ATP-dependent chromatin remodeling complexes, including BAF (BRG1 associated factor) and PBAF (Polybromo-associated BRG1 associated factor), play major roles in development and disease. These complexes contain multiple subunits and utilize ATP to remodel nucleosomes and, as a result, affect downstream processes. Subunits of the BAF and PBAF complexes are frequently mutated in various cancers, including in acute myeloid leukemia. However, the role of distinct members of these complexes in leukemogenesis is poorly understood. BAF and PBAF complexes are distinguished from each other by unique AT-rich interacting domain (ARID) subunits, ARID1A/B and ARID2, respectively. Recently, Arid2 was suggested to have tumor suppressive functions in MLL-AF9 leukemogenesis, however the role of Arid1b is unknown. In this study, we further evaluated the tumor suppressor roles of Arid2 and Arid1b in MLL-AF9 leukemogenesis.

To study the effect of the loss of Arid2 and Arid1b on MLL-AF9 leukemia in vitro, hematopoietic stem and progenitor cells from Arid1bfl/fl, Arid2fl/fl and wild type (control) mice were transformed by retroviral expression of MLL-AF9. Arid1b or Arid2 was then deleted by retroviral delivery of Cre-recombinase and cellular phenotypes were characterized by flow cytometry. In vitro characterization revealed Arid1b-/-and Arid2-/-MLL-AF9 cells have similar growth rates to control cells and no changes in cell cycle status and apoptosis.

To study leukemogenesis in vivo, two hematopoietic specific conditional knockout mouse models, Vav1Cre, a constitutively active Cre-recombinase, and Mx1Cre, an inducible Cre-recombinase, were used. Primary and secondary transplant recipient mice were assessed for survival, disease burden, and leukemia stem cell frequency. In primary transplantation, leukemic initiation was not altered by the loss of the Arid2. Overall survival of primary recipients and disease burden in moribund mice (percentage of GFP positive MLL-AF9 cells in bone marrow/spleen and spleen weight) were similar to control counterparts. However, in limiting dilution secondary transplantation of Arid2fl/flVav1Cre (Arid2-/-) and wild type Vav1Cre (control) MLL-AF9, median survival of the Arid2-/- cohort decreased 1 to 2 weeks in comparison to the control cohort and there was a two-fold increase in leukemia initiating cell frequency upon deletion of Arid2. Together these results support the previous indication that Arid2 is a tumor suppressor in MLL-AF9 leukemic initiation.

To gain insight into the function of Arid1b in MLL-AF9 leukemogenesis, we performed primary transplantation of Arid1bfl/flcells transduced with MLL-AF9. Our data showed that loss of Arid1b enhanced leukemic initiation in the MLL-AF9 leukemia model. One-month post induction with pIpC, Arid1bfl/fl Mx1Cre MLL-AF9 cells accumulated faster in the peripheral blood and primary recipients exhibited decreased median survival compared with wild type Mx1Cre controls. Percentage of GFP+ leukemia cells was higher in the spleen of moribund mice in the Arid1bfl/fl Mx1Cre cohort, compared with the control cohort. Taken together, our data suggests Arid1b, in addition to Arid2, acts as a potential novel tumor suppressor in MLL-AF9 leukemogenesis.

To study the mechanism of Arid1b and Arid2's function in MLL-AF9 leukemogenesis, RNA-sequencing of leukemic bone marrow cells was conducted. Consistent with the previous study, Arid2 null leukemia cells had increased Gata2 expression and decreased p57 expression. Gene Set Enrichment Analysis of Arid1b null cells showed enrichment of oncogenic signatures including ATM, Cyclin D and macrophage development as potentially contributing to the underlying tumor suppressive mechanism of Arid1b.

In summary, our study uncovered Arid1b, in addition to Arid2, as a potential novel tumor suppressor in MLL-AF9 leukemia. Our study, together with published studies on the requirement of other subunits (SMARCA4, SMARCD2 and DPF2) of the BAF and PBAF complexes in MLL-AF9 leukemia, reveal opposing roles of the components of the BAF and PBAF complexes in leukemogenesis, the mechanism of which require further investigation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.