In this study, we characterized a PhiC31-IMSI system in cancer cell lines that allows for the control and precise insertion of a transgene in a predetermined genomic locus. To broaden the utility of this system, we combined it with a doxycycline-inducible expression system of the transgene as well as a luciferase reporter that features in vivo live imaging in animal models.
The phage serine recombinase PhiC31 is adopted in this system in order to facilitate the integration of a transgene into the host genome. First, a docking site carrying an attP site upstream of a promoter-less neoR gene is randomly inserted as a single copy into the host genome by means of selection for puromycin resistance. The incoming vector, carrying an attB site in front of a promoter, is subsequently inserted via PhiC31-mediated integration into the genome (Figure 1A). The incoming vector also carries the transgene under the control of the desired promoter. In addition, using the Flp recombinase and thymidine kinase for negative selection, the neo and puro selection markers can be excised leaving behind only a clean integration of the transgene . However, this aspect was not examined in the present study since maintenance of the selection markers offers the advantage for easy isolation of pure populations of cancer cells from tumour models that, for example, resist or develop resistance to various therapeutic schemes. In order to further advance the IMSI, we introduced the expression of luciferase that allows for in vivo live imaging and monitoring of tumour growth and metastasis. We implemented this feature by constitutively expressing luciferase as a fusion transgene with the PAC gene of the docking site (Figures 1 and 2). As far as we know, this is the first luciferase-containing IMSI.
Using this system, we were able to show that all of the resistant colonies in two different cancer cell lines, Pc-3 and SKOV-3, were a result of the integration of the incoming vector in the desired genomic locus. The DU145 cancer line was not permissive to PhiC31-mediated integration. Our unpublished data (K. Nishino, A. Nagy) indicate that this might be either the result of methylation of the Att P site or due to the fact that the Att P is missing as a result of endogenous nuclease activity. Taken our combined experiences with IMSI in cancer and Embryonic Stem cell lines [13, 39] we do not expect considerable limitations in the use of this method. In agreement with previous studies describing pseudo-attP sites in human cell lines, we did not observe any multiple integrations in the same clone [35, 38]. It is possible that integration at pseudo-attP sites may occur yet G418 selection selects against these events.
Insertion of various transgenes in the same genomic locus resulted in similar expression levels when isogenic clones were compared (Figure 4). Such reproducible expression allows for various applications, such as the characterization of methylation and transgene orientation on transgene expression, as well as the characterization of promoter and enhancer elements [46–49]. In this study, we compared the levels of EGFP expressed either directly under the control of a strong promoter or as a bicistronic gene under the control of the same promoter. The GFP levels were different in the two cases, with the levels dropping an order of magnitude when EGFP was expressed in a bicistronic fashion under the control of an IRES sequence (Figure 4 and Table 1). We were also able to show that all the aspects of expression levels characterized in vitro were maintained after establishment of in vivo xenografts, implying that this system is robust and can be utilized for in vivo studies (Figure 4).
Many studies involving transgenes require temporal control of their expression. Thus, we coupled our system with the doxycycline inducibility. In addition to the temporal expression, this system also allows for control of the expression levels of the transgene . After creating stable lines of the docking site, we derived sublines constitutively expressing rtTA-Advance . In combination with incoming vectors containing the TRE promoter , we showed that this system allows for tight and inducible regulation of EGFP. Furthermore, in a proof-of-principle experiment, we derived stable lines expressing soluble VEGF trap, Flt1-Fc, in a doxycycline inducible manner [43, 44]. We showed that the expression of Flt1-Fc was tightly regulated in vitro. Furthermore, this tight regulation was also maintained in xenograft assays, where the transgene (in our case Flt1-Fc) could exert its biological activity (Figure 5).
Finally, in an attempt to make this system more flexible, we combined it with Gateway technology (Figure 1 and Table 2). All incoming vectors were made Gateway compatible, which allows for fast and reliable generation of incoming-expression libraries. This, in combination with the fact that this system can be used in a 96-well set-up, could allow for high-throughput generation of isogenic stable lines for the expression of gene libraries.