Interleukin 8 (IL-8) and vascular endothelial growth factor (VEGF)promote tumor angiogenesis, growth, and metastasis and are coexpressed by human head and neck squamous cell carcinomas (HNSCCs) and a variety of other cancers. The promoters of the IL-8 and VEGF genes contain different recognition sites for transcription factors nuclear factor (NF)-κB and activator protein-1 (AP-1), which we showed previously are coactivated in HNSCCs. NF-κB and AP-1 may be modulated by the inhibitor κB kinase (IKK)and mitogen-activated protein kinase (MAPK) signal pathways, but the contribution of these pathways to expression of IL-8 and VEGF and as potential targets for antiangiogenesis therapy in HNSCC is not known. In this study, we examined the effects of modulation of the MAPK and IKK pathways on expression of IL-8 and VEGF by UM-SCC-9 and UM-SCC-11B cell lines. Interruption of IKK-mediated activation of NF-κB by expression of an inhibitor κBα mutant (IκBαM) in UM-SCC-9 cells resulted in partial inhibition of expression of IL-8 but not VEGF. Analysis of possible alternative pathways for induction of these genes revealed activation of the MAPK extracellular signal-regulated kinase(ERK1/2) in cell lines UM-SCC-9 and UM-SCC-11B. Basal and tumor necrosis factor-α-inducible phosphorylation of ERK1/2 and secretion of IL-8 and VEGF could be specifically inhibited by a MEK inhibitor,U0126. Expression of IL-8 and VEGF in the cell lines was associated with coactivation of both NF-κB and AP-1, and U0126 inhibited both NF-κB and AP-1 reporter activity in UM-SCC-9 and UM-SCC-11B cells. The ERK pathway appears to contribute to expression of IL-8 and VEGF and transactivation of NF-κB as well as AP-1 in HNSCC. Combined inhibition of both MAPK and IKK pathways may be needed for suppression of the signal transduction mechanism(s) regulating VEGF and IL-8 secretion and angiogenesis by human HNSCC.
Increased expression of proinflammatory and proangiogenic factors are associated with aggressive tumor growth and decreased survival of patients with HNSCC3(1, 2, 3, 4, 5, 6). We showed recently that HNSCCs express cytokines in vivo, including VEGF and IL-8, which have known angiogenesis and tumor growth-promoting activity (1). The promoter region of VEGF contains several binding sites for AP-1 (7), whereas the promoter region of IL-8contains binding sites for AP-1, NF-κB, and NF-IL6 (8, 9, 10). We showed that these transcription factors are coactivated in HNSCC and that differences in constitutive activation of NF-κB and AP-1 contribute to the differences in expression of these cytokines (11). Transcription factors NF-κB and AP-1 may be activated by several common and distinct signal transduction pathways, but those pathways involved in activation of NF-κB and AP-1 in squamous cell carcinoma have not yet been defined.
Several growth factor and cytokine signal transduction pathways have been reported to contribute to activation of NF-κB (12, 13, 14). IKK mediates the convergence of signals from the pathways that activate NF-κB by phosphorylation of IκB (13) and by direct phosphorylation of the RelA p65 subunit of the NF-κB/RelA (p50/p65) heterodimer (15, 16, 17). Signal-induced translocation of NF-κB from the cytoplasm to the nucleus and DNA binding requires the degradation of IκB, and transactivation of genes by NF-κB in some systems has been shown to be dependent on additional phosphorylation of p65 and other cofactors (15, 16, 17). We showed previously that expression of a dominant-negative mutant of IκBα (IκBαM) strongly but incompletely inhibited activation of NF-κB; expression of cytokines IL-1α, IL-6, IL-8, and granulocyte/macrophage-colony stimulating factor; and survival and growth of HNSCCs (18).
AP-1 may be activated by signal transduction events and cellular responses through various MAPK signal transduction pathways. MAPK activation can also result from overexpression of several receptors or growth factors that have been detected in HNSCCs and other cancers (19, 20, 21). The three major groups of MAPKs include ERKs,JNKs, and p38 kinases (also known as stress-activated protein kinases;Refs. 22 and 23). These kinases are activated by distinct extracellular stimuli through different signaling cascades (24, 25). The best described of these three pathways is the MAPK-ERK cascade (25). The ERK pathway is strongly activated by growth factors and mitogenic stimuli, such as EGF and phorbol esters (26, 27). Activation of the MAPK-ERK pathway results in the translocation of dually phosphorylated ERK (24) into the nucleus to activate transcription factors,protein kinases, and protein phosphatases that regulate proliferation,differentiation, and migration (28, 29).
