EGFR/JIP-4/JNK2 Signaling Attenuates Cetuximab-Mediated Radiosensitization of Squamous Cell Carcinoma Cells

The EGF receptor (EGFR) belongs to the ErbB family of transmembrane receptor tyrosine kinases (RTK; ref. 1). Because of increased activity and/or expression in human malignancies and its role in, for example, cell proliferation, migration, and invasion, EGFR emerged as a potent molecular target (2–4). On binding of ligands, such as EGF or amphiregulin, EGFR dimerization results in RTK autophosphorylation and channeling of mitogenic signals predominantly via the Ras-extracellular signal–regulated kinase (ERK), Akt, and c-jun-NH₂-kinase (JNK) pathway (1).

Preclinical work showed the effectiveness of pharmacologic inhibitors and antibodies against EGFR in tumor cells from head and neck, lung, colon, and other cancers (3, 5–10). Cetuximab, for example, is approved for use in combination with radiotherapy for treatment-naïve patients with locally advanced squamous cell carcinoma of the head and neck (HNSCC) and as monotherapy for recurrent or metastatic HNSCC (11, 12). These trials showed significantly increased rates of both locoregional control and overall survival.

Despite this clinical success, escape- or efficacy-attenuating mechanisms in response to cetuximab treatment have been widely neglected. Recent studies suggested EGFR ubiquitination, cooperative EGFR-Src family kinase interactions, ErbB2 signaling, and p53 are involved in reduced cetuximab efficacy in colorectal or non–small cell lung cancer cell lines (13–16). As our understanding of EGFR signaling in cancer is still marginal and obviously tumor type-dependent, it is of high clinical relevance to elucidate escape mechanisms of cetuximab in HNSCC by taking into consideration how HNSCC cells grow in vivo. Our own work and work from others have shown that 3-dimensional (3D) growth conditions are more physiologically relevant with regard to cell morphology, gene and protein expression patterns, protein–protein interactions, and intracellular signaling as well as response to irradiation and chemotherapeutics relative to conventional 2D monolayer cell cultures (17–26).

In the present study, we sought to elucidate the nature of EGFR signaling in more detail and evaluate whether cetuximab mediates differential cytotoxicity and radiosensitization in squamous cell carcinoma (SCC) cells grown in 3D versus 2D. In a systematic comparison of different growth conditions, we show significant similarity of EGFR, ERK1/2, and Akt phosphorylation in SCC cells grown in 3D extracellular matrix (ECM) and as tumor xenografts in contrast to conventional monolayer cell cultures. In line with in vivo studies assessing cetuximab plus radiotherapy (7, 9), cetuximab radiosensitized cells in a 3D, but not 2D, environment. Using phosphoproteome arrays, we observed strong induction of JNK2 phosphorylation potentially resulting from an interaction between cetuximab-inhibited EGFR and JNK-interacting protein-4 (JIP-4), which was identified in an immunoprecipitation-mass spectrometric approach. Inhibition of this signaling bypass by JIP-4 or JNK2 knockdown or pharmacologic JNK2 inhibition enhanced cetuximab efficacy and tumor cell radiosensitivity.

Please see Supplementary Experimental Procedures for the following sections: “Measurement of apoptosis” and “Immunoprecipitation of EGFR in untreated cells.”

Antibodies used for Western blotting include EGFR Y1173, EGFR Y1068, JNK T183/Y185 (Invitrogen), EGFR, Akt, Akt S473, Akt T308, ERK1/2, ERK1/2 T202/Y204, MEK1/2, MEK1/2 S217/221, JNK, c-Jun, c-Jun S63, JIP-4, caspase-3, phosphotyrosine, phosphothreonine (Cell Signaling), β-actin (Sigma), horseradish peroxidase–conjugated donkey anti-rabbit, and sheep anti-mouse antibodies (Amersham). Antibodies for immunofluorescence staining include EGFR (Abcam), JIP-4, c-Jun S63, c-Jun (Cell Signaling), Alexa594 phalloidin, Alexa488 anti-rabbit, and Alexa594 anti-mouse antibodies (Invitrogen). Antibodies for immunoprecipitation include EGFR (Invitrogen) and mouse immunoglobulin G (IgG; Santa Cruz Biotechnology). Antibodies were purchased as indicated.

