AFFILIATIONS
1 Hannover Medical School, Hannover, Germany
2 German Cancer Research Center (DKFZ), Heidelberg, Germany
3 Oregon Health & Science University, Portland, United States
4 Hannover Medical School, Hannover, Germany
5 Hannover Medical School, Hannover, Germany
6 Hannover Medical School, Hannover, Germany
7 Hannover Medical School, Hannover, Germany
8 Oregon Health & Science University, Portland, United States
9 Hannover Medical School, Hannover, Germany
10 Hannover Medical School, Hannover, Germany
11 Hannover Medical School, Hannover, Germany
12 Hannover Medical School, Hannover, Germany
13 Oregon Health & Science University, Portland, United States
14 University Medicine Greifswald, Greifswald, Germany
15 University Hospital Halle, Halle, Germany
16 Charité -Universitätsmedizin Berlin, Berlin, Germany
17 Charité -Universitätsmedizin Berlin, Berlin, Germany
18 Charité -Universitätsmedizin Berlin, Berlin, Germany
19 Charité -Universitätsmedizin Berlin, Berlin, Germany
20 Charité -Universitätsmedizin Berlin, Berlin, Germany
21 Broad Institute, Cambridge, United States
22 Oregon Health & Science University, Portland, United States
23 Hannover Medical School, Hannover, Germany
24 German Cancer Research Center (DKFZ), Heidelberg, Germany
25 Hannover Medical School, Hannover, Germany
26 Hannover Medical School, Hannover, Germany
27 Oregon Health & Science University, Portland, United States
28 Hannover Medical School, Hannover, Germany
ABSTRACT
Y-box binding protein 1 (YBX1) is a cold-shock domain protein and a m⁵C reader that stabilizes oncogenic transcripts and represents a genetically validated yet clinically intractable vulnerability in acute myeloid leukemia (AML). SU056, an azopodophyllotoxin that binds YBX1 RNA-binding domain and induces its proteasomal degradation, has recently enabled pharmacologic YBX1 targeting, but its activity in hematologic malignancies and the mechanistic basis of YBX1-dependent epitranscriptomic regulation remain undefined. Here, we delineate the therapeutic and mechanistic implications of YBX1 inhibition in AML.
Combined transcriptome and proteome profiling of MOLM-13 cells treated with SU056 (500nM, 96h) revealed 2647 differentially abundant proteins, of which 195 were altered at the protein but not the transcript level. These included oncogenic effectors governing signaling (FLT3, STAT5B, MAPK), transcription (KMT2A, DOT1L, HOXA10, PBX3, MEF2C), DNA-damage response (PARP1/2), cell cycle (CDK6), and proteostasis (MDM2, PSMB2). A YBX1-focused CRISPR base-editor screen (1266 guide RNAs) identified loss-of-function residues Tyr72, Ile75, and Phe85 within the cold-shock domain of YBX1. Phe85, the top-scoring residue (two independent guides, β-score < −1), corresponds to the reported m⁵C-reader site of YBX1 being essential for RNA binding. Its editing (3 independent sgRNAs) impaired proliferative capacity of AML cells (growth reduction to sgLuc control: 65-74%, p<0.0001) relative to sgRNAs editing surrounding residues (such as Asn78-Asp79, Asp83, His87-Gln88; growth reduction to sgLuc control: 0-4%, not significant).
Although YBX1 does not directly recognize m6A-marked mRNAs, it has been proposed to associate with m6A-modified transcripts through members of the IGF2BP family. To investigate if inhibition of a bona fide m⁵C reader influences m⁶A-marked mRNAs, global m6A abundance was assessed in leukemic cells following inactivation of YBX1 by SU056.
SU056 significantly reduced global m⁶A deposition (colorimetric m6A RNA methylation quantification assay, p=0.0047; confirmed by m6A-specific antibody dot blotting) and produced dose-dependent loss of METTL-complex members RBM15, METTL3, and METTL14 as seen by western blot. Genetic and pharmacologic METTL3 inactivation phenocopied the SU056 phenotype, reducing MDM2, MEF2C, c-MYC, and FLT3 protein abundance (MDM2, p=0.0338; MEF2C, p=0.0485; c-MYC, p=0.0012; FLT3, p=0.0401). Oxford Nanopore sequencing of METTL3-inhibited AML cells (STM2457, 30 μM, 24h) confirmed loss of m⁶A modifications on these transcripts. METTL3 overexpression rescued MDM2 and c-MYC expression in SU056-treated AML cells, whereas genetic or pharmacologic METTL3 inhibition synergized with SU056 (250 nM, 9d) to suppress proliferation, suggesting that YBX1 sustains oncogenic translation by stabilizing METTL-complex transcripts and thereby reinforcing m⁶A deposition.
To define functional dependencies of SU056 activity, a genome-wide CRISPR-knockout screen was performed in MOLM-13 treated with 500 nM SU056 versus vehicle for 21 days. Genes whose depletion sensitized cells to SU056 included FLT3, CDK6, PARP1/2, BCL2, and JMJD1C, concordant with the proteomic data. Here, the MDM2/TP53 axis emerged as the principal determinant of response, scoring as both resistance mediators (TP53, ATM, TP73) and sensitizers (MDM2, CDK2).
In humanized AML patient-derived xenograft models, treatment with SU056 (20 mg/kg) combined with Navtemadlin (100 mg/kg) sustainably reduced circulating blasts over 16 weeks, significantly prolonged survival (SU056 97.5 days, Navtemadlin 112 days; combination not reached; p<0.0001), and reduced disease penetrance by 85.8%. Human CD45⁺ blast infiltration in spleen and bone marrow decreased from a median of 98.4% (SU056) and 82.9% (Navtemadlin) to 14.1% with the combination. Limiting-dilution serial transplantation (2×10⁶, 2×10⁵, 2×10⁴ cells) revealed a profound, several-fold reduction in leukemia-initiating cell frequency (Diluent 1/1, SU056 1/1444931, Nav 1/1286296, combination Inf) and prolonged secondary-recipient survival (median not reached) following combination therapy.
Taken together, pharmacologic YBX1 inhibition disrupts a previously unrecognized m⁵C-m⁶A epitranscriptomic circuit that sustains oncogenic translation in AML and synergizes with MDM2 blockade to eradicate leukemia-initiating cells, providing a mechanistic and translational rationale for combined YBX1-MDM2 targeting in refractory AML.
DISCLOSURES
Tina M. Schnoeder: No relevant conflicts of interest to declare.