Current understanding of the pathogenesis of rheumatoid arthritis has provided evidence that therapeutic benefit can be achieved by using antagonists targeted to the inflammatory cytokines involved, mainly tumour necrosis factor-alpha and interleukin-1. Gene delivery of antagonists, which can inhibit the production or action of these cytokines and other mediators, has been achieved in experimental animal models. This new method of delivery can produce therapeutic effects at lower concentrations and in a local environment, overcoming the adverse effects that often accompany protein therapy. However, several technological and biological restraints preclude the immediate adaptation of this method to human treatment. Based on the experimental evidence, possible target therapeutic genes, cell types and vector systems that could be used are discussed in this article.