Oncogenic KRAS desensitizes colorectal tumor cells to epidermal growth factor receptor inhibition and activation

Abstract

Epidermal growth factor receptor (EGFR)-targeting therapeutics have shown efficacy in the treatment of colorectal cancer patients. Clinical studies have revealed that activating mutations in the KRAS protooncogene predict resistance to EGFR-targeted therapy. However, the causality between mutant KRAS and resistance to EGFR inhibition has so far not been demonstrated. Here, we show that deletion of the oncogenic KRAS allele from colorectal tumor cells resensitizes those cells to EGFR inhibitors. Resensitization was accompanied by an acquired dependency on the EGFR for maintaining basal extracellular signal-regulated kinase (ERK) activity. Deletion of oncogenic KRAS not only resensitized tumor cells to EGFR inhibition but also promoted EGF-induced NRAS activation, ERK and AKT phosphorylation, and c-FOS transcription. The poor responsiveness of mutant KRAS tumor cells to EGFR inhibition and activation was accompanied by a reduced capacity of these cells to bind and internalize EGF and by a failure to retain EGFR at the plasma membrane. Of 16 human colorectal tumors with activating mutations in KRAS, 15 displayed loss of basolateral EGFR localization. Plasma membrane localization of the EGFR could be restored in vitro by suppressing receptor endocytosis through Rho kinase inhibition. This caused an EGFR-dependent increase in basal and EGF-stimulated ERK phosphorylation but failed to restore tumor cell sensitivity to EGFR inhibition. Our results demonstrate a causal role for oncogenic KRAS in desensitizing tumor cells not only to EGFR inhibitors but also to EGF itself.

Figure 1

Deletion of oncogenic KRAS sensitizes colorectal tumor cells to EGFR inhibition. (A) HCT116 and HKH2 cells were seeded in 96-well plates and were treated with cetuximab (20 µg/ml), erlotinib (5 µM), or gefitinib (2 µM) for four consecutive days in triplicate. Mitochondrial activity was determined byMTT assays. (B) The experiment was performed as in panel A, using cetuximab (20 µg/ml) to treat DLD1 and DKO4 cells. (C) The experiment was performed as in panel A, using gefitinib (2 µM) to treat CT26 cells expressing luciferase-targeting shRNA (CT26) and CT26 cells in which endogenous Kras^D12 is stably suppressed by RNAi [19]. (D) The experiment was performed as in panel A, using cetuximab (20 µg/ml) or gefitinib (2 µM) on L145 cells. L145 cells were freshly established from a liver metastasis harboring a KRAS^D12 mutation. Primary human epithelial cells (HIEC [20]) express only wild-type KRAS. *Statically significant differences, P < .05.

Figure 2

EGFR activity is required formaintenance of ERK phosphorylation in wild-type KRAS cells but not in mutant KRAS cells. (A) HCT116 cells and HKH2 cells were treated overnight with the indicated EGFR inhibitors, and phosphorylated and total ERK levels were assessed by Western blot analysis. Bar diagrams represent means and SEM of three independent experiments. (B) As in panel A, using CT26 cells expressing luciferase-targeting shRNA (CT26) and CT26-KrasKD. (C) As in panel A, using L145 cells and HIEC. *Statically significant differences, P < .05.

Figure 3

Oncogenic KRAS suppresses EGFR signaling. (A) Serum-starved HCT116 and HKH2 cells were stimulated with 20 ng/ml EGF for 0, 30, and 60 minutes. c-FOS messenger RNA levels were determined using reverse transcription-polymerase chain reaction. (B) Serum-starved HCT116, DLD1, and CT26 cells and their isogenic derivatives lacking oncogenic KRAS (HKH2, DKO4, and CT26-KrasKD) were stimulated with 20 ng/ml EGF for 5minutes. The levels of phosphorylated and total ERK were determined by Western blot analysis. (C) Cells were cultured and stimulated as in panel B. Ras activity assays were performed using the Ras-binding domain (RBD) of Raf1 fused to glutathione-S-transferase immobilized on glutathione-sepharose. Lysates and RBD-Raf1-bound proteinswere analyzed for the presence of NRAS and KRAS by Western blot analysis. (D) Cells were cultured as previously mentioned and stimulated with the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (5 nM) for 5minutes. The levels of phosphorylated and total ERK were determined by Western blot analysis. (E) Cells were cultured as previously mentioned and stimulated with 20 ng/ml EGF for 5 minutes. The levels of phosphorylated and total AKT were determined by Western blot analysis.

