A TGF-ß-induced cis-regulatory long non-coding RNA eSNAI1 activates SNAI1 enhancer and stimulates epithelial-to-mesenchymal transition of cancer cells [CRISPR screen, activation]. A TGF-ß-induced cis-regulatory long non-coding RNA eSNAI1 activates SNAI1 enhancer and stimulates epithelial-to-mesenchymal transition of cancer cells [CRISPR screen, activation]

Transforming growth factor-β (TGF-β) signaling stimulates cell movement and plasticity by inducing epithelial-to-mesenchymal transition (EMT). This process is aberrantly activated in cancer and facilitates tumor cell migration, invasion, metastasis, and therapy resistance. In this study, we identified the lncRNA eSNAI1 (SCREEM2) as a potent activator of TGF-β/SMAD signaling and SNAI1 expression. eSNAI1 depletion reduces TGF-β-induced EMT, migration, in vivo extravasation, stemness, and chemotherapy resistance of breast cancer cells. eSNAI1 promotes TGF-β/SMAD signaling by inhibiting TGF-β type I receptor polyubiquitination and proteasomal degradation. eSNAI1 stimulates the expression of the nearby gene SNAI1, which encodes the EMT transcription factor SNAI1 that facilitates TGF-β/SMAD signaling by retaining the inhibitory SMAD7 in the nucleus. Mechanistically, eSNAI1 acts as an enhancer RNA (eRNA) to potentiate SNAI1 expression through in-cis regulation. Furthermore, we uncovered that eSNAI1 interacts with and reinforces the binding of the co-activator bromodomain-containing protein 4 (BRD4) to the H3 lysine 27 acetylation (H3K27ac)-enriched SNAI1 enhancer region. Our findings identify eSNAI1 as a potent activator of TGF-β signaling and a promising therapeutic target to mitigate overactive TGF-β signaling and EMT in cancer cells. Overall design: CRISPR-activation screen for a custom set of lncRNAs using the SAM system. MDA-MB231 expressing a TGFbeta inducible GFP reporter and dCas9-VP64 were transduced with a guide library. An early time point (ETP) control sample was taken at day 5 post-transduction. AT day 7 post transduction, cells were treated with TGF-beta for 24 hours. An unsorted control sample was taken, and the 15% highest and lowest GFP expressors were sorted. The screen was performed in two replicates