Analysis of Hierarchical Organization in Gene Expression Networks Reveals Underlying Principles of Collective Tumor Cell Dissemination and Metastatic Aggressiveness of Inflammatory Breast Cancer

Clusters of circulating tumor cells (CTCs), despite being rare, may account for more than 90% of metastases. Cells in these clusters do not undergo a complete epithelial-to-mesenchymal transition (EMT), but retain some epithelial traits as compared to individually disseminating tumor cells. Determinants of single cell dissemination versus collective dissemination remain elusive. Inflammatory breast cancer (IBC), a highly aggressive breast cancer subtype that chiefly metastasizes via CTC clusters, is a promising model for studying mechanisms of collective tumor cell dissemination. Previous studies, motivated by a theory that suggests physical systems with hierarchical organization tend to be more adaptable, have found that the expression of metastasis-associated genes is more hierarchically organized in cases of successful metastases. Here, we used the cophenetic correlation coefficient (CCC) to quantify the hierarchical organization in the expression of two distinct gene sets, collective dissemination-associated genes and IBC-associated genes, in cancer cell lines and in tumor samples from breast cancer patients. Hypothesizing that a higher CCC for collective dissemination-associated genes and for IBC-associated genes would be associated with retention of epithelial traits enabling collective dissemination and with worse disease progression in breast cancer patients, we evaluated the correlation of CCC with different phenotypic groups. The CCC of both the abovementioned gene sets, the collective dissemination-associated genes and the IBC-associated genes, was higher in (a) epithelial cell lines as compared to mesenchymal cell lines and (b) tumor samples from IBC patients as compared to samples from non-IBC breast cancer patients. A higher CCC of both gene sets was also correlated with a higher rate of metastatic relapse in breast cancer patients. In contrast, neither the levels of CDH1 gene expression nor gene set enrichment analysis (GSEA) of the abovementioned gene sets could provide similar insights. These results suggest that retention of some epithelial traits in disseminating tumor cells as IBC progresses promotes successful breast cancer metastasis. The CCC provides additional information regarding the organizational complexity of gene expression in comparison to GSEA. We have shown that the CCC may be a useful metric for investigating the collective dissemination phenotype and a prognostic factor for IBC