Abstract
Hematopoietic stem cells (HSCs) and multipotent progenitors (MPPs) reside in multipotent states and are capable of giving rise to myeloid, lymphoid, megakaryocytic, and erythroid lineages. While gene regulatory programs that guide multilineage hematopoietic cell fate decisions have been heavily scrutinized, we have limited insights into how specific epigenetic regulators are deployed to maintain multipotency and instruct HSC/MPP fate decisions. Mll3 and Mll4 are two highly homologous genes encoding epigenetic regulators that arbitrate hematopoietic cell fate decisions. Both MLL3 and MLL4 nucleate large chromatin-bound complexes that belong to the Complex of Proteins Associated with SET1 (COMPASS) family. The proteins share highly conserved N-terminal PHD and C-terminal SET domains, and each interacts with a common set of binding partners, including UTX, PTIP, PA1, NCOA6, ASH2L, DPY30, RBBP5 and WDR5. MLL3/4 COMPASS complexes are generally thought to bind enhancer elements, though promoter interactions have also been demonstrated. MLL3 and MLL4 monomethylate histone H3, lysine 4 (H3K4) via their SET domains, and UTX removes repressive H3K27 trimethylation marks via its JmjC domain.
Despite biochemical and structural similarities between MLL3 and MLL4 proteins, mouse genetic studies have identified distinct, antagonistic functions for Mll3 and Mll4 in HSCs. Mll3 promotes HSC differentiation into MPPs, and it suppresses HSC mobilization and myeloid progenitor expansion in response to granulocyte-colony stimulating factor. In contrast, Mll4 sustains HSC self-renewal by opposing myeloid differentiation. Deficiencies in COMPASS co-factor genes, such as Utx and Ptip, do not fully phenocopy effects of Mll3 or Mll4 loss in the hematopoietic system, suggesting that MLL3 and MLL4 proteins can act independently of COMPASS co-factors. These observations raise questions as to how Mll3 and Mll4 interact in HSCs, MPPs and committed progenitor populations to adjudicate competing hematopoietic cell fate decisions.
In this study, we conditionally deleted Mll3, Mll4 or both genes together to elucidate the role of each gene in specifying differentiated cell fates. We found that interactions between Mll3 and Mll4 were far more complex than a simple antagonism model would suggest, as the genes had unanticipated redundant roles in maintaining HSC/MPP identity, myeloerythroid potential, and B-cell maturation. In the absence of both Mll3 and Mll4, hematopoietic progenitors arrested at a pre-B-cell-like stage without self-renewal or mature B-cell potential. While MLL3 and MLL4 were dispensable for B-cell priming, they were required for B-cell maturation. In contrast, MLL3 and MLL4 became dispensable for late stages of myeloid differentiation once myeloid identities were established. MLL3 and MLL4 did not require catalytically active SET domains to license HSC/MPP identity or myeloerythroid potential, nor did they require the COMPASS catalytic co-factor UTX.
At a molecular level, loss of either Mll4 alone or Mll3/4 together led to widespread but distinct changes in enhancer and superenhancer activity within HSC/MPP cells. Mll4 deficiency caused precocious activation of myeloid enhancers whereas Mll3/4 deficiency caused ectopic activation of B-lymphoid enhancers. Mll3/4 deficiency led to near total loss of H3K27 acetylation (H3K27ac) at many documented HSC/MPP superenhancers while promoter elements for B-cell master regulator genes, such Pax5 and Ebf1, acquired ectopic H3K27ac marks. Several MLL4 target genes, such as Gata2 and Gpx4, exhibited further reduced expression in the absence of both MLL3 and MLL4 compared to MLL4 loss alone, consistent with a redundant mode of regulation.Altogether, our data reveal a previously unappreciated but critical role for MLL3 and MLL4 in maintaining HSC/MPP multipotency and licensing myeloerythroid potential. The data show that MLL3 and MLL4 holds myeloid and lymphoid programs in tension against one another within HSC/MPPs. Thus, in the absence of both MLL3 and MLL4, HSC/MPPs rapidly convert to a B-cell-like default state defined by ectopic expression of B-cell regulators and collapse of HSC/MPP superenhancer networks. Our data further illustrate how key functions of MLL3 and MLL4 can be decoupled from their SET methyltransferase function and from other COMPASS co-factors, such as UTX.
