Many diseases, such as Huntington's disease and most cancers, occur through the expression of mutant genes, causing the formation of proteins whose actions harm the host. If the formation of these proteins could somehow be stopped, the diseases they cause would most likely cease to exist.

According to the central dogma of molecular biology, proteins are made in two steps. The first step, transcription, copies genes from double-stranded deoxyribonucleic acid (DNA) molecules to mobile, single-stranded ribonucleic acid (RNA) molecules called messenger RNA (mRNA). In the second step, translation, the mRNA is converted to its functional protein form. Since there are two steps to making a protein, there are two ways of preventing one from being made. Scientists have made exciting progress in blocking the protein synthesis through the second step, translation. One way they have accomplished this is by inserting synthetic molecules that trigger a cellular process called RNA interference (RNAi).

RNAi is a natural phenomenon believed to occur in the nematode Caenorhabditis elegans, in the fruit fly Drosophila melanogaster, and in some plant species (Alvarado 1999). It most likely serves to protect organisms from viruses, and suppress the activity of transposons, segments of DNA that can move from one location to another, sometimes causing abnormal gene products. Recent research has shown that an intermediate in the RNAi process, called short-interfering RNA (siRNA), might be effective in degrading mRNA in mammalian cells. siRNA therefore carries the potential to specifically degrade mRNA that corresponds to mutant genes involved in disease, shutting off the harmful effects of the proteins they encode.

To understand how siRNA can be used, we must first understand how RNA interference works. The currently favored mechanism for RNAi is a complicated process (Figure 1). Long molecules of double-stranded RNA (dsRNA) trigger the process. The dsRNA comes from virus and transposon activity in natural RNAi processes, while it can be injected into cells in experimental processes (Elbashir 2001). The strand of the dsRNA that is identical in sequence to a region on a target mRNA molecule is called the sense strand, and the other strand, which is complementary, is termed the antisense strand. Overhanging nucleotides on the 3' ends of the dsRNA help two different processing proteins recognize the dsRNA. An enzyme, called "dicer" in D. melanogaster (Baulcombe 2001), thought to be similar to RNase III then cleaves the dsRNA into 21-23 nucleotide fragments, called short interfering RNAs (siRNAs), which remain in double-stranded duplexes with very short 3' overhangs (Elbashir 2001).

At this point, the identity of the processing proteins becomes important. While these proteins have not been characterized, scientists hypothesize that only one of these proteins helps mediate cleavage of target mRNA. Duplexes associated with that particular processing protein form small interfering ribonucleoprotein complexes (siRNPs), which have endonuclease activity allowing them to cut nucleic acids. Since the hypothetical processing proteins can associate with either strand of the original dsRNA, the siRNPs either contain sense- or antisense-siRNA. The siRNPs that contain sense-siRNA are called antisense-cleaving, since they hybridize to complementary antisense mRNA and cleave it, while those containing antisense-siRNA are called sense-cleaving. To silence a gene, the target mRNA transcribed from the gene is cleaved through an antisense siRNA, preventing cells from translating it into a protein (Elbashir 2001).

Through this mechanism, scientists hoped the introduction of dsRNA to mammalian cells would prevent expression of the genes they correspond to. This task proved to be complicated by a second RNAi pathway. Introducing dsRNA longer than 30 base pairs (bp) into mammalian cells leads to an overall shutdown in protein synthesis, mediated by protein kinase C (PKC). Therefore, dsRNA shorter than 30 bp must be introduced to elicit the specific degradation of mRNA and specific gene silencing (Bass 2001). One way to do this uses the siRNA fragments that result from cleavage of the dsRNA. Unlike the longer dsRNA, which causes a nonspecific response, siRNA has been shown to specifically silence targeted genes in mammalian and invertebrate cell cultures (Caplen 2001).

Short-interfering RNA could provide medical researchers new hope in using gene silencing for therapeutic purposes. Until now, another gene silencing technique, using antisense oligonucleotides, had been the main hope for clinical application. Antisense oligonucleotides are short pieces of DNA or RNA complementary to sequences on mRNA. They are believed to work by hybridizing to the mRNA, creating a double stranded stretch, which slows down ribosome transcription. Antisense DNA creates RNA-DNA duplexes that are most likely recognized by RNase H, an enzyme that cuts double-stranded molecules containing one DNA and one RNA strand, which cleaves the mRNA (Crooke 1999). Since the mRNA is cut, it cannot be translated into a functional protein product. While this process can be triggered somewhat effectively in vitro, antisense technology has not completely lived up to expectations due to difficulties in delivering oligonucleotides to cells in vivo and problems with the accessibility of specific sites on mRNA (Caplen 2001).

siRNA, on the other hand, might prove to be more effective. It has more potential for success since it seems to be more stable than single-stranded antisense molecules, making cellular delivery easier. So far, all siRNAs tested in mammalian cells have inhibited expression of the target genes, which is an encouraging sign (Caplen 2001). If siRNA can indeed be an effective gene silencer, it could serve many important uses. siRNA could serve as a functional genomics tool, because it can be used to knock out specific genes to see what effect their absence has on the cell (Fire 1999). Of a more therapeutic interest, siRNA could possibly be used to silence oncogenes, which cause cancer when mutated. More than one oncogene must usually be mutated to cause cancer, but theoretically, if all oncogene activity can be turned off, the cancer will cease to exist. Antisense molecules have already been used in cancer therapies (Xu 2001), but since it is hard to get them into cells and to cleave their targets efficiently, scientists are searching for better methods. While siRNA has not been studied enough for any conclusions to be made, perhaps its stability will prove to make it more efficient at getting to and eliminating its targets than the antisense oligonucleotides tried so far.

Even if siRNA does not prove effective in mammals, it can still have therapeutic uses. It is effective against parasites, so perhaps it can be used to silence parasitic genes (Fire 1999) or used against other pathogens to benefit host organisms like humans. Further studies need to be carried out before we will know if siRNA will live up to its potential. But for now, siRNA brings the possibility of specific gene-silencing through mRNA degradation, something its precursor, dsRNA, cannot do, while possibly being more versatile than less-stable single-stranded antisense oligonucleotides.