The upstream activation of NF-κB and AP-1 regulated by IKK and MAPKs provide potential targets for inhibiting the coactivation of these signal pathways and cytokines in HNSCCs. In this study, we examined the potential contribution of MEK and IKK on AP-1- and NF-κB-dependent activation of VEGF and IL-8 in HNSCC cell lines, using the dominant-negative mutant IκBαM and a MEK inhibitor, U0126. We report that interruption of IKK-mediated activation of NF-κB by expression of an IκBαM in UM-SCC-9 cells resulted in partial inhibition of expression of IL-8 but not VEGF. U0126 inhibited constitutive and TNF-α-inducible ERK1/2 phosphorylation by MEK and constitutive and TNF-α-inducible IL-8 and VEGF expression. Inhibition of both MEK and IKK pathways may be necessary for suppression of the signal transduction mechanism(s) regulating VEGF and IL-8 secretion and angiogenesis by human HNSCCs.
MATERIALS AND METHODS
Cell Lines and Tissue Culture.
UM-SCC-9 and UM-SCC-11B cell lines were derived from SCCs of the upper aerodigestive tract at the University of Michigan (Ann Arbor, MI) and kindly provided by Dr. T. E. Carey. The derivation and characterization of stable transfectants of UM-SCC-9 expressing IκBαM (UM-SCC-9 I-11) or vector alone (UM-SCC-9 C-11) has been described elsewhere (18). The expression of cytokines and activation of transcription factors NF-κB and AP-1 of the cell lines used in the study were characterized previously (11, 18). The cell lines were maintained as monolayer cultures in Eagle’s MEM supplemented with 10% fetal bovine serum, 50 μg/ml penicillin/streptomycin, and 2 mm glutamine at 37°C.
The MEK inhibitor U0126 was purchased from Promega Corp. (Madison, WI). Human IL-8 and VEGF ELISA kits were purchased from R&D Systems(Minneapolis, MN). Human TNF-α was kindly provided by Knoll Pharmaceutical Company (Whippany, NJ). The BCA Protein Assay and Super Signal West Pico Chemiluminescent Detection kits were obtained from Pierce Corp. (Rockford, IL). Phospho- and non-phospho-specific antibodies and control proteins for Erk1/2, JNK, and p38 were purchased from New England Biolabs (Beverly, MA).
Cytokine Quantitation by ELISA.
HNSCCs were plated overnight at 5 × 104cells/well/ml in a sterile 24-well culture plate. The culture medium was removed, and the cells were preincubated ± U0126 (1 h) and then stimulated ± 1000 units/ml TNF-α (24 h). Supernatants were harvested, centrifuged at 12,000 × g for 5 min at 4°C, and stored at −80°C. ELISA for human IL-8 and VEGF was performed according to the manufacturer’s instructions.
Preparation of Nuclear Extracts and Gel Shift Analysis.
UM-SCC-9 and UM-SCC-11B cell lines were grown to 60–90% confluence in sterile 100 × 15-mm polystyrene tissue culture dishes. The cells were preincubated ± U0126 (1 h) and then stimulated ± 1000 units/ml TNF-α (12 h). Nuclear extracts were harvested, and gel shift assays were conducted as described previously (11). The relative binding of AP-1 and NF-κB was determined using NIH Image 1.62 and normalized to OCT-1 binding.
Isolation of Whole-Cell Lysates and Western Blot Analysis.
