Oncogenic KRAS Impairs EGFR Antibodies' Efficiency by C/EBPβ-Dependent Suppression of EGFR Expression^,

Abstract

Oncogenic KRAS mutations in colorectal cancer (CRC) are associated with lack of benefit from epidermal growth factor receptor (EGFR)-directed antibody (Ab) therapy. However, the mechanisms by which constitutively activated KRAS (KRAS^G12V) impairs effector mechanisms of EGFR-Abs are incompletely understood. Here, we established isogenic cell line models to systematically investigate the impact of KRAS^G12V on tumor growth in mouse A431 xenograft models as well as on various modes of action triggered by EGFR-Abs in vitro. KRAS^G12V impaired EGFR-Ab-mediated growth inhibition by stimulating receptor-independent downstream signaling. KRAS^G12V also rendered tumor cells less responsive to Fc-mediated effector mechanisms of EGFR-Abs—such as complement-dependent cytotoxicity (CDC) and Ab-dependent cell-mediated cytotoxicity (ADCC). Impaired CDC and ADCC activities could be linked to reduced EGFR expression in KRAS-mutated versus wild-type (wt) cells, which was restored by small interfering RNA (siRNA)-mediated knockdown of KRAS4b. Immunohistochemistry experiments also revealed lower EGFR expression in KRAS-mutated versus KRAS-wt harboring CRC samples. Analyses of potential mechanisms by which KRAS^G12V downregulated EGFR expression demonstrated significantly decreased activity of six distinct transcription factors. Additional experiments suggested the CCAAT/enhancer-binding protein (C/EBP) family to be implicated in the regulation of EGFR promoter activity in KRAS-mutated tumor cells by suppressing EGFR transcription through up-regulation of the inhibitory family member C/EBPβ-LIP. Thus, siRNA-mediated knockdown of C/EBPβ led to enhanced EGFR expression and Ab-mediated cytotoxicity against KRAS-mutated cells. Together, these results demonstrate that KRAS^G12V signaling induced C/EBPβ-dependent suppression of EGFR expression, thereby impairing Fc-mediated effector mechanisms of EGFR-Abs and rendering KRAS-mutated tumor cells less sensitive to these therapeutic agents.

Introduction

The epidermal growth factor receptor (EGFR) is among the most widely expressed tumor-related antigens [1] and is successfully targeted in patients by both tyrosine kinase inhibitors (TKIs) and monoclonal antibodies (mAbs) [2–4]. For the last years, important progress has been made in identifying molecular biomarkers that predict response or resistance to EGFR inhibitors in colorectal cancer (CRC) [5]. Resistance to EGFR inhibitors has been associated with activating mutations of KRAS [6–9] and other mediators of downstream signaling such as v-raf murine sarcoma viral oncogene homolog B1 (B-RAF) [6,8], phosphoinositide-3-kinase, phosphatase and tensin homolog (PTEN), and others [5,10]. Nevertheless, the expression of the EGFR ligands epiregulin and amphiregulin was associated with improved response rates [11]. However, the molecular mechanisms underlying these clinical observations are incompletely understood.

KRAS belongs to the family of three RAS proto-oncogenes encoding five small monomeric guanosine triphosphatases (GTPases) (NRAS, HRAS1, HRAS2, KRAS4a, and KRAS4b). RAS proteins have intrinsic GTPase activity—enabling them to switch between inactivated, guano-sine diphosphate-bound and activated, GTP-bound states. Thereby, they mediate ligand-induced signal transduction by receptor tyrosine kinases like the EGFR [12,13]. Importantly, distinct somatic point mutations in RAS genes (commonly restricted to codon 12, 13, or 61) turn these proto-oncogenes into oncogenes by affecting their intrinsic GTPase activity—thereby forcing constitutive activation of RAS proteins and stimulating downstream signaling pathways such as mitogen-activated protein kinases (MAPKs) and phosphoinositide-3-kinase [14].

Recent studies investigated the frequency of RAS gene point mutations in distinct solid tumors and demonstrated the KRAS gene to be more frequently mutated than NRAS or HRAS—mainly in tumors of the pancreas, the lung, or the colon [15]. Furthermore, clinical trials studying therapeutic efficacy of EGFR-directed antibodies like cetuximab or panitumumab showed that patients experiencing KRAS-mutated colorectal tumors do not respond to these agents [6,8,9]. This led to label restrictions for EGFR-Abs in CRC patients bearing wild-type (wt) KRAS tumors [16].

EGFR-Abs are able to elicit distinct effector mechanisms for tumor cell destruction: Fab-mediated effects comprise inhibition of ligand-binding or tumor cell growth, apoptosis induction, as well as EGFR down-modulation, whereas Fc-mediated effector mechanisms are triggered through the Fc region by binding either complement component C1q to induce complement-dependent cytotoxicity (CDC) or Fc receptors on effector cells to trigger Ab-dependent cell-mediated cytotoxicity (ADCC) or phagocytosis [4]. Both Fab- and Fc-mediated effector mechanisms have been suggested to be important for therapeutic outcome of EGFR-Abs [3]—with recent studies demonstrating a strong impact of Fc-mediated effector mechanisms on the efficacy of EGFR-Abs [17]. Furthermore, distinctgenetic polymorphisms that determine the binding affinity and ADCC efficacy for FcγRIIa (131 H/R) and FcγRIIIa (158 V/F) have been described and were linked to clinical outcome of therapeutic EGFR-Abs in metastatic CRC (mCRC) [18]. Considering these findings, it might be hypothesized that Fc-mediated mechanisms of EGFR-Abs play important roles in tumor cell destruction. However, whether oncogenic mutations in the KRAS gene directly affect ADCC or CDC activity, as described for Fab-mediated effector mechanisms [19], has—to our knowledge—not been previously investigated.

In this study, we observed that oncogenic KRAS^G12V signaling is accompanied by down-regulation of EGFR transcript and protein levels in a C/EBPβ-dependent manner. Decreased EGFR cell surface expression was accompanied by diminished ADCC as well as CDC and might in part explain the lack of efficacy of EGFR-Abs in the therapy for KRAS-mutated tumors.

Materials and Methods

Blood Donors

Experiments reported here were approved by the Ethics Committee of the Christian-Albrechts-University, Kiel, Germany, in accordance with the Declaration of Helsinki. Blood donors were randomly selected from healthy volunteers, who gave written informed consent before analyses.

Cell Lines and Antibodies

Human epidermoid carcinoma cell line A431 (DSMZ, Braunschweig, Germany) was kept in RPMI 1640 medium, whereas A549 (ECACC, Salisbury, UK), HCT-116 (DSMZ), and A1207 (originally established by Dr Aaronson, National Cancer Institute, National Institutes of Health, Bethesda, MD [20]) cell lines were cultivated in Dulbecco modified Eagle medium. Cell culture media were supplemented with 10% (vol/vol) heat-inactivated fetal calf serum, 100 U/ml penicillin, and 100 µg/ml streptomycin.

