METTL14 Inhibits Hematopoietic Stem/Progenitor Differentiation and Promotes Leukemogenesis via mRNA m6A Modification

Abstract

N⁶-methyladenosine (m⁶A), the most prevalent internal modification in eukaryotic messenger RNAs (mRNAs), plays critical roles in many bioprocesses. However, its functions in normal and malignant hematopoiesis remain elusive. Here, we report that METTL14, a key component of the m⁶A methyltransferase complex, is highly expressed in normal hematopoietic stem/progenitor cells (HSPCs) and acute myeloid leukemia (AML) cells carrying t(11q23), t(15;17), or t(8;21) and is downregulated during myeloid differentiation. Silencing of METTL14 promotes terminal myeloid differentiation of normal HSPCs and AML cells and inhibits AML cell survival/proliferation. METTL14 is required for development and maintenance of AML and self-renewal of leukemia stem/initiation cells (LSCs/LICs). Mechanistically, METTL14 exerts its oncogenic role by regulating its mRNA targets (e.g., MYB and MYC) through m⁶A modification, while the protein itself is negatively regulated by SPI1. Collectively, our results reveal the SPI1-METTL14-MYB/MYC signaling axis in myelopoiesis and leukemogenesis and highlight the critical roles of METTL14 and m⁶A modification in normal and malignant hematopoiesis.

Keywords: HSC; HSPCs; LSCs/LICs; METTL14; MYB; MYC; N(6)-methyladenosine modification; SPI1; acute myeloid leukemia; myeloid differentiation block.

Figure 1. Impact of METTL14 on normal myeloid differentiation

(A,B) Expression of individual m⁶A modifiers (A) and global m⁶A levels in mRNA (B) in c-Kit⁺ and c-Kit⁻ BM cells from wildtype C57BL/6 mice (n=3), as detected by qPCR and LC-MS/MS, respectively. (C) Relative expression of Mettl14 in different sub-populations of BM cells from wildtype C57BL/6 mice as detected by qPCR. Expression of Mettl14 in HPCs was set as 1. (D) C57BL/6 Lin⁻ HSPCs were co-cultured with OP9 cells in vitro for 5 days and subjected to flow cytometric analysis. (E) OP9 co-cultured cells were subjected to qPCR (left) and western blot (right) analysis for the expression of Mettl14. (F) Schematic diagram showing the procedure of inducing human CD34⁺ HSPCs toward monocyte/macrophage differentiation. Note that the day when M-CSF was first added to induce differentiation was set as day 0. (G) qPCR showing changes of METTL14, CD11b, CSF1R, and GATA1 during differentiation in the control (shNS) and METTL14-depleted (shM14-#2) groups. (H) Western blot showing expression changes of METTL14 during differentiation. Signal of METTL14 was quantified and normalized to that of GAPDH. (I) In vitro induced human CD34⁺ cells were stained at the indicated time points and subjected to flow cytometric analysis (left). Wright-Giemsa staining of cells collected at day 11 was shown on right. (J) In vivo competition assays. Mean±SD values are shown for Figures 1A, B, C, E (left panel), G, and J. *, p < 0.05; **, p <0.01; ***, p < 0.001; N.S., non-significant; t-test. See also Figure S1.

Figure 2. METTL14 expression is upregulated in AML and negatively regulated by SPI1

(A) Expression levels of METTL14 in primary AML patients with various chromosomal translocations relative to that in BM mononuclear cells (MNCs) from healthy donors (NC) as detected by qPCR. (B) qPCR showing expression of METTL14 in leukemia cell lines as compared to MNCs or different fractions (CD34⁺ and CD34⁻) of MNCs from healthy donors. (C) Lin⁻ BM of wildtype mice were transduced with MSCVneo empty vector or AML fusion genes and seeded for CFA assays. Cells were harvested after two rounds of plating and subjected to qPCR analysis for expression of Mettl14 and Mettl3. (D) Change of METTL14 expression (upper) and global m⁶A level (lower) in NB4 cells after treatment with 500 nM ATRA for 72 hours or 0.5 ng/mL PMA for 48 hours as compared to DMSO-treated cells as detected by qPCR and LC-MS/MS, respectively. (E,F) MM6 cells (E) and CD34+ HSPCs (F) were transduced with shRNAs targeting SPI1 or SPI1 overexpression vector and subjected to western blot analysis (upper) and qPCR analysis (lower). (G) ChIP-qPCR assays showing direct binding of SPI1 on METTL14 promoter. Five potential binding sites upstream of the transcription start site (TSS, indicated as 0) and one downstream of TSS were predicted by JASPAR. The regions containing these sites for qPCR were shown. NC, negative control site. (H,I) Expression of SPI1 in BM MNCs from healthy donors (NC) or primary AML patients with various chromosomal translocations as detected by qPCR (H), and the Pearson correlation of SPI1 with METTL14 in expression across these samples (I). Mean±SD values are shown (except for Figures 2E (upper panel), F (upper panel), and I). *, p < 0.05; **, p <0.01; ***, p < 0.001; t-test. See also Figure S2.

