FIGURE 1 from D-2-HG Inhibits IDH1mut Glioma Growth via FTO Inhibition and Resultant m6A Hypermethylation

IDH1mut production of D-2-HG induces RNA m6A hypermethylation in gliomas. A,IDH1mut patient gliomas manifested RNA m6A hypermethylation, IDH1mut (n = 12) versus IDH1wt (n = 12; P ≤ 0.04). B,IDH1mut gliomaspheres exhibited increased m6A content, IDH1mut (n = 4) versus IDH1wt (n = 7; P ≤ 0.04). In NHA lines, U87 cells, and IDH1wt gliomaspheres, IDH1mut forced expression elevates intracellular D-2-HG content (n = 1; C) and induce m6A enrichment [ANOVA, F(1,27) = 115.0, P ≤ 0.0001; asterisks indicate post hoc Newman–Keuls comparison with vector for each cell type] (D). E, Octyl-D-2-HG treatment (0.5 mmol/L) of IDH1wt gliomaspheres increased m6A content compared with PBS treatment controls (n = 6; paired t test; P ≤ 0.01). F, D-2-HG content in HK385, U87, and NHA cells following treatment with IDH1mut inhibitors (C35 = 2 µmol/L; AG881 = 1 µmol/L), quantified using D-2-HG enzymatic assay. G, Effect of IDH1mut inhibitor on m6A content in HK385, U87, and NHA cells with IDH1mut forced expression [ANOVA, IDH1mut inhibitor treatment: F(8,32) = 12.9, P ≤ 0.0001; vector vs. IDH1mut: F(1,32) = 79.22, P ≤ 0.0001; inhibitor and IDH1 status interaction: F(8,32) = 7.3, P ≤ 0.0001; asterisks indicate post hoc Newman–Keuls comparison with vector or between the groups as indicated by horizontal bars]. Pooled data from native IDH1mut gliomaspheres (HK252, HK211) demonstrating a reduction in D-2-HG (one-tailed paired t test; P ≤ 0.04; H) and m6A content (paired t test; P ≤ 0.003; I) following IDH1mut inhibitor AG881 treatment (1.0 µmol/L) in comparison with DMSO control. J, Volcano plot showing the RADAR log₂ fold change in m6A expression against −log₁₀ converted P values between IDH1mut and IDH1wt U87, NHA, and gliomasphere cells. Transcript sites with an absolute value log₂ fold change of less than 0.5 are excluded (void region in between vertical, dashed red lines). The horizontal, dashed red line indicates the demarcation of RADAR P values <0.01. K, Venn diagram showing the total number of genes fitting TCGA criteria (TCGA P value <1.0 × 10⁻⁶; absolute value log₂ fold change in RNA expression of 1.0 or greater between IDH1mut and IDH1wt patients) and RADAR criteria (P < 0.01) for all transcripts. Note that the total RADAR gene number is slightly larger here than the totals in J, because this list includes all P < 0.01 transcripts (including those with low magnitude fold changes). L, Supervised hierarchical clustering heat map of intersectional TCGA and RADAR data between U87, NHA, and gliomasphere cell lines that conformed to m6A and RNA expression conventions (i.e., an increase in RADAR m6A expression corresponded to a decrease in TCGA RNA expression). Heat map colors represent sample (columns) natural log converted normalized RADAR m6A values as a difference from the transcript (rows) average across genotype. A transcript of interest, ATF5, is identified. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; and ****, P ≤ 0.0001 compared with relevant controls. Unless otherwise stated, P values indicate unpaired Student t test comparisons with the control, or between two groups as indicated by the horizontal line.