EGFR-directed antibodies used in this study: C225 (cetuximab; chimeric immunoglobulin [Ig] G1; Erbitux; Merck, Darmstadt, Germany), 2F8 (zalutumumab; human IgG1; Genmab, Utrecht, the Netherlands), and HuMab 003 (human IgG1; Genmab). VH and VL regions of the humanized EGFR-directed IgG1 antibody (Ab) H425 (source: PCT application WO 92/15683) were cloned into an expression vector harboring an IgG1 Fc variant with amino acid exchanges K326A and E333A in the CH2 domain (=H425-E3), known to increase C1q binding [21]. An irrelevant human IgG1 Ab (Alpha Diagnostic Intl. Inc, San Antonio, TX) or an irrelevant mouse Ab served as control antibodies.

Plasmid Construction

Expression plasmid for KRAS4b^G12V was generated as described previously [19]. The wt promoter region of EGFR, spanning nucleotides -620 to -1000, as well as a mutated version lacking the predicted C/EBP binding site (ATTGG and GCAAT) were synthesized by Entelechon GmbH (Regensburg, Germany). For bioinformatic analysis of the EGFR promoter sequences regarding C/EBP binding, the TFSEARCH software (Computational Biology Research Center, AIST, Japan http://www.cbrc.jp/research/db/TFSEARCH.html) was used. The promoter regions were inserted into the pGL3Enhancer vector (Promega, Madison, WI) by using the restriction sites KpnI and BglII. The resulting constructs were termed pGL3E-EGFR prom wt and pGL3E-EGFR prom mut.

Immunoprecipitation

Activated RAS^GTP was precipitated as previously described [19].

RNA Isolation and Reverse Transcription-Polymerase Chain Reaction

Total RNA was isolated using the RNeasy kit (Qiagen, Inc, Valencia, CA) following the manufacturer's instructions. Reverse transcription was performed by using iScript complementary DNA (cDNA) Synthesis Kit (Bio-Rad Laboratories, Munich, Germany). The expression of EGFR, KRAS4b, β-actin, or G3PDH was assayed using standard semiquantitative reverse transcription-polymerase chain reaction (RT-PCR) procedures and following sequence specific primers: EGFR sense 5′-GTGAGTTGATCATCGAATTCTC-3′ and antisense 5′-CATGCTCCAATAAATTCACTGC-3′, KRAS4b sense 5′-ATGACTGAATATAAACTTGTGG-3′ and antisense 5′-CCATCTTTGCTCATCTTTTC-3′, β-actin sense 5′-GATGGTGGGCATGGGTCAG-3′ and antisense 5′-CTTAATGTCACGCACGATTTCC-3′, and G3PDH sense 5′-TGAAGGTCGGAGTCAACGGATTTGGT-3′ and antisense 5′-CATGTGGGCCATGAGGTCCACCAC-3′.

SDS-PAGE and Immunoblot Analysis

Whole protein extracts were prepared by lysing cell pellets in denaturing lysis buffer containing 1% SDS, 10 mM Tris (pH 7.4), and 1% protease inhibitor mixture (Complete Protease Inhibitor Cocktail; Roche Applied Science, Mannheim, Germany). Nuclear protein extracts were prepared from 1 x 10⁷ cells according to the manufacturer's instructions using the Nuclear Extraction Kit (Signosis, Inc, Sunnyvale, CA). Ten micrograms of protein extracts was separated by denaturing SDS-PAGE and transferred onto polyvinylidene fluoride membranes (Millipore, Billerica, MA). After blocking, membranes were probed with specific primary Ab, washed, and incubated with HRP-conjugated IgG as secondary Ab. Proteins were visualized by chemiluminescence (Thermo Fisher Scientific, Rockford, IL). To determine even transfer and equal loading, membranes were stripped and reprobed with antibodies specific for nonphosphorylated protein or β-actin/G3PDH.

Immunofluorescence Microscopy

A431-control-vector or A431-KRAS^G12V cells were seeded on sterile coverslips. Next day, cells were washed, fixed in 4% paraformaldehyde-phosphate-buffered saline and blocked for 1 hour in 0.75% bovine serum albumin-phosphate-buffered saline. Cells were stained for EGFR cell surface expression using C225 (2 µg/ml) and goat phycoerythrin-conjugated F(ab′)₂ fragments against human IgG (Jackson ImmunoResearch, Suffolk, England) as secondary Ab. 4′,6-diamidino-2-phenylindole (DAPI) was used for DNA counterstaining. Coverslips were mounted onto glass slides and examined using a Zeiss AxioImager.Z1 apotome fluorescence microscope and the AxioVision Imaging software (Carl Zeiss MicroImaging, Inc, Thornwood, NY).

Flow Cytometric Analyses

For indirect immunofluorescence, cells were stained as described previously [19]. Relative fluorescence intensity (RFI) was calculated with the following formula: mean fluorescence intensity (MFI) EGFR-Ab / MFI control Ab. Quantitative surface EGFR expression was determined using a murine EGFR-Ab, a murine IgG1 Ab as a control Ab, and the QIFIKIT (Dako, Glostrup, Denmark) according to the manufacturer's instructions except for using PerCP-conjugated (Fab′)₂ fragments (Jackson ImmunoResearch, England). Samples were analyzed on a flow cytometer (Epics Profile; Beckman Coulter, Fullerton, CA).

Determination of Viable Cell Mass (MTS Assay)

The CellTiter 96 nonradioactive cell proliferation assay was performed using A431-control-vector or A431-KRAS^G12V cells (5 x 10³ cells per well in 96-well microtiter plate) according to the manufacturer's instruction (Promega).

Cytotoxicity Assays

As previously described, human effector cells such as peripheral blood mononuclear cells (MNCs), NK cells, monocytes, or polymorphonuclear cells (PMNs) were isolated from peripheral blood drawn from healthy volunteers and assayed in ADCC experiments [22].

CDC assays were performed as described in Dechant et al. [20] using freshly drawn human plasma (25% vol/vol), anticoagulated with 100 U/ml heparin, and the Ab H425-E3 alone or HuMab 003 (final concentration of 5 µg/ml) in combination with C225, 2F8, or control Ab at various concentrations was added to microtiter plates.

Percentage of cytotoxicity was calculated using the following formula: percent specific lysis = [(experimental cpm - basal cpm) / (maximal cpm ? basal cpm)] x 100.

Small Interfering RNA-Mediated Knockdown Experiments

Cells were seeded at a density of 1 x 10⁵/well in six-well plates. Next day, transfection of small interfering RNA (siRNA) was carried out using Lipofectamine2000 (Invitrogen, Carlsbad, CA). For all procedures, standard protocols were used according to the manufacturers' manuals.

Synthetic siRNAs targeting KRAS4b or C/EBPβ were purchased from Applied Biosystems/Ambion (Carlsbad, CA). Target sequences were as follows: KRAS siRNA 1 (ID s7939) sense 5′-CUAUGGUCCUAGUAGGAAAtt-3′ and antisense 5′-UUUCCUACUAGGACCAUAGgt-3′; KRAS siRNA 2 (ID s7940) sense 5′-GCCUUGACGAUACAGCUAAtt-3′ and antisense 5′-UUAGCUGUAUCGUCAAGGCac-3′; and C/EBPβ (ID s2892) sense 5′-GGCCCUGAGUAAUCGCUUAtt-3′ and antisense 5′-UAAGCGAUUACUCAGGGCCcg-3′. As a control, unspecific siRNA Negative Control 1 (Applied Biosystems/Ambion) was used.