Figure 3. METTL14 plays a critical role in AML development and maintenance

(A) Lin⁻ BM cells from Mettl14^fl/fl-CRE^ERT mice were transduced with MSCVneo-MLL-AF9 (MA9), -MLL-AF10 (MA10), or -AML1-ETO9a (AE9a) retroviruses and seeded onto methylcellulose medium for CFA assays. 4-OHT at a final concentration of 1 µM was added to the methylcellulose medium at the first round of plating. N.D., non-detectable. (B) Lin⁻ BM cells from wildtype CD45.1 mice were transduced with MSCVneo-MLL-AF9 (MA9) plus MSCV-PIG (PIG), MSCV-PIG-METTL14-WT (M14-WT), or MSCV-PIG-METTL14-R298P (M14-R298P) retroviruses, and seeded for CFA assays. (C) Cells collected from the 2^nd round of plating in Figure 3B were subjected to mRNA extraction following by LC-MS/MS for detection of global m⁶A changes. (D) Kaplan–Meier curves showing the effect of METTL14 knockout on MLL-AF9-induced primary leukemogenesis. BMT recipient mice were treated with vehicle (CT) or tamoxifen (TAM) for five consecutive days as indicated with arrows. (E) Wright–Giemsa staining of peripheral blood (PB) and BM, and hematoxylin and eosin (H&E) staining of spleen (SP) and liver (LV) of the primary fl/fl BMT recipient mice at the end point. Bar= 20 µm (for PB and BM) or 200 µm (for SP and LV). (F) LC-MS/MS detection of global m⁶A changes in mRNA of BM cells (a mixture of fully or partially Mettl14 depleted AML cells and normal BM cells) isolated from fl/fl recipient mice in Figure 3D. (G) Percentage of LSK cells in the BM of primary leukemic fl/fl BMT mice treated with vehicle (CT) or tamoxifen (TAM) as determined by flow cytometry. (H) CFA assays of BM cells harvested from primary leukemic fl/fl BMT mice (2 mice/group). Representative images of the colonies after the first round of plating were displayed on top, while the numbers of colonies were shown at the bottom. (I) Limiting dilution assays using BM leukemia cells from primary fl/fl BMT mice. The estimated LSC/LIC frequency was shown on the plot. Dose, number of donor cells; tested, total number of mice used as BMT recipients in the assay; response, mice developed leukemia within 4 weeks post BMT. (J) BM cells from MA9, AE9a, and PML-RARa leukemia mice were transduced with Mettl14 shRNAs and subjected to CFA assays. (K) Kaplan–Meier curves showing the effect of Mettl14 knockdown on the maintenance/progression of MA9-induced AML in secondary BMT recipient mice. (L,M) Effect of Mettl14 knockout on the progression of MA9-induced AML in secondary BMT recipient mice. Recipient mice were divided randomly into two groups and treated with vehicle (CT) or tamoxifen (TAM) at the indicated time points (see arrows). Kaplan–Meier curves were shown in (L). Wright-Giemsa stained PB and BM, and H&E stained spleen and liver of the secondary BMT recipient mice at the end point were shown in (M). Bar= 20 µm (for PB and BM) or 200 µm (for SP and LV). Mean±SD values are shown for Figures 3A, B, C, F, G, H (lower panel), and J. *, p < 0.05; **, p <0.01; ***, p < 0.001; N.S., non-significant; t-test (for Figures 3B, 3G, 3H, and 3J) or log-rank test (for Figures 3D, 3I, and 3L). See also Figure S3.

Figure 4. METTL14 blocks myeloid differentiation of human AML cells

(A,B) Effect of METTL14 knockdown (A) or overexpression (B) on AML cell growth/proliferation was analyzed. The knockdown efficiency of shRNAs were confirmed by western blot and shown on the left (A). (C,D) Flow cytometric analysis of CD11b⁺ and/or CD14⁺ cell populations (C) or Wright-Giemsa staining of cytospin slides (D). Arrows indicate differentiated cells. Bar= 20 µm. (E,F) Effects of METTL14 knockdown on cell proliferation (E) and differentiation (F) of leukemia blasts (CD34⁺) from 3 primary AML patients. (G–I) Effects of METTL14 knockdown on progression of human AML cells in NSGS (Wunderlich et al., 2010) mice. MM6 cells were transduced with shMT14-#2 or scramble shRNA. After selected with puromycin for two passages, cells were injected into NSGS mice (0.3×10⁶ cells/mouse) via tail vein. Kaplan–Meier curves were shown in (G), engraftment of MM6 cells in BM and PB of NSGS mice was shown in (H), while representative images of liver tissues were shown in (I). (J) MM6 and NB4 cells with or without METTL14 knockdown were treated with DMSO or ATRA (100 nM) for 72 hours and subjected to flow cytometric analysis. The percentages of each cell population were summarized on right. (K) AML cells were transduced with empty vector or METTL14 overexpression vectors and treated with ATRA (20 nmol/L for 24 hours for NB4, 200 nmol/L for 48 hours for U937) before subjected to flow cytometric analysis. Percentages of cells with CD11b staining in each group were shown. *, p < 0.05; **, p <0.01; ***, p < 0.001; t-test (for Figures 4A, 4B, 4C, 4E, 4F, and 4K; Mean±SD values are shown) or log-rank test (for Figure 4G). See also Figure S4.

