A previously unrecognized molecular mechanism in an aggressive blood cancer subtype may open a new class of therapeutic targets — one that challenges a foundational assumption about what RNA-binding proteins actually do inside a cell. The implications extend beyond leukemia into basic cancer biology and transcriptional regulation.
Working from the DepMap functional genomics dataset combined with combinatorial CRISPR screens, the investigators identified MBNL1 — classically understood as an RNA-processing protein — as a critical driver of FLT3 oncogene expression specifically in KMT2A-rearranged (KMT2A-r) leukemias. KMT2A rearrangements affect roughly 10% of all acute leukemias and carry a poor prognosis, partly because FLT3 overexpression fuels uncontrolled proliferation. Crucially, MBNL1 does not act through RNA at all here: its zinc finger domains and an unstructured carboxyl-terminal region physically bind a structured single-stranded DNA (ssDNA) element in the FLT3 enhancer — specifically a five-consecutive-guanine motif. ChIP-seq and KAS-seq data confirmed this protein-ssDNA interaction occurs in primary KMT2A-r tumor cells, and ablating either key MBNL1 amino acids or the guanine-rich ssDNA sequence abolished the interaction and suppressed leukemia cell survival.
This finding is potentially paradigm-shifting for two reasons. First, it expands the functional repertoire of RNA-binding proteins into direct transcriptional regulation via enhancer DNA — a role not previously attributed to MBNL1 and largely unanticipated for this protein class. Second, it identifies a targetable protein-DNA interface that is contextually specific to KMT2A-r leukemia, meaning therapeutic disruption could be selective. Limitations include the predominantly cell-line and ex vivo primary tumor scope; in vivo efficacy data in animal models and eventual human trials are needed before clinical translation. Still, as an mechanistic discovery bridging RNA biology and chromatin regulation, this represents a meaningful conceptual advance rather than incremental refinement.