Making better transgenic models: conditional, temporal, and spatial approaches

Over the last decade transgenic mouse models have become a common experimental tool for unraveling gene function. During this time there has been a growing expectation that transgenes resemble the in vivo state as much as possible. To this end, a preference away from heterologous promoters has emerged, and transgene constructs often utilize the endogenous promoter and gene sequences in BAC, PAC and YAC form without the addition of selectable markers, or at least their subsequent removal. There has been a trend toward controlled integration by homologous recombination, either at a characterized chromosomal localization or in some cases within the allele of interest. Markers such as green fluorescent protein (GFP), beta-galactosidase (LacZ), and alkaline phosphatase (AP) continue to be useful to trace transgenic cells, or transgene expression. The development of technologies such as RNA interference (RNAi), are introducing new ways of using transgenic models. Future developments in RNAi technology may revolutionize tissue specific inactivation of gene function, without the requirement of generating conditionally targeted mice and tissue specific recombinase mice. Transgenic models are biological tools that aid discovery. Overall, the main consideration in the generation of transgenic models is that they are bona fide biological models that best impart the disease model or biological function of the gene that they represent. The main consideration is to make the best model for the biological question at heart and this review aims to simplify that task somewhat. Here we take a historical perspective on the development of transgenic models, with many of the important considerations to be made in design and development along the way.