RNA-binding proteins of KHDRBS and IGF2BP families control the oncogenic activity of MLL-AF4

Abstract

Chromosomal translocation generates the MLL-AF4 fusion gene, which causes acute leukemia of multiple lineages. MLL-AF4 is a strong oncogenic driver that induces leukemia without additional mutations and is the most common cause of pediatric leukemia. However, establishment of a murine disease model via retroviral transduction has been difficult owning to a lack of understanding of its regulatory mechanisms. Here, we show that MLL-AF4 protein is post-transcriptionally regulated by RNA-binding proteins, including those of KHDRBS and IGF2BP families. MLL-AF4 translation is inhibited by ribosomal stalling, which occurs at regulatory sites containing AU-rich sequences recognized by KHDRBSs. Synonymous mutations disrupting the association of KHDRBSs result in proper translation of MLL-AF4 and leukemic transformation. Consequently, the synonymous MLL-AF4 mutant induces leukemia in vivo. Our results reveal that post-transcriptional regulation critically controls the oncogenic activity of MLL-AF4; these findings might be valuable in developing novel therapies via modulation of the activity of RNA-binding proteins.

Fig. 1. MLL-AF4 translation is inhibited by post-transcriptional regulation.

a Colony-forming units (CFUs) of various MLL fusions per 10,000 cells at third- and fourth-round passages under myeloid conditions (n = 9: Vector, MLL-ΔFP, MLL-AF4, MLL-mAf4; n = 7: MLL-ENL; n = 6: MLL-AF10). Hoxa9 expression normalized to that of Gapdh of first round colonies is shown as the relative value of the vector control (set to 1) (n = 12, Vector, MLL-ΔFP, MLL-AF4, MLL-mAf4; n = 8: MLL-ENL; n = 7: MLL-AF10). b Leukemogenic potential of MLL-AF4 and MLL-mAf4 in vivo. Murine HSPCs were transduced with MLL-AF4 constructs and transplanted into syngeneic mice (mock, n = 6; MLL-AF4, n = 10; MLL-mAf4, n = 30; MLL-mAf4 secondary transplantation n = 5). c mRNA expression of the MLL fusion genes in 293 T cells analyzed using a qPCR probe for the EPS region in the pMSCV neo plasmids (n = 3). d Western blotting of MLL fusions in 293 T cells transfected with the MLL fusion expression vectors in c. Anti-beta tubulin (TUBB) (an internal standard) and neomycin phosphatase II (NPTII) antibodies (a gene transduction control) were included for comparison. e Virus particle production by the MLL fusion expression vectors was quantitated using qRT-PCR with the MSCV-EPS probe and absolute quantification methods (n = 4). f Transduction of recombinant viruses carrying various MLL fusion genes in murine HSPCs were determined using qPCR probes for MSCV-EPS and the murine Gapdh locus (n = 7: Vector, MLL-ENL, MLL-AF4, MLL-mAf4, Mock; n = 6: MLL-ΔFP; n = 4: MLL-AF10). g Cell numbers of murine HSPCs infected with various retroviruses carrying MLL fusion genes after 5 days of selection with G418 (n = 3). h Relative mRNA expression of MLL fusion genes after antibiotic selection using qRT-PCR with the MSCV-EPS probe (n = 3). i Western blotting of HSPCs transduced with retroviruses carrying MLL fusion genes cultured in G418 for 5 days, as shown in d. Data are presented as the mean ± SD of biologically independent replicates (a, c, e, f, g, h). P-value was calculated by one-way ANOVA followed by Tukey’s test (a, c, e, f, g). Western blotting was performed on two biological replicates (d, i). See also Supplementary Fig. 1. Source data are provided as a Source Data file.