Figure 4

Aberrant EGFR localization and reduced EGF internalization in HCT116 cells. (A) Serum-starved HCT116 and HKH2 cells were stimulated with 20 ng/ml EGF for 0 or 5 minutes. Total and phosphorylated EGFR levels were determined by Western blot analysis. (B) Serumstarved HCT116 and HKH2 cells were stimulated with Alexa 488-conjugated EGF (30 ng/ml; 20 minutes) in the presence of LysoTracker. The uptake of fluorescent EGF and its trafficking to lysosomes were analyzed by live cell imaging. Final images are shown. (C) HCT116 cells were grown on glass coverslips, and EGFR localization was studied by immunofluorescence analysis. (D) Serum-starved HCT116 cells were stimulated with 20 ng/ml EGF (0 or 20 minutes). EGFR (green) and F-actin (red) distributions were then analyzed by immunofluorescence. (E) HCT116 cells were incubated overnight with EGF (20 ng/ml) or cetuximab (20 µg/ml) under serum-free conditions. EGFR and actin levels were determined by Western blot analysis.

Figure 5

Aberrant EGFR localization and reduced EGF internalization in CT26 cells. (A) CT26 control cells and CT26-KrasKD cells were grown on glass coverslips. Cells were serum-starved overnight and were subsequently stimulated with 20 ng/ml EGF for 20 minutes. Coverslips were then stained for EGFR and were analyzed by immunofluorescence. (B) CT26 and CT26-KrasKD cells were stimulated with Alexa 488-conjugated EGF (30 ng/ml; 20 minutes). Cells were then fixed and analyzed for the uptake of fluorescent EGF by confocal microscopy.

Figure 6

Aberrant localization of the EGFR in colorectal tumors expressing oncogenic KRAS. A TMA containing a panel of colorectal tumors with known KRAS mutation status was used to study EGFR localization. (A) We distinguished three types of staining. 1) Basolateral and membranous; examples are shown in the left upper and left lower images. 2) Negative; an example is shown in the right upper image. 3) Diffuse throughout the tumor cells with negative membrane staining; an example is shown in the right lower image. (B) The staining coefficient was determined as the product of the staining intensity on a 0 to 4 scale (with 0 = negative, 1 = weak, 2 = moderate, 3 = strong, 4 = very strong) and the percentage positive cells on a 0 to 3 scale (with <1% = 0,1%–25% = 1, 25%–50% = 2, >50% = 3). The staining scores for all tumors (with a maximum score of 12) were then plotted. The tumors with activating mutations in KRAS are circled in red.

Figure 7

ROK inhibition restores EGFR localization and signaling but not EGFR dependency. (A) HCT116 cells were serum-starved overnight in the presence of 20 µM Y27632, 10 mM Rac1-Inh, 10 µM U0126, or 10 µM LY294002 or in the absence of inhibitors (control). HKH2 cells served as a positive control. The cells were then stimulated with 20 ng/ml EGF for 5 minutes (5) or were left unstimulated (0). ERK1/2 and AKT phosphorylation were then determined by Western blot analysis. (B) The intensities of the pERK1/2 and pAKT signals before and after EGF stimulation were measured using Quantity One software. The percentage of signal intensities (n = 3) in treated versus untreated cells was then plotted. (C) HCT116 cells were serum-starved overnight in the presence or absence of 20 µM Y27632. The cells were then stimulated with 20 ng/ml EGF for 5 minutes (5) or were left unstimulated (0). EGFR expression and phosphorylation were determined by Western blot analysis. (D) HCT116 cells were grown on glass coverslips in the presence or absence of 20 µM Y27632. EGFR localization was then determined by immunofluorescence. (E) HCT116 cells were treated for 4 days with 20 µg/ml cetuximab or 2 µM gefitinib either alone or in combination with 20 µM Y27632. Mitochondrial activity was then assessed by MTT assays. All data points represent means of triplicates ± SEM. (F) CT26 cells were serum-starved overnight in the presence or absence of 20 µM Y27632. The cells were then stimulated with 20 ng/ml EGF for 5 minutes (5) or were left unstimulated (0). ERK expression and phosphorylation were then determined by Western blot analysis. (G) CT26 cells were grown on glass coverslips in the presence or absence 20 µM Y27632. EGFR localization was then determined by immunofluorescence. (H) CT26 cells were treated for 4 days with 2 µM gefitinib either alone or in combination with 20 µM Y27632. Mitochondrial activity was then assessed by MTT assays. All data points represent means of triplicates ± SEM.