The coexpression of EphB4 and EphrinB2 is associated with poor prognosis in HER2-positive breast cancer

Abstract

Objective: HER2 overexpression is associated with aggressive phenotypes in breast cancer, including increased tumor proliferation, greater invasiveness, and reduced overall survival. The overall response rate to HER2-targeted therapies remains <30%. There is an urgent need for the identification of efficient markers to predict patients with a poor prognosis. This study was designed to investigate the correlation between EphB4 and EphrinB2 expression and the clinicopathological characteristics of HER2-positive breast cancer.

Materials and methods: A total of 111 primary HER2-positive breast cancer patients were enrolled in this study (diagnosed since December 2005 to November 2010 from the Second Hospital of Dalian Medical University). The protein expression of EphB4 and EphrinB2 was examined by immunohistochemistry using paraffin-embedded tumor tissues.

Results: There was a significant correlation between EphB4 and EphrinB2 expression (P=0.013, r=0.255). Kaplan-Meier analysis showed that the prognosis of patients with a high expression of both EphB4 and EphrinB2 was significantly worse than the prognosis of patients with either EphB4 or EphrinB2 expression and patients with negative expression (hazard ratio [HR] =1.935, P=0.0224). However, high expression of EphB4 or EphrinB2 alone was not an independent prognostic factor to predict worse overall survival. To summarize, HER2-positive breast cancer patients with overexpression of both EphB4 and EphrinB2 were associated with the worst prognosis.

Conclusion: High expression of EphB4 and EphrinB2 correlated with poor overall survival, which can serve as an independent prognostic indicator in primary HER2-positive breast cancer patients.

Keywords: EphB4; EphrinB2; breast carcinoma; prognosis.

Figure 1

The interaction between EphB4 and EFNB2 and the distribution of tissue samples. Notes: (A) There is an interaction between EphB4 and EFNB2 based on genetic interactions in the budding yeast Saccharomyces cerevisiae (reported by BioGRID database). (B and C) The gross photos of tissue microarray and immunohistochemical staining. Scale bars, 400 µm. (D) The tissue samples were arranged as shown in the form N, normal tissue. I, II, III, and IV, pTNM staging; ×, the tissue samples not meeting the inclusion criteria are excluded. Abbreviation: pTNM, pathological tumor–node–metastasis.

Figure 2

EphB4 and EFNB2 are highly expressed in ER-negative breast cancer and their overexpression strongly correlates in HER2-positive subtype. Notes: (A, B) Amplification of EphB4 and EFNB2 gene is recurrent in breast carcinoma based on cBioPortal database. (C, D, G, H, K, L) The expression levels of EphB4 and EFNB2 tightly correlate in ER-negative breast cancers in the three independent cohorts. P-values were calculated with one-way ANOVA. (E, I, M) EphB4 expression correlates with EFNB2 in the three independent cohorts, especially (F, J, N) in the HER2-positive subtype. The linear regression Pearson’s correlation coefficient (R) and its P-value are indicated. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. Abbreviations: ER, estrogen receptor; ANOVA, analysis of variance.

Figure 3

Evaluation of immunohistochemical staining for (A, B) EphB4 and (C, D) EphrinB2 in breast cancer tissue. Notes: (A) and (C) represent negative; (C) and (D) represent positive. Scale bars, 200 µm.

Figure 4

Kaplan–Meier analysis of overall survival in HER2-positive breast cancer Notes: (A, B) Kaplan–Meier analysis of overall survival of HER2-positive breast cancer patients. Samples were divided into two groups with high and low expression levels of EphB4 and EphrinB2. (C, D) Kaplan–Meier analysis of overall survival through different subgroups. (E) Patients were divided as in (D). P-values are obtained from the log-rank test. Hazard ratio (HR) is calculated using GraphPad Prism. *P<0.05 (Student’s t-test). Abbreviation: HR, hazard ratio.