ALBERT

Description: Rearrangement of the immunoglobulin heavy chain locus (IgH) through non-homologous end-joining (NHEJ)-mediated VDJ recombination is a requisite step in normal lymphocyte development. This process is initiated through cleavage of DNA by RAG1/2 recombinases and repair by classical NHEJ pathway factors. Loss or defects of these factors can impair end-processing (DNA-PKcs, ARTEMIS) or end-ligation (KU70/80, XRCC4, XLF, LIGIV) of rearrangement and are drivers of disease, notably, severe combined immunodeficiency (SCID). In addition to these core NHEJ factors, other proteins, such as ATM kinase, have been shown to be critical for the stabilization of the post-cleavage VDJ intermediates to ensure proper signal joint formation. While epigenetic regulation of this process has been described through the maintenance of accessibility for transcription factors and RAG1/2, very little is known about the role of chromatin modifications on NHEJ-mediated VDJ recombination. Histone H3 lysine-36 tri-methylation (H3K36me3) is a histone modification associated with actively transcribed genes and has been shown to be critical for many biological processes, including transcription, mismatch repair and homologous recombination. Tri-methylation is catalyzed by the non-redundant histone methyltransferase, SETD2. Loss of function mutations in SETD2 or dominant negative onco-histone mutations in the H3K36 residue have been described in a broad array of tumors, including several hematopoietic malignancies. To study the role of Setd2 in hematopoiesis, we generated conditional knockout mice, where loss of Setd2 results in the ablation of the H3K36me3 mark. While heterozygous mice had no overt phenotype, mice with homozygous loss of Setd2(Setd2KO) displayed aberrant hematopoiesis, with a significant depletion of lymphoid primed progenitor population and B/T-cells in hematopoietic tissues. Setd2 loss early in hematopoiesis lead to a profound block of B-cell development at the proB cell stage and an early block in T-cell development, in a compartment where TCRβ rearrangements initiate. In Setd2KO mice, V to DJ rearrangements of the IgH locus were impaired, suggesting that the block at the proB stage of development was due to defective VDJ recombination. To assess this, we crossed our knockout mice with an IgHelMD4 mouse strain, and found that the block at the proB stage could be bypassed by the expression of a transgenic, fully rearranged IgH locus. We confirmed global loss of H3K36me3 in proB cells from our Setd2KO mice by ChIP-Seq and found that the ablation of this mark did not affect germline IgH locus accessibility (by ATAC-seq) or local chromatin architecture (H3K4me3 or H3K9ac) at a well-studied critical regulatory region near the Eμ enhancer. Loss of Setd2 and H3K36me3 only mildly effected sterile transcription of IgH genes and did not effect the overall cellular proliferative rate or the recruitment of NHEJ factors. Furthermore, loss of Setd2 did not cooperate to generate B-cell lymphomas in a p53-/- background. However, sequencing of VDJ recombination products from knockout mice indicated not only decreased clonality of products and variable gene usage but also resulted in alterations of complementarity defining region-3 (CDR3) hypervariable sequences. These results phenocopy defects or loss of NHEJ end-ligation factors and suggest that Setd2 and H3K36me3 could play a pivotal role in the end-ligation steps of NHEJ. Defects in end-ligation and resolution of VDJ intermediates have previously been shown to result in the formation of aberrant coding-signal hybrid joints. To see if this was occurring with loss of Setd2, we generated knockout murine proB v-Abl transformed lines where we could induce κ light chain rearrangement upon addition of an Abl kinase inhibitor. Remarkably, upon induction in Setd2KO v-Abl lines, we detected the formation of hybrid joints by both PCR and by the loss of a GFP-based coding-joint formation signal from a reporter expressed in these cells. We additionally could detect hybrid joint formation in the endogenous κ light chain locus of Setd2KO splenocytes, further indicating a role for Setd2/H3K36me3 in the end-ligation process. These studies demonstrate a novel and critical role for SETD2 and H3K36me3 in preventing aberrant, non-functional VDJ end joining of NHEJ-induced DNA breaks during lymphocyte development. Disclosures No relevant conflicts of interest to declare.