HNSCC cell lines were grown to 60–90% confluence in sterile 100 × 15-mm polystyrene cell culture dishes. The cells were preincubated ± U0126 (1 h) and stimulated ± 1000 units/ml TNF-α (15 min). The cells were rinsed once with ice-cold PBS,scraped, and lysed in 250 μl of Western lysis buffer [1% Triton X-100, 150 mm NaCl, 10 mm Tris-HCl (pH 7.4), 1 mm EDTA, 1 mm EGTA, 0.5% NP40, 0.2 mm Na3VO4, and 0.2 mm phenylmethylsulfonyl fluoride), and transferred to Eppendorf tubes. The lysates were passed three times through a 23-gauge needle, centrifuged at 12,000 × g for 5 min at 4°C,and the supernatants were stored at −80°C. Protein concentrations were determined using the Pierce BCA Protein Assay. Each sample (40μg) was mixed with Laemmli loading buffer containingβ-mercaptoethanol and boiled for 5 min at 100°C. The samples were electrophoresed through 10% Tris-Glycine precast gels (Novex, San Diego, CA) at 120 V and transferred to nitrocellulose using the Novex Gel Blot Module for 90 min at 20 V. Ponceau-S (Sigma Chemical Co., St. Louis, MO) was used to determine transfer efficiency. Immunoblotting was performed according to the manufacturer’s protocol (New England Biolabs, Beverly, MA).
Transient Transfection and Reporter Assay.
The AP-1 and NF-κB reporter constructs have been described previously (30). UM-SCC-9 and UM-SCC-11B were seeded at 1 ×104 cells/well in sterile 96-well culture plates. The cells were transfected with either AP-1 or NF-κB luciferase reporter plasmids using LipofectAMINE Plus transfection reagent according to the manufacturer’s protocol (Life Technologies, Inc.,Grand Island, NY). The cells were preincubated ± U0126 (1 h) and then stimulated ± 1000 units/ml TNF-α (12 h). The cells were lysed, and luciferase activity was measured using the Dual Light Reporter Gene assay (Tropix, Bedford, MA) and a Wallac Victor2 1420 Multilabel Counter (EG&G Wallac,Gaithersburg, MD) according to the manufacturer’s instructions.
Expression of IκBαM Partially Inhibits IL-8 but not VEGF Secretion by UM-SCC-9 Cells.
IKK phosphorylation of IκB at S32 and S36 has been shown to be required for ubiquitination and degradation of IκB and activation of NF-κB (31). We previously established UM-SCC-9 cells stably transfected with a dominant-negative phosphorylation mutant of IκBα (IκBαM; UM-SCC-9 I-11) and showed that IκBαM inhibited activation of NF-κB, survival, cytokine expression, and growth of UM-SCC-9 in vivo (18). We examined the effect of inhibition of IKK-mediated activation of NF-κB on IL-8 and VEGF production by comparing secretion of these factors by UM-SCC-9 and UM-SCC-9 C-11 cells transfected with an empty vector control and UM-SCC-9 I-11 cells transfected with the dominant-negative mutant IκBαM, established previously (18). Fig. 1 shows that the three UM-SCC-9 cell lines secrete IL-8 and VEGF, and that secretion of both factors is inducible by TNF-α. Consistent with previous results (18), UM-SCC-9 I-11 cells expressing IκBαM secrete decreased concentrations of IL-8 relative to the control vector-transfected cells (UM-SCC-9 C-11) and the parental UM-SCC-9 cells (Fig. 1,A). However, cells expressing IκBαM showed no decrease in secretion of VEGF (Fig. 1 B). The partial inhibition of IL-8 and lack of inhibition of VEGF are consistent with the possibility that one or more additional mechanisms contribute to expression of these factors in UM-SCC-9.
MAPK ERK 1/2 Is Constitutively Phosphorylated and MAPKK-MEK Inhibitor U0126 Inhibits Activation of ERK1/2, IL-8, and VEGF Production in UM-SCC-9 and UM-SCC-11B.
