Figure 2.

A comparison of models of miRNA targeting and how each relates to the potential for ceRNA crosstalk. In the nonhierarchical model, miRNA molecules bind target transcripts independently of their affinity for their miRNA binding sites. As a result, a ceRNA has to contribute an equivalent number of miRNA binding sites to those already present in the transcriptome before significant derepression of endogenous miRNA target transcripts will be observed. Due to such a high requirement for additional miRNA binding sites, the potential for ceRNA crosstalk is low. In the hierarchical model, miRNA molecules preferentially bind higher affinity sites (8mers) before spreading across low affinity sites. A ceRNA with a high affinity miRNA binding site therefore only has to contribute miRNA binding sites at a number similar to the miRNA molecule count before significant derepression of targets will be observed. Therefore, there is potential for ceRNA crosstalk provided that the miRNA is not highly abundant in comparison to the number of its high affinity binding sites. In the preferential targeting model, certain transcripts are preferentially targeted and repressed by miRNA molecules. In this model, the potential for ceRNA crosstalk is high if the ceRNA is a preferentially targeted transcript. However, it is currently unclear what factors may contribute to preferential targeting (see color version of this figure at www.tandfonline.com/ibmg).