As the catalog of oncogenic driver mutations is expanding, it becomes clear that alterations in a given gene might have different functions and should not be lumped into one class. The transcription factor GATA3 is a paradigm of this. We investigated the functions of the most common GATA3 mutation (X308_Splice) and five additional mutations, which converge into a neoprotein that we called “neoGATA3,” associated with excellent prognosis in patients. Analysis of available molecular data from >3000 breast cancer patients revealed a dysregulation of the ER-dependent transcriptional response in tumors carrying neoGATA3-generating mutations. Mechanistic studies in vitro showed that neoGATA3 interferes with the transcriptional programs controlled by estrogen and progesterone receptors, without fully abrogating them. ChIP-Seq analysis indicated that ER binding is reduced in neoGATA3-expressing cells, especially at distal regions, suggesting that neoGATA3 interferes with the fine tuning of ER-dependent gene expression. This has opposite outputs in distinct hormonal context, having pro- or anti-proliferative effects, depending on the estrogen/progesterone ratio. Our data call for functional analyses of putative cancer drivers to guide clinical application.

1As the catalogue of oncogenic driver mutations is expanding, it is becoming clear that alterations 2 in a given gene should not be lumped into one single class, since they might have different functions. The 3 transcription factor GATA3 is a paradigm of this. Here, we address the functions of the most common 4 GATA3 mutation (X308_Splice) which generates a neoprotein that we designate as neoGATA3, associated 5 with good patient prognosis. Based on extensive analyses of molecular and clinical data from 6 approximately 3000 breast cancer patients, supported by mechanistic studies in vitro, we show that 7 neoGATA3 interferes with the transcriptional programs controlled by estrogen and progesterone 8 receptors, without fully abrogating them. This has opposite outputs in the pre-or post-menopausal 9 hormonal context, having pro-or anti-proliferative effects, respectively. NeoGATA3 is an example of a 10 context-and stage-dependent driver mutation. Our data call for functional analyses of putative cancer 11 drivers to guide clinical application. 12 13Several evidences indicate that GATA3 is involved in the activation of the mammary 1 differentiation program: 1) in normal tissue, it is necessary for the formation of the luminal compartment 2 [9]; 2) GATA3 expression in BC strongly correlates with estrogen receptor (ER) expression [10]; 3) GATA3 3 functions in a complex with FOXA1 and ER to recruit RNA polymerase and enhance transcription of ER-4 responsive genes [11]; and 4) ectopic expression in GATA3-negative basal-like BC cells is sufficient to 5 induce luminal differentiation and inhibit tumor dissemination [12]. Consistent with this function, GATA3 6 expression decreases during the progression to metastatic BC [13]. The high frequency of GATA3 7 mutations in BC supports the idea that they are driver mutations, but whether they result in loss-of-8 function (LOF) or gain-of-function (GOF) is not fully clear. Most GATA3 mutations are rare or unique 9 frameshift indels (insertion/deletions) distributed along the 3' end of the gene ( Figure 1A), consistent with 10