Inhibition of protein-protein interaction of HER2-EGFR and HER2-HER3 by a rationally designed peptidomimetic

Abstract

Protein-protein interactions (PPI) play a crucial role in many biological processes and modulation of PPI using small molecules to target hot spots has therapeutic value. As a model system we will use PPI of human epidermal growth factor receptors (EGFRs). Among the four EGFRs, EGFR-HER2 and HER2-HER3 are well known in cancer. We have designed a small molecule that is targeted to modulate HER2-mediated signaling. Our approach is novel because the small molecule designed disrupts dimerization not only of EGFR-HER2, but also of HER2-HER3. In the present study we have shown, using surface plasmon resonance analysis, that a peptidomimetic, compound 5, binds specifically to HER2 protein extracellular domain and disrupts the dimerization of EGFRs. To evaluate the effect of compound 5 on HER2 signaling in vitro, Western blot and PathHunter assays were used. Results indicated that compound 5 inhibits the phosphorylation of HER2 kinase domain and inhibits the heterodimerization in a dose-dependent manner. Molecular modeling methods were used to model the PPI of HER2-HER3 heterodimer.

Figure 1

Sensograms of A) compound 5 interaction with immobilized HER2-ECD protein. The sensograms show dose-dependent binding of compound 5 injected at concentrations of a) 0.0, b) 5, c) 10, d) 25, e) 50, f) 75, and g) 100 µM. Injections were performed at the rate of 50 µL/ 5 min. A dose-dependent binding of compound 5 onto HER2-ECD can be observed. B) HER2 ECD interaction with immobilized compound 5. The sensograms show dose-dependent binding of HER2 ECD injected at concentrations of a) 0.0, b) 5, c) 10, d) 20, e) 40, f) 80, and g) 160 µM. Injections were performed at the rate of 30 µL/min. The kinetic constant was estimated by global fitting analysis of the titration curve to the 1:1 Langmuir interaction model and the K_d value calculated was around 1.1 µM. C) the specific binding of compound 5 (100 µM) to HER2 ECD and control peptidomimetic (100 µM) that does not bind to HER2 ECD. Refer to Table I for the structure of the control peptidomimetic.

Figure 1

Sensograms of A) compound 5 interaction with immobilized HER2-ECD protein. The sensograms show dose-dependent binding of compound 5 injected at concentrations of a) 0.0, b) 5, c) 10, d) 25, e) 50, f) 75, and g) 100 µM. Injections were performed at the rate of 50 µL/ 5 min. A dose-dependent binding of compound 5 onto HER2-ECD can be observed. B) HER2 ECD interaction with immobilized compound 5. The sensograms show dose-dependent binding of HER2 ECD injected at concentrations of a) 0.0, b) 5, c) 10, d) 20, e) 40, f) 80, and g) 160 µM. Injections were performed at the rate of 30 µL/min. The kinetic constant was estimated by global fitting analysis of the titration curve to the 1:1 Langmuir interaction model and the K_d value calculated was around 1.1 µM. C) the specific binding of compound 5 (100 µM) to HER2 ECD and control peptidomimetic (100 µM) that does not bind to HER2 ECD. Refer to Table I for the structure of the control peptidomimetic.

Figure 2

Sensograms showing the A) interactions of compound 5 with HER2-ECD, EGFR-ECD, and HER3-ECD. Binding of compound 5 with HER2-ECD (40RU) can be seen whereas no binding can be seen with EGFR and HER3-ECD. Compound 5 specifically binds to HER2-ECD but not to the homologues receptors EGFR and HER3-ECD. B) interaction of compound 5 with domain IV of HER2 immobilized on a CM5 sensor chip (48 RU). C) interaction of trastuzumab with domain IV of HER2 immobilized on a CM5 sensor chip (850 RU).

Figure 2

Sensograms showing the A) interactions of compound 5 with HER2-ECD, EGFR-ECD, and HER3-ECD. Binding of compound 5 with HER2-ECD (40RU) can be seen whereas no binding can be seen with EGFR and HER3-ECD. Compound 5 specifically binds to HER2-ECD but not to the homologues receptors EGFR and HER3-ECD. B) interaction of compound 5 with domain IV of HER2 immobilized on a CM5 sensor chip (48 RU). C) interaction of trastuzumab with domain IV of HER2 immobilized on a CM5 sensor chip (850 RU).

Figure 3

Sensogram showing the A) heterodimerization of EGFR-HER2 and its blockade by compound 5. Compound 5 (100 µM) binds to HER2 (38 RU). In the presence of compound 5 the interaction of EGFR-ECD with HER2-ECD was blocked whereas the activated EGFR-ECD (200 nM) binds to HER2-ECD in the absence of compound 5 (12 RU). B) the heterodimerization of HER2–HER3 and its blockade by compound 5. Compound 5 (50 µM) binds to HER2 ECD (14 RU). In the presence of compound 5 the interaction of HER3-ECD with HER2-ECD was blocked whereas the activated HER3 (200 nM) binds to HER2-ECD in the absence of compound 5 (23 RU).

Figure 4

Western blot analysis of phosphorylated HER2. A) in BT-474 cells that overexpress HER2 protein upon treatment with compound 5 (0.4 µM) and AG-825 (0.35 µM). The visualization of β-actin was used to ensure equal sample loading in each lane. Phosphorylation was detected using p-HER2 antibody. Total HER2 protein is also shown. B) in MCF-7 cells that do not overexpress HER2 protein upon treatment with compound 5 (0.4 µM) and AG-825 (0.35 µM).

Figure 5

Inhibition of heterodimerization of HER2–HER3 in HER2–HER3-transfected U2OS cells by compound 5 at different concentrations using PathHunter assay. Dose-response curve for agonist NRG1 and induction of dimerization is shown (filled circles). Dose-response curve for inhibition of heterodimerization by compound 5 (filled triangles) in the presence of 0.3 µM NRG1.

Figure 6

Proposed model for HER3 extracellular domain showing domains I to IV in open conformation.

Figure 7

A) Proposed model for PPI of extracellular domains of HER2 and HER3. Notice the absence of ligand for HER2. C-terminal portions of domain IV of HER2 and HER3 interact, stabilizing the heterodimer apart from domain II interactions. B) HER2 and HER3 domain II-interactions with residues in domain II that are important for dimerization are shown (magenta-HER2, blue-HER3).

Figure 7

A) Proposed model for PPI of extracellular domains of HER2 and HER3. Notice the absence of ligand for HER2. C-terminal portions of domain IV of HER2 and HER3 interact, stabilizing the heterodimer apart from domain II interactions. B) HER2 and HER3 domain II-interactions with residues in domain II that are important for dimerization are shown (magenta-HER2, blue-HER3).

Figure 8

Proposed dimerization interface/binding site of compound 5 (shaded area) on domain IV of HER2 protein.