FTO, RNA epigenetics and epilepsy.

Several recent landmark papers describing N(6) -methyladenosine (m(6) A) RNA modifications have provided valuable new insights as to the importance of m(6) A in the RNA transcriptome and in furthering the understanding of RNA epigenetics. One endogenous enzyme responsible for demethylating RNA m(6) A, FTO, is highly expressed in the CNS and is likely involved in mRNA metabolism, splicing or other nuclear RNA processing events. microRNAs (miRNAs), a family of small, non-coding transcripts that bind to target mRNAs and inhibit subsequent translation, are highly expressed in the CNS and are associated with several neurological disorders, including epilepsy. miRNAs frequently bind to recognition sequences in the 3'UTR, a region that is also enriched for m(6) A. Certain specific miRNAs are upregulated by neuronal activity and are coupled to epileptogenesis; these miRNAs contain a consensus m(6) A site that if methylated could possibly regulate miRNA processing or function. This commentary highlights aspects from recent papers to propose a functional association between FTO, RNA epigenetics and epilepsy.

FTO is Important in the CNS

The gene fat mass and obesity-associated (FTO) was first identified in mice as one of the genes encoded by the 1.6 Mb deletion that produced a phenotype with partial syndactyly of forelimbs and extensive thymic hyperplasia. Subsequently, a common variant in the FTO gene was identified as a risk allele for type 2 diabetes and increased body mass, and many studies of FTO have focused on the association with diabetes, obesity and metabolism. In the course of conducting these large scale genotyping studies, the role of FTO in human disorders was expanded to include the central nervous system (CNS). Variants have been found to be associated with neurological disease conditions including depression and Alzheimer disease. FTO is highly expressed in brain tissue, and is essential for normal development of the CNS in human. The generation of mice that were specifically deleted only for neuronal FTO had a similar phenotype of growth retardation as the whole body FTO deletion, suggesting that a major function of FTO occurs in the brain. The obesity-associated risk allele has been shown to have a potential pathological effect on brain volume: healthy elderly subjects with the risk allele had brain volume deficits (average differences of 8% in frontal lobes and 12% in occipital lobes) compared with non-carriers. Recently, brain derived neurotrophic factor (BDNF) was identified as a candidate gene for functional coupling to FTO, leading the authors to speculate on a role of FTO in neuronal plasticity possibly via interaction with CCAAT/enhancer binding protein β. These data provide strong evidence that FTO has a functional role in the CNS and, by implication, to CNS disorders.

Fto Demethylates m6A RNA

Most studies involving FTO have focused at the genome level and on correlation of variants with phenotypes. Evidence for the molecular action of the expressed protein (Fto) is more limited. Fto has been shown to localize to the nucleus and to catalyze the Fe(II) and 2-oxoglutarate-dependent demethylation of 3-methylthymine in ssDNA. One year later it was shown to catalyze the demethylation of 3-methyluracil in ssRNA with slightly higher efficiency over that of 3-methylthymine in ssDNA. However, a recent article provides the strongest evidence to date on the enzymatic activity of Fto. Jia et al. provide evidence that Fto strongly prefers to demethylate N6-methyladenosine (m6A) in ssRNA (Fig. 1). By direct comparison with other substrates these authors conclude that m6A in ssRNA is the best substrate discovered so far for Fto, having a greater than 50-fold preference for m6A over 3-methyluracil. Expected changes in levels of m6A in mRNA were found when human cells were manipulated to either overexpress Fto (which caused a decreased level of m6A) or underexpress Fto (which caused an increased level of m6A). These authors further showed that Fto partially co-localizes with nuclear splicing speckle factors (SART1 and SC35) and with RNA polymerase II phosphorylated at Ser2, but not with markers for other nuclear subregions such as telomeres, replication site, Cajal body, cleavage body or P-body. These authors conclude that m6A in nuclear RNA is the physiological substrate of Fto, and that the function of Fto likely affects the processing of pre-mRNA and/or other nuclear RNAs. This report provides strong evidence that Fto is involved in the processing of nuclear RNAs and in RNA epigenetic modifications. As such, Fto is the first m6A RNA demethylase that exhibits significant regulatory functions.

Figure 1.FTO catalyzes the conversion of N6-methyladenosine in RNA to adenosine.

m6A and RNA Epigenetics

m6A is the most common mRNA modification in eukaryotes and also in the RNA of viruses that replicate in eukaryotes. The modification is catalyzed by the methyltransferase like 3 (METTL3) enzyme, which is thought to be one component of a multi-component complex. A degenerate methylation consensus sequence, purine-purine-m6A-C-[A/C/U], has been known for many years. The presence of this consensus sequence does not guarantee methylation, suggesting that this process is regulated. There is also a report of another m6A methylase activity (toward U6 snRNA) that does not utilize this consensus sequence and appears to be a separate enzyme. Two recent independent studies utilizing m6A-specific antibodies and next generation sequencing provide a transcriptome-wide assessment of mRNA m6A methylation, substantially increasing knowledge of this modification., These papers clearly demonstrated that m6A methylation is a very prominent mRNA modification, identifying more than 7,000 genes that contain m6A. They were in general agreement on a recognition consensus sequence for the adenosine that is methylated, the overall distribution of m6A sites along the length of the transcripts, and in the high conservation between human and mouse of major elements of this common RNA modification. These papers provide valuable new insights of m6A in the RNA transcriptome and further the understanding of RNA epigenetics. One of these studies identified potential m6A-specific binding proteins, which may have functional significance. Evidence was also presented that m6A affects RNA splicing. Using conditions that focused on differentially expressed isoforms, a positive relationship was seen between m6A and isoform switching. Further, differentially spliced exons and introns were significantly enriched with m6A. These data indicate a role for m6A in splicing, further supporting the role of the m6A demethylase, Fto, in this process.