Fig. 2. The post-transcriptional regulatory sequence is responsible for MLL-AF4 translation.

a Domain swapping mutants of various MLL-AF4 mutants are shown in blue (mouse), white (human), and green (synonymous mutations). RNA and amino acid sequences of the post-transcriptional regulatory sequence (PTRS) of human AF4 (hPTRS) and its corresponding sequences of mouse AF4 (mPTRS) and the synonymous mutant (sPTRS). The minimum PTRS is indicated with a red rectangle. b Western blotting of MLL-AF4 mutants in 293 T cells, as described in Fig. 1d. c Transforming ability of MLL-AF4 mutants under myeloid conditions, as described in Fig. 1a (n = 8: Vector, MLL-ΔFP, MLL-AF4, MLL-mAf4; n = 5: MLL-AF4 sPTRS; n = 4: MLL-AF4 mPTRS, MLL-mAf4 hPTRS). Hoxa9 expression normalized to that of Gapdh of first round colonies is shown as the relative value of the MLL-mAf4 (set to 100) (n = 9, Vector, MLL-AF4, MLL-mAf4; n = 8: MLL-ΔFP; n = 6: MLL-AF4 sPTRS; n = 5: MLL-AF4 mPTRS, MLL-mAf4 hPTRS). d Virus particle production by the MLL-AF4 mutant expression vectors. Virus particle production was quantitated as described in Fig. 1e (n = 4). e Relative transduction units of retroviruses carrying various MLL-AF4 mutant genes in murine HSPCs were determined as described in Fig. 1f (n = 7, Vector, MLL-AF4, MLL-mAf4; n = 6: MLL-ΔFP, Mock; n = 3: MLL-AF4 sPTRS). f Cell numbers of MLL-AF4-transduced murine HSPCs after 5 days of G418 selection (n = 3). g Transforming ability of MLL-AF4 mutants under an ex vivo lymphoid culture condition. Cell numbers per 10,000 cells at the third passage is shown (n = 3). h Western blotting of the various MLL-AF4 synonymous mutants in 293 T cells, as described in Fig. 1d. i Transforming ability of various MLL-AF4 synonymous mutants under myeloid conditions. CFUs per 10,000 cells in third- and fourth-round culture is shown (n = 4). Hoxa9 expression normalized to Gapdh is shown as the relative value of MLL-mAF4 (set to 100) (n = 4). Data are presented as the mean ± SD of indicated biologically independent replicates (c, d, e, f, g, i P-value was calculated by one-way ANOVA followed by Tukey’s test (c, d, e, f, i). Western blotting was performed on two biological replicates (b, h). See also Supplementary Fig. 2. Source data are provided as a Source Data file.

Fig. 3. A synonymous mutant of MLL-AF4 induces leukemia.

a Leukemogenic potential of MLL-AF4 sPTRS in vivo. Murine HSPCs were transduced with retrovirus for the MLL-AF4 sPTRS construct and transplanted into syngeneic mice. MLL-AF4 sPTRS, n = 30. The B/M-MPAL and T-ALL cases were excluded as retrovirus genome integration was not detected (Supplementary Fig. 3). The MLL-AF4 and MLL-mAf4 data (Fig. 1b) are shown for comparison. b The immunophenotype of MLL-AF4 sPTRS-mediated leukemia. Bone marrow cells from mice with MLL-AF4 sPTRS-mediated leukemia were analyzed via flow cytometry for the indicated markers. Unstained control is indicated in blue and stained sample indicated in red. c The expression of myeloid markers, Gr1 and Cd11b, in AML cells of MLL-AF4 sPTRS-leukemia. MLL-mAF4- and MLL-ENL-mediated AMLs were included for comparison. d qRT-PCR using bone marrow cells from mice with MLL-AF4 sPTRS-mediated leukemia. Expression levels normalized to Gapdh are shown relative to those of HSPCs. Data are presented as the mean ± SD of PCR triplicates. See also Supplementary Fig. 3. Source data are provided as a Source Data file.