A431 and FaDu cells were purchased from the American Type Culture Collection. The SAS, UTSCC14, and UTSCC5 cells were a generous gift from Dr. R. Grenman (Turku University Central Hospital, Turku, Finland). Asynchronously and exponentially growing cells were used in all experiments. In both 2D and 3D cell-culture conditions, cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) containing GlutaMAX-I supplemented with 10% fetal calf serum (except for experiments shown in Supplementary Fig. S1, which were conducted under serum starvation) and 1% nonessential amino acids (PAA) at 37°C in a humidified atmosphere containing 7% CO₂. Conventional tissue culture plates (BD Biosciences) were used for 2D cell culture. For 3D cell culture, plates were coated with 1% agarose (Sigma) to prevent cell attachment to the bottom of the well. Cultrex 3D Culture Matrix (lrECM; Trevigen) was added to the cell-culture medium to obtain a final concentration of 0.5 mg/mL. Thereafter, 3D cultured cells were covered with supplemented DMEM.

Irradiation was delivered at room temperature using single doses of 200 kV X-rays (Yxlon Y.TU 320; Yxlon) filtered with 0.5-mm Cu. The absorbed dose was measured using a Duplex dosimeter (PTW). The dose-rate was approximately 1.3 Gy/min at 20 mA, and applied doses ranged from 0 to 6 Gy.

The 2D colony formation assay was conducted as published previously (27). Asynchronously and exponentially growing cells were plated as single cells in 6-well cell-culture plates. After 24 hours, cetuximab was added to the medium to final concentrations of 0 to 5 μg/mL. After 24 hours, cells were either left unirradiated or were irradiated (2–6 Gy) and then cultured for 9 days (A431 and UTSCC14) or 11 days (FaDu). Cetuximab remained in the supernatant for the entire culture period. After fixation with 80% ethanol and Coomassie staining (Merck), cell colonies with more than 50 cells were counted microscopically (Fig. 2A). Plating efficiencies were calculated as follows: number of colonies formed/number of cells plated. Surviving fractions (SF) were calculated as follows: numbers of colonies formed/(numbers of cells plated (irradiated) × plating efficiency (unirradiated). Each point on survival curves represents the mean surviving fraction from at least 3 independent experiments.

Clonogenic survival under 3D conditions was determined in a 3D colony formation assay as published previously (26, 27). Asynchronously and exponentially growing single cells were mixed with 0.5 mg/mL lrECM and placed in 96-well plates. LrECM was covered with cell-culture medium. After 24 hours, cetuximab was added to the medium to final concentrations of 0 to 5 μg/mL. After 24 hours, irradiation (0–6 Gy) was carried out. Cetuximab remained in the cell-culture medium for the entire growth period. Incubation of cells with the JNK inhibitor SP600125 (Santa Cruz Biotechnology) was carried out for 26.5 hours, whereby the inhibitor was added 30 minutes before cetuximab and was withdrawn 2 hours after irradiation by washing 5 times using DMEM. Cells were cultured for 9 days (A431 and UTSCC14), 11 days (FaDu), or 14 days (UTSCC5). Cell clusters (with the minimum size of a cell cluster containing 50 cells) were counted microscopically without fixation or staining (see Fig. 2A). Plating efficiencies and surviving fractions were calculated as described earlier. Synergism was calculated as recently published (28) using the following formula: S = r(a, b) − r(a₀, b) × r(a, b₀).