Transcription Factor Activity Profiling Assay

Analysis of activity of 48 transcription factors (TFs) was performed according to the manufacturer's instructions using the TF Activation Profiling Plate Array I (Signosis). Ten micrograms of nuclear protein extracts was assayed per sample. TFs were selected by the fold-change method (≤-1.5 or ≥1.5) and the rank sum difference (≥7) of A431-KRAS^G12V cells compared with A431-control-vector cells.

Dual Luciferase Reporter Gene Assay

To quantify C/EBP activation, a dual-luciferase reporter gene assay (Promega) was used according to the manufacturer's instructions. A431-control-vector or A431-KRAS^G12V cells were seeded at a density of 1.5 x 10⁴ cells per well and grown until 50% to 70% confluence on 96-well plates. The next day, cells were transfected with 120 ng/well of cis-reporting pC/EBP_Luc plasmid (Agilent Technologies, La Jolla, CA), 40 ng/well of pRL-TK (Promega) reference plasmid, and 40 ng/well of pSec empty vector for 24 hours. C/EBP activation was calculated as relative light units (RLU) by normalizing firefly luciferase activity to reference activity (Renilla luciferase). For quantification of EGFR promoter transactivation, A431-control-vector or A431-KRAS^G12V cells were seeded at a density of 1.5 x 10⁴ cells per well and grown until 50% to 70% confluence on 96-well plates. Next day, cells were transfected with 40 ng/well of pGL3E, pGL3E-EGFR prom wt, or pGL3E-EGFR prom mut and 20 ng/well of pRL-TK reference plasmid for 48 hours. Dual luciferase assay was performed using the Dual-Luciferase Reporter Assay System (Promega) and Tecan GeniosPro microplate reader (Tecan Trading AG, Crailsheim, Germany).

TaqMan Real-time PCR

For real-time PCR experiments, cDNA derived from four colorectal carcinoma biopsy samples (nos. 3–6: no. 3 = human colon 2 matched cDNA pair [Takara Bio Europe/Clontech, Saint-Germain-en-Laye, France] and nos. 4–6 = RNA [AMS Biotechnology Europe Ltd, Bioggio-Lugano, Switzerland]) as well as two biopsy samples from normal colon tissue (nos. 1 and 2: no. 1 = human MTC panel II [Takara Bio Europe/Clontech] and no. 2 = human colon 2 matched cDNA pair [Takara Bio Europe/Clontech]) were included in this study. All six samples were analyzed by Sanger sequencing experiments regarding KRAS status (Figure 7A).

Figure 7.

EGFR expression is decreased in KRAS-mutated but not in KRAS wt CRC biopsy samples. (A and B) EGFR mRNA expression in colonic biopsies from healthy and tumor tissue was determined by real-time PCR. In comparison to biopsies from healthy tissues, KRAS-mutated CRC sample revealed significant down-regulation of EGFR expression, whereas an increase in EGFR expression was determined in two of three KRAS wt CRC samples. Means ± SEM are presented from quintuplicates. ^#P ≤ .05, CRC samples versus normal colon tissue; *P ≤ .05, no. 5 or 6 versus no. 3 or 4. (C and D) EGFR-IHC on paraffin-embedded colon tumors. Paraffin-embedded CRC samples (n = 7 KRAS codon 12 mutated; n = 10 KRAS wt) were stained using the EGFR pharmDx kit and evaluated by IHC scoring (0–300; formula for calculation is depicted in the Materials and Methods section) with parameters as follows: 0 = no staining of cells, 1+ = percentage of cells stained weakly, 2+ = percentage of cells stained moderately, 3+ = percentage of cells stained strongly. Left y axis = frequency of staining (= percentage of stained cells); right y axis = calculated IHC score. (E) Model of oncogenic KRAS^G12V signaling and its impact on the efficiency of EGFR-directed antibodies. In comparison to wt KRAS expressing cells, oncogenic KRAS^G12V-expressing cells display strong down-regulation of EGFR expression and decreased activity of the family of C/EBP TF. The family member C/EBPβ is upregulated by constitutively activated KRAS^G12V signaling via MAPK p44/42 and could be shown to be involved in EGFR down-regulation. It might be hypothesized that the induction of the inhibitory isoform C/EBPβ-LIP in KRAS^G12V-expressing cells is involved in the suppression of C/EBP TF activity. Because the family of C/EBP TF is involved in the regulation of EGFR promoter activity, suppression of C/EBP activity is accompanied by down-regulation of EGFR transcription. Hence, decreased expression of EGFR molecules on the cell surface impairs EGFR-directed Ab-mediated effector mechanisms.

EGFR messenger RNA (mRNA) transcript levels were measured using quantitative real-time PCR. cDNA was arrayed on 384-well plates using the expression assay Hs01076078_m1 (Applied Biosystems, Foster City, CA) on the ABI Prism 7900HT Sequence Detection System (Applied Biosystems) according to the manufacturer's instructions. Relative transcript levels were determined using G3PDH (Hs99999905_m1) as the endogenous control gene.

Immunohistochemical Analysis of Human Colorectal Carcinoma Samples

Paraffin-embedded human tumor samples were provided by the Pathology “Nordhessen,” Kassel, Germany, in accordance with local ethical guidelines. KRAS codon 12 mutation status was determined by melting point analysis using the cobas z 480 system (Roche). EGFR-immunohistochemistry (IHC) was performed using the EGFR pharmDx kit on the Link 48 Autostainer (both by Dako) according to the manufacturer's protocol. Slices were cut at 4 µm, and incubation times were as follows: peroxidase blocking, 5 minutes; proteinase K digestion, 5 minutes; primary Ab (clone 2-18C9), 30 minutes; HRP-conjugated secondary Ab, 30 minutes; Tris-buffered saline supplemented with 0.05% (vol/vol) Tween-20 (TBST) 0.05%, 5 minutes; DAB chromogen, 5 minutes; hematoxylin counterstain, 5 minutes. Intensity evaluation by IHC scoring (score 0–300) was done as described in various studies on EGFR expression [23,24]. By integration of the data relating to the intensity and frequency of staining, the IHC score was calculated with the formula: 1 x (percentage of cells staining weakly [1+]) + 2 x (percentage of cells staining moderately [2+]) + 3 x (percentage of cells staining strongly [3+]). To avoid bias, the scoring pathologist was blinded about the KRAS status of the samples.

