p107 during G0 in MEFs. MYC is a direct target of E2F1 in human fibroblasts. E2F3, E2F5 and TFDP1, which have higher expression in ER2 tumors, are all induced by MYC in MCF-7 cells. Zeller et al. found the E2F1 binding motif is enriched 16-fold in clusters of MYC-binding genes, and 37-fold within the subset of E-box containing 
genes. In our study, the genes shared between gene sets relating to MYC- and E2F-action were predominantly proliferation-associated genes. In this context, we believe that dysregulation or constitutive activation of processes regulated by MYC in conjunction with increased E2F activity may lead to uncontrolled cell cycle progression and proliferation, such as we see in ER2 tumors and particularly in the basal subtype within the ER2 tumors. Kreike et al. showed that the basal subtype may be further divided into five subgroups on the basis of gene expression profiles. A group of proliferation-associated genes was among those contributing to the clustering. The enrichment of MYC- and E2F-regulated genes in the genes partitioning the basal tumors in the cohort of Kreike et al. may prove a valuable avenue of investigation. Our findings may also have important therapeutic implications. A recent bioinformatic study indicated a significant association between the set of genes activated by MYC in HMECs, and genes repressed on treatment of MCF-7 or HL-60 cells with the PI3K signaling pathway inhibitors wortmannin or LY-294002. MYC-activated genes had lower expression in cells treated with wortmannin or LY-294002, suggesting that PI3K inhibitors repress MYC-activation. Our study provides strong evidence for MYC-action in basal tumors, suggesting that PI3K inhibitors, and other potential repressors of MYC-action, should be investigated as therapeutic candidates for these tumors which have no targeted therapies at present. We anticipate that the biological insights generated by this study will prove valuable in the development of new therapeutic strategies for the poor prognosis ER2 tumors and in particular the basal subtype. We conclude that over-expression or constitutive activation of MYC, possibly in conjunction with elevated E2F activity, may contribute to increased proliferation in ER2 breast tumors, particularly in the basal subgroup. Supporting Information Dataset S1 GSEA results from first screen. The file contains a order Thiazovivin zipped 19478133 archive of the HTML results of the second GSEA screen. Reports for enriched or depleted datasets may be accessed from the index.htmllink within the my_analysis.GseaPreranked.1219975950625.rpt.zip ZIP archive. Enrichment plots are available for each of the 50 most enriched and depleted gene sets. Found at: doi:10.1371/journal.pone.0004710.s001 MYC and E2F ER_Transcript”, PGR_Transcript”, and ERBB2_Transcriptas above; and Tumor_Grade- Elston Ellis grading. Two major clusters of cell-cycle-associated genes were observed in all three heatmaps; in each dataset, the highest expression of cell-cycle genes was observed in 22440900 the basal samples, as marked. Found at: doi:10.1371/journal.pone.0004710.s004 Gene Name, Gene Symbol, Z score, BY-adjusted P-value, status as an ERA gene, transformed Weighted Average Ratio, as well as whether they have an association with, E2-action, MYC-action, E2F-action, the cell cycle or significantly enriched or depleted chromosomal position in several different datasets. A negative Overall Z scoreindicates a probe set with higher intensity in ER2 tumors. Found at: doi:10.1371/jour