Asynchronously and exponentially growing 2D cell cultures were rinsed with ice-cold 1× PBS before harvesting total proteins by scraping with modified radioimmunoprecipitation assay (RIPA) buffer [50 mmol/L Tris–HCl (pH 7.4), 1% Nonidet-P40, 0.25% sodium deoxycholate, 150 mmol/L NaCl, 1 mmol/L EDTA, complete protease inhibitor cocktail (Roche), 1 mmol/L NaVO4, and 2 mmol/L NaF]. In 3D, asynchronously and exponentially growing cells were lysed within the lrECM using modified RIPA buffer. Homogenization of lysates was accomplished by 4 passages through a 25-gauge needle followed by centrifugation. Samples were stored at −80°C. Protein concentration of 2D lysates was measured using the BCA assay (Thermo Scientific). SDS-PAGE, transfer of proteins onto nitrocellulose membranes (Schleicher und Schuell), probing, and detection of specific proteins with indicated antibodies and SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) was carried out as described previously (27).

MEK1 siRNA (sequence: 5′-GGAGCUAGAGCUUGAUGAGtt-3′), JNK2 siRNA (sequence: 5′-GGGAUUGUUUGUGCUGCAUtt-3′), EGFR siRNA (sequence: 5′-GGCACGAGUAACAAGCUCAtt-3′), and JIP-4 siRNA (sequence: 5′-GCAGUAUCGUCAGGUAAAAtt-3′) were obtained from Applied Biosystems. The nonspecific control siRNA (sequence: 5′-GCAGCUAUAUGAAUGUUGUtt-3′) was from Eurofins MWG Operon. siRNA transfection was conducted as published (17). Twenty-four hours after delivery of 20 nmol/L siRNA using oligofectamine (Invitrogen), cells were plated on 2D cell-culture plastic or embedded in 3D lrECM. A 1- or 24-hour incubation with cetuximab (0–5 μg/mL) was followed by irradiation. Colony formation assays and Western blotting were carried out. Efficient MEK1, JNK2, EGFR, and JIP-4 knockdown was confirmed by Western blotting.

Immunofluorescence stainings of 3D cultured cells were carried out as recently published (8). Three-dimensionally cultured cells were treated with cetuximab (0–5 μg/mL) for 1 hour and then fixed with 3% formaldehyde/PBS in lrECM for 15 minutes. After transfer to a centrifugation tube, cells were permeabilized with 0.25% Triton X-100/PBS for 10 minutes, washed with PBS, and blocked with 1% BSA/PBS for 30 minutes. Staining of c-Jun, c-Jun S63, EGFR, and JIP-4 was carried out with specific antibodies for 2 hours. After washing with 1× PBS, secondary antibodies and/or Alexa594 phalloidin were added for 1 hour. Cells were washed and transferred onto microscopic slides before covering with Vectashield/DAPI mounting medium (Alexis) and cover slip. Images were acquired using a confocal laser scanning microscope LSM 510meta (Zeiss). Colocalization was measured using the LSM510 software (Zeiss).

The experiments were carried out using 7- to 14-week-old male and female NMRI (nu/nu) mice obtained from the pathogen-free animal breeding facility (Experimental Center, Medical Faculty, Technische Universität Dresden, Germany). Animal facilities and experiments were approved according to the institutional guidelines and the German animal welfare regulations. For the experiments, small tumor pieces were transplanted subcutaneously into the right hind-leg as published previously (7, 26, 27). At a tumor size of 7 mm in diameter, mice were treated with 1 mg cetuximab (Erbitux) intravenously in a volume of 0.2 mL. Please see Supplementary Experimental Procedures for detailed description.

The Phospho Explorer Antibody Microarray was conducted by Full Moon BioSystems Inc. Whole-cell lysates from 2D and 3D FaDu cell cultures treated with cetuximab (5 μg/mL; or untreated controls) were harvested after 1 hour using Protein Extraction Buffer (Full Moon BioSystems Inc.) and transferred to Full Moon BioSystems Inc., on dry ice. The array consists of 1,318 phospho-specific antibodies. In brief, proteins were labeled with biotin and placed on preblocked microarray slides. After washing, detection of total and phosphorylated proteins was conducted using Cy3-conjugated streptavidin. Expression of phosphorylated proteins was normalized to corresponding total protein expression. Fold change was calculated as follows: phosphorylation of cetuximab-treated cells/phosphorylation of untreated cells. This experiment was carried out once. Where indicated, protein phosphorylation data were confirmed by Western blotting.

