I am facing problems in preparing egfp-tagged mouse cells lines while I could easily get human cell line transformed by the same construct. I am using pLJM1-EGFP (Addgene) along with psPAX2 (Addgene) and pMD2.G (Addgene). I used lipofectamine 3000 for the transfection in HEK293FT cells. The MDA-MB231 human TNBC cell line was showing stable GFP signals which increased to 100% after puromycin selection. However, the mouse cell lines py230 and 4T1 showed initial GFP signals that faded away with time. After a few rounds of puromycin-based selection very few cells with GFP signals were seen but with a changed star-shaped morphology.

Could anyone kindly suggest a better way to have GFP tagged mouse cell lines (py230 & 4T1)?