A novel embryonic plasticity gene signature that predicts metastatic competence and clinical outcome

Currently, very few prognosticators accurately predict metastasis in cancer patients. In order to complete the metastatic cascade and successfully colonize distant sites, carcinoma cells undergo dynamic epithelial-mesenchymal-transition (EMT) and its reversal, mesenchymal-epithelial-transition (MET). While EMT-centric signatures correlate with response to therapy, they are unable to predict metastatic outcome. One reason is due to the wide range of transient phenotypes required for a tumor cell to disseminate and recreate a similar histology at distant sites. Since such dynamic cellular processes are also seen during embryo development (epithelial-like epiblast cells undergo transient EMT to generate the mesoderm, which eventually redifferentiates into epithelial tissues by MET), we sought to utilize this unique and highly conserved property of cellular plasticity to predict metastasis. Here we present the identification of a novel prognostic gene expression signature derived from mouse embryonic day 6.5 that is representative of extensive cellular plasticity, and predicts metastatic competence in human breast tumor cells. This signature may thus complement conventional clinical parameters to offer accurate prediction for outcome among multiple classes of breast cancer patients.

9 Determining metastatic competence of commonly used prostate cancer cell lines based on concordance with the E6.5 gene expression signature.
Suppl Figure 2: Kaplan-Meier curves of recurrence-free survival (RFS) in breast cancer patients stratified by expression of the E6.5-, the adult-, or the EMT gene expression signature, were plotted using patient data from four independent studies 5,6,7,8 [Data from one of the cohorts is presented in Fig. 5, the remaining are presented here]. Log-rank tests were used to assess statistical significance of differences in RFS based on concordance with each signature. p<0.05 was considered as the necessary cutoff for statistical significance. , VIM and TWIST1, were plotted using patient data from three independent studies [GSE7390 3 , GSE11121 4 , GSE20685 1 ]. Log-rank tests were used to assess 2 statistical significance of differences in DMFS based on concordance with expression of each gene. p<0.05 was considered as the necessary cutoff for statistical significance.

Suppl Figures 6, 7: Kaplan-Meier curves of recurrence-free survival (RFS) in breast cancer patients stratified by expression of CDH1, PRRX1, TCF-8, SNAI2
, VIM and TWIST1, were plotted using patient data from three independent studies [GSE12276 6 , GSE4922_UPPSALA 7 , and GSE1653 8 ]. Log-rank tests were used to assess statistical significance of differences in RFS based on concordance with each signature. p<0.05 was considered as the necessary cutoff for statistical significance.
Suppl Figure 8: Well-characterized EMT-and stem cell-markers used for comparative analyses with the newly identified E6.5 gene expression signature. 39 genes were selected on the basis of their established EMT/stem cell-related functions. Panel A shows the full list of compared genes. The scatter plot in panel B shows compiled Kaplan-Meier curves of distant metastasis-free survival (DMFS) in breast cancer patients stratified by expression of EMT-and stem cell-markers listed in A, plotted using patient data from four independent studies 1, 2, 3, 4 . The scatter plot in panel C shows compiled Kaplan-Meier curves of recurrence-free survival (RFS) in breast cancer patients stratified by expression of EMT-and stem cell-markers listed in A, plotted using patient data from four independent studies 5, 6, 7, 8 .

Adult Up
Suppl