Abstract
Triple negative breast cancer (TNBC) is a heterogeneous disease with diverse transcriptomic and genomic characteristics, however many of the deregulated signalling pathways underlying these tumours are poorly understood. Based on an integrative approach of diverse molecular features, we aimed to identify TNBC subgroups of clinical relevance.
We applied COpy Number and EXpression In Cancer (CONEXIC), a gene module-based method, to integrate Affymetrix SNP6.0 copy number and microarray-based gene expression data from TNBCs of our in-house Guy’s (n=88) and METABRIC (n=112) cohorts. Concurrent gene modules with the same driver gene (called modulator) were assessed for enrichment of biological processes. TNBC subgroups, as defined by CONEXIC-modelled decision trees, were molecularly characterised by their levels of genomic instability, TNBC subtypes, gene signatures and clinical relevance.
A total of 88 independent gene modules with an average size of 193 genes (range 12 – 463) were identified in Guy’s-TNBCs and METABRIC-TNBCs by CONEXIC. Two Guy’s-TNBC modules (Guy’s-Module 27 and 33) and one METABRIC-TNBC module (METABRIC-Module 15) had EXO1 as one of their modulators and showed a significant enrichment in overlapping genes. Classifying TNBCs based on the CONEXIC-modelled EXO1-driven decision trees resulted in six subgroups. In both Guy’s-Module 33 and METABRIC-Module 15, the subgroups with the highest EXO1 expression displayed the highest level of genomic instability, defined by CIN70, but the lowest expression of MAP-kinase inactivating genes. In both cohorts this subgroup was enriched for the basal-like 1 Pietenpol-TNBC subtype. Applying the same decision tree clustering to transcriptomic data of external TNBC cohorts recapitulated the associations between EXO1, high genomic instability and low MAP-kinase inactivation.
In conclusion, we have established a classification of TNBCs based on an integrative approach, which identified a subgroup with specific molecular features. The clinical relevance of this TNBC subtype is currently under investigation.
We applied COpy Number and EXpression In Cancer (CONEXIC), a gene module-based method, to integrate Affymetrix SNP6.0 copy number and microarray-based gene expression data from TNBCs of our in-house Guy’s (n=88) and METABRIC (n=112) cohorts. Concurrent gene modules with the same driver gene (called modulator) were assessed for enrichment of biological processes. TNBC subgroups, as defined by CONEXIC-modelled decision trees, were molecularly characterised by their levels of genomic instability, TNBC subtypes, gene signatures and clinical relevance.
A total of 88 independent gene modules with an average size of 193 genes (range 12 – 463) were identified in Guy’s-TNBCs and METABRIC-TNBCs by CONEXIC. Two Guy’s-TNBC modules (Guy’s-Module 27 and 33) and one METABRIC-TNBC module (METABRIC-Module 15) had EXO1 as one of their modulators and showed a significant enrichment in overlapping genes. Classifying TNBCs based on the CONEXIC-modelled EXO1-driven decision trees resulted in six subgroups. In both Guy’s-Module 33 and METABRIC-Module 15, the subgroups with the highest EXO1 expression displayed the highest level of genomic instability, defined by CIN70, but the lowest expression of MAP-kinase inactivating genes. In both cohorts this subgroup was enriched for the basal-like 1 Pietenpol-TNBC subtype. Applying the same decision tree clustering to transcriptomic data of external TNBC cohorts recapitulated the associations between EXO1, high genomic instability and low MAP-kinase inactivation.
In conclusion, we have established a classification of TNBCs based on an integrative approach, which identified a subgroup with specific molecular features. The clinical relevance of this TNBC subtype is currently under investigation.
| Original language | English |
|---|---|
| Journal | NCRI Conference |
| Publication status | Accepted/In press - 2016 |