Description: The post-translational modification consisting of trimethylated Histone H3 lysine-36 (H3K36me3) is a chromatin mark associated with actively transcribed genes and implicated in many biological processes, including transcription initiation and elongation, splicing, and DNA damage repair response via homologous recombination. H3K36me3 is catalyzed by the non-redundant histone methyltransferase, SETD2. Loss of function mutations in SETD2 or dominant negative "onco-histone" mutations in the H3K36 residue have been described in a broad array of solid tumors. More recently, SETD2 has been found to be commonly mutated in acute and chronic leukemias (B-ALL, ETP-ALL, CLL, AML), indicating its potential role as a tumor suppressor in hematopoietic malignancies. In order to study the role of Setd2 in normal and malignant hematopoiesis, we generated Vav1-cre and Mx1-cre Setd2 conditional knockout mice. Expression of Cre-recombinase resulted in loss of Setd2 and ablation of H3K36me3. While heterozygous mice had no overt phenotype, homozygous loss of Setd2(Setd2KO) in early hematopoietic development resulted in pancytopenia, splenomegaly, and overall bone marrow hypocellularity. Significant reduction in the total number of hematopoietic stem cells (HSCs) was also observed. Hematopoietic reconstitution assays in competitive and non-competitive transplant settings additionally revealed a qualitative defect in Setd2KO HSCs. Within the LSK-defined HSC fraction (Lin-Kit+Sca1+), a significant depletion of the lymphoid primed progenitor population (MPP4) was observed while the erythroid primed progenitor (MPP2) population was significantly expanded. Interestingly, while myelopoiesis was largely unaffected (in number of progenitors and hematopoietic output), B-cells (B220+) and T-cells (CD4+CD8+) were significantly depleted in Setd2KO mice (8 and 10-fold respectively). In both the bone marrow and spleen, a significant expansion of early erythroid progenitors was observed. Gene set enrichment analysis of RNA-Sequencing of the LSK HSC fraction of Setd2KO mice revealed strong correlations to signatures of HSC differentiation, particularly significant enrichments in erythroid priming signatures, (eg. Gata1) and signatures related to proliferation and loss of stem cell quiescence. Altogether, these data indicate that Setd2is important in regulating normal HSC self-renewal and lineage commitment, with differential effects in each lineage. We wanted to further explore the defect in lymphopoiesis and observed a profound block of B-cell development at the pro-B cell stage (B220+CD43+IgM-), with a 25-fold depletion of pre-B cells (B220+CD43-IgM-) and an almost complete ablation of immature IgM+ B-cells in the bone marrow of our Setd2KO mice. Interestingly, Mb1-cre and Cd19-cre Setd2KO mice had no arrest in early B cell development, indicating that loss of Setd2 prior to pro-B cell stage was necessary for the block in differentiation. In Setd2KO mice, V to DJ rearrangements of immunoglobulin heavy-chain locus (IgH) were not detectable by a non-quantitative PCR method, suggesting that the block at the pro-B stage of development in our Setd2KO mice was due to the impairment of complete VDJ recombination at this locus. To prove this, we crossed our knockout mice with an IgHelMD4+ mouse strain and found that the block at the pro-B stage could be bypassed by the expression of a transgenic, fully rearranged IgH locus in Setd2KO mice, indicating the requirement of Setd2 and H3K36me3 in VDJ recombination. Additionally, we observed that loss of Setd2 later in B cell development impairs class switch recombination by LPS and IL-4 co-stimulation of Setd2KO Cd43- splenocytes of Mb1cre Setd2KO mice. This defect in VDJ recombination as a result of loss of H3K36me3 is associated with impaired recruitment of NHEJ machinery to the IgH locus. Similarly, Setd2KO mice exhibited significantly reduced thymic size and a comparable block in T-cell development at the DN3 stage, where TCRb rearrangements initiate. Altogether our studies demonstrate not only a critical role for Setd2 in the HSC compartment in determining hematopoietic differentiation, but also the requirement of Setd2 and H3K36me3 in lymphopoiesis, specifically for fully competent NHEJ-driven VDJ recombination and class switch recombination. Disclosures No relevant conflicts of interest to declare.