miRNA and Epilepsy

Recently, a connection has been made between RNA processing and epilepsy. MicroRNA (miRNA) is a major RNA regulatory gene family in eukaryotes of which hundreds have been identified. Mature miRNA forms base pairs with mRNA, often in the 3′UTR, to induce mRNA degradation, translational repression, or both. Since a single miRNA can target multiple mRNAs, and multiple miRNAs can act on a single mRNA, miRNAs are thought to operate highly complex regulatory networks that silence targeted genes. A recent review highlights the functions of miRNAs in CNS development and provides multiple examples in which misregulation of CNS miRNAs is associated with neurological disorders, including epilepsy. Specific miRNAs are also being investigated in the context of epilepsy. miRNA-146a has been shown to be upregulated in an animal model of temporal lobe epilepsy (TLE), as well as in hippocampal tissue from patients with TLE and hippocampal sclerosis. Similarly, miRNA-134 has been shown to be upregulated in an experimental model of epilepsy involving status epilepticus and also in temporal neocortex tissue from patients with TLE. These reports suggest that these miRNAs are neuronally activated and are coupled to epileptogenesis. Upregulation of miRNA-134 had a suppressive effect on a known target, LIM kinase 1 (Limk1). Antagomir silencing of miRNA-134 after status epilepticus caused a substantial reduction in the number of subsequent seizures, and the rescue of Lim1k was implicated in this protection. While the mechanism of this dramatic seizure suppression could not be absolutely defined, an antiepileptogenic effect was a considered possibility. This paper clearly shows that silencing miRNA-134 has a neuroprotective and seizure-suppressive effect and provides another avenue in which to study epileptogenesis.

miRNA and m6A

The methylation status of miRNA has been shown to affect stability and turnover. As mentioned above, the location of m6A residues in the 3′UTR of the transcriptome has been linked with miRNA binding sites. These authors found that 67% of 3′UTRs that contain m6A peaks also contained at least one predicted miRNA binding site, significantly greater than expected by chance alone. They also showed that highly expressed brain miRNAs had a significantly greater percentage of target transcripts that contain m6A than those of lowly expressed miRNAs, further supporting a link between the CNS, miRNAs, and m6A. Based on the evidence that miRNA-134 and miRNA-146a are involved in the pathogenesis of epilepsy, we searched the sequence of these miRNAs for a m6A consensus site. We found that both miRNA-134 and miRNA-146a contain a potential m6A site in the seed region, which is thought to play an important role in miRNA recognition of target mRNA (Fig. 2). This suggests that m6A methylation of miRNAs could interfere with binding to target mRNAs, in a manner analogous to that which is known for m6A DNA methylation to inhibit the binding of certain restriction endonucleases. The occurrence of m6A in dsRNA is also known to lower melting temperature and destabilize base-paired duplex structures: a single m6A modification of an octamer is estimated to increase (destabilize) ΔG by about +1 to 1.2 kcal/mol. m6A modification of miRNAs may therefore affect the processing of primary miRNAs to mature miRNAs, which involves duplex formation, and/or the binding of miRNA to mRNA, which also involves a dsRNA component. The confirmation of the m6A modification on these miRNAs remains to be determined, although m6A is known to occur in 3′UTRs of target mRNA. The presence of m6A in either the miRNA or in the 3′UTR in mRNA would affect the subsequent interactions, likely via destabilization of duplexes. Furthermore, the presence of m6A in miRNA or mRNA may affect the interaction with RNA-binding proteins., These effects suggest an important additional layer of control over miRNA actions to impact neurological development and disease states.

Figure 2. A. Pre-miRNA-134. Mature miRNA-134 is in red, the potential m6A methylation residue is in cyan (UGAC, from ref. 18), and the seed region of the mature miRNA is in the box. B. Pre-miRNA-146a. Mature miRNA-146a is in red, the potential m6A methylation residue is in cyan (Pu-Pu-AC[A/C/U]), and the seed region of the mature miRNA is in the box. Sequences were obtained from the microRNA database, www.mirbase.org/index.shtml

Summary

Based on recent evidence, we have provided a possible link between Fto, RNA epigenetics and epilepsy. Fto is highly expressed in the CNS and demethylates m6A in ssRNA. Two recent, independent studies provide ample evidence that m6A is a prevalent and important RNA epigenetic mark, thus implicating Fto in the regulation of these RNA epigenetic events. Reversible RNA methylation involving Fto suggests that dynamic methylation states of mRNA or miRNA may modulate protein expression and cell status, analogous to reversible DNA and histone methylations. miRNAs play important roles in CNS physiology and pathology, including epilepsy, and several have been implicated in epileptogenesis. Certain m6A sites may play a role in miRNA stability, processing or regulation. Thus, the recently discovered m6A RNA demethylation activity of Fto places it in the heart of RNA epigenetics with subsequent regulatory roles on critical CNS processes.