CDK4/6 or MAPK blockade enhances efficacy of EGFR inhibition in oesophageal squamous cell carcinoma

Abstract

Oesophageal squamous cell carcinoma is a deadly disease where systemic therapy has relied upon empiric chemotherapy despite the presence of genomic alterations pointing to candidate therapeutic targets, including recurrent amplification of the gene encoding receptor tyrosine kinase epidermal growth factor receptor (EGFR). Here, we demonstrate that EGFR-targeting small-molecule inhibitors have efficacy in EGFR-amplified oesophageal squamous cell carcinoma (ESCC), but may become quickly ineffective. Resistance can occur following the emergence of epithelial-mesenchymal transition and by reactivation of the mitogen-activated protein kinase (MAPK) pathway following EGFR blockade. We demonstrate that blockade of this rebound activation with MEK (mitogen-activated protein kinase kinase) inhibition enhances EGFR inhibitor-induced apoptosis and cell cycle arrest, and delays resistance to EGFR monotherapy. Furthermore, genomic profiling shows that cell cycle regulators are altered in the majority of EGFR-amplified tumours and a combination of cyclin-dependent kinase 4/6 (CDK4/6) and EGFR inhibitors prevents the emergence of resistance in vitro and in vivo. These data suggest that upfront combination strategies targeting EGFR amplification, guided by adaptive pathway reactivation or by co-occurring genomic alterations, should be tested clinically.

Figure 1. Amplified EGFR is a putative target in ESCC cell line models.

(a) Integrative Genomics Viewer (IGV) screenshots of chromosome 7p12.3-p12.1 and the EGFR locus in ESCC patients from The Cancer Genome Atlas (TCGA). The broader view shows chromosome 7p in 90 ESCC samples with the inset image focussed in at the EGFR locus in patients with copy-number gains. Red colour means copy-number gain and blue colour means copy-number loss (x axis: chromosomal coordinates; y axis: individual cases). (b) Single-nucleotide polymorphism (SNP) array inferred copy-number and immunoblots showing basal level of phosphorylation and total EGFR protein expression in a panel of ESCC cell line models and normal oesophageal squamous epithelial cell EPC. (c) Plots showing the in vitro sensitivity of a panel of ESCC cell line models to distinct EGFR inhibitors erlotinib and afatinib. Cell viability at distinct doses relative to vehicle-treated controls is shown. (d) Immunoblots evaluating the biochemical response to erlotinib and afatinib in representative EGFR inhibitor-sensitive cell line models. Cells were harvested at the indicated time points after treatment with 1 μM erlotinib or 100 nM afatinib. (e) Plots show analysis of cell cycle arrest after 48 h of inhibitor treatment with 1 μM erlotinib or 100 nM afatinib. (f) Plots show analysis of apoptosis after 72 h of treatment with 1 μM erlotinib or 100 nM afatinib. All experiments were performed in triplicate for each condition and repeated at least twice. All error bars represent s.d., n≥3. Student's t-test was used for statistical analysis. *P<0.05.

Figure 2. OE21 acquired resistance to EGFR inhibition in vitro with induction of an EMT phenotype.

(a) The EGFR-amplified ESCC OE21 cell-resistant variant was generated by the indicated culture with gradually increasing concentrations of erlotinib with resistance as confirmed by dose–response curve using Cell-Titer GLO. OE21 parental cells and erlotinib-resistant cells were treated with erlotinib and afatinib at indicated concentrations for 72 h and relative cell growth was quantified using the Cell-Titer-Glo assay and plotted as a percentage of growth relative to dimethylsulfoxide (DMSO)-treated control cells. Data points are represented as mean±s.d. of three independent experiments. (b) Signalling responses of OE21 parental and resistant cells with increasing concentrations of afatinib. Cells were harvested 6 h after drug treatment. (c) Representative images of parental and erlotinib-resistant cells demonstrating apparent mesenchymal morphology in the resistant models. Scale bar, 100 μm. (d) Immunoblot measurement of candidate EMT markers of OE21 parental, erlotinib-resistant and afatinib-resistant cell lines, and TE8, KYSE520 cell lines. (e) Immunoblots evaluating distinct isolated OE21ER subclones for EGFR and downstream pathway phosphorylation and EMT marker expression. All experiments were performed in triplicate for each condition and repeated at least twice. All error bars represent s.d., n=3.

Figure 3. ERK reactivation following EGFR tyrosine kinase inhibitor (TKI) treatment facilitates resistance to EGFR inhibition.

(a) Immunoblots demonstrating the biochemical effects of distinct EGFR kinase inhibitors in OE21 cell line to erlotinib and afatinib at multiple time points after onset of therapy. (b) Immunoblots showing biochemical responses of OE21 cells to erlotinib alone or in combination with multiple doses of the MEK inhibitor trametinib. For erlotinib and trametinib combination, cells were treated with erlotinib first and were allowed to grow for 72 h. Then, erlotinib was added to the culture 6 h before protein harvest. (c) Plots depict the growth of OE21 and KYSE140 treated in vitro with either 1 μM erlotinib or 100 nM trametinib alone or in combination. All data are expressed as the percentage of growth relative to that of vehicle-treated control cells. (d) Plots depict the percentage of cells in G0/G1 following treatment with dimethylsulfoxide (DMSO), 1 μM erlotinib, 100 nM trametinib or a combination thereof for 48 h, with cell cycle status then assessed by flow cytometry. (e) Plots representing the induction of apoptosis, as measured by flow cytometry, in cells after 72 h of in vitro treatment with vehicle, 1 μM erlotinib and 100 nM trametinib or combination. For sequential strategy (c–e), cells were treated with one drug first and were allowed to grow for 24 h. Then, the second drug was added to the culture for another 24 h (for cell cycle analysis) and 48 h (for growth curve and apoptosis analysis). (f) Images show representative results of focus formation assays where cells were grown in culture in 6-well plates and treated with DMSO, erlotinib 1 μM, afatinib 100 nM and trametinib 10 nM weekly, and then fixed and stained with crystal violet solution after 4 weeks of treatment. All experiments were performed in triplicate for each condition and repeated at least twice. All error bars represent s.d., n≥3. Student's t-test was used for statistical analysis (*P<0.05).

Figure 4. Blockade of CDK4/6 with EGFR prevents the emergence of resistance to EGFR inhibition.

(a) Integrated view of genomic aberrations of genes encoding cell cycle regulatory proteins in The Cancer Genome Atlas (TCGA) data in EGFR-amplified tumours from TCGA. Each column denotes an individual tumour, and each row displays a gene. Mutations are colour coded by the type of mutation, and amplifications are depicted as red outlines. (b) In vitro growth inhibition of OE21 and KYSE140 cell lines following treatment with erlotinib 1 μM, palbociclib 1 μM or a combination thereof for 72 h or sequential treatment. The flow cytometry (c,d) as well as crystal violet (e) assays for these combination treatment studies are shown as in Fig. 3. All experiments were performed in triplicate for each condition and repeated at least twice. Student's t-test was used for statistical analysis (*P<0.05). All error bars represent s.d., n≥3.

Figure 5. CDK4/6 or MEK inhibition improves erlotinib response in ESCC xenografts.

(a) Curves plot the growth of OE21 xenograft tumours were treated with vehicle control, erlotinib, trametinib, palbociclib, erlotinib+trametinib or erlotinib+palbociclib (mean±s.e.m., 5 mice per condition). (b) Waterfall plot showing the percentage change in volume for the individual tumours in each arm at treatment day 28. Tumour volumes were normalized individually to their volumes at treatment day 1. Student's t-test was used for statistical analysis (*P<0.05).