The MAPK ERK1/2 is an important upstream activator of c-Fos and Fos-related antigen-1 (Fra-1), a component of AP-1 activated in UM-SCC cell lines (11). To explore the activation and function of the MAPK ERK1/2 in UM-SCC cell lines, we examined two cell lines,UM-SCC-9 and UM-SCC-11B, which were found previously to exhibit low and high AP-1 and IL-8 activation among a panel of HNSCC lines (11). Immunoblot analysis for the phosphorylated form of ERK1/2 demonstrated constitutive activation of ERK1/2 in both UM-SCC-9 and UM-SCC-11B, and the intensity of the phospho-ERK and total ERK signal was found to be greater in the UM-SCC-11B than UM-SCC-9 (Fig. 2,A, and a repeat experiment, data not shown), consistent with differences in AP-1 and cytokine activation (11). TNF-αtreatment further induced ERK1/2 activation. To determine whether ERK activation could be inhibited by an antagonist of the upstream MEK, we determined whether inhibitor U0126 at 1 and 10μ m concentration could block ERK phosphorylation. Preincubation with the MEK inhibitor U0126 inhibited both the constitutive and TNF-α-inducible phosphorylation of ERK1/2 but did not affect total ERK1/2 protein levels nor phosphorylation of JNK or p38 (Fig. 2, B and C), indicating the specificity of U0126 for MEK inhibition. U0126 in the 1 and 10μ m concentration range appeared to more completely inhibit phospho-ERK in the UM-SCC-9 cell line, which exhibits lower ERK activity than the UM-SCC-11B cell line. U0126 in the 1 and 10 μm concentration range demonstrated no significant effect on survival or proliferation of UM-SCC-9 or UM-SCC-11B cells, as determined in a 4-day 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay(data not shown).
To determine whether ERK1/2 activation contributes to constitutive and inducible IL-8 and VEGF production in HNSCC, UM-SCC-9 and UM-SCC-11B cells were exposed to MEK inhibitor U0126 at 1 and 10 μmconcentration or control medium for 24 h, and the effect on production of these factors was determined by ELISA. Fig. 3,A shows that U0126 inhibited constitutive VEGF production in a dose-dependent manner with estimated IC50s of 0.2 and 3.48 μm for UM-SCC-9 and UM-SCC-11B,the low and high cytokine producers, respectively. TNF-α-inducible VEGF production was similarly inhibited by U0126 with IC50s of 1.35 μm for UM-SCC-9 and 2.81 μm for UM-SCC-11B. Comparable results were observed for IL-8 production (Fig. 3 B). U0126 inhibited the already low-level production of IL-8 in UM-SCC-9, whereas an IC50 of 7.35 μm was calculated for UM-SCC-11B. As observed with inducible VEGF expression,U0126 inhibited TNF-α-inducible IL-8 production with IC50s of 1.38 and 9.27 μmfor UM-SCC-9 and UM-SCC-11B, respectively. The decrease in production of IL-8 and VEGF was not attributable to an effect on survival or proliferation of UM-SCC-9 or UM-SCC-11B cells, as determined in a 4-day 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay(data not shown). The results indicate that higher concentrations of U0126 are needed to inhibit expression of IL-8 and VEGF in the UM-SCC-11B cell line, which exhibits higher activation of ERK, AP-1,and cytokine production.
Constitutive and Inducible Cytokine Expression and Transcription Factor Activation in UM-SCC-9 and UM-SCC-11B.
We reported previously that the production of cytokines such as VEGF and IL-8 is associated with coactivation of transcription factors AP-1 and NF-κB in HNSCC cell lines (11). Fig. 4,A shows a comparison of the differences in IL-8 and VEGF secretion between cell lines UM-SCC-9, the lower cytokine producer, and UM-SCC-11B, the high cytokine producer. Fig. 4,B confirms that the differences in constitutive production of these cytokines correspond to differences in levels of constitutive NF-κB and AP-1 DNA binding activity in nuclear extracts observed previously (11). TNF-α treatment further induced cytokine production (Fig. 4,A) and NF-κB and AP-1 binding (Fig. 4 B). The observed correlation between VEGF and IL-8 expression and constitutive and inducible coactivation of AP-1 and NF-κB in UM-SCC-9 and UM-SCC-11B is consistent with previous observations in a wider panel of HNSCC lines (11).
U0126 Inhibits Activation of AP-1 and NF-κB.
The effect of MEK inhibition by U0126 on inducible AP-1 and NF-κB transcriptional activity was examined by luciferase reporter assay(Fig. 5). TNF-α was used as a control for induction of AP-1 and NF-κB activity. A dose-dependent inhibition of TNF-α-inducible AP-1 activity was observed (Fig. 5,A). Of interest, a similar inhibition of NF-κB activity was also noted (Fig. 5 B). This effect on both pathways was observed in two additional experiments after 8 and 18 h of TNF-α treatment (data not shown). U0126 has been reported to inhibit inducible but not constitutively activated forms of AP-1 (32).