Mouse Tumor Xenograft Models

SCID mice (C.B.-17/IcrCrl-scid/scid) were purchased from Charles River (Maastricht, the Netherlands). All experiments were performed with 8-to 12-week-old female mice. Mice were housed in a barrier unit of the Utrecht University Central Laboratory Animal Facility (Utrecht, the Netherlands) and kept in filter-top cages with water and food provided ad libitum. Mice were checked at least twice a week for clinical signs of disease and discomfort. All experiments were approved by the Utrecht University animal ethics committee. Subcutaneous tumors were induced by subcutaneous inoculation of 5 x 10⁶ A431 cells or 5 x 10⁶ A431-KRAS4b^G12V cells in the right flank of mice. Tumor volumes were calculated from digital caliper measurements as 0.52 x length x width² (mm³). For the A431 cells, the 95% confidence interval (CI) of the tumor development was calculated over seven independent experiments, comprising a total of 53 mice. Tumor growth data from each mouse were fit to a monoexponential curve, determined by y⁰ and k. Subsequently, the average y⁰, k, and the 95% CI were calculated. These data were plotted against the average tumor growth of 5 x 10⁶ A431-KRAS4b^G12V cells in five mice.

Data Processing and Statistical Analyses

Data are displayed graphically and were statistically analyzed using GraphPad Prism 4.0 (GraphPad Software, La Jolla, CA). Curves were fitted using a nonlinear regression model with a sigmoidal dose response (variable slope). Statistical significance was determined by the Student's t test (paired, 2-tailed; Figures 1C, 4C, and 5B) or by the two-way analysis of variance repeated-measures test with the Bonferroni post test (Figures 2, 3, 4B, 5D, and 6). The respective results were displayed as mean ± SEM of at least three independent experiments. P values were calculated, and null hypotheses were rejected when P ≤ .05.

Figure 1.

KRAS^G12V-transfected cells display increased MAPK activation and cell growth. (A) Human KRAS4b in complex with GTP. The presented models illustrate the position of the G12V mutation, which is located in the GTP-binding pocket—ranging from amino acids 10 to 16. Human KRAS4b crystal structure pdb file 2PMX. The picture was generated using three-dimensional Mol Viewer software (a component of Vector NTI Advance 10.3.0; Invitrogen): top (I) and bottom (II) views. Blue indicates backbone (upper panels) or calculated surface (lower panels); red, GTP analog; green, amino acid substitution G12V analyzed in the present study. (B) To ensure that over-expression of KRAS^G12V is accompanied by expression of constitutively activated KRAS, activated RAS was precipitated and analyzed by SDS-PAGE and immunoblot analysis against RAS and Myc-tag (left panel). Whole-protein lysates served as expression control for endogenous RAS and transfected KRAS^G12V (right panel). (C) Overexpression of KRAS^G12V was accompanied by constitutive activation of MAPK p44/42. Whole protein extracts were prepared from A431-control-vector or A431-KRAS^G12V cells and analyzed by SDS-PAGE and immunoblot analysis against phosphorylated and unphosphorylated p44/42. Immunoblot analysis against G3PDH served as loading control. (D) Overexpression of oncogenic KRAS^G12V induced enhanced cell growth rate in 72 hours MTS assays. Data are presented as mean ± SEM of three independent experiments. * P ≤ .05 for A431-KRAS^G12V cells versus A431-control cells. (E) Overexpression of oncogenic KRAS^G12V induced enhanced tumor growth in vivo. Data for A431-KRAS^G12V represent an average and SD (n = 5). Data for A431 cells represent an average and 95% CI (n = 53) and calculated as described in the Materials and Methods section. The A431-KRAS^G12V growth rate lies outside the 95% CI of the A431 cells, indicating that the difference is significant.

Figure 4.

EGFR expression is downregulated in KRAS^G12V-transfected cells. (A) EGFR transcripts were downregulated in A431-KRAS^G12V cells (upper panel). Total RNA was isolated from A431-control-vector and A431-KRAS^G12V cells. mRNA expression of EGFR and G3PDH was assayed by RT-PCR and visualized by agarose gel electrophoresis. Diminished EGFR mRNA expression was associated with lower EGFR protein levels (lower panel). Membranous protein extracts were prepared from A431-control-vector and A431-KRAS^G12V cells. Protein extracts were separated by SDS-PAGE and immunoblotted against EGFR. Loading of equal protein amounts was controlled with β-actin Ab. (B) EGFR protein expression is also downregulated in A431 xenograft models. Whole protein extracts were prepared from explanted A431 wt and A431-KRAS^G12V tumors, separated by SDS-PAGE and immunoblotted against EGFR and β-actin as the loading control. (C and D) Diminished EGFR cell surface expression in A431-KRAS^G12V cells is reflected by lower C225 and 2F8 binding levels at Ab saturating concentrations. (C) For indirect immunofluorescence, A431-control-vector and A431-KRAS^G12V cells were incubated with increasing concentrations of C225, 2F8 or a control Ab and stained with polyclonal rabbit PerCP-conjugated F(ab′)₂ fragments against human IgG. Each data point represents the RFI of C225 or 2F8 versus control IgG1 Ab. (D) A431-control-vector cells or A431-KRAS^G12V cells were fixed, incubated with C225 and goat phycoerythrin-conjugated F(ab′)₂ fragments against human IgG and counterstained using DAPI. (E) Quantitative analyses of EGFR molecules per cell were performed using QIFIKIT. Cells were incubated with a murine EGFR or control Ab and stained with goat PerCP-conjugated F(ab′)₂ fragments against murine IgG. (F) siRNA-mediated knockdown of KRAS4b in nontransfected A431 cells. Cells were transfected with two different KRAS4b-specific or control siRNA. After 72 hours, cDNA was prepared, and mRNA expression of KRAS4b and G3PDH was assayed by RT-PCR. (G and H) Knockdown of KRAS4b induced up-regulation of EGFR. A431-KRAS^G12V cells were transfected with KRAS4b-specific or control siRNA for up to 72 hours and analyzed by RT-PCR (G) and Western blot (H). Down-regulation of KRAS4b mRNA level (G) as well as RAS protein level (H) was detected, whereas EGFR mRNA and protein expressions were time-dependently upregulated. Means ± SEM of three independent experiments are presented in C and E. *P ≤ .05 for A431-KRAS^G12V cells (C225) versus A431-control cells (C225); ^#P ≤ .05 for A431-KRAS^G12V cells (2F8) versus A431-control cells (2F8).

Figure 5.

The family of C/EBP transcription factors displays decreased activity in A431-KRAS^G12V cells and is involved in the suppression of the EGFR promoter. (A) Table providing an overview of significantly regulated TF in A431-KRAS^G12V cells versus A431-controlvector cells determined by TF activity profiling assay. (B) A reporter gene assay for C/EBP TF family activation revealed decreased C/EBP activation in A431-KRAS^G12V cells. A431-control-vector or A431-KRAS^G12V cells were cotransfected with a C/EBP-dependent luciferase reporter gene construct and a pRL-TK Renilla plasmid for normalization. Luciferase activity was measured and presented as RLU. (C) Model of EGFR promoter constructs. EGFR promoter sequences representing wt or mutated sequences are presented. The mutated EGFR promoter sequence lacks putative C/EBP DNA-binding sites. For generation of EGFR promoter constructs, a sequence of 381 nucleotides (-620 to -1000 nt; presented are -620 to -754 nt) upstream of the EGFR coding sequence was inserted into the Firefly luciferase reporter gene plasmid pGL3Enhancer. The resulting constructs were termed EGFR prom wt and EGFR prom mut. (D) The family of C/EBP TF inhibited EGFR promoter activity. Dual luciferase assays with reporter gene constructs representing the wt or mutated EGFR promoter were used to quantify promoter activity. A431-control-vector or A431-KRAS^G12V cells were transfected with EGFR promoter constructs and a pRL-TK Renilla plasmid for normalization. Luciferase activity was measured and calculated as RLU. Data are presented as mean ± SEM of triplicate wells from at least three independent experiments. *P ≤ .05.

