IGF2BP3 redirects glycolytic flux to promote one-carbon metabolism and RNA methylation

Abstract

Insulin-like growth factor 2 mRNA-binding protein 3 (IGF2BP3), an oncofetal RNA-binding protein and a non-canonical reader of N6-methyladenosine (m⁶A) mRNA modifications, is known to be critical for leukemogenesis. To understand how the oncogenic function of IGF2BP3 impacts metabolism, we performed metabolic profiling and observed changes in glycolytic flux and one-carbon metabolism, including the biosynthesis of S-adenosyl methionine (SAM), a key substrate for methylation reactions within the cell. Using enhanced crosslinking immunoprecipitation (eCLIP) and polyribosome profiling, we found that IGF2BP3 promotes translation of the regulatory subunit of the methionine adenosyltransferase complex (MAT2B), which is involved in SAM production. Remarkably, IGF2BP3 promotes and alters the level and pattern of m⁶A modifications on mRNA. Taken together, these data suggest the intriguing hypothesis that IGF2BP3 rewrites the epitranscriptome in leukemia cells. Furthermore, this work highlights an interconnection between oncogenic metabolism and RNA modifications, suggesting that pervasive gene expression changes necessary for oncogenesis may be perpetuated by post-transcriptional gene regulation.

Keywords: B-acute lymphoblastic leukemia; CP: Metabolism; MLL-rearranged leukemia; RNA-binding protein; epitranscriptomics.

Figure 1.. IGF2BP3 impacts glycolytic metabolism in B-acute lymphoblastic leukemia cells

(A) Western blots for IGF2BP3-depleted (I3sg2 and I3sg5) SEM, NALM6, and Lin− MLL-Af4 murine cells. (B) Seahorse XF extracellular acidification rate (ECAR) kinetic trace in control and IGF2BP3-depleted SEM cells (I3sg2). (C) Aggregate lactate efflux rates from Seahorse XF analysis in control (NT and NT2) versus IGF2BP3-depleted (I3sg2 and I3sg5) SEM, NALM6, and Lin− MLL-Af4 murine cells (n = 2). (D) Pyruvate and lactate amounts measured by GC/MS in control versus IGF2BP3-depleted (I3sg2) SEM cells. (E) Incorporation of carbon from ¹³C-labeled glucose into pyruvate and lactate, measured as mole percent enrichment (MPE) from GC/MS experiments. All data are n > 3 replicates, represented as mean ± standard deviation (SD), compared by two-sided unpaired t test; *p < 0.05, **p < 0.01, and ***p < 0.001. All experiments were repeated at least twice for consistency. Refer also to Figure S1.

Figure 2.. IGF2BP3 supports one-carbon metabolism pathways that serve as methyl donors

(A) Heatmap depicting significantly altered metabolites from control versus IGF2BP3-depleted SEM cells, as indicated, using targeted analysis of polar central carbon metabolites by LC/MS. Shown are metabolites with a consistent change in both IGF2BP3-depleted lines, I3sg2 (purple) and I3sg5 (sky blue), with their respective p value significance; n = 3. (B) Schematic of metabolites that are produced in one-carbon metabolism. Metabolites reduced after IGF2BP3 depletion are marked in pink. (C–J) Intracellular abundance and steady-state incorporation of carbon from ¹³C-labeled glucose, measured as mole percent enrichment (MPE), into one-carbon pathway metabolites serine, glycine, S-adenosyl-methionine (SAM), and glutathione (GSH) in control versus IGF2BP3-deleted SEM cells. All data are n = 3 replicates represented as mean ± standard deviation (SD), compared by two-sided unpaired t test; *p < 0.05, **p < 0.01, and ***p < 0.001. All experiments were repeated at least twice for consistency. Refer also to Figure S2.

