The Oncogenic Role and Potential Treatment of a Novel CPSF6-Rarg Fusion in Myeloid Transformation

Acute promyelocytic leukemia (APL) is characterized by PML-RARA, a translocation-driven oncoprotein. While recently, several other types of X-RARA/RARB/RARG fusions have been described in acute myeloid leukemia (AML) resembling APL. The novel recurrent CPSF6-RARG fusion was identified in our previous study, and the patients showed insensitive to ATRA and arsenic therapy. Here we investigated the oncogenic role of CPSF6-RARG fusion in AML transformation and explore the potential treatment.

Two respectively fusions CPSF6-RARG-L (CRL) and CPSF6-RARG-S (CRS) were used in this study. We first assessed the effect of CRL or CRS on proliferation, differentiation of human cord blood hematopoietic stem/ progenitor cells (HSPCs). Expression of CRL or CRS enhanced proliferation, blocked myeloid differentiation in vitro and in the xenograft model. Moreover, the transcriptome of CRL or CRS showed an enrichment of AML gene sets. These data indicated that CRL or CRS could transform myeloid cells into a pre-leukemic state.

We further confirmed the CRL or CRS function by generating a myeloid conditional knock in mice. The mice showed a myeloid proliferative phenotype without leukemia onset. The competition assay and serial transplantation assay indicated the consistent self-renewal capacity with wild type (WT) mice. The hyperplastic neutrophils suggested an enrichment of APL gene sets by single cell RNA sequencing.

Since RAS gene mutation was detected in the AML Patient with CPSF6-RARG fusion, we then transduced NRAS G12D mutation into c-kit⁺ cells from CRL or CRS knock in mice. All the mice transplanted with the cells developed myeloid leukemia. Although WT cells expressed NRAS G12D also developed myeloid leukemia in the transplantation model, expression of CRL or CRS accelerated leukemia genesis by almost 10 days.

To elucidate the underlying molecular mechanisms of CPSF6-RARG fusion, we performed chromatin immunoprecipitation sequencing using kasumi-1 cells expressing CRL or CRS. CRL or CRS protein showed more bound regions than RARG group. In combination with this ChIP-seq and RNA-seq data using CRL or CRS expressing CB HSPCs, we identified CRL and CRS common gene sets with 177 activated and 212 repressed targets. Gene set enrichment analysis revealed that repressed targets were more likely downregulated in CPSF6-RARG patients, while activated genes tended to be enriched among genes highly expressed in CPSF6-RARG patients. We then validated that MYC was the active target and PU.1 was the repressed target of the fusion by ChIP-PCR, luciferase reporter assay and RT-qPCR.

Interestingly, we observed HDAC3 also bound to DNA regions of the fusion targets. HDAC3 inhibitors could partly release the transcriptional repression of PU.1 and abolish the transcriptional activation of MYC. In the CPSF6-RARG and NRAS G12D leukemic model, chidamide (HDAC Ⅰ inhibitor) or RGFP966 (HDAC3 inhibitor) treatment prolonged the mice survival by eliminating the leukemic burden, and the residual leukemic cells were differentiated to GR1⁺MAC1⁺ cells.

Taken together, our study demonstrated that CPSF6-RARG fusion promoted myeloid proliferative through unique transcriptional gene targets including MYC and PU.1, and the fusion accelerated NRAS G12D driven myeloid leukemogenesis. HDAC inhibitors could eliminate leukemic cells by releasing the myeloid differentiation arrest.

No relevant conflicts of interest to declare.