Cetuximab-treated (5 μg/mL) or untreated 3D grown cells were lysed with cell lysis buffer (Cell Signaling) supplemented with complete protease inhibitor cocktail as described (26, 27). Protein-G-Agarose beads (Sigma) were incubated with nonspecific mouse IgG antibody overnight. Protein lysates were added to the beads and incubated overnight. Cetuximab-bound EGFR was pulled down by interaction of cetuximab with the beads, and not by the nonspecific mouse IgG. After washing of the beads, SDS-PAGE and Western blotting were carried out as described earlier.

Three-dimensionally grown cells were lysed with cell lysis buffer (Cell Signaling) supplemented with complete protease inhibitor cocktail (Roche). Protein-G-Agarose beads (Sigma) were incubated with EGFR antibody or nonspecific mouse IgG antibody overnight. Protein lysates were added to the beads and incubated overnight. After SDS-PAGE, gel fixation, and Coomassie (Merck) staining, mass spectrometric analysis was carried out at the Max Planck Institute of Molecular Cell Biology and Genetics Mass Spectrometry (MPI-CBG MS) Facility (Dresden, Germany) as published (26, 27). Proteins were first in-gel digested with trypsin as described in ref. 29 and the liquid chromatography/tandem mass spectrometric (LC/MS-MS) analysis of extracted peptide mixtures was conducted on an Ultimate nanoLC systems (Dionex) interfaced online to a LTQ Orbitrap hybrid mass spectrometer (Thermo-Fisher Scientific) via a robotic nanoflow ion source TriVersa (Advion BioSciences Ltd.) as described in ref. (29). Acquired spectra were searched against human international protein index (IPI) protein databases. Spectra were matched by MASCOT software (Matrix Science, v.2.2.0) under the following settings: mass tolerance was set as ±10 ppm for precursor and ±0.5 Da for fragmentation spectra; variable modifications: propionamide (C), N-acetylation (Protein N-terminus), oxidation (M); enzyme: trypsin; and 1 missed cleavage allowed. The result was evaluated by Scaffold software (Proteome Software) under the following settings: minimal number of peptides was 2; minimal peptide and protein probabilities were 95% and 99%, respectively.

Means ± SD of at least 3 independent experiments were calculated with reference to untreated controls defined in a 1.0 scale. To test statistical significance, Student t test was conducted using Microsoft Excel 2003. Results were considered statistically significant if a P value of less than 0.05 was reached. Densitometry of Western blot analyses was conducted with ImageJ analysis software (http://rsbweb.nih.gov/ij).

We first characterized and compared EGFR expression and phosphorylation of 3 SCC cell lines in 3D lrECM versus conventional monolayer cultures versus tumor xenografts (Fig. 1A). Intriguingly, patterns of EGFR expression and EGFR tyrosine (Y)1068 and Y1173 as well as ERK1/2 threonine (T)202/Y204 phosphorylation showed great similarity in FaDu and UTSCC14 cells grown in 3D lrECM and in vivo (Fig. 1B and C). Akt T308 phosphorylation was higher in 3D than in 2D, but only matched with in vivo levels in FaDu cells (Fig. 1B and C). The extreme levels of EGFR phosphorylation in A431 cells might be causative for only small detectable differences of proteins of interest under the different growth conditions (Fig. 1B and C). To our knowledge, this is the first comparative analysis of EGFR signaling in HNSCC cells growing in 3D, in vivo, or as a monolayer, providing strong support for 3D lrECM-based cell cultures as correlates to in vivo growth conditions.

