To the editor—DNA vaccination has entered clinical trials, and may ultimately be used on a large scale. It involves the injection of naked DNA that encodes antigens under the control of a eukaryotic promoter, and results in strong and sustained humoral and cell–mediated responses. The generic nature of the technology will facilitate the development of new vaccines; DNA vaccines are stable and will not require continuous cold storage. These properties are likely to lead to extremely widespread clinical and agricultural use in both the developed and developing world. In the developing world, inadequate health care resources will make post–vaccination control difficult. In this context, safety must be a particular concern.

After injection, some DNA may persist and reach distant sites. Although the amount of DNA uptake by distant cells is not large, it is unlikely to be zero. Moreover, DNA injected intravenously into pregnant mice reaches fetuses1. If after vaccination DNA is taken up by fetal or germ–line cells, immunological tolerance may be induced in the progeny (and descendants) of the vaccinated individual. In hepatitis B patients infected at birth, neonatal tolerance plays an important part in virus persistence2. Transgenic mice are often tolerant to transgene–encoded proteins; for example, expression of Schistosoma mansoni glutathione S–transferase leads to tolerance and to more severe experimental infections3. However, in mice that express the transgene extrathymically and relatively late, tolerance may not be induced. Current approaches to DNA vaccination use the strongest available promoters, usually the cytomegalovirus enhancer–promoter, to drive expression of antigen genes. In transgenic mice, this promoter is active during development4 and in the thymus5.

Tolerance resulting from DNA vaccination would produce consequences both for the individuals and for the population, as tolerant individuals are expected to be more susceptible to infection and/or they may become carriers, a potentially much more serious problem. Possible solutions include the use of promoters that are both muscle–specific and completely inactive during pre–natal and early post–natal life, and systems that rely on the expression of a ligand–regulatable transactivator, in which case tolerance induction would require the expression of transactivator and presence of the ligand as well as the plasmid that encodes the antigen. Whatever the magnitude of this potential problem, it should be taken into account in the design and testing of vaccines before DNA vaccination becomes widespread.