Figure 2.

Oncogenic KRAS renders cells unresponsive to EGF stimulation and impairs proliferation inhibition by EGFR mAb. (A) A431 cells were seeded into 96-well microtiter plates, incubated in the presence or absence of rhEGF (left panel, 0.001–100 ng/ml for 72 hours; right panel, 10 ng/ml for 24–72 hours), and analyzed by MTS assays. Data are presented as mean ± SEM from at least three independent experiments. *P ≤ .05 stimulated cells versus unstimulated cells, ^#P ≤ .05 for A431-KRAS^G12V cells versus A431-controlvector cells. (B) EGFR Ab-mediated growth inhibition was impaired in KRAS^G12V-transfected cells. A431 cells were incubated in the presence or absence of C225 (left), 2F8 (middle), or control Ab (right; 0.0128–40 µg/ml). After 72 hours, cells were analyzed by MTS assays. Data are presented as mean ± SEM from at least three independent experiments. *P ≤ .05 C225/2F8 versus control Ab, ^#P ≤ .05 for A431-KRAS^G12V cells versus A431-control-vector cells.

Figure 3.

Tumor cell lysis mediated by complement or ADCC is diminished in KRAS^G12V-transfected cells. For cytotoxicity experiments, A431 control or KRAS^G12V cells were incubated with increasing concentrations of C225, 2F8 or control Ab and (A) NK cells (E/T ratio 10:1), (B) monocytes (E/T ratio 40:1), (C) PMN (E/T ratio 80:1), or (D) plasma (25% vol/vol). For CDC experiments, a second noncompeting EGFR-Ab was added at 5 µg/ml. Data are presented as mean ± SEM from at least three independent experiments with different donors. * P ≤ .05 C225/2F8 versus control Ab, ^#P ≤ .05 for A431-KRAS^G12V cells versus A431-control-vector cells.

Figure 6.

C/EBPβ is induced by oncogenic KRAS and is involved in the regulation of EGFR expression. (A) Expression of oncogenic KRAS was accompanied by regulation of C/EBP TF family members' nuclear protein expression. Nuclear protein fractions were prepared from A431-KRAS^G12V and A431-control-vector cells. Ten micrograms of protein extracts was separated by SDS-PAGE and immunoblotted against C/EBPβ/δ//γ/ζ or β-actin—serving as a protein loading control. Arrows indicate the predicted molecular weights of respective TF. One of several independent experiments is presented. (B) Densitometric analysis of C/EBPβ specific immunoblot presented in A was performed using ImageJ software. (C) C/EBPβ expression was controlled by KRAS. A431-KRAS^G12V cells were transfected with 25 nM KRAS4b-specific or control siRNA for 72 hours. Nuclear protein extracts were separated by SDS-PAGE and immunoblotted against C/EBPβ. (D and E) RNAi-induced knockdown of C/EBPβ enhanced EGFR expression. A431-KRAS^G12V (upper panel) or A549 (lower panel) cells were transfected with 25 nM control siRNA or C/EBPβ siRNA. After 24 and 48 hours, cDNA was prepared and mRNA expression of EGFR, β-actin, and G3PDH was assayed by RT-PCR. (F) Whole protein extracts of A431-KRAS^G12V cells, transfected with C/EBPβ-specific siRNA, were prepared after 48 hours of transfection, separated by SDS-PAGE and immunoblotted against C/EBPβ, EGFR, and β-actin. (G) RNAi-induced knockdown of C/EBPβ enhanced EGFR-directed Ab-mediated cytotoxicity. A431-control-vector or A431-KRAS^G12V cells were transfected with 25 nM control siRNA or C/EBPβ-specific siRNA for 48 hours. Plasma served as effector source in ⁵¹Cr release assays using H425-E3 Ab (0–1 µg/ml). (G and H) A431 KRAS^G12V cells were transfected with a plasmid coding for human EGFR for 24 hours by lipofection. After transfection, cells were harvested and either analyzed by SDS-PAGE and immunoblot experiments (G) or assayed in 3 hours of ⁵¹Cr release assays using plasma and the indicated antibodies (both at 1 µg/ml; H). Data are presented as mean ± SEM of triplicate wells from at least three independent experiments with different blood donors. *P ≤ .05.

Densitometric analysis of immunoblots received from the detection of proteins by chemiluminescence was performed using the non-commercial software ImageJ (http://rsb.info.nih.gov/ij/).

Results

Constitutively Activated KRAS^G12V Stimulates MAPK p44/42 and Increases Viable Cell Mass

Intrinsic GTPase activity of KRAS4b facilitates the switch between activated (binding of GTP) and inactivated state (binding of guanosine diphosphate). The amino acid substitution G12V impairs GTPase activity—leading to the constitutive activation of KRAS4b (Figure 1A). To ensure that overexpression of oncogenic KRAS4b^G12V in A431 cells led to constitutive activation of KRAS and downstream signaling pathways, RAS and MAPK p44/42 activation was analyzed by immunoprecipitation and immunoblot experiments. The expression of exogenous KRAS4b^G12V led to constitutive activation of KRAS4b^G12V (Figure 1B), which triggered enhanced phosphorylation of MAPK p44/42 (Figure 1C) that has been also detected in KRAS-mutated CRC cell lines such as SW480 and HCT-116 (data not shown). Furthermore, A431-KRAS^G12V cells displayed significantly increased cell growth rates when compared with A431-control-vector cells in vitro (∼45% increase; Figure 1D). The growth rate of A431-KRAS^G12V cells in vivo as a xenograft was also significantly increased compared with A431-WT cells (average A431-KRAS^G12V growth rate lies outside the 95% CI of A431-WT; Figure 1E).

KRAS^G12V Prevents EGF-Induced Reduction of Cell Viability and Impairs Growth Inhibition by EGFR Antibodies

A431 cells have been shown to undergo cell death when stimulated with high concentrations of EGF [25]. To analyze whether constitutively activated KRAS signaling in A431-KRAS^G12V cells impaired ligand-induced signal transduction, A431-control-vector and A431-KRAS^G12V cells were stimulated with recombinant human (rh) EGF at increasing concentrations (0.001–100 ng/ml; 72 hours), for different time intervals (24, 48, and 72 hours; 10 ng/ml), or were left untreated (Figure 2A). After incubation, viable cell mass was analyzed by MTS assay. Viable cell mass was significantly reduced by EGF stimulation in a concentration- and time-dependent manner in A431-control-vector cells (∼70% at 100 ng/ml, 72 hours) but not in A431-KRAS^G12V cells.