The inhibitory potential of cetuximab under serum starvation and subsequent EGF stimulation was conducted. EGF exposure resulted in enhanced EGFR Y1173 and Y1068 phosphorylation in FaDu and UTSCC14 cells (Supplementary Fig. S1). Next, we analyzed the effects of cetuximab on clonogenic survival of 3D- and 2D-grown SCC cell lines and found great intercell line variances independently of 3D or 2D culture settings (Fig. 2A and B; Supplementary Table S1). In contrast to monolayer cell cultures (Fig. 2A and C), 3D lrECM cell cultures were significantly (P < 0.01) and concentration-dependently radiosensitized by cetuximab (Fig. 2A and D), whereas apoptosis induction was negligible in this context (Supplementary Figs. S2 and S3). From these results, we conclude cetuximab efficacy alone to be independent of growth conditions in contrast to combined cetuximab/irradiation treatment regimens in SCC cell lines.

Next, we sought to identify the modifications of the intracellular signaling network resulting from cetuximab treatment by using a phosphoproteome array technique. This array contains 1,318 well-characterized phospho-specific antibodies. EGFR and ERK1/2-, Akt-, and JNK-related signaling pathways were selected from the complete data set (data not shown). Only proteins with modified phosphorylation of 67% or less and 150% or more from cetuximab-treated 3D FaDu cell cultures were listed according to pathway after normalization to untreated controls (Figs. 3A, 4A). In general, the data showed reduced phosphorylation at kinase-stimulating sites as well as increased phosphorylation at kinase-inhibitory sites in response to cetuximab (Figs. 3A, 4A). Detection of total phosphotyrosine and phosphothreonine levels supported this observation (Supplementary Fig. S4). For EGFR, we found reduced phosphorylation at 5 of 10 examined EGFR sites, that is, Y1068/1069 and Y1172/1173; borderline changes at 4 sites; and induction at EGFR Y1016 (Fig. 3A). Selective confirmatory Western blotting analysis of basal and cetuximab-related EGFR phosphorylation patterns showed approximately 70% reduction in EGFR phosphorylation at particular sites independent from growth conditions; however, for 3D cell culture and tumor xenograft, the reduction was significant (Fig. 3B and C). Moreover, inhibition of EGFR resulted in decreased phosphorylation levels of downstream mitogen-activated protein kinase (MAPK) signaling pathways including MAPK kinase [MAP—ERK kinase (MEK); ∼45% reduction in Western blot analysis], ERK1/2 (∼80% reduction in Western blot analysis), E twenty-six-like transcription factor 1 (Elk1), MAPK-interacting kinase 1 (MNK1), and cAMP response element-binding protein (CREB; Fig. 3). Interestingly, changes in growth factor receptor–bound protein 2 (Grb2), Ras, and Raf were not observed. EGFR blocking also led to significant dephosphorylation of phosphoinositide 3-kinase (PI3K) and Akt in 3D and in vivo by 80% and 70%, respectively, in contrast to 2D (Fig. 4; according to www.phosphosite.org). Besides kinase-deactivating phosphorylations, we detected hyperphosphorylation of Rac-MEK4/7-JNK-c-Jun signaling components in both 3D HNSCC cell cultures and tumor xenografts upon cetuximab treatment relative to untreated controls (Figs. 4 and 5A–C). Intriguingly, phosphorylation of JNK1/2 and c-Jun S63 was highly significantly induced upon cetuximab treatment by approximately 600% and 250%, respectively, in 3D and in vivo as compared with 2D (Fig. 4B and C). The presented phosphoproteome data in 3D lrECM HNSCC cell cultures suggest cetuximab causes a distinct EGFR dephosphorylation pattern that translates into selective inhibition of proteins associated with the MEK1/2-ERK1/2 and PI3K/Akt signaling cascade. Of specific interest is the similarity of signal transduction events in these 2 cascades in HNSCC cells grown in 3D lrECM and as xenograft tumors. These findings support our supposition that 3D lrECM-based cell cultures better reflect in vivo growth conditions compared with 2D monolayers.

