Micro in size but not in function: microRNAs in axonal survival and regeneration

Successful axon regeneration after injury to neurons in the peripheral nervous system is accomplished by a sophisticated coordination of gene expression in both the cell body and axon proximal to the site of injury. Although the detailed molecular regulatory mechanism of the regenerative responses in gene expression alterations is not well understood, one of the most remarkable findings in the experiments on axon regeneration is the temporally and spatially differential expression of microRNAs (miRNAs) that regulate target mRNA expression at the post-transcriptional level. Studies utilizing high-throughput technologies, such as microarray and next-generation deep sequencing, have shown that numerous miRNAs are highly neuron-specific or enriched in the nervous system, and their endogenous levels are dynamically altered following the onset of various types of injuries, including traumatic events, non-traumatic events and degenerative injuries [1]. These findings suggest that miRNAs are closely associated with specific functions in pathophysiological processes of axon regeneration in neurons.

A large number of spatial and temporal profiling studies of miRNAs in the nervous systems have been carried out and have predicted putative target mRNAs that are associated with axon regeneration by utilizing single platform bioinformatic algorithms (e.g., TargetScan, PicTar, and miRanda) or system approaches with integrative miRNA and gene expression data to predict the functional implication of miRNAs [2]. However, all these studies suffer from the most common and vital limitation that a great number of potential targets might be false-positive, as these targets cannot be experimentally validated in all cases. To overcome this obstacle, increasingly sophisticated analytical tools with improved efficiency are being generated and some are already available [2]. Another limitation in analyzing the biological functions of altered miRNA profiles in axon regeneration is the presence of cell type- and/or subcellular region-specific isoforms of mRNAs that are generated through, for example, alternative splicing and polyadenylation for short or longer 3’ untranslated regions (UTRs), that complicates the prediction of functional miRNA targets.

Genes can give rise to multiple mRNA isoforms with alterations in the coding regions and 3’UTRs, and the number of mRNA isoforms for any given gene continues to increase with sequencing depth. Interestingly, the isoforms of a gene with 3’UTR variants could function differently through differential expression in specific cell types or distinct distribution in subcellular regions of a cell [3]. Although it remains largely unexplored how differential expression/distribution of mRNA isoforms is achieved at post-transcriptional level, miRNAs would be able to discriminate one isoform from another. For example, two 3’UTR variant isoforms of BDNF mRNA, previously inferred by their differential localization in dendrites and cell body of hippocampal neurons [3], are differentially regulated by miRNA-206, in that only the longer 3’UTR variant of BDNF mRNA is specifically regulated by miR-206 (Shrestha and Yoo, unpublished data). Therefore, data analysis of the biological importance of the differentially expressed miRNAs and target gene interaction at the isoform level will be an exciting and essential task for fully understanding the molecular mechanism that involves in axon survival and regeneration following injury.

A recent study published in FEBS Letters by Motti et al. [4] found 49 differentially expressed miRNAs in dorsal root ganglion (DRG) neurons responding to nerve injury during regeneration and utilized integrated analysis of the miRNA and mRNA isoform expression data to derive putative functional networks. Profiles of differentially expressed miRNAs reported in injured and regenerating neurons by recent studies [5-7], including this new study by Motti et al., were distinct — presumably due to differences in the types of injuries and developmental or regenerative stages. However, the results of miRNA–mRNA target interaction regulatory network analyses were all comparable in that miRNAs had important regulatory roles that were directly associated with axon regeneration. Furthermore, Motti et al. showed that, when individually over-expressed in cultured DRG neurons, 23 of 49 differentially expressed miRNAs resulted in a significant phenotypic alteration (greater than two-fold changes) in neurite length. Despite the fact that the expression vector-based system for overexpression of miRNAs they used cannot be employed to specifically target one mRNA, it is unlikely that the phenotypic change in neurite outgrowth of miRNA-overexpressing DRG neurons in culture is due to off-target effects or non-physiological levels of miRNAs, as the scrambled seed region of the corresponding miRNAs has been included and used as a control for comparisons.

Lastly, it is worth mentioning that, as distinct populations of mRNAs are differentially localized within neuronal cells and intra-axonal translation is essential for neurophysiological functions of the distal axons in elongation, regeneration and viability [8], the potential miRNA regulators would also be differentially localized and enriched in distal axons, allowing for the rapid and precise control of local mRNA translation autonomously. Interestingly, recent studies have clearly demonstrated that selective miRNA precursors and mature miRNAs are differentially targeted to axons and dendrites [9-11]. For example, miR-338 is localized into axons of superior cervical ganglion neurons and regulates the expression of Cytochrome C Oxidase IV (COXIV) and ATP synthase 5 gamma 1 (ATP5G1) in mitochondria [12,13]. Although the localization mechanism of axon-enriched miRNAs remains largely unknown, one would expect to find active transport and local processing of precursor miRNAs that eventually result in temporally and spatially differential levels of miRNAs to differentially regulate mRNA isoforms precisely and autonomously, particularly in distal subcellular regions of neurons.