Quantitative proteomics with siRNA screening identifies novel mechanisms of trastuzumab resistance in HER2 amplified breast cancers

Abstract

HER2 is a receptor tyrosine kinase that is overexpressed in 20% to 30% of human breast cancers and which affects patient prognosis and survival. Treatment of HER2-positive breast cancer with the monoclonal antibody trastuzumab (Herceptin) has improved patient survival, but the development of trastuzumab resistance is a major medical problem. Many of the known mechanisms of trastuzumab resistance cause changes in protein phosphorylation patterns, and therefore quantitative proteomics was used to examine phosphotyrosine signaling networks in trastuzumab-resistant cells. The model system used in this study was two pairs of trastuzumab-sensitive and -resistant breast cancer cell lines. Using stable isotope labeling, phosphotyrosine immunoprecipitations, and online TiO(2) chromatography utilizing a dual trap configuration, ~1700 proteins were quantified. Comparing quantified proteins between the two cell line pairs showed only a small number of common protein ratio changes, demonstrating heterogeneity in phosphotyrosine signaling networks across different trastuzumab-resistant cancers. Proteins showing significant increases in resistant versus sensitive cells were subjected to a focused siRNA screen to evaluate their functional relevance to trastuzumab resistance. The screen revealed proteins related to the Src kinase pathway, such as CDCP1/Trask, embryonal Fyn substrate, and Paxillin. We also identify several novel proteins that increased trastuzumab sensitivity in resistant cells when targeted by siRNAs, including FAM83A and MAPK1. These proteins may present targets for the development of clinical diagnostics or therapeutic strategies to guide the treatment of HER2+ breast cancer patients who develop trastuzumab resistance.

Fig. 1.

Altered tyrosine phosphorylation patterns between trastuzumab-sensitive and trastuzumab-resistant cell lines. A, immunoblot of HER2 protein expression and phosphorylation in whole cell lysates (WCL) from SkBr3/SkBr3^R and BT474/BT474^R pairs of cell lines. SkBr3^R and BT474^R denote the respective trastuzumab-resistant cell lines. B, lysates from SkBr3, SkBr3^R, BT474, and BT474^R cells were subjected to phosphotyrosine immunoprecipitation (using 4G10 antibody) and then immunoblotted with the same phosphotyrosine antibody. Asterisks represent differences in phosphoprotein bands seen between the respective trastuzumab-sensitive and -resistant cell lines. C, experimental workflow for SILAC labeling and LC-MS/MS.

Fig. 2.

Distribution of protein ratios from identified proteins in the SkBr3/SkBr3^R and BT474/BT474^R SILAC experiments. A–D, volcano plots showing the protein ratios (in log₂) as a function of the log intensity for each inferred protein. Green circles are proteins with a p value of <0.001 as determined by the Perseus Significance B calculation, yellow diamonds are proteins with a p value between 0.001 and 0.01, red squares represent p values between 0.01 and 0.05, and blue crosses are proteins whose fold change is not significant (p > 0.05). Protein ratios are resistant:sensitive (heavy:light), and the log intensity of the proteins is the sum of peptide ion intensities for that protein. E, comparison of biological replicates (rep.) 1 and 2 in the SkBr3/SkBr3^R and BT474/BT474^R datasets. F, comparison of all proteins identified from each dataset. G, comparison of proteins with significantly increasing/decreasing ratios in both datasets.

Fig. 3.

Validation of SILAC ratios with Western blots. A, validation of EGFR and HER4 ratios in BT474 whole cell lysate (WCL) and phosphotyrosine (pTyr) IP. B, EGFR expression in BT474 WCL and EGFR phosphorylation by EGFR IP followed by Western blot (WB) for pTyr. C, validation of CDCP1 and FAK in SkBr3 WCL and pTyrIP. D, Western blot for Paxillin (PXN) and PXN pY 118 in WCL and pTyr IP in SkBr3 cells.

Fig. 4.

siRNA screen results. A, selected siRNA results for SkBr3^R cells (n = 3) showing increasing magnitude of resensitization to trastuzumab from left to right. Error bars represent the 95% confidence limits. B, selected siRNA screen results for BT474^R cells (n = 3). Data in both A and B are representative of two independent experiments. C, responses of SkBr3R cells (n = 3) to individual siRNA targeting MAPK1. Asterisks denote significant difference in cell viability between siRNA only and siRNA + trastuzumab (Trstzmb). D, response of BT474 cells (n = 3) to individual siRNA targeting FAM83A. Full results for the siRNA screen are available in supplemental Table S4.