Figure 3.. IGF2BP3 regulates N6-methyladenosine marks on RNA

(A) Western blot analysis of histone methylation (H3K4me1 and H3K4me4) in SEM and Lin− MLL-Af4 cells, control or depleted for IGF2BP3; n = 3. (B) Dot blot analysis of m⁶A modification (left) and methylene blue staining in SEM cells, control or depleted for IGF2BP3. (C) ELISA measurement of m⁶A modification on RNA isolated from SEM, Lin− MLL-Af4, and NALM6 cells (n = 4 for SEM and Lin− MLL-Af4, n = 5 [sg5 = 3] for NALM6). (D) Bar plot (left) and pie chart (right) depicting the m⁶A peak distribution across genomic locations from the m⁶A-eCLIP data in SEM control and IGF2BP3-depleted cells. (E) Metagene plots depicting the changes in the m⁶A peak coverage across the transcriptome in SEM control and IGF2BP3-depleted cells. (F) Volcano plot (top) for genes showing differential m⁶A RNA methylation after IGF2BP3 depletion and IGF2BP3 targets defined by eCLIP analysis. Gray dashed lines indicate the significant cutoffs for differential expression (±1) and the adjusted p value (0.05). Hypomethylated genes are highlighted in blue, while hypermethylated genes are highlighted in red. IGV browser snapshots (bottom) of m⁶A-eCLIP depicting the coverage and change in the peak height between the NT and IGF2BP3-depleted cells for MAT2A 3′ UTR are shown. All data are n ≥ 3 replicates represented as mean ± standard deviation (SD), compared by two-sided unpaired t test; *p < 0.05, **p < 0.01, and ***p < 0.001. In case of missing or outlier values, the replicate was not reported. All experiments were repeated at least twice for consistency. All the western blots were repeated at least three times to report the changes, if any. Refer also to Figure S3.

Figure 4.. IGF2BP3 regulates translation of metabolic genes

(A) Volcano plot showing differentially expressed transcripts after IGF2BP3 depletion and IGF2BP3 targets defined by eCLIP analysis (dots exceeding the thresholds depicted by dashed lines). Gray dashed lines mark the significant cutoffs for differential expression (−1/1) and the adjusted p value (0.05). Putative IGF2BP3 targets identified using Skipper (see Boyle et al.) are highlighted as green triangles. (B) MetaboAnalyst-based pathway enrichment analysis of consistently differentially regulated metabolites after IGF2BP3 loss with IGF2BP3 eCLIP targets in SEM cells. (C) Genome browser snapshots of eCLIP read coverage across some putative IGF2BP3 target genes. Depicted are the genes with key roles in glycolysis and one-carbon metabolism, and they map to the enriched terms in (B). (D) Western blot analysis of key genes in metabolic pathways (left) and simplified schematic depiction of genes that control metabolic pathways altered in IGF2BP3-depleted cells. Numbers alongside the bars represent p values; n = 3. (E) 10%–45% sucrose gradient fractionation of cytosolic extracts from control or IGF2BP3-depleted SEM cells, along with the EDTA control. MAT2B mRNA distribution was measured by RT-qPCR and represented as mean ± standard deviation (SD) (n = 3). All data are n = 2 replicates. Data are represented as mean ± SD, compared by two-sided unpaired t test; *p < 0.05, **p < 0.01, and ***p < 0.001. All experiments were repeated at least twice for consistency. Refer also to Figures S4 and S5.

Figure 5.. IGF2BP3 loss of function impacts glycolytic metabolism and m⁶A RNA modifications in vivo

(A) ELISA measurement of m⁶A modification from murine bone marrow isolated following transplantation with Lin− MLL-Af4 bone marrow (see Lin et al.) (n = 6, NT; n = 7, I3sg2; one-way ANOVA with Bonferroni’s test; *p < 0.05). (B) Chemical structure of I3IN-002. (C) Representative Seahorse XF extracellular acidification rate (ECAR) kinetic trace in SEM cells treated with vehicle or I3IN-002, a small-molecule inhibitor of IGF2BP3 (n = 4). (D) Aggregate lactate efflux rates from Seahorse XF analysis in SEM cells treated with the vehicle of I3IN-002 (n = 3). (E) ELISA measurement of m⁶A RNA modifications in SEM cells treated with vehicle, STM2457, and I3IN-002; n = 3. (F) Heatmap depicting significantly altered metabolites from DMSO-treated versus I3IN-002-treated SEM cells by LC/MS. The common metabolites that are downregulated both after IGF2BP3 deletion and after treatment with I3IN-002 are marked in purple; n = 3. (G) MetaboAnalyst-based pathway enrichment analysis of differentially regulated metabolites in the DMSO-treated versus I3IN-002-treated SEM cells. The common pathways that change after IGF2BP3 deletion and after treatment with I3IN-002 are marked in purple. Unless otherwise noted, all data are represented as mean ± standard deviation (SD), compared by two-sided unpaired t test; *p < 0.05, **p < 0.01, and ***p < 0.001. All experiments were repeated at least twice for consistency. Refer also to Figure S6.

