RNA interference (RNAi) — the process in which double-stranded RNA (dsRNA) binds to specific messenger RNA and silences gene expression — has generated great excitement, as dsRNA could theoretically be created to target any disease-causing gene more quickly, cheaply and effectively than any present method. Preliminary in vivo work seems to be maintaining this promise. Earlier this year it was shown that RNAi can silence the expression of the gene encoding the Fas receptor, protecting mice from fulminant hepatitis. Now, Kay and colleagues report for the first time in Nature Biotechnology that RNAi can prevent viral replication in live mammals.

The authors used a mouse model of hepatitis B virus (HBV) infection, in which a plasmid carrying the entire HBV genome was transfected into the mouse liver. In these mice, HBV proteins are expressed and most of the stages of viral replication take place. Kay and colleagues co-transfected the mice with a second plasmid that expressed small fragments of dsRNA called short hairpin RNAs (shRNAs), designed to target specific HBV mRNA sequences. This caused a dramatic decrease in viral gene expression at the levels of RNA and protein, and strongly inhibited the replication of the virus.

One of the concerns about testing RNAi in whole animals is that any observed effect could result, fully or in part, from the foreign shRNAs triggering the immune response. But by comparing mice that lack a functional immune system with wild-type animals, the authors showed that the HBV inhibition was a specific effect of targeting viral sequences, rather than being due to a generalized immune-based attack.

At present, there are few effective treatments that can treat hepatitis B once infection is established. The findings not only show the therapeutic promise of RNAi in hepatitis B but also indicate that this strategy could treat other forms of viral infection. But many challenges remain. First, RNAi-based inhibition of infection by intact HBV needs to be proved. Second, new methods of delivery will need to be developed, as the one used in this study is unsuitable for a clinical setting. And finally, successfully silencing the HBV genes of interest without affecting the expression of any host genes must be ensured.