RNAi and microarrays reveal biological pathways: the combination of RNAi with microarrays has enormous potential for elucidating biological pathways. However, before this potential can be fulfilled, important questions need to be answered to ensure the proper interpretation of gene silencing results. (Keywords: biochemistry * interfering RNA * gene analysis).
In mammalian cells, RNAi is induced by the introduction or expression of small interfering RNAs (siRNAs), which are double-stranded RNAs, 19 to 23 nucleotides in length. The widespread use of RNAi technology in mammalian cells is largely due to the relative ease in obtaining siRNA molecules, which effectively reduce target gene expression. An additional advantage is the low cost of these experiments in relation to older techniques, such as the creation of gene knockouts in animals.
Another powerful technique for functional genomic studies is global gene expression profiling with DNA microarrays. The marriage of RNAi with microarrays, which provides a snapshot of gene expression within a cell, may provide a powerful method for analyzing biological pathways. Such experiments could be performed by silencing a single gene and then evaluating the downstream effects on other genes by array analysis. However, before realizing the potential of these two technologies, several issues need to be resolved, the most critical of which is confirming the specificity of siRNA-induced gene silencing.
Genome-wide RNAI effects
A number of researchers have already begun addressing the issues resulting from the joining of RNAi and microarrays by examining the effects of RNAi induction on a genome-wide scale. As a result, it has been shown that siRNAs directed toward an exogenous green fluorescent protein (GFP) gene, which acts as a reporter gene, specifically suppressed GFP expression without significantly affecting the expression levels of ~36,000 human genes. This and other initial experiments by multiple groups indicate that siRNAs, when well designed and used at relatively low concentrations, are remarkably specific. However, higher concentrations of siRNA (~100 nanoMole) have been demonstrated to induce nonspecific changes in gene expression profiles.
In contrast to the previous findings, one group of scientists has shown that siRNAs can induce off-target silencing if the off-target genes contain 14 to 15 contiguous nucleotides of sequence identity. In such cases, high concentrations of siRNA reduced expression of messenger RNAs (mRNAs) with as few as 9 contiguous nucleotides of sequence identity to the 3'-end of the siRNA sense strand.
The varying data obtained by the genome-wide studies of the effects of RNAi induction may be in part due to the differences in experimental design. Such differences arise from the specific cell lines, transfection agents, gene targets, analysis time points, and RNA extraction procedures used. It should be noted, however, that other researchers have found siRNA-induced gene silencing to be remarkably specific. In several instances, a single mismatch between the siRNA and its target mRNA was enough to abolish silencing.
All these variations in results point to two critical considerations prior to the design of any siRNA experiment:
* First, the siRNA sequence has to be carefully selected. The chosen siRNAs should have limited sequence similarity to genes other than their intended target to minimize the chance of off-target effects. Moreover, since it is not difficult to identify active siRNA, very stringent design procedures can be used without eliminating all functional siRNAs for a particular target gene.
* Second, the minimum siRNA concentration required to achieve silencing should be used. This is necessary because higher siRNA concentrations lead to nonspecific changes in gene expression.
Controlling gene silencing
Going forward, it is also critical to control the many variables that can affect the interpretation of gene silencing. One such variable is the method of siRNA production. Currently, there are five methods for producing siRNA: chemical synthesis, in vitro transcription, preparation of siRNA population by digestion, in vivo expression of hairpin siRNA from an expression vector, and in vivo expression of siRNA from a PCR-derived expression cassette. However, before any conclusions can be made regarding the effect of experimental designs on siRNA generation, further studies are needed to compare the various methods of siRNA preparation for their specificity of gene silencing.
Translational effects, cell line, and analysis procedures may also change gene silencing results. In the case of proteins, it may be important to monitor changes in their levels to detect silencing at the translational rather than post-transcriptional level. This is important because protein profiles of cells may not correlate with the RNA profiles if siRNA affects translation. Cell lines, on the other hand, are significant because the results obtained using one cell line, particularly if it is a tumor or transformed cell, may not correspond to those produced by the same siRNA when non-transformed or primary cell lines are employed. Thus, monitoring gene expression changes in multiple cell types in response to a specific siRNA could be a method for understanding the specific effects of siRNA induced gene suppression.
Finally, in analysis techniques, the amount of time elapsed between siRNA introduction (time of transfection) and analysis (time of RNA extraction) may have a substantial effect on the results obtained. Time course experiments, where the analysis is done at several time points after siRNA introduction, may be important to completely understand the temporal effects of siRNA-mediated gene repression. In addition, for array experiments, variables for RNA preparation and labeling should be eliminated and replicates should be performed to ensure that the differences in gene expression profiles are due only to RNAi.
siRNA experimental controls
Recently, interesting suggestions for siRNA experiment controls were proposed to help produce meaningful results that are consistent across various samples. These controls can be used with virtually any gene silencing experiment. Some of the recommended procedures include the use of scrambled or mismatched siRNA controls. Such siRNA controls can be employed to discount any gene expression profile changes due to the siRNA delivery method.
"Rescue" experiments can also be conducted so a target gene can be engineered to have a mismatch to an effective siRNA before being introduced into cells and expressed. If designed well, the gene will express a functional protein but will be refractory to RNAi induced by the mismatched siRNA. If its expression reverses the phenotypic changes induced by the siRNA, one can be confident that the gene expression changes induced by the siRNA are due to specific suppression of the gene of interest.
Lastly, the use of multiple siRNAs directed to a single target can be employed. Although the rescue experiment appears to be an excellent control, a much simpler control may actually be sufficient. For instance, the use of several siRNAs directed to a single target could be used. In such a case, the various siRNAs should all induce the same changes in the gene expression profile.
The use of RNAi to analyze gene function, particularly when coupled to genome-wide analysis of gene expression, has enormous potential for elucidating biological pathways, and the interaction between different, pathways. Although many siRNAs appear to be extremely specific for their target gene, others have been found to induce off-target effects. As more is learned about siRNA design, improvements will be made to experimental methodologies. We believe that the rules for designing very specific siRNAs will soon be refined, and then siRNAs can be used to their fullest potential.
Ambion, Inc., 512-651-0200, www.ambion.com
Lance Ford, senior scientist, and Kathy Latham, product manager at Ambion, Inc., Austin, Texas.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||RNA interference|
|Author:||Ford, Lance; Latham, Kathy|
|Publication:||R & D|
|Date:||Jul 1, 2003|
|Previous Article:||How to choose a turbopump. (Vacuum Tech).|
|Next Article:||CDS/LIMS interfaces: easier to implement and use: this widespread use of chromatography as a primary form of analysis, with chromatography data...|