Messenger RNA methylation regulates glioblastoma tumorigenesis

Abstract

Messenger RNA modification provides an additional layer of gene regulation in cells. In this issue of Cancer Cell, Zhang et al. report that ALKBH5, a demethylase of the mRNA modification N⁶-methyladenosine, regulates proliferation and self-renewal of glioblastoma stem-like cells by modulating pre-mRNA stability and expression of the FOXM1 gene.

Glioblastoma (GBM), the most prevalent primary intrinsic brain tumor, is fatal, with patient median survival time of only ~14 months. Clinically, these tumors display devastating features such as therapeutic resistance, promotion of angiogenesis, immune modulation, and propensity to invade normal brain tissue. GBM is highly heterogeneous and displays cellular hierarchies related to tumorigenic capability, with glioblastoma stem-like cells (GSCs) representing the top of the hierarchy. While standard therapies kill bulk cancer cells in GBM tumors, they typically are ineffective in eliminating the GSC population, leading to treatment resistance and tumor recurrence. As yet, few effective therapeutic approaches targeting GSCs exist, highlighting the need to understand regulatory mechanisms governing GSC proliferation and self-renewal, in order to expand the therapeutic toolbox.

In eukaryotes, post-transcriptional messenger RNA (mRNA) modification is an emerging gene regulatory mechanism. Among five reported internal mRNA modifications, methylation at the N⁶ position of adenosine (N⁶-methyladenosine, or m⁶A) is the most abundant and tags over 10,000 mRNAs in mammalian cells. Although discovered in early 1970s, the biological significance of m⁶A mRNA modification has only been appreciated recently, owing to advances in techniques to locate m⁶A in the transcriptome and the discovery of m⁶A-specific methylases and demethylases. m⁶A modification is reversible. In terms of methyltransferases, a heterodimer consisting of methyltransferase-like protein 3 (Mettl3) and methyltransferase-like protein 14 (Mettl14) is the core subunit of a m⁶A methyltransferase complex (Wang and Zhao, 2016). By contrast, proteins from the AlkB family of Fe(II)-α-ketoglutarate-dependent dioxygenases, including ALKBH9 (FTO) and ALKBH5, have been described as m⁶A demethylases (Wang and Zhao, 2016). A recent study reported that FTO primarily demethylates cap methylation (N⁶,2′-O-dimethyladenosine, or m⁶Am), while ALKBH5 specifically demethylates m⁶A (Mauer et al., 2017). In mice, loss of function studies of genes encoding methylases or demethylases demonstrate that m⁶A mRNA modification is essential for normal development: Mettl3 knockout mice exhibit early embryonic lethality (Geula et al., 2015) and Alkbh5 deletion impairs fertility in male mice (Zheng et al., 2013). As yet, little is known about m⁶A dysfunction in diseases such as cancer.

In the current issue of Cancer Cell, Zhang et al. (Zhang et al., 2017) searched existing datasets for levels of m⁶A regulators and discovered that the expression of m⁶A demethylase ALKBH5 in tumor cells correlates with poor clinical outcome for GBM patients. The authors performed ALKBH5 immunostaining on primary patient GBM bulk tumor samples and detected upregulated ALKBH5 expression in cells expressing the GSC markers Sox2 and Nestin, suggesting that ALKBH5 regulates GSC activity. Using a knockdown approach in cultured human GSCs, the authors showed that ALKBH5 loss deceases GSC proliferation and reduces expression of the stemness markers Nestin, Sox2, Nanog, and Oct4, which are normally expressed in GSCs. Importantly, these phenotypes were rescued by wildtype but not catalytic inactive ALKBH5. Furthermore, mice injected with ALKBH5 knockdown GSCs displayed extended survival with fewer tumors than did mice injected with control GSCs. These results suggest ALKBH5 is required for GSC proliferation and self-renewal and suggest a role for m⁶A in glioblastoma formation.

