Intentionally introduced dsRNA must first be processed into small interfering RNAs (siRNAs) which are 22nt in length and analogous to miRNA. Fire and Mello directly injected dsRNA into a nematode and saw a response, but sophisticated mechanisms targeting the long RNA duplex made it ineffective in mammalian cells.
In 2003, Gregory Hannon and coworkers reported in Nature Genetics a very successful method to apply RNAi to an in vivo mammalian system. They developed a system of short hairpin RNAs (shRNAs) expressed from viral vectors whose stem-loop structure allowed them to effectively enter the miRNA pathway in the same way as pre-miRNA. In order to test the efficacy of shRNAs in mice, the authors targeted Trp53, a previously identified tumor suppressor gene. In transgenic mice, the deletion or inactivation of Trp53 will eliminate p53 expression and accelerate tumor formation. This approach represents a straightforward loss-of-function experiment to compare gene silencing by shRNA to standard gene deletions known as knockouts.