice and treated with 2 mg/dose/mouse gefitinib or vehicle alone for 3 weeks.D, in vivo phosphorylated GSK-3h was determined by Western blot on tumor cell lysates sampled after the first week of treatment.Vinculin served as loading control.E, tumor weights at the end of treatment are expressed as percentiles from control vehicle-treated tumors.Columns, averages of 12 values; bars, SD.P < 0.01.F, schematic representation of the signal transduction events in sensitive and resistant bladder cancer phenotypes.EGFR (gray and blue ) or PDGFR (gray and red) may stimulate MAPK-dependent cell cycle progression by increasing the availability/activity of cyclin D1.Inactivation (red arrows ) of EGFR (in EGFR sensitive) or PDGFR (in EGFR insensitive) bladder cancer cells may lead to GSK-3h activation, cyclin D1 degradation, and G 0 -G 1 cell cycle arrest.k, kinase domains; regulatory phosphorylation sites are marked inside ovals attached to the respective proteins.

Table 1 .
Dose-response activity of Gefitinib on DNA synthesis