Open Access

One of the hallmarks of cancer is the escape of the transformed cells from apoptosis. Therefore, the identification of survival genes, allowing cancer cells to circumvent programmed cell death, could provide new diagnostic markers as well as targets for therapeutic intervention. A well known transcription factor regulating the balance between pro- and anti- apoptotic factors is NF-kappaB, which is strongly induced by tumor necrosis factor alpha (TNFalpha). When cells are stimulated by TNFalpha their response is biphasic with an initial NF-kappaB induction of survival genes which is overridden by the subsequent activation of initiator caspases triggering apoptosis. By combining gene trap mutagenesis with site specific recombination a strategy was developed, which enriches for genes induced by TNFalpha in the human breast cancer cell line MCF-7. The strategy relies on a one way gene expression switch based on Cre/loxP mediated recombination, which uncouples the expression of a marker gene from the trapped cellular promoter thereby enabling the recovery of genes that are only transiently induced by TNFalpha. The marker gene used in these experiments was a dominant negative variant of the TNFalpha-receptor associated protein FADD (dnFADD), which blocks the apoptotic branch of the TNFalpha induced signaling pathway. Initial experiments indicated that MCF-7 cells expressing high levels of dnFADD were insensitive to TNFalpha induced apoptosis and therefore suitable for the installment of a one way gene expression switch susceptible to Cre/loxP mediated recombination. A MCF-7 reporter clone harboring the recombinase dependent gene expression switch was infected with the gene trap retrovirus U3Cre, which inserts the Cre recombinase gene into a large collection of chromosomal sites. Insertion of Cre downstream of an active cellular promoter induces dnFADD expression from the gene expression switch enabling the cells to block TNFalpha triggered apoptosis. From a gene trap integration library containing approximately 2000000 unique proviral integrations, 69 unique TNFalpha inducible gene trap insertion sites were recovered in a two step selection procedure. Sequencing of the genomic regions adjacent to the insertion sites, which were obtained by inverse PCR (gene trap sequence tags, GTSTs), and data base analysis revealed that 42% of the GTSTs belonged to annotated genes, 13% to known cDNAs with open reading frames, 17% to Genscan predicted genes, 9% to ESTs, 9% to repetitive sequences and 10% to unannotated genomic sequence. Overall, 44% of the annotated genes recovered in this screen were directly or indirectly related to cancer, indicating that the gene trap strategy developed here is suitable for the identification of cancer relevant genes. Analysis of the expression patterns of the trapped and annotated genes in wild type cells revealed that 19 out of 24 genes were either up- or down- regulated by a factor of at least 1.45 by TNFalpha. A large fraction of the gene trap insertions were located upstream, in introns or in opposite orientation to annotated transcripts, indicating that the strategy efficiently recovers non-coding RNAs (ncRNAs). While the biological significance of these transcripts still needs to be elucidated, they fall into two main categories. The first category includes gene trap insertions upstream of genes, which could either represent regulatory RNAs interacting with promoter elements or transcripts driven by bidirectional promoters. The second includes inverse orientation gene trap insertions in introns of annotated genes suggesting the presence of natural antisense transcripts (NATs). Interestingly, more than 50% of all antisense integrations are located downstream of transcription start sites predicted by different algorithms supporting the existence of RNAs transcribed from the corresponding genomic regions. Intronic integrations on the coding strand could be derived from cryptic splicing, alternative promoter usage or additional, so far uncharacterized transcripts. Preliminary functional analysis of two genes recovered in this screen encoding the transcription factor ZFP67 and the FLJ14451 protein revealed that FLJ14451 but not ZFP67 inhibited anchorage independent growth in soft agar, suggesting that FLJ14451 might have some tumor suppressor functions. In summary, besides identifying a putative tumor suppressor protein, the present experiments have shown that gene trapping is useful in identifying non-coding transcripts in living cells and may turn out to be the method of choice in characterizing these transcripts whose functions are still largely unknown.