BackgroundTo obtain a deep understanding of the mechanism by which breast cancer develops, the genes involved in tumorigenesis should be analyzed in vivo. Mouse mammary gland can regenerate completely from a mammary stem cell (MaSC), which enables us to analyze the effect of gene expression and repression on tumorigenesis in mammary gland regenerated from genetically manipulated MaSCs. Although lentiviral and retroviral systems have usually been applied for gene transduction into MaSCs, they are associated with difficulty in introducing long, repeated, or transcriptional termination sequences. There is thus a need for an easier and quicker gene delivery system.MethodsWe devised a new system for gene delivery into MaSCs using the piggyBac transposon vectors and electroporation. Compared with viral systems, this system enables easier and quicker transfection of even long, repeated, or transcriptional termination DNA sequences. We designed gene expression vectors of the transposon system, equipped with a luciferase (Luc) expression cassette for monitoring gene transduction into regenerative mammary gland in mice by in-vivo imaging. A doxycycline (Dox)-inducible system was also integrated for expressing the target gene after mammary regeneration to mimic the actual mechanism of tumorigenesis.ResultsWith this new gene delivery system, genetically manipulated mammary glands were successfully reconstituted even though the vector size was > 200 kb and even in the presence of DNA elements such as promoters and transcription termination sequences, which are major obstacles to viral vector packaging. They differentiated correctly into both basal and luminal cells, and showed normal morphological change and milk production after pregnancy, as well as self-renewal capacity. Using the Tet-On system, gene expression can be controlled by the addition of Dox after mammary reconstitution. In a case study using polyoma-virus middle T antigen (PyMT), oncogene-induced tumorigenesis was achieved. The histological appearance of the tumor was highly similar to that of the mouse mammary tumor virus-PyMT transgenic mouse model.ConclusionsWith this system, gene transduction in the mammary gland can be easily and quickly achieved, and gene expression can be controlled by Dox administration. This system for genetic manipulation could be useful for analyzing genes involved in breast cancer.Electronic supplementary materialThe online version of this article (10.1186/s13058-018-1086-8) contains supplementary material, which is available to authorized users.

Screening for oncogenes has mostly been performed by in vitro transformation assays. However, some oncogenes might not exhibit their transforming activities in vitro unless putative essential factors from in vivo microenvironments are adequately supplied. Here, we have developed an in vivo screening system that evaluates the tumorigenicity of target genes. This system uses a retroviral high-efficiency gene transfer technique, a large collection of human cDNA clones corresponding to ~70% of human genes and a luciferase-expressing immortalized mouse mammary epithelial cell line (NMuMG-luc). From 845 genes that were highly expressed in human breast cancer cell lines, we focused on 205 genes encoding membrane proteins and/or kinases as that had the greater possibility of being oncogenes or drug targets. The 205 genes were divided into five subgroups, each containing 34-43 genes, and then introduced them into NMuMG-luc cells. These cells were subcutaneously injected into nude mice and monitored for tumor development by in vivo imaging. Tumors were observed in three subgroups. Using DNA microarray analyses and individual tumorigenic assays, we found that three genes, ADORA2B, PRKACB and LPAR3, were tumorigenic. ADORA2B and LPAR3 encode G-protein-coupled receptors and PRKACB encodes a protein kinase A catalytic subunit. Cells overexpressing ADORA2B, LPAR3 or PRKACB did not show transforming phenotypes in vitro, suggesting that transformation by these genes requires in vivo microenvironments. In addition, several clinical data sets, including one for breast cancer, showed that the expression of these genes correlated with lower overall survival rate.