Figure 5. Transcriptome-wide identification of METTL14 targets

(A) Venn diagram showing numbers of genes with significant changes in expression (p<0.05; fold change ≥1.2) upon METTL14 knockdown. (B) Expression changes of mRNA transcripts with decreased m⁶A abundance (m⁶A-hypo) upon METTL14 knockdown in METTL14 silenced (shM14-#1 and shM14-#2) cells relative to control cells. Note that the fold changes (FC) were log2 transformed. (C) The m⁶A abundances on MYB and MYC mRNA transcripts in METTL14-knockdown and control MM6 and NB4 cells as detected by m⁶A-seq. Solid triangles indicate regions for qPCR in (D) and (E). (D) Reduction of m⁶A modification in specific regions of MYB and MYC transcripts upon METTL14 knockdown as tested by gene-specific m⁶A-qPCR assay in MM6 cells. (E) CLIP-qPCR showing the association of MYB and MYC transcripts with METTL14 in MM6 cells. (F) qPCR (upper) and western blot (lower) showing decrease of MYB and MYC expression after knockdown of METTL14 in AML cell lines. (G) qPCR (left) and western blot (right) showing increase of MYB and MYC expression in U937 cells with ectopic expression of wildtype (M14-WT) but not mutated (M14-R298P) METTL14. (H,I) Gene-specific m⁶A-qPCR (H) and qPCR/western blot (I) showing decrease of m⁶A modification in specific region of Myb or Myc transcripts and expression changes of Myb or Myc mRNA/protein, respectively, in c-Kit⁺ BM of Mettl14^fl/fl-CRE^ERT mice treated with TAM (KO) as compared to those treated with oil (CT). (J) Expression changes of Myb and Myc mRNA in c-Kit⁺ BM of leukemic fl/fl BMT mice from Figure 3D. (K) Dual luciferase reporter assays showing the effect of METTL14 on MYB/MYC reporters with either wild-type or mutated m⁶A sites. Mean±SD values are shown for Figures 5D–K. *, p < 0.05; **, p <0.01; ***, p < 0.001; t-test. See also Figure S5.

Figure 6. METTL14 regulates mRNA stability and translation of MYB and MYC

(A) The mRNA half-life (t_1/2) of MYB or MYC transcripts in MM6 cells with (shM14-#1 or shM14-#2) or without (shNS) METTL14 depletion. (B) CLIP-qPCR showing decreased association of EIF3A with MYB and MYC transcripts in MM6 cells with (shM14-#1 or shM14-#2) or without (shNS) METTL14 depletion. (C) Ribosome profiling assays. Fractionations of MM6 cell lysates were shown on top. RNAs in different fractions of ribosome were extracted and subjected to qPCR analysis and shown at the bottom. Mean±SD values are shown for Figures 6A–C. *, p < 0.05; **, p <0.01; ***, p < 0.001; t-test. See also Figure S6.

Figure 7. MYB and MYC are critical targets of METTL14 that mediates myeloid differentiation block and AML cell proliferation

(A, B) Effect of MYB or MYC silencing on cell growth (A) and differentiation (B) of AML cell lines. (C) BM leukemic cells from the fl/fl-CT primary BMT mice were transduced with pmiRA1 empty vector or MYB or MYC expression vector and seeded for colony-forming assay with or without 4-OHT (1 µM). Colony numbers were counted and compared. (D) MM6 cells were transduced with shNS or shM14-#2, and with or without MYB or MYC encoding lentivirus as indicated. Western blots showing knockdown of METTL14 as well as ectopic expression of MYB or MYC in the corresponding groups. (E) MYB or MYC overexpression rescues terminal myeloid differentiation of MM6 cells induced by METTL14 knockdown. Representative images of flow cytometric analysis of CD11b and CD14 staining were shown on left and the mean percentages of each population were shown on right. (F) Wright-Giemsa staining of MM6 cells showing reduced differentiation in cells with MYB (shM14+MYB) or MYC (shM14+MYC) overexpression as compared to those with empty vector (shM14+vector) when METTL14 was knocked down. Bar= 20 µm. (G) Proposed model depicting regulation and role of METTL14 in normal and malignant hematopoiesis. Mean±SD values are shown for Figures 7A and 7C. **, p <0.01; ***, p < 0.001; t-test. See also Figure S7.