Figure 6.. IGF2BP3 promotes glycolytic metabolism and m⁶A RNA modifications in vivo

(A) Western blot analysis of Lin− cells from Igf2bp3^del/del mice. Briefly, cells were isolated from mice with a germline deletion of Igf2bp3, transformed with MLL-Af4, and then subjected to transduction with MSCV-based constructs carrying the wild-type murine Igf2bp3. Proteins that were analyzed are Igf2bp3, Mat2a, Mat2b, and actin. (B) Cell growth, measured by CellTiter-Glo, over 4 days in Igf2bp3^del/del Lin− MLL-Af4 cells with enforced IGF2BP3 expression as above. Viability has been normalized to control cells; mean ± standard deviation (SD) (n = 5); one-way ANOVA followed by Bonferroni’s multiple comparisons test; ****p < 0.0001. (C) Representative Seahorse XF extracellular acidification rate (ECAR) kinetic trace in cells described above (n = 4). (D) Aggregate lactate efflux rates from Seahorse XF analysis in cells described above; two-sided unpaired t test; *p < 0.05 (n = 4). (E) Colony formation assays from Lin− MLL-Af4 cells as described above; two-sided unpaired t test; **p < 0.01 (n = 2). (F) ELISA measurement of m⁶A modification on RNA isolated from Igf2bp3^del/del Lin− MLL-Af4 cells with enforced IGF2BP3 expression as above; two-sided unpaired t test; *p < 0.05 (n = 3). (G) Percentage engraftment of CD45.2 Lin− cells in bone marrow from Igf2bp3^del/del mice transduced with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (H) Quantitation of bone marrow count in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (I) Spleen weights of mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (J) Quantitation of spleen cell count in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (K) Quantitation of bone marrow CD11b+ cell count in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (L) Quantitation of bone marrow Lin− cell count along with representative fluorescence-activated cell sorting (FACS) plots in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (M) Quantitation of bone marrow CD11b+cKit+ cell count in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (N) Quantitation of bone marrow LSK (Lin− cKit+Sca1−) cell count in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (O) Quantitation of bone marrow CD11b+Sca1− (potential LIC) cell count in mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. (P) Seahorse XF ECAR kinetic trace for bone marrow cells isolated from the empty vector (Ctrl) or IGF2BP3 re-expression group at 6 weeks (n = 4, each group; for representation n = 2). (Q) Aggregate lactate efflux rates from Seahorse XF analysis in cells described above; reported as mean ± SD (n = 4). (R) ELISA measurement of m⁶A RNA modifications in splenic tumors isolated from mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks; reported as mean ± SD; 8 mice/group. (S) H&E staining of spleen of mice transplanted with mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups at 6 weeks. Scale bar: 100 μm. (T) Overall survival of mice transplanted with MLL-Af4 re-expressing empty vector (Ctrl) or IGF2BP3 in the two groups (representative graph cumulative of two experiments, 8 mice/group; the experiment was terminated after 12 weeks; Kaplan-Meier method with log rank test was used to report the results). The animal experiments were repeated twice. All the western blots were repeated at least three times to report the changes, if any. Data in this figure are represented as mean ± SD with n = 8 mice per group. Statistical tests were performed using two-sided unpaired t test with significance levels as indicated; *p < 0.05, **p < 0.01, and ***p < 0.001. Refer also to Figures S7 and S8.