Based on these observations, the authors then searched for ALKBH5 downstream targets. Coupling differential m⁶A methylome analysis in control vs. ALKBH5 knockdown GSCs with investigation of differential gene expression, the authors identified FOXM1 gene as a direct ALKBH5 target. Since FOXM1 reportedly functions at a critical signaling node in GSCs (Gong et al., 2015), the authors performed mechanistic analysis to determine how m⁶A loss might increase FOXM1 expression. m⁶A regulates diverse mRNA activities. Among them, several groups report m⁶A-regulated mRNA decay mediated by multiple mechanisms. Some propose that m⁶A enhances mRNA binding to the cytoplasmic protein YTH domain protein 2 (YTHDF2), which promotes mRNA degradation either by pulling mRNAs into p-bodies or by deadenylating mRNA (Wang and Zhao, 2016). Others suggest that m⁶A disrupts interaction between RNA and the well-established RNA stabilizer HuR (Human Antigen R or ELAVL1) to facilitate microRNA targeting and mRNA decay (Wang et al., 2014). The new study by Zhang et al. supports the latter mechanism as governing decreased FOXM1 expression in ALKBH5-depleted GSCs (Figure 1). Interestingly, the authors show that m⁶A regulates stability of nascent FOXM1 transcripts rather than the fully processed mature mRNA. m⁶A is known to regulate pre-mRNA splicing by interacting with the nuclear m⁶A binding protein YTHDC1 (Xiao et al., 2016), and published investigations of m⁶A-regulated mRNA stability have mostly focused on mature mRNAs. Therefore, the work by Zhang et al. is the first to link m⁶A to pre-mRNA stability. That finding raises several interesting questions. How does HuR regulate pre-mRNA stability, as microRNAs reside primarily in the cytoplasm? Can YTHDC1 regulate pre-mRNA stability in addition to splicing? And how is Mettl3/14 and ALKBH5 activity coordinated to methylate and demethylate pre-mRNAs?

Figure. 1. FOXM1-AS long noncoding RNA facilitates interaction between ALKBH5 and FOXM1 pre-mRNA resulting in FOXM1 pre-mRNA demethylation and stabilization.

A. In ALKBH5 knockdown cells, FOXM1 pre-mRNA shows increased m⁶A levels at the 3’-UTR that block HuR binding to pre-mRNA. This promotes FOXM1 pre-mRNA decay, down-regulated FOXM1 expression, and decreased GSC proliferation. Dotted line: mRNA with reduced stability. B. In GSCs that express high levels of ALKBH5, a long non-coding RNA transcribed in the FOXM1 antisense orientation facilitates interaction of ALKBH5 and with nascent FOXM1 transcripts. FOXM1 pre-mRNAs lacking m⁶A are stabilized by increased interaction with HuR, promoting FOXM1 overexpression and production of GSCs with enhanced tumorigenic activity in vivo.

To explain how ALKBH5 is recruited onto nascent FOXM1 transcripts, the authors report that a chromatin-bound long non-coding RNA (lncRNA) transcribed in the antisense orientation of FOXM1 (named FOXM1-AS) facilitates nuclear interaction of ALKBH5 and FOXM1 pre-mRNA (Figure 1). Importantly, FOXM1-AS knockdown in GSCs phenocopied depletion of either ALKBH5 or FOXM1, supporting the idea that FOXM1-AS is critical for GSC maintenance. This observation raises the exciting possibility that nuclear lncRNAs function to define m⁶A targets. As yet, mechanisms underlying m⁶A target specificity remain unclear. Although the short nucleotide sequence RRACH (R= G or A, H=A, C, or U) reportedly serves as an m⁶A motif, only a small fraction of these motifs in the transcriptome are methylated, suggesting other uncharacterized regulatory mechanisms. Antisense transcript-mediated interaction of pre-mRNA and m⁶A enzymes could represent one such mechanism and warrants future investigation.

Together with a recent study reporting that ALKBH5 is required for stem cell phenotypes in hypoxia-induced breast cancer (Zhang et al., 2016), the new Cancer Cell study begins to define m⁶A function in cancer stem cell formation or maintenance. Nonetheless, we still do not know whether or how other components of the m⁶A machinery, such as m⁶A methyltransferases or m⁶A binding proteins, regulate cancer stem cell activity. Also, due to tumor cell heterogeneity, it is necessary to investigate m⁶A function in a panel of cancers to define genes/pathways regulated by m⁶A, as well as in cancer type-specific ones. Furthermore, as ALKBH5 demethylase activity requires a non-heme iron moiety and iron plays an important role in proliferation of some cancer stem cells, such as GSCs (Schonberg et al., 2015), it would be interesting to investigate how m⁶A mRNA methylation functions in cancer cell metabolism. Lastly, the list of other types of mRNA modification is expanding and the role of these activities in cancer remains unexplored. Thus, although we are only beginning to understand the relationship of mRNA modification to cancer, it is possible that studies such as the new one by Zhang et al. will lead to novel cancer therapeutics to target mRNA modification.