We performed EMSA to compare the effect of U0126 on AP-1 and NF-κB DNA binding activity. Fig. 6 confirms that U0126 inhibits inducible but not basal AP-1 binding activity. Although U0126 inhibited NF-κB transactivation in Fig. 5,no significant inhibitory effect on NF-κB DNA binding activity was observed over the time course studied in either UM-SCC-9 or 11B (Fig. 6). Thus, the inhibitory effects of U0126 on production of IL-8 and VEGF correlate primarily with inhibition of inducible AP-1 and NF-κB reporter activity and transactivation in UM-SCC-9 and UM-SCC-11B.
In this study, we investigated the upstream signaling events involved in the expression of IL-8 and VEGF and their relationship to coactivation of NF-κB and AP-1 in HNSCC lines UM-SCC-9 and UM-SCC-11B. We obtained evidence that IKK-dependent activation of NF-κB contributes to expression of IL-8 but not VEGF, and that the MAPKK MEK activates MAPK ERK1/2 and contributes to expression of both IL-8 and VEGF in the HNSCC lines. MEK inhibitor U0126 inhibited ERK1/2 phosphorylation and expression of IL-8 and VEGF. As observed in previous experiments with a wider panel of HNSCC lines (1, 2, 33), we confirmed that UM-SCC-9 and UM-SCC-11B produced IL-8 and VEGF at different levels, and that these differences in cytokine expression were associated with differences in transcription factor AP-1 and NF-κB DNA binding activity. U0126 inhibited TNF-α-induced transactivation of both NF-κB and AP-1 reporters but had no effect on basal DNA-binding activity of these transcription factors. Thus, the MAPK-ERK pathway appears to contribute to expression of IL-8 and VEGF and transactivation of NF-κB as well as AP-1 in HNSCCs. Combined inhibition of both MAPK and IKK pathways may be needed for suppression of the signal transduction mechanism(s) regulating VEGF and IL-8 secretion and angiogenesis by human HNSCCs.
In our previous study, we showed that expression of IL-8 and growth in vivo could only be inhibited partially by expression of a dominant-negative mutant of IκBα suppressing activation of NF-κB (18). In the present study, we show that interruption of the IKK signal pathway can inhibit IL-8 but not VEGF production by UM-SCC-9 cells. These results are consistent with the presence of NF-κB sites in IL-8 and apparent lack of NF-κB sites in the VEGF genes, except the presence of AP-1 sites in both IL-8 and VEGF, which could be induced through activation of the MAPK pathways (7, 34). It remains to be determined whether activation of NF-κB may affect activation of VEGF through occult regulatory mechanisms in other HNSCCs or other cell types.
Analysis of possible alternative pathways for induction of these proangiogenic factors revealed activation of the MAPK ERK1/2 in cell lines UM-SCC-9 and UM-SCC-11B. Basal and TNF-α-inducible phosphorylation of ERK1/2 and secretion of IL-8 and VEGF could be inhibited by the MEK inhibitor U0126. We found that U0126 inhibited IL-8 and VEGF cytokine expression in UM-SCC-9 and UM-SCC-11B in a dose-dependent manner. UM-SCC-9 was much more sensitive to U0126 compared with UM-SCC-11B. Because there was no effect on cell density or growth observed using U0126 in this concentration range, this may be attributable to the fact that UM-SCC-11B has a significantly higher constitutive activation of MEK, ERK, and expression of these factors than UM-SCC-9 to block (11).