Owing to the observed unresponsiveness of A431-KRAS^G12V cells to EGF-induced cell death at high EGF concentrations, further experiments were performed to study the impact of proliferative signaling cascades triggered by oncogenic KRAS on EGFR-Ab-mediated cell growth inhibition. Hence, cells were incubated for 72 hours in the presence or absence of increasing concentrations (0–40 µg/ml) of C225, 2F8, or an irrelevant control human IgG1 Ab and analyzed by MTS assays (Figure 2B). Whereas growth of A431-control-vector cells was significantly inhibited by C225 (∼25%) or 2F8 (∼25%) in a concentration-dependent manner when compared with control Ab-treated cells, only modest—albeit also statistically significant—inhibition was detected for A431-KRAS^G12V cells by C225 (∼13%) or 2F8 (∼11%).

ADCC as well as Complement-Mediated Tumor Cell Lysis Is Diminished in A431 KRAS^G12V Cells

In mouse A431 wt and KRAS^G12V xenograft models, also used in the present study, activation of ADCC has been proposed to be an important mechanism of action of EGFR-mAb such as zalutumumab (2F8) in early stages of tumor development [26]. However, the relative contribution of different effector cell populations remains to be unraveled. Therefore, A431-KRAS^G12V and A431-control-vector cells were analyzed regarding Fc-mediated effector mechanisms induced by C225 and 2F8. ADCC experiments were performed with isolated NK cells, monocytes, or PMN effector cells. A slight but significant (only for 2F8) decrease in ADCC activity was detected with NK cells (A431-KRAS^G12V maximum lysis at 10 µg/ml [ML]: C225 ≈ 38%, 2F8 ≈ 34%; A431-control-vector cells ML: C225 ≈ 45%, 2F8 ≈ 46%; Figure 3A). Furthermore, significantly diminished monocyte- and PMN-mediated tumor cell lysis was observed in A431-KRAS^G12V cells (Figure 3, B and C; ML monocytes: C225 ≈ 23%, 2F8 ≈ 18%; ML PMN: C225 ≈ 14%, 2F8 ≈ 18%) when compared with A431-control-vector cells (ML monocytes: C225 ≈ 35%, 2F8 ≈ 35%; ML PMN: C225 ≈ 25%, 2F8 ≈ 38%; Figure 3, B and C). A431-KRAS^G12V cells also displayed significantly impaired cell lysis compared with A431-control-vector cells in CDC experiments with C225 and 2F8 using a second noncompeting EGFR-mAb and human plasma as effector source. Whereas strong CDC activity was induced in A431-control-vector cells by C225 (ML C225 ≈ 35%) and 2F8 (ML 2F8 ≈ 26%), only low killing of A431-KRAS^G12V cells was observed (ML C225 ≈ 11%; 2F8 ≈ 6%; Figure 3C). Because CDC activity depends on the expression of complement-inhibitory proteins like CD46 or CD59, we quantified cell surface expression of these molecules on A431-KRAS^G12V and A431-control-vector cells. In A431-KRAS^G12V cells, significantly lower levels of CD46 and CD59 cell surface expression (CD46 0.8 x 10⁵ ± 0.02 x 10⁵ molecules per cell; CD59 3.1 x 10⁵ ± 0.15 x 10⁵ molecules per cell) were detected compared with A431-control-vector cells (CD46 1.9 x 10⁵ ± 0.43 x 10⁵ molecules per cell; CD59 4.7 x 10⁵ ± 0.15 x 10⁵ molecules per cell; Figure W1). Hence, it may be concluded that diminished CDC activity in A431-KRAS^G12V cells is not the result of enhanced expression of complement-inhibitory proteins.

Expression of Oncogenic KRAS^G12V Is Accompanied by Down-regulation of EGFR

KRAS is an established downstream mediator of EGFR signal transduction and overexpression of constitutively activated mutant (KRAS^G12V) hampered cell growth inhibition as well as ADCC and CDC induced by EGFR-Ab. Because Ab-mediated cytotoxicity correlated with antigen cell surface expression levels [27], we investigated whether expression of EGFR was altered in KRAS^G12V-transfected cells. Interestingly, EGFR transcript and protein levels were downregulated in A431-KRAS^G12V cells when compared with A431-control-vector cells (Figure 4A). As shown in Figure 4B, decreased EGFR expression was accompanied by decreased binding of EGFR-Abs. Significantly lower fluorescence signal intensities for C225 and 2F8 were detected by indirect immunofluorescence in A431-KRAS^G12V cells (RFI_{C225; 200 µg/ml} = 13.7; RFI_{2F8; 200 µg/ml} = 13.9) compared with A413-control-vector cells (RFI_{C225: 200 µg/ml} = 89.1; RFI_{2F8: 200 µg/ml} = 95.3; Figure 4B). To determine EGFR cell surface expression, A431-KRAS^G12V or A431-control-vector cells were examined by immunofluorescence microscopy (Figure 4C) as well as by quantitative flow cytometry (Figure 4D). A significant decrease in EGFR molecules per cell (∼50%) was detected in KRAS^G12V-transfected cells (1.4 x 10⁶ ± 0.2 x 10⁶ molecules per cell) compared with control cells (2.8 x 10⁶ ± 0.1 x 10⁶ molecules per cell; Figure 4D). To confirm results received from A431 cells in another cell line model, A1207 cells were stably transfected with KRAS4b^G12V (Figure W2A). Quantitative flow cytometry revealed statistically significant down-regulation of EGFR cell surface expression in A1207-KRAS^G12V cells (1.2 x 10⁶ ± 0.2 x 10⁶ molecules per cell) compared with control cells (2.8 x 10⁶ ± 0.2 x 10⁶ molecules per cell; Figure W2B).

To further elucidate the impact of oncogenic KRAS on EGFR expression in A431-KRAS^G12V cells, RNAi-induced knockdown experiments using KRAS4b-specific siRNAs were performed. A decrease in KRAS4b mRNA expression was detected after transfection of A431 cells for 72 hours with two different KRAS4b-specific siRNAs, whereas no regulation was observed using a negative control siRNA (Figure 4E). In contrast to control siRNA-transfected A431-KRAS^G12V cells, RNAi-induced knockdown of KRAS4b in A431-KRAS^G12V cells led to time-dependent up-regulation of EGFR mRNA (Figure 4F) as well as protein expression levels (Figure 4G).