Because of the context dependency of JNK with regard to anti- and prosurvival effects, we next investigated whether the cetuximab-induced JNK2 hyperphosphorylation modifies cetuximab efficacy (Figs. 4 and 5A). JNK2 targeting using siRNA or the pharmacologic inhibitor SP600125 resulted in significant cytotoxicity in 3D FaDu cell cultures (Fig. 5B and C; Supplementary Fig. S5). While JNK2 depletion significantly sensitized 3D-grown FaDu cells to X-rays, combined cetuximab plus JNK2 depletion showed additive effects in contrast to cetuximab plus MEK1 knockdown (Fig. 5B; Supplementary Table S3). Similar results were obtained on SP600125 treatment (Fig. 5C). EGFR knockdown caused higher radiosensitization than cetuximab alone and remained unchanged by additional cetuximab administration (Fig. 5D). Notably, EGFR depletion prevented cetuximab-induced JNK2 hyperphosphorylation, and JNK inhibition abrogated cetuximab-dependent activation of c-Jun (Fig. 5D; Supplementary Fig. S6). In cetuximab-resistant UTSCC5 cells, supportive data were generated delineating JNK2 to be key for radiosensitization by cetuximab (Supplementary Fig. S7 and Supplementary Table S3). These observations show a novel link between EGFR and JNK2 signaling and suggest this interaction plays a critical role in the self-attenuating process that reduces the radiosensitizing potential of cetuximab.

By carrying out mass spectrometry on EGFR-immunoprecipitates, we identified a variety of interacting proteins (Supplementary Table S2). The most potent candidate found in these immunoprecipitates from cetuximab-treated 3D FaDu cell cultures that connects EGFR to JNK signaling was JIP-4 (Fig. 6A; Supplementary Fig. S8). The radiosensitization by JIP-4 depletion was superimposable to that of cetuximab, and cetuximab-induced JNK2 hyperphosphorylation was prevented (Fig. 6B; Supplementary Table S3). An enhanced recruitment of JIP-4 to cetuximab-inhibited EGFR was visualized by immunofluorescence staining and colocalization measurement in 3D FaDu cell cultures as compared with controls (Fig. 6C). These data show a functional role of JNK2 in the radiation survival response of 3D lrECM grown cells and suggest JNK2 activation under cetuximab-dependent EGFR inhibition to be transduced by interaction of EGFR with JIP-4 (Fig. 6D).

Understanding EGFR signaling in cancer is essential for treatment optimization and individualization. In addition to a plethora of novel molecular biology technologies, 3D cell cultures have been postulated to better reflect in vivo growth condition and, therefore, might be highly valuable to support this endeavor. Here, we present data showing great similarity in EGFR signal transduction events between 3D lrECM HNSCC cell cultures and HNSCC tumor xenografts in contrast to monolayer cell cultures. Effective EGFR inhibition by cetuximab elicited concentration-dependent cytotoxicity and radiosensitization. In our phosphoproteome analyses, we observed a strong induction of JNK2 phosphorylation potentially resulting from cetuximab-inhibited EGFR. Suppressing bypass signaling via JIP-4 or JNK2 inhibition enhanced cetuximab efficacy and tumor cell radiosensitivity. These findings add new facets to EGFR signaling with particular focus on resistance pathways, which limit tumor cell eradication and radiosensitization.

As emphasized earlier (8, 18, 30–34), growth conditions critically impact on growth factor receptor activity and downstream signaling offering a certain degree of plasticity for bypass signaling. In cancer cells, these mechanisms may considerably contribute to radio- and chemoresistance as well as resistance to novel molecular agents. Our study presents compelling data on the resemblance of EGFR-associated signaling in 3D lrECM cell cultures and corresponding tumor xenografts. About the differences in EGFR phosphorylation in 3D versus 2D, the similarity in the cytotoxicity profiles of cetuximab was unexpected. Importantly, radiosensitization of 3D SCC cell cultures corroborated with tumor sensitization of SCC tumor xenografts (7, 9) and improved locoregional control and overall survival of patients with head and neck cancer treated with combined radiotherapy and cetuximab (11, 12). A mechanistic explanation for cetuximab-mediated radioprotection of 2D-grown cells is currently lacking but should take into consideration signal transduction events, gene expression, and cell cycling (20, 21, 25).

