Kinase Plasticity in Response to Vandetanib Enhances Sensitivity to Tamoxifen in Estrogen Receptor Positive Breast Cancer

Abstract

Resistance to endocrine therapy (ET) is common in estrogen receptor (ER) positive breast cancer. Multiple studies have demonstrated that upregulation of MAPK signaling pathways contributes to ET resistance. Herein we show that vandetanib treatment enhances sensitivity to ET in ET-sensitive and -resistant ER+ breast cancer models. Vandetanib treatment alters the gene expression program of ER+ breast cancer cells resulting in a less proliferative and more estrogen responsive Luminal-A like character. Tyrosine kinase network reprogramming was assessed using multiplexed kinase inhibitor beads-mass spectrometry (MIB/MS) assay to identify adaptive resistance mechanisms to vandetanib treatment, including upregulation of HER2 activity. Co-treatment to inhibit HER2 with lapatinib enhanced sensitivity to vandetanib, demonstrating biologic activity of HER2 upregulation. Using a CRISPR knockout model, we demonstrate that vandetanib effects are partially mediated by RET receptor tyrosine kinase. Finally, we use our operating room-to-laboratory assay that measures drug response in individual primary tumor cells in short term cultures to demonstrate conserved gene expression changes, including increased HER2 activity signatures, in vandetanib treated cells, and identify features associated with vandetanib response. These results support future investigation of RET targeting strategies considering reprogrammed networks, such as activated HER2, in patients with ET resistant ER+ breast cancer.

Keywords: Breast cancer; endocrine resistance; functional proteomics; targeted therapeutics; tyrosine kinase inhibition.

Figure 1.

(A) Treatment of MCF-7 and tamoxifen resistant MCF7 (MCF7-TAM) and exemestane resistant MCF7 (MCF7-EXE) reduces pERK/ERK in all 3 cell lines. (B) Vandetanib dose dependent reduction in cell viablilty in MCF7, MCF7-TAM, and MCF7-EXE. (C) Treatment with Vandetanib results in a left shift of the tamoxifen sensitivity curve and reduction in IC50 in all 3 cell lines. (D) Synergy plots demonstrate synergistic activity of tamoxifen and vandetanib. *p<0.05, ***p<0.01 by t-test with Welch correction and analysis of the slope difference in the nonlinear regression model.

Figure 2.

Heatmap and enrichment analysis of differentially expressed genes in (A) MCF7, (B) MCF7-TAM, (C) MCF7-EXE. KRAS activity signatures were depleted in all 3 cell lines by vandetanib treatment. (D) Vandetanib treatment increases association with the more indolent, endocrine therapy responsive Luminal A subtype, measured by correlation to Luminal A defining genes (centroid). (E) Overlap of up- and down- regulated genes by cell line. (F) Signatures of overlapping up-and down regulated genes are associated with recurrence free and overall survival in ER+ breast cancer patients in METARIC and SCAN-B studies and (G) downregulated genes are enriched in clinically ET-resistant ER+ tumors. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by t-test with Welch correction for Wilcoxon test.

Figure 3.

(A) Multiplex inhibitor bead mass spectrometry assay work flow. (B) Effects of vandetanib on enrichment and depletion of known vandetanib targets. Baseline expression of vandetanib targets in tamoxifen and exemestane resistant cell lines compared to WT MCF7. (C) Heatmap showing enriched and depleted proteins after vandetanib treatment. (D) Overlap analysis of up- and downregulated proteins between the 3 cell lines when treated with vandetanib. (E) Pathway analysis of the downregulation of the ERK cascade and upregulation of the PI3K/AKT pathway.

Figure 4.

(A) Heatmap showing enrichment levels of druggable kinases of sensitive and resistant MCF7 cell lines before and after vandetanib treatment. Of note, ERBB2 (encoding HER2) was enriched in all 3 cell lines after treatment and RET was dpeleted in MCF7 and MCF7-TAM. (B) Sensitivity plots of co-treated cells with vandetanib and HER2 inhibitor lapatinib demonstrating enhanced lapatinib sensitivity with vandetanib treatment. *p<0.05, ****p<0.0001 by t-test with Welch correction.

Figure 5.

(A) Co-treatment with vandetanib and the selective RET inhibitor BLU667 reduced vandetanib sensitivitiy in MCF7 and MCF7-TAM, but not MCF7-EXE with RET inhibition consistent with activity depletion in Fig 4. (B) Western blot demonstrating CRISPR knockout of RET resulting in >90% reduction in RET protein levels. (C) RET knockout cells have reduced sensitivity to vandetanib. (D) RET knockout cells have reduced sensitivity to tamoxifen. (E) ESR1 RNA levels encoding the estrogen receptor are decreased in RET KO cells, consistent with reduced sensitivity to tamoxifen. (F) Heatmap demonstrating differentially expressed genes with RET KO in MCF7. Enrichment analysis demonstrates reduction in hallmark estrogen response and RAS signaling in RET KO cells. RET KO cells have lower proliferation score. (G) RET knockout cells have reduced proliferation compared to WT. (. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 by t-test with Welch correction for the bar graphs, and Wilcoxon test.

Figure 6.

(A) Single-cell RNA sequencing uniform manifold projection (UMAP) annotated by cell type within a primary ER+ human breast tumor. (B) Tumor cells were computationally isolated as having significant inferred copy number changes and plotted by treatment condition. (C) Heatmap showing differentially expressed genes compared between vandetanib treated vs non-treated tumor cells. (D) Enriched and depleted pathways of vandetanib treated vs non-treated tumor cells. (E) ERBB2 activity signature score was increased in vandetanib treated tumor cells as seen in cell lines. Down regulated genes in vandetanib treated cell lines are depleted in vandetanib treated primary tumor cells, and LumA correlation increased in vandetanib treated tumor cells as was seen in cell lines, and depleted KRAS activity signature demonstrates on target effects. Significant signatures filtered by < 0.1 FDR on GSEA analysis and significant genes filtered by adjusted p value <0.05 (Wilcoxon test).