High Expectations for Lower Organisms: Off-target Effects
Since the discovery that long dsRNA introduced into the nematode C. elegans could successfully target and knockdown mRNA, a new field of genomics has emerged - RNA interference (RNAi) [1]. Now that the genomes for a number of organisms has been discovered, we are poised to create functional maps that will expand our understanding of biologic processes and uncover strategies to combat many diseases.
When applying this revolutionary technology to functionally interpret the information on a whole-genome scale, the reliability of the research tools used is paramount. In particular, the functionality or activity and associated specificity of any RNAi-based application will ensure that firstly a true knockdown event has occurred and it is as specific an action as possible. This specificity will lead to the correct interpretation of data regarding the true functional consequence of the mRNA knockdown. Moreover, a functional and specific RNAi application will enable researchers to quickly confirm, validate and develop those results that are real.
One approach to identify those target genes that have a causal relationship with a particular biologic event (i.e. phenotype) is to use RNAi to screen lower organisms such as Drosophila melangaster, the fruit fly, thus identifying potential orthologues in higher organisms such as mammalian systems. However, regardless of the organism in which the RNAi screens are performed, the underlying necessity for highly functional and specific reagents is essential due to a non-specific effect termed off-target effects (OTEs). The generation of OTEs can create data unrelated to the true function of a particular mRNA target which is then very difficult to interpret.
The application of long dsRNA for RNAi in Drosophila melangaster is well tolerated due to the lack of an interferon response in this organism. The long dsRNA is processed (by DICER) in many lower organisms creating an endogenous pool of siRNAs that differ in both length, functionality and specificity [2]. Two recent publications describing the use of RNAi in Drosophila melangaster has brought into question the long standing assumption that these systems are devoid of OTEs.
1) Kulkarni MM, Booker M, Silver SJ, Friedman A, Hong P, Perrimon N, Mathey-Prevot B (2006) Evidence of off-target effects associated with long dsRNAs in Drosophila melangastor cell-based assays. Nature Methods, 3(10): 833-8
In this study the authors conducted a retrospective analysis of 30 RNAi screens performed in Drosophila melangasterthat revealed a high degree of sequence-dependent off-targets hitherto an unanticipated consequence within this model system.
In the process of this retrospective analysis, the group discovered a limited redundancy in the original RNAi libraries used in these screens. This serendipitous observation meant that there were actually two siRNAs designed against the same open reading frame for >1300 gene targets. A direct comparison of the resulting phenotype produced by these pairs of siRNAs targeting the same mRNA was conducted. This comparison led the authors to conclude that in a substantial number of cases there are pairs of siRNAs that are designed to knockdown the same mRNA result in disparate phenotypes.
As a result of describing OTEs as a consequence of complimentarity to other targets, the group found that a perfect 19 nt target match (sufficient for a cleavage event) was a reasonable threshold to limit inadvertent recognition of unwanted gene targets. Using this 19 nt threshold for determining off-target recognition of siRNAs in the Drosophila melangaster RNAi library the authors found that 32% of dsRNAs contained 1-10 off-targets and 7% of dsRNAs contain >10 off-targets. Furthermore, these off-targets were not correlated with high dsRNA concentration nor was there any obvious bias for genes having a general metabolic function.
The group then chose 7 dsRNAs that consistently resulted in positive ‘hits’ in both primary and secondary screens involving activation of MAPK. Each of these 7 dsRNA sequences contained anywhere from 48-138 off-targets. For each of these 7 dsRNAs, two additional sequences were redesigned that were devoid of any predicted off-targets (based on the 19 nucleotide threshold). When these new sequences were compared to the original group of 7 dsRNAs and verified for functionality, none were found to affect the activation of MAPK. The authors went on to create a custom microarray to measure all of the predicted 135 off-targeted mRNAs from a number of siRNAs. However, the expression levels of 50 (~37%) of the 135 predicted off-target mRNAs were unchanged following the application of these siRNAs, suggesting that prediction of off-targets based merely on sequence identity is not straightforward.
2) Ma Y, Creanga, A, Lum, L, Beachy, PA (2006) Prevalence of off-targets effects in Drosophila RNA interference screens. Nature, 443(7109): 359-63
In this study the authors used a library of over 20,000 dsRNAs and employing a luciferase reporter to detect responses in the Wingless pathway. Wingless is a member of the Wnt gene family, and is essential for segmentation in Drosophila melangaster, as well as many other patterning events [3].
The initial screen identified 5 known pathway components as well as seven novel regulators of the Wingless pathway. Of these seven putative regulators, the cDNAs for six were readily available and subsequently used to confirm the role of these genes in the Wingless pathway. Double stranded RNAs spanning the entire length of these cDNAs were tested to ascertain if these results were indeed target-dependant events and thus true regulators of the pathway. This test led the investigators to identify a short region (16 bp) of homology to armadillo, a member of the Wingless pathway that was common to all seven candidate genes from the initial screen. This observation suggested that these effects were sequence-dependant events and the corresponding results were off-target effects, mediated through the cleavage of armadillo mRNA rather than a true regulatory effect on the Wingless pathway.
Moreover, 19 of 22 dsRNAs that caused a 1.5 or greater fold reduction of the reporter system contained short homologies to known components of the Wingless pathway. The remaining 3 dsRNAs, previously reported to be growth and viability-related genes, could potentially affect the Wingless reporter system in a non-specific manner [4]. Therefore, because “these non-overlapping dsRNAs did not reproduce the inhibitory effects of their library dsRNAs in a Wg reporter assay”, it is possible to inadvertently enrich for off-target effects.
A retrospective analysis of a previous RNAi screen for the Wingless pathway indicated that approximately 60% of putative negative regulators contained a tandem repeats of the sequence CAN. The inadvertent enrichment of such repeats in addition to significant complimentarity to other gene targets further increase the propensity of off-target effects making the true interpretation of the functional consequence of any gene problematic.
Conclusions
These two publications bring into focus the issue of OTEs assumed to be a component of mammalian systems only. As pointed out “one vexing issue is that we cannot predict a priori which of the possible siRNAs will be generated with a given dsRNA by Dicer” leading to an unpredictable functionality and specificity of the reagent. Therefore, the previous assumption that the desirable siRNAs in an uncontrolled endogenous pool would always prevail over those undesirable siRNAs that might otherwise lead to sequence-specific OTEs has been severely challenged.
Moreover, even the use of individual siRNAs has been shown to lead to OTEs in the Drosophila melangaster model. This further raises “…a cautionary note about interpreting results based on the use of a single dsRNA per gene” even in this model organism previously assumed to be resistant to such OTEs.
- Fire, A., et al., Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature, 1998. 391(6669): p. 806-11.
- Zamore, P.D., et al., RNAi: double-stranded RNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell, 2000. 101(1): p. 25-33.
- Baker, N.E., Embryonic and imaginal requirements for wingless, a segment-polarity gene in Drosophila. Dev Biol, 1988. 125(1): p. 96-108.
- Boutros, M., et al., Genome-wide RNAi analysis of growth and viability in Drosophila cells. Science, 2004. 303: p. 832-835.
^ back to top
