• circRNAs
• melanoma
• transposon

Circular RNAs (circRNAs) are a novel class of RNA molecules characterized by circular structures, meaning they lack 3’ end and 5’ ends, polyadenylated tails, and 5’ CAPs. Because of their structure they are more resistant to miRNAs and exonucleases, with a median half-life ranging from 18.8 to 23.7 hours, compared with 4.0–7.4 hours for their cognate linear RNAs. In contrast to linear transcripts that are produced by canonical splicing, circRNAs are generated from linear pre mRNAs by “back-splicing”, a phenomenon where the downstream donor splice site is linked to an upstream acceptor splice site.

Many recent studies now indicate that circRNAs have important roles in many biological processes, including in disease such as cancer. The aim of our project is to identify and characterize novel circRNAs, since we hypothesize that they could be involved in melanoma.

We performed RNA-seq analysis to identify 42 significantly differentially expressed circRNAs comparing Hermes1, Hermes2, Hermes3, Hermes4 human melanocytes cell lines, 1205Lu, WM164, WM793, WM35 and SKMel28 human melanoma cell lines and Hermes1-BRAFV600E mutated cell line. We validated 22 circRNAs through qRT-PCR and Sanger sequencing over the back-splice junction. By comparing circRNAs expression in Hermes1, Hermes2 and Hermes3 melanocytes cell lines, Hermes1-BRAFV600E mutated cell line and WM164, WM793, 1205Lu and A375 melanoma cell lines through qRT-PCR we selected two candidates for further characterization: circDNAJC2 and circPMS1, both upregulated in melanoma. We knocked them down through siRNAs and showed that their silencing affects the proliferation of WM164 and A375 melanoma cell lines. In parallel, we developed the EF1alpha-splitGFP-ZKSCAN-Blast transposon for a stable in vitro and in vivo circRNAs overexpression. To date, with this construct we overexpressed the splitGFP reporter in vitro, to confirm that the circRNAs overexpression is doable without recurring neither to lentiviral vectors (that often have packaging issues) or plasmids (that do not allow us to have a stable circRNAs overexpression). Instead for a global and tissue-specific in vivo circRNAs expression we made constitutive EF1alpha-ZKSCAN-splitGFP construct and dox-inducible cTRE-ZKSCAN-splitGFP construct, respectively. To make these melanoma mouse models we used the faster and cost-effective ES cell-genetically engineered mouse model (ESC-GEMM) platform, that relies on a panel of ES cells derived from existing GEMMs harboring 12 different combinations of the LSL-BrafV600E, LSL-NrasQ61R, PtenFlox, and Cdkn2aFlox driver alleles. These ESCs are modified in vitro and used to generate chimeric mice. The ES cell-derived melanocytes in ESC-GEMM chimeras harbor all alleles necessary to initiate melanomagenesis and modulate the expression or function of the gene(s)-of-interest.

We will use these tools to functionally analyze our circRNAs candidates in vitro and in vivo.