Reactivation of ERK signaling causes resistance to EGFR kinase inhibitors

Abstract

The clinical efficacy of epidermal growth factor receptor (EGFR) kinase inhibitors is limited by the development of drug resistance. The irreversible EGFR kinase inhibitor WZ4002 is effective against the most common mechanism of drug resistance mediated by the EGFR T790M mutation. Here, we show, in multiple complementary models, that resistance to WZ4002 develops through aberrant activation of extracellular signal-regulated kinase (ERK) signaling caused by either an amplification of mitogen-activated protein kinase 1 (MAPK1) or by downregulation of negative regulators of ERK signaling. Inhibition of MAP-ERK kinase (MEK) or ERK restores sensitivity to WZ4002 and prevents the emergence of drug resistance. We further identify MAPK1 amplification in an erlotinib-resistant EGFR-mutant non-small cell lung carcinoma patient. In addition, the WZ4002-resistant MAPK1-amplified cells also show an increase both in EGFR internalization and a decrease in sensitivity to cytotoxic chemotherapy. Our findings provide insights into mechanisms of drug resistance to EGFR kinase inhibitors and highlight rational combination therapies that should be evaluated in clinical trials.

Figure 1

WZ4002 resistant EGFR mutant Del E746_A750/T790M cells contain an amplification in MAPK1. A. PC9 GR4 (Del E746_A750/T790M) and DR1 (Del E746_A750/T790M amplification) cells were treated with WZ4002 at the indicated concentrations, and viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. WZR10 and WZR12 cells are resistant to WZ4002. Cells were treated with WZ4002 as in A. C. PC9GR4 and WZR12 cells were treated with WZ4002 at indicated concentrations for 6 hours. Cell extracts were immunoblotted to detect the indicated proteins. D. WZR cells contain an amplification in MAPK1. The PC9 WZR clones (right) were compared with the PC9 GR4 cells (first column). The blue curve on the right indicates degree of amplification of each SNP from 0 (left) to 8 (right). Left, genome wide view; Right, detailed view of chromosome 22. The genomic location of MAPK1 is indicated by an asterix. E. Metaphase FISH of PC9 GR4 and WZR10 cells using MAPK1 (red) and reference probe (green; RP11-768L22). Amplification of MAPK1 is observed in WZR10 cells (arrow).

Figure 2

Inhibition of ERK 1/2 signaling restores sensitivity to WZ4002 in PC9 WZR cells. A. PC9 GR4 or WZR10 cells were treated with the indicated concentrations of CI-104 for 6 hrs. Cell extracts were immunoblotted to detect the indicated proteins. B. PC9GR4 or WZR10 cells were treated with WZ4002 alone at the indicated concentrations or in combination with CI-1040 (3 µM). Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. C. PC9GR4 or WZR10 cells were treated with WZ4002 alone at indicated concentrations or with CI-1040 (3 µM) for 6 hours or 48 hours (PARP lane only). Cell extracts were immunoblotted to detect the indicated proteins. D. Cells were treated with WZ4002 at the indicated concentrations or in combination with Cmp 11E (WZR10 only). Viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. E. WZ4002 (6 hours) inhibits p90RSK phosphorylation in PC9 GR4 but not WZR10 cells. F. Cmp 11E (6 hours) inhibits p90RSK phosphorylation in both PC9 GR4 and WZR10 cells. G. WZ4002 treatment (6 hours) leads to BIM upregulation in PC9 GR4 but not WZR10 cells.

Figure 3

WZ4002 resistant H1975 cells exhibit persistent ERK activation and MEK inhibition restores WZ4002 sensitivity. A. WZ4002 resistant H1975 clones. Cells were treated with WZ4002 at the indicated concentrations, and viable cells were measured after 72 hours of treatment and plotted relative to untreated controls. B. WZ4002 treatment does not fully inhibit ERK1/2 phosphoryaltion in WZR6 cells. H1975 and H1975 WZR6 cells were treated with increasing concentrations of WZ4002. Cell extracts were immunoblotted to detect the indicated proteins. C. CI-1040 (1 µM) restores sensitivity to WZ4002 in the WZR6 cells. D. CI-1040 (1 µM) restores the ability of WZ4002 to fully inhibit ERK 1/2 phosphorylation in the WZR6 cells. E. Comparison of expression profiles of H1975 and WZR 6 cells (left panel). Hierarchical clustering of the differentially expressed genes (p<0.0025, fold change (FC) >3.9) was performed using GENE-E. DUSP6 (asterix) ranks among the top differentially expressed genes (right panel). F. Quantitative PCR of genes in MEK/ERK transcriptional output in H1975 and H1975 WZR cells. The data are normalized to the H1975 cells. Error bars denote standard deviation. G. PC9 GR4 cells were treated with gefitinib (1 µM) or WZ4002 (100 nM) following transfection with control (NT) or DUSP6 siRNA and viable cells were measured after 48 hours of treatment and plotted (mean +/− SD) relative to untreated controls. *; p < 0.05 DUSP6 vs. NT