Description: Childhood B-cell acute lymphoblastic leukemia (B-ALL) that harbor a translocation of the MLL1 and AF4 genes are considered high-risk with poor prognosis (event-free survival (EFS) of 35%-50%), especially when compared to non-MLL-rearranged (MLL-R) childhood ALL (EFS 〉85%). An important obstacle to developing new therapeutic approaches for this patient population is the lack of models that faithfully recapitulate the short latency and aggressiveness of this disease. Recently, whole genome sequencing of patient childhood MLL-R leukemias revealed that activating mutations of the proto-oncogenes involved in signaling, most prominently, N or K-RAS were found in nearly 50% of patients. Patients with these co-occurring mutations have an even poorer overall survival rate, indicating that a model harboring both mutations is of extreme interest. Here, we report the generation of a highly aggressive, serially transplantable B-ALL by the retroviral overexpression of activating N-RasG12D mutant in bone marrow of an inducible knock-in Mll-Af4 murine model that we have previously published. Recipient mice injected with Mll-Af4/N-RasG12Dpre-leukemic bone marrow cells developed an acute B-ALL (B220+CD43+IgM-) with short latency to development of disease (median 35 days). Furthermore, the resultant primary B-ALL was serially transplantable into sub-lethally irradiated recipients with accelerated latency to secondary and tertiary disease developing at a median of 20 and 12 days, respectively. As our model includes an activating mutation in N-Ras, we wanted to see if the cells would be sensitive to small molecule inhibitors of downstream effectors of Ras. Pre-leukemic Mll-Af4/N-RasG12D cells were sensitive to two different MEK inhibitors, Trametinib or PD901, in vitro. Furthermore, in vivo treatment of tertiary B-ALL mice with Trametinib showed significant reduction in leukemia burden after 7 days of treatment, as well as increase in survival, compared to vehicle controls. However, prolonged in vivo treatment with Trametinib eventually led to loss of sensitivity and development of B-ALL in our mouse model, suggesting that Trametinib alone is insufficient to prevent leukemia progression. As single agent MEK inhibition was insufficient to generate long-term durable responses, we conducted RNA-Sequencing of primary Mll-Af4/N-RasG12D leukemias to discover pathways amenable for therapeutic intervention. Gene set enrichment analysis suggested that targeting the DNA damage response (DDR) pathway as an attractive therapeutic opportunity. We were able to demonstrate an increased basal level of replicative stress in our Mll-Af4/NrasG12D pre-leukemic cells and sensitivity to small molecule inhibition of ATR, a master regulator of the G2 to M transition of cell cycle progression, with AZ20, a selective ATR inhibitor. In vitro and in vivo treatment with AZ20 led to increased leukemia cytotoxicity. However, similar to Trametinib treatment, tertiary B-ALL mice eventually succumbed to disease with prolonged AZ20 treatment in vivo. Since neither single agent MEK nor ATR inhibition could prevent leukemic progression in vivo, we tested the combination and found increased cytotoxicity and cell cycle arrest in vitro at concentrations well below the IC50, as compared to single agent treatment. In vivocombination treatment also demonstrated decreased leukemia burden and significant prolonged survival compared to either AZ20 or Trametinib alone. Lastly, we tested out the efficacy of combination therapy in human B-ALL patient derived xenograft harboring both MLL-AF4 and activating N-RASmutations. 250,000 human primary leukemic blasts were transplanted into non-irradiated immune-compromised mice and treated with vehicle, single agent, or the combination for 14 days. Similar to the results seen in our mouse model, combination treatment with Trametinib and AZ20 led to significant reductions in leukemic burden. In summary, our model of B-ALL faithfully recapitulates the short latency and aggressiveness of this disease and was predictive of response in human patient samples harboring MLL-AF4 and activating N-RAS mutations to small molecule inhibitor therapy to MEK and DDR pathways. In the future, this model can be used as a platform to not only better understand the molecular events governing and sustaining leukemogenesis, but also as a discovery platform for novel therapeutic combinations. Disclosures Armstrong: Epizyme, Inc: Consultancy; Vitae Pharmaceuticals: Consultancy; Imago Biosciences: Consultancy; Janssen Pharmaceutical: Consultancy.