The Role of H2afy in Normal and Malignant Hematopoiesis

Myelodysplastic syndromes (MDS) are the most common myeloid cell malignancy in adults in the US, characterized by ineffective hematopoiesis, cytopenias, dysplasia, and an increased risk of developing acute myeloid leukemia. Recent sequencing studies have identified spliceosome gene mutations in ~50% of MDS patients. Among the commonly mutated spliceosome genes, we and others have demonstrated that mutations in a core splicing factor gene U2AF1 (found in 11% of MDS patients) alter RNA splicing and hematopoiesis in mice. Whether splicing alterations induced by mutant U2AF1 have functional consequences on hematopoiesis is unclear. We hypothesize that alternative splicing induced by mutant U2AF1(S34F) contributes to hematopoietic phenotypes in U2AF1(S34F) mice and potentially to MDS pathogenesis in patients. To address this hypothesis, we sought to identify a functional downstream target of mutant U2AF1(S34F). We previously identified 13 alternatively spliced genes shared across three RNA-seq datasets (human acute myeloid leukemia and mouse and human hematopoietic cells) by comparing the transcriptome of mutant U2AF1-expressing cells with that of wild-type U2AF1 cells. Validation of these alternatively spliced targets using an orthogonal Nanostring assay revealed that the most robust splicing change induced by U2AF1(S34F) expression occurred in a histone H2A variant gene, H2AFY. H2AFY (also known as macroH2A1) encodes two splice isoforms, H2AFY1.1 and 1.2. Expression of U2AF1(S34F) reduces expression of the H2AFY1.1 isoform that has been implicated in transcriptional gene regulation and DNA damage response. H2AFY is also commonly deleted in MDS patients that harbor an interstitial deletion of chromosome 5q, occurring in ~20% of MDS patients. Collectively, the data led us to hypothesize that alternative splicing of H2afy induced by U2AF1(S34F) expression contributes to hematopoietic phenotypes observed in mutant U2AF1(S34F) expressing mice. We first characterized H2afy knock-out mice that lack expression of both H2afy1.1 and 1.2, and then tested whether re-expression of H2AFY1.1 can revert the hematopoietic phenotypes in U2AF1(S34F) or H2afy knock-out mice. H2afy loss in hematopoietic cells led to reduced B-cells in peripheral blood, bone marrow, and spleen in a gene dose-dependent manner. Colony forming units of pre-B-cells (CFU pre-B) also showed reductions in H2afy+/- and H2afy-/- pre-B colonies ex vivo. Transplanted H2afy knock-out bone marrow stem cells had a competitive disadvantage compared to wild-type bone marrow cells largely due to reduced B-cells. Immunophenotypic analysis of B-cell development revealed reductions in H2afy-/- pro-B-cells through mature B-cells in the bone marrow of mice, while all lymphoid progenitors (ALPs) or B-cell-biased lymphoid progenitors (BLPs) showed no differences. RNA-seq analysis of BLPs identified reduced expression of early B-cell factor 1 (Ebf1), a master transcription factor that is necessary for early B-cell development, in H2afy-/- cells compared to wild-type cells. As U2AF1(S34F) mice also develop similar B-cell defects, we further characterized B-cell development in U2AF1(S34F) mice. U2AF1(S34F) mice also had reductions in pro-B-cells through mature B-cells, CFU pre-B counts, and Ebf1 expression, similar to H2afy-/- mice. Finally, flow cytometric analysis of human MDS bone marrow samples revealed reductions in pro-B-cells and EBF1 expression in U2AF1-mutated MDS samples compared to healthy donor bone marrow samples. Collectively, the data indicate that alternative splicing of H2afy may contribute to the B-cell defects in U2AF1(S34F) mice. Next, we investigated the effects of H2AFY1.1 re-expression on H2afy-/- or U2AF1(S34F) hematopoietic cells. Lentiviral expression of H2AFY1.1 in H2afy-/- bone marrow cells led to increased frequencies of functional CLPs and peripheral and bone marrow B-cells in vivo. Similarly, expression of H2AFY1.1 in U2AF1(S34F) bone marrow cells led to increased CFU pre-B counts, peripheral blood B-cells and Ebf1+ B-cells in vivo. This suggests that H2AFY1.1 is important for normal B-cell development, and its reduction induced by U2AF1(S34F) expression contributes to B-cell deficiency in U2AF1(S34F) mice. Finally, we asked whether H2AFY haploinsufficiency can contribute to MDS pathogenesis. Consistent with a heterozygous deletion of H2AFY in del(5q)-associated MDS, we determined that reduced mRNA expression of H2AFY occurs in del(5q)-associated MDS samples when compared to non-del(5q) MDS or healthy donor controls. H2afy+/- and H2afy-/- mice had reduced overall survival and higher incidence of hematopoietic malignancies relative to wild-type littermates. The malignant phenotypes observed in H2afy+/- and H2afy-/- mice included MDS, erythroid leukemia, and T-cell leukemia. A subset of these tumors were lethal in secondary recipients. Collectively, our data demonstrate that H2afy, in particular H2afy1.1, is important for B-lymphopoiesis, and haploinsufficiency of H2afy promotes hematologic malignancies in mice