Figure 4

Development of in vivo resistance to WZ4002 in genetically engineered mouse models of EGFR T790M. A. Kaplan Meier survival curves of L858R/T790M or DelE746_A750/T790M mice treated with erlotinib, vehicle or WZ4002. Treatment with WZ4002 significantly prolongs survival compared to erlotinib (p = 0.0015 (L858R/T790M) and p = 0.0064 (Del E746_A750/T790M); both log-rank test). B. Change in tumor volume over time in EGFR L858R/T790M mice (n = 5) treated with WZ4002. Each curve represents and individual mouse. C. MRI images from mouse 7907 from B. T = Tumor. D. Immunohistochemical analyses using indicated antibodies of tumors from an EGFR L858R/T790M untreated and WZ4002 treated (10 weeks) mice. EGFR phosphorylation but not ERK 1/2 phosphorylation is inhibited. H&E; hematoxylin and eosin E. Immunohistochemical analyses using indicated antibodies of tumors from EGFR Del E746_A750/T790M untreated and WZ4002 treated (24 hours and 26 weeks) mice. WZ4002 treatment inhibits both EGFR and ERK 1/2 phosphorylation at 24 hrs but ERK 1/2 phosphorylation returns following 26 weeks of therapy. F. Serial MRI images of an EGFR L858R/T790M mouse treated with WZ4002 alone or with the combination of WZ4002 and GSK1120212. T = Tumor. G. Tumor volume measurements from mouse in F. Blue arrow, treatment with WZ4002; red arrow; treatment with both WZ4002 and GSK1120212. Numbers indicate weeks of treatment. H. Waterfall plot of tumor volume from 3 L858R/T790M mice following treatment with WZ4002 and GSK1120212. I. Long term treatment of PC9 GR4 cells with CI-1040 alone, WZ4002 alone or the combination of both drugs. The percent of resistant colonies relative to control are plotted from 2 independent experiments. J. Long term treatment of H1975 cells with CI-1040 alone, WZ4002 alone or the combination of both drugs. The percent of resistant colonies relative to control are plotted from 2 independent experiments.

Figure 5

MAPK1 amplification is present in erlotinib resistant EGFR mutant NSCLC. FISH analysis of a baseline (left) and a post-erlotinib treated tumor (right). There is evidence of MAPK1 amplification (arrows) in the relapsed tumor. MAPK1 (red), 22q13.33 reference probe (green; RP11-768L22). Scale bar; 5 µm

Figure 6

MAPK1 amplification alters EGFR internalization. A. PC9 GR4 or WZR10 cells are treated with increasing concentrations of WZ4002 for 6 hrs. Cell extracts from whole lysates (top) or following immunopercipitation with an EGFR DelE746_A750 antibody (bottom) were immunoblotted to detect the indicated proteins. B. Internalization rate constants (k_e) for ¹²⁵I labeled EGF in different cell lines. The k_e is significantly greater for the WZR cells compared to the PC9 GR4 cells. C. PC9 GR4 and WZR 10 cells treated with WZ4002 alone or in combination with CI-1040 for 6 hours. Cell extracts following immunopercipitation with an EGFR DelE746_A750 antibody were immunoblotted to detect the indicated proteins. D. Internalization rate constants (k_e) for 125I labeled EGF in following treatment with CI-1040 (3 mM) for 24 hrs. There is a significant reduction in k_e in the WZR10 cells with CI-1040 treatment. E. EGFR phosphorylation at Thr-669 is markedly increased in WZR10 compared to GR4 cells. CI-1040 alone inhibits phosphorylation at Thr-669 but not Tyr-1068.

Figure 7

EGFR TKI resistant PC9 cells are also more resistant to cytotoxic chemotherapy. A. Cell viability experiments following 24 h exposure to staurosporine (1 µM), paclitaxel (1 µM) or etoposide (100 µM). DMSO is used as a control. The mean values and standard deviation from 3 independent experiments is shown. B. Mitochondrial depolarization response to BH3 peptides for PC9, PC9 GR4 and WZR10 cells. The mean values and standard deviation from 3 independent experiments are shown.