Fig. 4. Unique RNA-binding proteins specifically bind to RNA containing the PTRS.

a Silver staining of purified factors associated with the minimum PTRS of AF4. b Heat map of proteins co-purified with the PTRS RNAs with the relative scores. Proteins were identified using mass spectrometry. c Immunoprecipitation-western blotting of proteins associated with the PTRS RNAs using 293 T cells transiently expressing FLAG-tagged proteins. d CISBP-RNA database analysis of the RNA sequences specifically recognized by various RBPs in the minimum PTRS of human AF4 (http://cisbp-rna.ccbr.utoronto.ca/index.php). The nucleotide bases in the three AU-rich sites are marked in blue letters, whereas the different bases from the human sequence in the three AU-rich sites are in red. e Transforming ability of MLL-AF4 mutants. Various MLL-AF4 constructs carrying synonymous mutations on the AU-rich sites were examined for transformation of HSPCs under an ex vivo myeloid condition, as shown in Fig. 1A (n = 8: Vector, MLL-AF4, MLL-mAf4, MLL-AF4 sPTRS; n = 4: the others). Hoxa9 expression normalized to Gapdh is shown as the relative value of MLL-mAF4 (set to 100) (n = 5). Data are presented as the mean ± SD of indicated biologically independent replicates. P-value was calculated by one-way ANOVA followed by Tukey’s test. f Western blotting of MLL-AF4 mutants in 293 T cells transfected with the MLL fusion expression vectors shown in (e) (as in Fig. 1d). g Association of endogenous RBPs with the minimum PTRS of AF4. Immunoprecipitation-western blotting was performed using 293 T cells. Endogenous proteins were detected using specific antibodies. Silver staining (a) and Western blotting (c, f, g) were performed on two biological replicates. See also Supplementary Fig. 4. Source data are provided as a Source Data file.

Fig. 5. IGF2BP3 is responsible for the post-transcriptional inactivation of MLL-AF4.

a Transforming ability of MLL-AF4 sAU13 in the absence of Igf2bp3. The MLL-AF4 construct carrying synonymous mutations at AU1 and 3 was examined for transformation of HSPCs under an ex vivo myeloid condition with co-transduction of a knockout construct for Igf2bp3. CFUs are shown as in Fig. 1a, with images of representative colonies (n = 4). Hoxa9 expression during fourth-round passage is shown (n = 3). b Western blotting of MLL-AF4 sAU13 mutant and endogenous murine IGF2BP3 in immortalized cells. Western blotting was performed on two biological replicates. c Inhibition of MLL-AF4-immortalized cell proliferation by rescuing human IGF2BP3 expression. Cells immortalized by MLL-AF4 sAU13 were transduced with a lentivirus carrying IGF2BP3 and GFP. The ratio of GFP-positive cells was monitored on days 3 and 7 (n = 4). Data are presented as the mean ± SD of indicated biologically independent replicates (a, c). P-value was calculated by two tailed T-test (c). See also Supplementary Fig. 5. Source data are provided as a Source Data file.

Fig. 6. RBP complex inhibits translation of MLL-AF4.

a Schematic representation of the reporter constructs for post-transcriptional regulation. b Subcellular localization of GFP and RFP reporters in 293 T cells. c Relative fluorescent signals of GFP and RFP reporters. The fluorescent intensities were analyzed using ImageJ software and expressed as the GFP/RFP ratio (n = 7: WT, hPTRS, mPTRS, sPTRS; n = 4: hPTRS sAU123, mPTRS hAU123, STOP-hPTRS). Data are presented as the mean ± SD of indicated biologically independent replicates. P-value was calculated by one-way ANOVA followed by Tukey’s test. d Western blotting of the GFP/RFP reporter proteins in 293 T cells transiently expressing the constructs. GFP and RFP proteins were detected using FLAG and HA antibodies, respectively. e Co-localization of GFP reporter and RNA-binding proteins. The GFP-tagged hPTRS reporter and mCherry-tagged RBPs constructs were co-transfected in 293 T cells. Nuclei were stained with Hoechst 33342. f Association of ribosomal proteins with GFP reporters. FLAG-tagged GFP reporters transiently expressed in 293 T cells were analyzed via immunoprecipitation-western blotting with anti-FLAG antibody. Endogenous proteins were detected using specific antibodies. Western blotting was performed on two biological replicates (d, f). See also Supplementary Fig. 6. Source data are provided as a Source Data file.