In line with current knowledge (1), reduced phosphorylation of EGFR downstream signaling mediators such as MEK, ERK1/2, and CREB as well as PI3K/Akt at kinase-activating amino acid residues occurred in response to cetuximab treatment. This compromised promitogenic and prosurvival signaling and translated into significant cytotoxicity and radiosensitization. Taking methodologic limitations into account, circumvention of EGFR proximal Grb2/Ras/Raf signaling was unexpected. Instead, hyperphosphorylation/activation of Rac-MEK4/7-JNK-c-Jun signaling components, acting in a context-dependent manner, was found in 3D HNSCC cell cultures and tumor xenografts on cetuximab treatment. By channeling into S63 phosphorylation of the JNK2 downstream target c-Jun, we hypothesize JNK2-c-Jun activation to critically cooperate in an efficacy-limiting stress response to cetuximab.

The JNK family members serve in a multifunctional, environment-driven, and cell line/type-dependent manner (35, 36). Among many factors, X-ray radiation is a well-known JNK activator (36, 37). To prove whether JNK2 is mechanistically involved in a cetuximab–self-attenuating reaction, siRNA-mediated JNK2 depletion and pharmacologic JNK inhibition (SP600125) were conducted. Although JNK2 inhibition conferred cytotoxicity and radiosensitization on its own, concomitant JNK2 knockdown plus cetuximab acted in an additive manner. By using a cetuximab-resistant cell line, that is, UTSCC5, we were able to confirm the importance of JNK2 for cetuximab-mediated radiosensitization.

One possible candidate connecting EGFR and JNK2 is JIP-4, which we found by mass spectrometry of cetuximab–EGFR immunoprecipitates. JIPs serve as scaffold proteins for interactions of different MAPK family members for distributing biochemical cues through specific signaling cascades (38, 39). By means of colocalization studies and gene silencing, we provide evidence that JIP-4 is more intensively recruited to cetuximab-inhibited, membranous EGFR and that JIP-4 depletion prevents cetuximab-mediated JNK2 and c-Jun hyperphosphorylation. Moreover, single JIP-4 knockdown provoked radiosensitization similar to cetuximab alone and, importantly, elicited higher cetuximab efficacy for radiosensitization in the combined JIP-4 knockdown/cetuximab regimen.

In summary, our findings show a high degree of similarity in EGFR downstream signaling in 3D lrECM cell cultures and corresponding tumor xenografts. Because tumor cells are thought to have multiple bypass possibilities under treatment, the presented data suggest participation of JNK2 in cetuximab-antagonizing signaling pathways via interaction between EGFR and the scaffold protein JIP-4, thus providing important novel insights into tumor cell responses and resistance mechanisms to a targeted therapy that has already reached the clinic.

No potential conflicts of interest were disclosed.

Conception and design: I. Eke, N. Cordes

Development of methodology: I. Eke, N. Cordes

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): I. Eke, L. Schneider, C. Förster, D. Zips

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): I. Eke, L. Schneider, C. Förster, D. Zips, N. Cordes

Writing, review, and/or revision of the manuscript: I. Eke, L.A. Kunz-Schughart, N. Cordes

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.A. Kunz-Schughart

Study supervision: N. Cordes

The authors thank A. Shevchenko (Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany) for mass spectrometry analysis, R. Grenman (Turku University Central Hospital, Finland) for the UTSCC14 and 5 cell lines, E. Hammond for critical reading of the article, and I. Lange for excellent technical assistance.

This work was supported in part by the Bundesministerium für Bildung und Forschung (BMBF-03ZIK041 to N. Cordes, BMBF-02NUK006B to N. Cordes, L. Kunz-Schughart, and D. Zips), the EFRE Europäische Fonds für regionale Entwicklung, Europa fördert Sachsen (100066308), and by the Medical Faculty Carl Gustav Carus, Dresden University of Technology, Germany (MeDDrive program to I. Eke).

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.