Future therapeutic applications of stem cells in regenerative medicine require efficient techniques for modulating gene expression. Conventionally, this is achieved through the use of recombinant DNA, which invariably leads to permanent genetic alteration to the cell. Overwhelming safety and ethical concerns are likely to preclude the application of genetically modified stem cells in human clinical therapy for the foreseeable near future. An alternative may be to adopt a "milieu-based" approach to influence gene expression, by exposing stem cells to a cocktail of exogenous cytokines, growth factors, and extracellular matrix. Nevertheless, the non-specific pleiotropic effects exerted by various cytokines, growth factors, and extracellular matrix would make this a relatively inefficient approach. Moreover, a "milieu-based" approach is likely to require extended durations of in vitro culture, which might delay autologous transplantation of adult stem cells to the patient and might alter their immunogenicity through prolonged exposure to xenogenic proteins within the culture milieu. The obvious solution would be to deliver proteins, RNA, or their synthetic analogs, such as peptide nucleic acid, directly into the cell to modulate gene expression. Currently, two promising delivery platforms are available: (1) protein transduction domains, and (2) immunoliposomes. Because such molecules have a limited active half-life in the cytosol and are obviously not incorporated into the genetic code of the cell, these would only exert a transient modulatory effect on gene expression. Nevertheless, a transient effect may be preferable for clinical therapy, since this would ultimately avoid permanent genetic alteration to the cell.