The pro-apoptotic paradox: the BH3-only protein Bcl-2 interacting killer (Bik) is prognostic for unfavorable outcomes in breast cancer

Abstract

Breast cancer is the leading cause of cancer-associated deaths in women worldwide. Clinical biomarkers give information on disease progression and identify relevant biological pathways. A confounding factor that uncouples markers from disease outcome is the ability of tumor cells to mutate and evade clinical intervention. Therefore, we focussed on apoptotic genes that modulate tumor regression. Using gene and tissue microarray analyses, we identified an association of Bcl-2 interacting killer (Bik) with poor breast cancer prognosis. Bik prognostic ability was independent of Estrogen Receptor/Progesterone Receptor and Her2 status. Additionally, Bik was independent of anti-apoptotic Bcl-2, Bcl-xL, Mcl-1 and Bcl-w suggesting a complex mechanism of tumor promotion identified by Bik high tumors. Bik also stimulates autophagy, which can contribute to enhanced tumor fitness. We found a significant association between the autophagy marker ATG5 and Bik. Combined high expression level of ATG5 and Bik was a stronger predictor of outcome than either alone. Thus, our study identifies Bik as a novel, independent prognostic biomarker for poor outcomes in breast cancer and suggests that Bik-mediated autophagy contributes to disease recurrence.

Keywords: BH3-only proteins; Bcl-2 interacting killer; BiK; autophagy; breast cancer.

Figure 1. Flow chart depicting selection of patients and experimental scheme

Two different patient cohorts were chosen for the study. Treatment-naïve tumor tissue was collected for gene expression studies (Dataset-1) and protein expression studies (Dataset-2). All patients subsequently received standard guideline based therapy.

Figure 2. Bik transcript levels are elevated in breast cancer patients with poor survival outcomes

Kaplan-Meier survival curves depicting A. five year disease-free (HR=1.78, 95% CI: 0.99 to 3.20) and B. overall survival (HR= 2.05, 95%CI: 0.96 to 4.37) outcomes of 175 patients based on Bik transcript levels. Patients with low levels of Bik transcript (n=143) were compared to patients with high levels of Bik transcript (n=32). The HR value of greater than 1.0 estimates the predicted risk of poor prognosis.

Figure 3. Bik protein levels are elevated in breast cancer patients with poor survival outcomes

A. Immunohistochemistry analysis of patient tumors stained with anti-Bik antibody. Score values are based on antibody staining intensity (brown) on a scale of 0 to 3. Images are representative of tumor cores with typical score values. Scale bars, 50μm. B. and C. Kaplan-Meier survival analysis depicting five year disease-free (HR=3.59, 95%CI: 1.32 to 9.82) (B) and overall survival (HR=3.40, 95%CI: 1.10 to 11.69) (C) outcomes of 152 patients based on Bik protein levels in tumor cores. Patients with low levels of Bik protein (n=99) were compared to patients with high levels of Bik protein (n=53). The associated HR value of greater than 1.0 estimates the predicted risk of poor prognosis.

Figure 4. Elevated expression of anti-apoptotic Bcl-2 family members is not associated with poor patient prognosis

A, B and C. Kaplan-Meier survival curves representing disease-free survival outcomes of 175 patients based on Bcl-2, Bcl-xL and Mcl-1gene expression levels respectively.

Figure 5. Bcl-2 protein levels are elevated in breast cancer patients with favourable survival outcomes

A. Immunohistochemistry analysis of patient tumors stained with anti-Bcl-2 antibody. Score values are based on antibody staining intensity (brown) on a scale of 0 and 1. Images are representative of tumor cores with typical score values. Scale bars, 50μm. B. and C. Kaplan-Meier survival analysis depicting 5 year disease-free (HR=0.13, 95%CI: 0.15 to 0.37) and overall survival outcomes (HR= 0.33, 95%CI: 0.15 to 0.55) of 149 breast cancer patients based on Bcl-2 protein levels in tumor cores. Patients with low levels of Bcl-2 protein (n=73) were compared to patients with high levels of Bcl-2 protein (n=76). The associated HR value of less than 1.0 estimates the predicted risk of poor prognosis.

Figure 6. ATG5 transcript levels are elevated in breast cancer patients with poor survival outcomes

A. and B. Kaplan Meier survival curves representing disease-free (HR=1.69, 95%CI: 1.11 to 2.65) survival and overall survival (HR=1.90, 95%CI: 1.11 to 3.41) outcomes of 175 BC patients based on ATG5 mRNA levels. C. Paired KM curves demonstrating individual or combined effects of high ATG5 and Bik expression levels on recurrence-free survival of patients. D. Scatter-plot comparing gene expression levels of ATG5 and Bik per patient with respect to disease recurrence. Horizontal and vertical dotted lines indicate dichotomizing score cut-points as determined by ROC analysis. Percentage values in rectangles indicate percent of patients that recurred within each quadrant.

Figure 7. Model describing relationships between Bik, Bcl-2 and ATG5 with respect to patient survival

We propose that Bik stimulates autophagy and Bcl-2 inhibits autophagy. Thus, Bik and Bcl-2 function in two separate yet converging pathways associated with poor patient prognosis.