Abstract
Acute myeloid leukemia (AML) with tandem duplications in the upstream binding transcription factor gene (UBTF-TD) has recently been described as a new AML entity in younger adults. The UBTF gene encodes the UBF protein, which facilitates transcription of 18S, 5.8S and 28S ribosomal RNAs (rRNAs) by forming a pre-initiation complex with RNA polymerase I and SL1, a selectivity factor required for rRNA transcription. In AML patients with UBTF-TD, these alterations frequently co-occur with WT1 mutations, internal tandem duplications of the FLT3 gene, and trisomy 8, but are mutually exclusive with other AML subtype-defining alterations such as NPM1 mutations and recurrent translocations. UBTF-TDs appear to be early clonal events associated with poor response to conventional chemotherapy, high variant allele frequencies, and stability throughout disease progression.
As an emerging AML entity, the molecular mechanisms by which the UBTF-TD promotes leukemogenesis remain incompletely understood. HOXA and HOXB gene overexpression has been identified as a predictive biomarker of sensitivity to BCL-2 inhibition, while UBTF-TD AMLs have recently been shown to respond to menin inhibition. In addition, as the UBTF gene encodes a nucleolar protein and has a well-established key role in rRNA biogenesis and nucleoli formation, RNA polymerase I inhibition might be beneficial in treating AML with UBTF-TD. CX-5461 is an anticancer drug that selectively inhibits RNA polymerase I-mediated transcription by disrupting rRNA synthesis initiation, and has received Fast Track Designation from the FDA due to its promising therapeutic potential.
In this study, we established xenograft models by injecting primary AML patient samples with UBTF-TD or UBTF wild-type into immunodeficient NSG mice. Following engraftment, mice were treated with vehicle, CX-5461, venetoclax plus azacitidine (VEN-AZA), or the combination of CX-5461 and VEN-AZA. The tumor burden was significantly reduced when CX-5461 was added to VEN-AZA. Flow cytometry revealed that CX-5461, alone or in combination with VEN-AZA, induced a near-complete shift toward a primitive CD34+ immunophenotype specifically in UBTF-TD PDX models, suggesting enhanced cellular plasticity. Single-cell sequencing (CITE-seq) of primary AML samples, including pre-treatment and relapse samples with UBTF-TD or UBTF wild-type controls, identified consistent transcriptional and surface marker dysregulation in UBTF-TD samples. This included downregulation of key cell cycle regulators (e.g., E2F targets and G2M checkpoint genes), consistent with reduced proliferation and a quiescent-like phenotype, alongside reduced TNF-α signaling. Analysis of paired diagnosis-relapse samples showed an increased stem-like cell population at relapse, mirroring the treatment-induced shift toward a primitive, stem-like state observed in PDX models. Further analyses to investigate the functional consequences of this pronounced primitive shift upon CX-5461 treatment are currently underway.
In conclusion, our findings demonstrate that VEN-AZA in combination with CX-5461 reduces tumor burden in preclinical models and induces a pronounced shift toward a primitive, quiescent, stem-like phenotype in UBTF-TD AML, reflecting significant cellular plasticity. These data support combined targeting of BCL-2 and RNA polymerase I as a promising therapeutic strategy for AML patients harboring UBTF-TD.
