Novel genes associated with lymph node metastasis in triple negative breast cancer

Abstract

Triple negative breast cancer (TNBC) is the most aggressive breast cancer subtype with the worst prognosis and no targeted treatments. TNBC patients are more likely to develop metastases and relapse than patients with other breast cancer subtypes. We aimed to identify TNBC-specific genes and genes associated with lymph node metastasis, one of the first signs of metastatic spread. A total of 33 TNBCs were used; 17 of which had matched normal adjacent tissues available, and 15 with matched lymph node metastases. Gene expression microarray analysis was used to reveal genes that were differentially expressed between these groups. We identified and validated 66 genes that are significantly altered when comparing tumours to normal adjacent samples. Further, we identified 83 genes that are associated with lymph node metastasis and correlated these with miRNA-expression. Pathway analysis revealed their involvement in DNA repair, recombination and cell death, chromosomal instability and other known cancer-related pathways. Finally, four genes were identified that were specific for TNBC, of which one was associated with overall survival. This study has identified novel genes involved in LN metastases in TNBC and genes that are TNBC specific that may be used as treatment targets or prognostic indicators in the future.

Figure 1. Unsupervised hierarchical clustering of TNBC specific genes in the second cohort and the TCGA cohort.

Through analysis of our second cohort (48 non-TNBC, 16 TNBC samples) we identified 28 TNBC specific genes, which were used for unsupervised hierarchical clustering of TNBC and non-TNBC samples (A). The TCGA breast cancer cohort (313 non-TNBC, 55 TNBC samples) identified 14 TNBC specific genes. The unsupervised hierarchical clustering of the 14 TNBC specific genes successfully clustered the TCGA breast cancer cohort in TNBC patients and non-TNBC patients (B). There are four common TNBC specific genes in both cohorts (our second cohort and the TCGA breast cancer cohort) – these four genes are ANKRD30A, ANP32E, DSC2, IL6ST. (C) shows the unsupervised hierarchical clustering of these four genes in the second cohort and (D) shows the clustering of these genes in the TCGA breast cancer cohort. TNBC samples are shown in light blue and non-TNBC samples are shown in red in the sample tree on the left (y-axis). Genes are clustered along the x-axis. Low gene expression is shown in blue, high expression is shown in red and equivocal expression is shown in grey.

Figure 2. Kaplan-Meier analysis of four validated TNBC specific genes.

Kaplan-Meier analysis was performed on the four validated TNBC specific genes (ANKRD30A, ANP32E, DSC2, IL6ST). The analysis of (A) ANKRD30A in TCGA TNBC cohort, Hazard ratio (HR) = 0.5326; p = 0.37. (B) ANP32E in TCGA TNBC cohort; HR = 1.137; p = 0.85. (C) DSC2 in TCGA TNBC cohort; HR = 0.4183; p = 0.21. (D) IL6ST in TCGA TNBC cohort; HR = 0.1921; p = 0.0421*. The Hazard ratio (HR) is smaller than 1 for three genes, which means down-regulated genes decrease overall survival in TNBC, whereas up-regulated genes would increase overall survival. A significant correlation to overall survival can be seen for IL6ST (p < 0.05). The p-value shows the significance of the difference between the survival curves.

Figure 3. Venn Diagram for identification of genes associated with lymph node metastasis.

All samples are categorised in lymph node positive primary tumours (IDC+), lymph node negative primary tumours (IDC−), lymph node metastases (LNmet) and matched normal adjacent tissue for lymph node positive cases (NAT+) and lymph node negative cases (NAT−). Gene expression was analysed in all three tumour categories compared to their matched NAT samples. Highlighted in red are the 83 genes that are differentially expressed in IDC+ and LNmet samples but not in IDC− samples.