Hoping someone can provide some additional insight or suggestions about a transgene expression problem in vivo. I have used piggyBac to re-express a wild-type transgene in murine tumor-derived cells in which both copies of the endogenous gene have been inactivated. Because this gene is generally growth suppressive, I used a lox-STOP-lox approach to overcome the initial selection pressure against the cells with an integrated transposon.

I used piggyBac to integrate the following into the parental cells:

EF1a | lox-STOP-lox | Transgene

(These are the "off" cells)

I then introduced Cre into those cells to get:

EF1a | Transgene

(These are the "on" cells)

In vitro the expression is beautiful. ~12X expression of the transgene in the "on" cells compared to the "off" cells, and very little leakage in the "off" cells compared to the parental cells. We have implanted these cells twice into syngeneic immunocompetent mice and after tumors form (~6 weeks later) there is no difference in transgene expression between the "off" and "on" tumors. The host mice are not deficient in the transgene, and there are no tags on the wild-type transgene; both the "off" and "on" cells do contain a puromycin selection marker, which could be potentially immunogenic, but to rule out any immune-mediated mechanism we grew tumors in NSG mice and got exactly the same result--no difference in expression.

The transgene is known to be downregulated during tumor progression via methylation of the endogenous promoter. My impression is that EF1a (compared to, say, CMV) is relatively insensitive to epigenetic silencing in vivo. I know that any transgene can be silenced by higher-order genome organization, but any idea why this would only be occurring in vivo?

Any ideas/insight/suggestions about troubleshooting this would be appreciated.