The recognition of non-protein-coding microRNAs (miRNAs) as key modulators of gene expression in recent years triggered an expansion of research efforts devoted to identifying miRNAs and their targets, as well as understanding their biogenesis, regulation, and function in physiological and pathological conditions. Found in plants and animals, miRNAs are single-stranded RNAs of about 22 nucleotides generally thought to repress gene expression by either inducing the degradation or blocking the translation of the complementary target mRNAs. They are encoded in the genome and transcribed as precursors that need to be processed to become mature miRNAs, which are then incorporated into the RNA-induced silencing complex to execute gene repression. As a class of gene regulators, it is perhaps not surprising that miRNAs influence a broad spectrum of cellular processes, ranging from cell survival/death to differentiation and cell migration, all of which are intimately tied to tumorigenesis (Esquela-Kerscher and Slack, 2006). Indeed, emerging evidence associates deregulated expression of certain miRNA families with various human cancers, including melanoma (Muller and Bosserhoff, 2009), underscoring the need to determine not only their targets but also how the expression and activity of miRNAs themselves are regulated.
A recent publication in Nature describes an endogenous mechanism that counteracts miRNA-mediated gene silencing through sequestration of miRNAs by pseudogene transcripts (Poliseno et al., 2010). This novel concept of pseudogenes acting as ‘miRNA decoys’ to control the activity of miRNAs appears strikingly obvious once revealed given our understanding of the mechanism of RNA interference. Of particular interest to this audience is the finding that such miRNA decoy strategy regulates the expression of PTEN, with a direct consequence on its tumor-suppressive activity, as well as the prediction that other driver genes of skin cancer, such as the RAS oncogene and BRAF, may also be regulated similarly.
The basis of this work stems from a finding made more than a decade ago that the tumor-suppressor gene PTEN on chromosome 10q23 has an intronless pseudogene counterpart PTENP1 (ΨPTEN) on chromosome 9p13. PTENP1 is highly homologous to PTEN, with a difference of only 18 nucleotides in the DNA coding sequence (Teng et al., 1997), and is actively transcribed, making up a significant portion of PTEN transcripts in various human tissues (Fujii et al., 1999). However, due to the absence of an initiation methionine, PTENP1 is not translated into a functional protein. It is likely because of this that PTENP1 was assumed to be biologically inactive and hence has escaped our attention since its initial characterization. Given that miRNAs can act by destabilizing complementary mRNAs, Poliseno et al. reasoned that PTENP1 transcripts could dilute the pool of available PTEN-targeting miRNAs to promote the expression of PTEN.
Using bioinformatics, the authors identified perfect seed matches for several miRNA families within a highly conserved region in the 3′UTRs of PTEN and PTENP1. Of these miRNAs, the miR-17 and miR-19 families were demonstrated to directly repress the transcript abundance of both PTEN and PTENP1 at the endogenous level in DU145 prostate cancer cells. The demonstration that both PTEN and PTENP1 are subjected to regulation by the same miRNAs suggests that manipulating the transcript abundance of one gene should have a direct impact on that of the cognate gene, by way of competitive binding to their common miRNAs. Indeed, overexpression of the 3′UTR of PTENP1 alone was sufficient to increase the amount of endogenous PTEN mRNA and protein with a corresponding inhibition on cell proliferation, whereas knockdown of PTENP1 reduced PTEN expression. Interestingly, the 3′UTR of PTENP1 had a far greater growth-suppressive effect than PTEN, which the authors attributed to the PTENP1-3′UTR acting as a decoy for miRNAs with multiple targets that antagonize proliferation, like p21. A direct correlation between PTEN and PTENP1 expression was then drawn in normal human tissues and a collection of human prostate tumor samples. Furthermore, focal loss of the PTENP1 locus on chromosome 9 was observed in sporadic colon cancer, as well as in acute lymphoblastic leukemia and breast cancer, accentuating the potential role of PTENP1 as a tumor suppressor.
In summary, this work identified the pseudogene PTENP1 as a biologically functional miRNA decoy for PTEN with the potential consequence of cancer formation when deregulated. An obvious question that follows is what other gene-pseudogene pairs are regulated in this manner. Poliseno et al. began to explore this uncharted territory with the identification of several candidate pairs that have well-conserved miRNA-binding sites, such as KRAS-KRAS1P and BRAF-ΨBRAF.
This study opens up yet another dimension in the already intricate network of gene regulation. An incredible coordination of gene expression could occur at the post-transcriptional level alone, considering that one gene could have multiple pseudogenes (like OCT4), and each gene/pseudogene could have binding sites to multiple miRNAs (like PTEN/PTENP1). Such combinatorial control enables fine-tuning of protein expression over a wide range. From this regulatory complexity emerge the following questions: in the case of pseudogenes that simultaneously bind several distinct miRNAs, do these miRNAs target genes that contribute to the same cellular process? In the case of genes with multiple pseudogenes, does the expression of each pseudogene exhibit tissue or developmental stage specificity?
Extending beyond the discovery of pseudogenes as gene regulatory modules is the implication of reciprocal regulation between miRNAs and their target mRNAs. Given that miRNAs are only about 22 nucleotides in length, it is conceivable that some pseudogenes may have undergone evolutionary trimming to retain only the minimal complementary sequences necessary to exert their function as miRNA decoys. The identification of these species will require genome-wide interrogation by computational methods.
Finally, with respect to melanoma biology, the inactivation of PTEN is proposed to involve epigenetic silencing, as somatic mutations in the phosphatase core are found in only 10–20% of the primary tumors (Palmieri et al., 2009). Findings from Poliseno et al. warrant a close examination of the status of the pseudogene of PTEN, as well as that of NRAS and BRAF, especially in melanomas lacking mutations that alter the function of these proteins. It would be of tremendous interest to explore the role of miRNAs in skin cancer biology, as the aberrant regulation of these species may provide a new molecular basis in disease development.