An enzyme called Dicer is known for its central role in RNA-controlled gene silencing. Mammalian Dicer1, however, also seems to have another function: it maintains visual health by degrading toxic RNA molecules. See Article p.325
As their name suggests, Dicer enzymes cut long double-stranded RNA molecules into shorter pieces1,2. They are therefore crucial for gene-silencing pathways that involve small RNAs such as short interfering RNAs or microRNAs — the most abundant classes of small RNAs in mammals3,4. Without functional Dicer, most of these RNAs cannot be generated5. On page 325 of this issue, Kaneko et al.6 uncover a function of mammalian Dicer that is independent of small-RNA generation but apparently equally essential.
The authors were interested in the molecular basis of geographic atrophy — an advanced form of an eye disease called age-related macular degeneration, which is a common cause of blindness in industrialized countries7. They found that patients with this condition have reduced levels of Dicer in their retinal pigment epithelium (RPE), an eye-specific tissue that is affected in geographic atrophy. Kaneko and colleagues also noted that in mice lacking Dicer selectively in the RPE, this tissue is severely degenerated, similar to the situation in geographic atrophy, and therefore further implicating reduced Dicer levels in the pathogenesis of the disorder (Fig. 1).
Loss of Dicer intuitively points to a role for microRNAs (miRNAs) in geographic atrophy, especially given that these short regulatory RNA sequences are highly abundant in mammalian cells. Surprisingly, however, Kaneko et al.6 could find no evidence to support this. For example, mice lacking two other enzymes required for miRNA processing (Drosha and DGCR8) did not show degeneration of the RPE, even though these animals showed severely impaired miRNA function. So how does Dicer deficiency contribute to geographic atrophy? Kaneko and colleagues describe an as-yet-unrecognized function of the enzyme.
Dicer processes double-stranded RNA, and the authors speculated that a class of double-stranded RNA other than miRNAs might be involved in the pathology of geographic atrophy. Indeed, they detected a marked increase in the levels of Alu RNA sequences in the eyes — and specifically in the RPE — of patients with this condition, but not in those of healthy individuals. And they noted that mice with Dicer deficiency in the RPE express increased levels of Alu RNAs.
Alu RNAs are transcripts of Alu elements — the most common non-coding, repetitive DNA sequences in the human genome. But how do they cause geographic atrophy? Kaneko et al. tackled this question from several angles.
Both enhancing the expression of Alu RNAs and their direct injection into the eyes of normal mice reduced cell viability in the RPE and led to degeneration of this tissue. But when the authors increased the expression of Dicer along with that of Alu RNAs, the toxic effects of Alu RNAs were no longer seen. This suggests that Dicer processes longer Alu RNAs into shorter, non-toxic sequences. To test this possibility, the authors injected long double-stranded Alu RNA, before and after treatment with Dicer, into the eyes of normal mice. Only the long unprocessed Alu RNAs caused RPE degeneration.
Kaneko et al.6 propose the following model for the pathogenesis of geographic atrophy. In the RPE, long double-stranded Alu RNA is constantly produced. In healthy individuals, Dicer processes these toxic RNAs into shorter, non-toxic versions. In patients with geographic atrophy, however, Dicer levels are greatly reduced, allowing the long Alu RNAs to cause cell death and RPE degeneration.
As Dicer normally produces functional classes of small RNAs from double-stranded RNA precursors, an intriguing question is whether the products of Alu RNA processing by Dicer have specific cellular functions. In support of this possibility, Dicer generates endogenous short interfering RNAs from repetitive elements in the mouse germ line8. Therefore, it could be that, in addition to the toxic long Alu RNAs, the absence of Dicer-generated small Alu RNAs contributes to features of geographic atrophy.
Previous work9 has shown that reduced Dicer levels can occur in many tissues and are associated with various diseases. But these studies did not analyse Alu RNAs. On the basis of Kaneko and co-workers' data, it is conceivable that the effects of toxic Alu RNAs are also widespread, as has always been expected. It remains unclear, however, why Dicer expression is reduced in the RPE, and not in other tissues. Are tissue-specific transcription factors also missing in geographic atrophy, or is Dicer degraded in the RPE after synthesis?
Another pertinent question is whether these results6 could further the search for new strategies for treating geographic atrophy. Kaneko et al. report that injecting antisense oligonucleotides into mouse eyes, which reduces Alu RNA levels, also decreases RPE degeneration. These are heartening findings. But before any therapy development can be contemplated, the molecular basis of Alu RNA detoxification by antisense oligonucleotide must be elucidated. Strategies to increase Dicer expression might also prove to be useful therapeutic approaches.
About this article
Loss of microRNAs in pyramidal neurons leads to specific changes in inhibitory synaptic transmission in the prefrontal cortex
Molecular and Cellular Neuroscience (2012)