Increased epithelial membrane protein 2 expression in glioblastoma after treatment with bevacizumab

Abstract

Background: Antiangiogenic therapy with bevacizumab has failed to provide substantial gains in overall survival. Epithelial membrane protein 2 (EMP2) is a cell surface protein that has been previously shown to be expressed in glioblastoma, correlate with poor survival, and regulate neoangiogenesis in cell lines. Thus, the relationship between bevacizumab and EMP2 was investigated.

Methods: Tumor samples were obtained from 12 patients with newly diagnosed glioblastoma at 2 time points: (1) during the initial surgery and (2) during a subsequent surgery following disease recurrence post-bevacizumab treatment. Clinical characteristics and survival data from these patients were collected, and tumor samples were stained for EMP2 expression. The IVY Glioblastoma Atlas Project database was used to evaluate EMP2 expression levels in 270 samples by differing histological areas of the tumor.

Results: Patients with high EMP2 staining at initial diagnosis had decreased progression-free and overall survival after bevacizumab (median progression-free survival 4.6 months vs 5.9 months; log-rank P = .076 and overall survival 7.7 months vs 14.4 months; log-rank P = .011). There was increased EMP2 staining in samples obtained after bevacizumab treatment in both unpaired (mean H-score 2.31 vs 1.76; P = .006) and paired analyses (mean difference 0.571; P = .019). This expression increase correlated with length of bevacizumab therapy (R ² = 0.449; Pearson P = .024).

Conclusions: Bevacizumab treatment increased EMP2 protein expression. This increase in EMP2 correlated with reduced mean survival time post-bevacizumab therapy. We hypothesize a role of EMP2 in clinical bevacizumab resistance and as a potential antiangiogenic therapeutic target in glioblastoma.

Keywords: EMP2; angiogenesis; bevacizumab; glioblastoma.

Figure 1.

Increased EMP2 is associated with decreased survival after bevacizumab therapy. The Cancer Genome Atlas (TCGA) database was queried for EMP2 RNA expression data and specimens were stratified by high or low EMP2 expression. Kaplan–Meier analysis using the TCGA dataset (A). The IVY Glioblastoma Atlas Project (IVY GAP) database was used to identify RNA expression through normalized z-score at different areas of the tumor as defined by tumor histology (B). Kaplan–Meier analysis using clinical data for progression-free survival (C), time to repeat surgery (D), and overall survival (E), after bevacizumab therapy was initiated.

Figure 2.

Bevacizumab therapy increases EMP2 levels. (A) Protein expression of EMP2 was visualized using standard IHC in clinical samples before and following bevacizumab treatment and eventual resistance. The expression pattern from 2 patients is shown. Magnification: 200×. Scale bar = 100 µm. (B and C) Protein expression of EMP2 in clinical samples increased after bevacizumab therapy in both unpaired (B) and paired (C) analyses. (D) Increase in protein expression was positively correlated with length of bevacizumab therapy.