Overexpression of KRAS^G12V Is Accompanied by the Inhibition of C/EBP Family Activity

As demonstrated above, overexpression of KRAS^G12V in A431 cells led to down-regulation of EGFR transcription, which then may impair Fab- and Fc-mediated effector mechanisms of EGFR-mAb. To obtain a more profound insight into mechanisms leading to the down-regulation of EGFR expression, activity of 48 TFs was analyzed in A431-KRAS^G12V and A431-control-vector cells using a TF activity profiling assay. Figure 5A represents data from 7 of 48 TFs that were significantly regulated (fold change ≥ 1.5 and ≤-1.5, rank sum difference ≥ 7) in A431-KRAS^G12V cells when compared with A431-control-vector cells. Among these, the C/EBP family of TFs, AP2, and ATF2 were the most suppressed ones. On the basis of these findings and previous studies, demonstrating a direct link between oncogenic RAS and the C/EBP family member C/EBPβ [28–31], C/EBP was chosen for further analyses. Confirmation of TF activation analysis was obtained from a reporter gene assay for C/EBP (Figure 5B). To further investigate whether the family of C/EBP TFs regulates EGFR promoter activity and controls transcription activity, luciferase reporter gene constructs harboring the wt EGFR promoter sequence or a mutated EGFR promoter sequence that lacks the C/EBP DNA-binding site were prepared (Figure 5C). Dual luciferase assays with reporter gene constructs representing wt or mutated EGFR promoter were performed and revealed significantly decreased activity of the wt EGFR promoter in A431-KRAS^G12V cells (RLU = 12.56 ± 0.80) compared with A431-control-vector cells (RLU = 24.86 ± 2.87). In A431-KRAS^G12V cells, the mutated EGFR promoter construct displayed significantly enhanced activity in comparison to the wt EGFR promoter construct (RLU_{EGFR prom mut} = 22.31 ± 3.95 vs RLU_{EGFR prom wt} = 12.56 ± 0.80), whereas no difference could be observed between both constructs in A431-control-vector cells (RLU_{EGFR prom mut} = 24.15 ± 4.84 vs RLU_{EGFR prom wt} = 24.86 ± 2.87; Figure 5D). Together, these data point to a C/EBP-dependent suppression of EGFR promoter activity in A431-KRAS^G12V cells.

C/EBPβ Is a Suppressor of EGFR Transcription in KRAS-Mutated Cells

The family of C/EBP TFs consists of six distinct TF, termed C/EBPα/β/δ/ε/γ/ζ [32]. To unravel which C/EBP family member might control EGFR expression in A431-KRAS^G12V cells, SDS-PAGE and immunoblot analysis experiments with nuclear protein extracts were performed. From the six known C/EBP family members, C/EBPβ, C/EBPδ, C/EBPγ, and C/EBPζ could be detected by immunoblot analysis using specific primary antibodies (Figure 6A), whereas C/EBPα and C/EBPε could not be detected in these experiments (data not shown). In A431-KRAS^G12V cells, the activating isoforms of C/EBPβ, namely liver-enriched activating proteins (LAP) 1 and 2, were found to be upregulated when compared with control vector cells. Furthermore, the inhibitory isoform of C/EBPβ, liver-enriched inhibitory protein (LIP), was exclusively expressed in A431-KRAS^G12V cells, but not in A431-control-vector cells—leading to a diminished LAP/LIP ratio (A431-control vector cells LAP/LIP ratio = 8.9, A431-KRAS^G12V cells LAP/LIP ratio = 4.4; data determined by densitometric analysis; Figure 6B). Decreased nuclear expression was found for C/EBPδ in A431-KRAS^G12V cells compared with A431-control-vector cells, whereas no regulation could be seen for C/EBPγ and C/EBPζ throughout several independent experiments (Figure 6A). Owing to our observations of strong regulation of C/EBPβ in A431-KRAS^G12V cells in combination with previous studies that demonstrated regulation of C/EBPβ activity by RAS signaling pathways [28,29,33], C/EBPβ was selected for further analyses. In line with these findings, siRNA-mediated knockdown of KRAS4b in A431-KRAS^G12V cells for 72 hours induced down-regulation of C/EBPβ nuclear expression and an increased LAP/LIP ratio (nontransfected LAP/LIP ratio = 8.8, control siRNA LAP/LIP ratio = 5.5, KRAS4b siRNA LAP/LIP ratio = 13.8; data determined by densitometric analysis; data not shown)—confirming a direct association between constitutively activated KRAS and nuclear expression of C/EBPβ (Figure 6C). Furthermore, siRNA-mediated knockdown of C/EBPβ in A431-KRAS^G12V cells led to an increase in LAP/LIP ratio (control siRNA LAP/LIP ratio = 1.0, C/EBPβ siRNA LAP/LIP ratio = 14.4; data determined by densitometric analysis; data not shown) and to time-dependent up-regulation of EGFR mRNA (Figure 6D, upper panel) as well as protein expression (Figure 6E) as also seen in KRAS4b-specific siRNA knockdown experiments (Figure 4, E and F). Up-regulation of EGFR mRNA expression by C/EBPβ-specific siRNA knockdown could be verified in A549 cells, endogenously expressing oncogenic KRAS^G12S (Figure 6D, lower panel). Up-regulation of EGFR expression by C/EBPβ-specific siRNA knockdown was accompanied by significantly increased CDC activity in A431-KRAS^G12V (C/EBPβ siRNA, lysis rate_{H425-E3; 1 µg/ml} = 12.0% ± 0.9%; control siRNA, lysis rate_{H425-E3; 1 µg/ml} = 6.5% ± 1.8%) but not in A431-control-vector cells (Figure 6F). Enhanced Ab-mediated cytotoxicity could also be observed in C/EBPβ knockdown experiments with the non-small cell lung carcinoma (NSCLC) cell line A549 as well as with the CRC cell line HCT-116 (both KRAS mutated; data not shown). To directly link upregulated EGFR cell surface expression to increased EGFR-Ab-induced cytotoxicity in A431-KRAS^G12V cells, human EGFR was overexpressed in A431-KRAS^G12V cells for 24 hours (Figure 6G). After transfection, cells were analyzed in ⁵¹Cr release assays, which revealed a significant increase in CDC activity only in A431-KRAS^G12V cells transfected with human EGFR but not with a mock vector (Figure 6H).

EGFR Expression Is Downregulated in KRAS-Mutated but Not in KRAS-wt CRC Samples

To support results received from experiments with KRAS-mutated A431 cells, four CRC biopsy samples and two normal colonic biopsy samples (Figure 7A) were analyzed by real-time PCR with regard to EGFR expression. In comparison to normal colonic tissue samples (nos. 1 and 2), two KRAS-wt expressing CRC samples (nos. 3 and 4) displayed significantly increased EGFR expression, whereas one KRAS-wt (no. 5) and one KRAS-mutated (KRAS4b^G12V_{heterogeneous}) CRC sample (no. 6) displayed significantly decreased EGFR expression (Figure 7, A and B). Furthermore, immunohistochemical analyses using the US Food and Drug Administration-approved EGFR pharmDx Kit (Dako) were performed with KRAS-mutated (only codon 12) and KRAS wt colorectal carcinoma biopsy samples. Interestingly, a moderate but not statistically significant trend toward lower EGFR expression in KRAS-mutated CRC samples (22.43 ± 7.48, n = 7) in comparison to KRAS wt CRC samples (43.50 ± 26.45, n = 10; Figure 7, C and D) could be detected. Together, these data suggest that the expression of EGFR in CRC tissue may be downregulated partially by, for example, oncogenic mutations such as KRAS^G12V and suggest an in vivo relevance of our data observed in stably transfected A431-KRAS^G12V cells.

Discussion

In patients with colorectal tumors, oncogenic mutations of KRAS are accompanied by a lack of benefit from therapeutic EGFR-Abs like cetuximab or panitumumab [6–9]. More recent studies demonstrated that this negative impact only applies for KRAS codon 12 but not for codon 13 mutations [34], which was predicted from preclinical studies demonstrating different functional characteristics of these mutations [35,36]. Here, we systematically investigated underlying mechanisms of resistance to EGFR antibodies in KRAS-mutated tumor cells by studying the impact of oncogenic KRAS^G12V on Fab- and Fc-mediated effector mechanisms of EGFR-Ab. Interestingly, KRAS^G12V was found to downregulate EGFR expression by a C/EBPβ-dependent mechanism.