Induced activation of ERK has been implicated in expression of VEGF and IL-8 in cultured cells exposed to pathogens and growth factors. Activation of ERK p44 by Rous sarcoma virus has been implicated in activation of IL-8 in A549 lung carcinoma cells (35). Activation of ERK p42/44 has been reported to be important in transcriptional regulation of VEGF in fibroblasts (36). Overexpression of ERK has been detected in a variety of cancers,including HNSCCs (37). The occurrence of ERK1/2 activation in colon cancer has been reported, and its role in growth in vitro and in vivo has recently been studied using another MEK inhibitor, PD 184352 (38). ERK activation was demonstrated in a variety of colon carcinomas in situ as well as in vitro. MEK inhibition resulted in cytostatic inhibition of growth in clonogenic assay and decreased spread and invasiveness in scatter and Matrigel assays in vitro. Tumor growth in vivo was inhibited by up to 50–80% for sensitive tumor lines. The effect on expression of proangiogenic cytokines was not evaluated, but the potential role of effects on ERK activation on tumor VEGF expression and in endothelial cells during angiogenesis was considered. We found that MEK inhibitor U0126 blocked IL-8 and VEGF at lower concentrations (1–10 μm) than those(30–60 μm) at which cytostatic inhibition of growth is observed in vitro.4Sustained activation of ERK in endothelial cells has also been shown to promote angiogenesis, raising the possibility that the host response to angiogenesis factor-producing tumors may also be subject to pharmacological MEK inhibitors (39).
Expression of IL-8 and VEGF in the cell lines was associated with coactivation of both NF-κB and AP-1, and U0126 inhibited both NF-κB and AP-1 reporter activity in UM-SCC-9 and UM-SCC-11B cells. The mechanism(s) of coactivation of IKK and MAPK in HNSCC cell lines remains to be determined. Ras activation can lead to MEK and NF-κB activation, but mutations in Ras occur in only 10%of HNSCCs (40, 41, 42). Up to 90% of epidermoid carcinomas have been reported to coexpress elevated levels of EGFR and/or its ligands (such as transforming growth factor-α, amphiregulin, and heparin-binding EGF; Refs. 20 and 43, 44, 45, 46). Inhibitors and antibodies against EGFR and ErbB-2/neu have been reported to lead to a decrease in VEGF production by human A431 human epidermoid carcinoma cells and SKBR-3 human breast cancer cells in vitro and in vivo (47). Activation of the EGFR can result in the activation of the multiple signal transduction cascades involving phospholipase Cγ, Ras, and phosphatidylinositol 3′-OH kinase (23). Overexpression and/or autophosphorylation of the EGFR in many tumors have been implicated in the constitutive activation of Ras(32). Ras complexes with and activates the serine/threonine kinase Raf-1. Raf-1 then phosphorylates MEK. MEK, a dual-specificity kinase, phosphorylates ERK on tyrosine and threonine,and ERK translocates into the nucleus to regulate transcription factors, such as c-Fos (48). The MEK inhibitor U0126 was identified as an antagonist of ERK activation that suppresses c-Jun and c-Fos mRNA expression and protein levels in activated cells (32).
We showed previously that c-Jun and Fra-1 are the most prevalent species activated in HNSCCs that express cytokines (11). ERK has been reported recently to contribute to transactivation of Fra-1 after transformation of murine epithelioid cells (49), but we have not yet determined whether Fra-1 is the primary target of ERK for activation of AP-1 in HNSCC. The mechanism by which ERK may contribute to NF-κB transactivation in HNSCCs also remains to be defined. ERK and AP-1 can induce expression of transforming growth factor-α, which may potentiate activation of EGFR and thereby NF-κB p65 and transactivation (50). Treatment of HNSCCs with U0126, IKK, or EGFR inhibitors could block the activation of AP-1 and NF-κB and inhibition of proinflammatory and proangiogenic cytokine gene expression. Therefore, it will be important to determine whether such inhibitors can inhibit angiogenesis and growth of HNSCC cells in vivo in future studies.
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.
Supported by National Institute on Deafness and Other Communication Disorders, Intramural Research Project Z01-DC-00016, and a Howard Hughes Medical Institute Scholarship (to C. P.).
The abbreviations used are: HNSCC, head and neck squamous cell carcinoma; VEGF, vascular endothelial growth factor; IL,interleukin; AP-1, activator protein 1; NF-κB, nuclear factor-κB;IKK, inhibitor κB kinase; IκB, inhibitor κB; MAPK,mitogen-activated protein kinase; ERK, extracellular signal-regulated kinase; JNK, c-Jun NH2-terminal kinase; EGF, epidermal growth factor; EGFR, EGF receptor; MEK, MAPK kinase; TNF, tumor necrosis factor.
C. Bancroft, data not shown.
We thank Drs. James F. Battey and Angelo Russo for review of the manuscript.