The ability to produce and express foreign genes in humans might cure or even prevent many important human diseases that are either treated poorly or are untreatable by present therapies. However, insufficient transgene expression in vivo has so far impaired the development of effective gene therapy. Thus , it is of interest to develop compositions and delivery methods for gene therapy that lead to high level transgene expression in a variety of cell and tissue types. Synthetic DNA delivery agents are of crucial interest for gene therapy as an alternative to viral vectors, since they potentially display fewer risks in terms of immunogenicity and propagation and are easier to produce under GMP conditions. In the last decade, a significant number of industrial and academic groups have emerged with a number of important patents claiming a variety of synthetic DNA delivery agents. Some of these agents were launched into clinical trials at very early stages of development leading to rather modest results. The lack of physiocochemical characterisation of the self-assembling complexes in those early clinical trials prevented interpretation, correlation and comparison of transgene expression and biodistribution with the supramolecular state of these self-assemblies. On the other hand, in the last two years, a significant body of information has emerged from academic and academic-industrial groups regarding the exhaustive physiocochemical characterisation of some complexes and the impact of these characteristics on transgene biodistribution and expression both in vitro and in vivo. Additionally, the intracellular fate of different self-assemblies has been studied by different groups. Together, these studies provide a rational basis for designing novel DNA delivery agents. Synthetic agents for gene delivery are classified into different chemical families. In this manuscript, we have focused on the cationic lipids family. We choose to show the approaches that introduce original elements into the backbone of the synthetic delivery agent. Data on the physiochemical characterisation of different synthetic agents for gene delivery are given when available. Clinical data (reported in other reviews) are beyond the scope of this article. Finally, a discussion on how to improve the results obtained so far in order to advance towards new human trials is presented.