Gasdermin B expression predicts poor clinical outcome in HER2-positive breast cancer

Abstract

Around, 30-40% of HER2-positive breast cancers do not show substantial clinical benefit from the targeted therapy and, thus, the mechanisms underlying resistance remain partially unknown. Interestingly, ERBB2 is frequently co-amplified and co-expressed with neighbour genes that may play a relevant role in this cancer subtype. Here, using an in silico analysis of data from 2,096 breast tumours, we reveal a significant correlation between Gasdermin B (GSDMB) gene (located 175 kilo bases distal from ERBB2) expression and the pathological and clinical parameters of poor prognosis in HER2-positive breast cancer. Next, the analysis of three independent cohorts (totalizing 286 tumours) showed that approximately 65% of the HER2-positive cases have GSDMB gene amplification and protein over-expression. Moreover, GSDMB expression was also linked to poor therapeutic responses in terms of lower relapse free survival and pathologic complete response as well as positive lymph node status and the development of distant metastasis under neoadjuvant and adjuvant treatment settings, respectively. Importantly, GSDMB expression promotes survival to trastuzumab in different HER2-positive breast carcinoma cells, and is associated with trastuzumab resistance phenotype in vivo in Patient Derived Xenografts. In summary, our data identifies the ERBB2 co-amplified and co-expressed gene GSDMB as a critical determinant of poor prognosis and therapeutic response in HER2-positive breast cancer.

Keywords: HER2-positive breast cancer; clinical behaviour; gasdermin B; predictive biomarker; resistance to therapy.

Figure 1. GSDMB over-expression is associated with poor prognosis in ERBB2-high breast cancers

Tumor samples with the top 25% expression levels of GSDMB gene (“high”) show significantly worse prognosis than the remaining tumors (“low”). (A) Disease-free survival and (B) Distant metastasis-free survival curves in breast cancers over-expressing (top 25%) ERBB2 from the Ur-Rehman dataset [35]. (C) Overall patient survival curves for ERBB2-high tumors in the TCGA dataset [36]. Kaplan-Meier survival curves were generated and differences in survival were assessed by log-rank test (p < 0.05 considered statistically significant) using GraphPad PRISM 6.0. (D) Survival analyses of 23 genes close to ERBB2 locus in breast cancers. Tumour samples with the top 25% expression levels of each gene (considered “high”) were compared to the remaining tumors (“low”). Kaplan-Meier survival curves were generated and differences in DFS and DMFS survival were assessed by log-rank Mantel-Cox test using GraphPad PRISM 6.0. Analyses were performed in the Ur-Rehman and the TCGA datasets with all tumours (n = 1, 102 and n = 493, respectively), and for HER2-high (HER2+) cancers only. The p value of each test was color-coded as indicated in the scale on the right. HR: Hazard ratio associated with GSDMB gene high expression in each dataset.

Figure 2. GSDMB gene amplification and protein expression in HER2-positive breast cancer

(A) Representative breast tumours with positive and negative HER2 and GSDMB amplification. Panels a, c, e: represent an example of breast carcinoma HER2-positive (red signal) without GSDMB amplification. Panels b, d, f: represent an example of case HER2-positive with GSDMB amplification (blue signal). The lower lane shows the merged image (e of a-c and f of b-d panels). Green signals represent CEP7. Magnification x 63x; insets x 120. (B) Immunohistochemical analysis of HER2 and GSDMB in normal tissue (panel a, b), and in breast carcinoma cases (panels c–f). Low or undetectable GSDMB staining was observed in normal breast tissue (b) or in lesions with GSDMB gene copy number of < 4 signals/nucleus (d) (considered non-amplified tumours). Intense GSDMB staining in a case with > 10 GSDMB copy number (f) considered as GSDMB-positive tumour. Scale bar measures 100 micrometers.

Figure 3. GSDMB gene amplification and/or protein expression associates with relapse in HER2-positive breast cancer samples under neoadjuvant and adjuvant treatment regimens

Relapse-free survival curves in HER2-positive breast cancer patients (n = 95) treated with neoadjuvant settings in relation to (A) GSDMB gene amplification (p = 0.0002) and (B) GSDMB expression scored by immunohistochemistry (IHC, p = 0.0004). (C) Relapse-free survival curves in HER2- positive breast cancer patients (n = 53) treated with adjuvant settings in relation to GSDMB protein expression measured by IHC (p = 0.007). Statistical differences, HR and p-value, were calculated via log-rank test. Red line: GSDMB amplified and expressed patients, green line: patients without GSDMB amplification and expression.

Figure 4. GSDMB expression increases survival to trastuzumab treatment

(A) Analysis of GSDMB expression by western blot in SKBR3 cells (empty vector, Control) and GSDMB-overexpressing cells (GSDMB). GAPDH was used as a loading control. (B) Analysis of cell growth (measured with Alamarblue) in control and GSDMB-overexpressing SKBR3 cells treated with 1 mg/ml trastuzumab for 72 h; (C) Reduction of cell growth in control (shNTC) and GSDMB-silenced (two shRNAs) HCC1954 cells treated with 5 mg/ml trastuzumab. (D)Analysis of cell death (measured with Annexin V and propidium iodide) in control and GSDMB-overexpressing SKBR3 cells 72 h after trastuzumab treatment. Bars represent mean + SEM of three independent experiments. (E) GSDMB and HER2 mRNA expression in Patient Derived Xenografts (PDX) from HER2-negative (PDX01, 02) and HER2-positivie breast carcinomas. PDXs were classified as trastuzumab sensitive (PDX118, 173) or resistant (PDX67, 288) according to in vivo patient response [42]. (F) GSDMB and HER2 mRNA expression in a trastuzumab-sensitive PDX (PDX118) and its trastuzumab-resistant derived PDXs (T-R1-R5), obtained by chronic in vivo treatment with trastuzumab. mRNA values were normalized using GAPDH as housekeeping gene. Bars represent mean + SEM expression of three experiments by quantitative PCR.