Activating mutations of KRAS have been demonstrated to overcome growth inhibition by EGFR antibodies [19]. As we demonstrate here, stable overexpression of oncogenic KRAS^G12V was also associated with impaired Fab-as well as Fc-mediated effector mechanisms of EGFR-Abs. Hence, CDC-, monocyte-, and PMN-mediated tumor cell lysis was significantly reduced in KRAS-mutated A431 cells, whereas NK cell-mediated ADCC was less affected in these cells. This observation might be explained by findings demonstrating that NK cell-mediated ADCC activity could be triggered by cetuximab and zalutumumab at lower EGFR expression, whereas PMN ADCC and CDC seem to require higher expression levels [37,38]. Reduced in vitro efficiency of EGFR-mAb in KRAS-mutated cells could be linked to significant down-regulation of EGFR expression, elicited by suppressed activity of C/EBP TFs—especially of C/EBPβ. As siRNA-mediated knockdown of C/EBPβ restored EGFR expression and thereby enhanced cytotoxic activity against KRAS-mutated cells, it could be hypothesized that EGFR cell surface expression impacts the efficiency of EGFR-mAb. Supporting data for oncogenic KRAS associated down-regulation of EGFR expression were presented in a recent study where down-regulation of EGFR mRNA expression was detected in a murine knock-in model for oncogenic KRAS [39]. In line with these findings, a marked decrease in cell surface expression of high affinity but not low affinity EGFR was also observed in stably KRAS4b^G12V-transfected human endometrial carcinoma cells [40]. Interestingly, van Houdt et al. [41] analyzed 55 human colorectal tumor samples for EGFR expression and cellular localization by immunostaining. They found a loss of basolateral EGFR localization in 15 of 16 tumor samples positive for KRAS mutations. Among these 15 tumor samples, 5 were negative for EGFR staining. Plasma membrane localization of EGFR could be restored in KRAS-mutated cells by Rho kinase inhibition. These observations strongly suggest that oncogenic KRAS leads to down-regulation of EGFR expression in vivo, but results from larger clinical studies need to confirm these results (see next paragraphs).

As shown in the present study, expression of oncogenic KRAS was accompanied by a strong up-regulation of activating isoforms (LAP) and the induction of the inhibitory (LIP) isoform of the TF C/EBPβ—leading to decreased LAP/LIP ratios and subsequently decreased TF activity of C/EBPβ as well as suppression of EGFR transcription. Decreased LAP/LIP ratios, thus increased expression of LIP, also have been detected in tumor tissue samples from breast cancer patients as well as in Duke's stage B tumor tissues from CRC patients in comparison to normal tissue samples and have been suggested to be associated with enhanced tumor invasiveness [42,43]. Previous studies also observed that C/EBPβ is a nuclear mediator of oncogenic RAS signaling transduction—promoting cell proliferation and transformation [28,29,44]. Interestingly, other studies found that the inhibitory isoform of C/EBPβ, LIP, is overexpressed in human breast cancer and has been shown to be associated with malignant transformation of epithelial cells [45–47]. Furthermore, when exogenously overexpressed in a human hepatoma cell line, C/EBPβ-LIP has been shown to suppress transcription in general and to be implicated in the induction of prosurvival functions [48]. Although it has been suggested in another study that no CCAAT binding sites are located in the sequence of the EGFR promoter between nucleotides -1 to -540 [49], our data show, for the first time, that the family of C/EBP TF is indeed involved in the regulation of the EGFR promoter (-620 to -754 nt) activity. Interestingly, EGFR expression levels were reported to also be downregulated by mechanisms targeting EGFR for degradation [50,51].

A negative impact of low EGFR expression levels on cetuximab-triggered NK cell-mediated ADCC activity has previously been demonstrated in other in vitro studies [27,37,52]. The observation that NK cell-mediated ADCC was only slightly impaired in A431-KRAS^G12V cells when compared with A431-control-vector cells in the present study may be explained by the fact that, despite of the significant down-regulation of EGFR expression in A431-KRAS^G12V cells, there is still a high cell surface expression of EGFR when compared with other tumor cell lines. Improving antibodies' Fc-mediated functions—such as ADCC and CDC—might be a promising approach to overcome inefficient tumor cell destruction in KRAS-mutated cells. In previous studies, our group demonstrated in MNC-mediated ADCC experiments that KRAS-mutated tumor cells, which were resistant to growth inhibition by EGFR-Abs, could be effectively eliminated by Fc-engineered Ab variants with increased FcγRIIIa binding affinity [19]. Recently, a phase 1 study with a glycoengineered EGFR-Ab (GA201) reported clinical responses in patients with KRAS-mutated tumors—suggesting that KRAS-mediated treatment resistance can indeed be overcome by improving ADCC efficacy [53].

The impact of EGFR expression levels on the clinical outcome of EGFR-mAb therapy is controversial. Whereas some studies on CRC patients did not find a correlation between the extent of EGFR expression, determined by fluorescent in situ hybridization or IHC, and the response rate to cetuximab or panitumumab [54,55], other studies detected a positive correlation [56–59]. Moreover, recent publications presented data suggesting that, within the population of mCRCs with wt KRAS, higher response rates to cetuximab-based treatment were observed in patients with tumors harboring high EGFR gene copy numbers [60,61]. Furthermore, recent analyses of the phase III FLEX study in NSCLC revealed a positive impact of high EGFR expression on the efficiency of chemotherapy plus cetuximab (including overall survival)—suggesting quantitative EGFR expression by IHC as a predictive biomarker for cetuximab therapy in NSCLC [62]. To the best of our knowledge, no data that compare EGFR expression levels by IHC, depending on the KRAS mutation status, are available.

Taken together, our data, although remaining to be verified in in-depth analyses using colorectal carcinoma cell line models, provide novel insights into signaling cascades triggered by oncogenic KRAS in the context of EGFR-Ab therapy. Besides the induction of proliferative processes, oncogenic KRAS has been demonstrated to induce expression of the TF C/EBPβ, especially the dominant-negative isoform C/EBPβ-LIP, that is implicated in the negative regulation of EGFR mRNA transcription—resulting in lower EGFR cell surface expression and impairing Fc-mediated functions of EGFR-Ab (Figure 7). Studying the complex interplay between oncogenic KRAS-associated signal transduction and EGFR expression provided novel insights into the resistance mechanisms of tumor cells to EGFR-Abs and suggested strategies how these antibodies can be optimized to overcome low EGFR expression in KRAS-mutated tumors.

Abbreviations

Ab

antibody

ADCC

antibody-dependent cell-mediated cytotoxicity

CDC

complement-dependent cytotoxicity

C/EBP

CCAAT/enhancer-binding protein

CRC

colorectal cancer

MNC

mononuclear cell

PMN

polymorphonuclear cell

TF

transcription factor