Chimeric antigen receptor T (CART) cell therapy has demonstrated clinical efficacy in hematologic malignancies; however, primary or secondary treatment failure remains a major obstacle to durable responses. Defining the optimal cellular composition of CART products therefore represents a critical unmet need. Here, we show that CARTs targeting acute leukemias achieve maximal efficacy when composed exclusively of CD4+ T cells. Pure CD4+ CARTs exhibit superior anti-tumor activity and proliferation compared with CD8+-containing CART products. To elucidate the molecular basis of this functional divergence, we applied a combinatorial exploratory approach integrating bulk RNA sequencing and quantitative proteomics. Transcriptomic analyses revealed that pure CD4+ CARTs adopt a highly proliferative state characterized by a cytotoxic effector-like polarization. On the protein level, we are demonstrating coordinated loss of cell-cycle machinery and induction of apoptotic pathways in CD4+ CART cells co-cultured together with CD8+ CART cells, while pure CD4+ CART cultures maintained a proliferative, cytotoxic phenotype. Using mechanistically discriminative co-culture systems, we demonstrate that CD8+ CART-mediated impairment of CD4+ CART functionality is primarily driven by competitive access to shared antigen. Collectively, these findings identify antigen competition between CART subsets as a previously unrecognized mechanism limiting CD4+ CART efficacy and provide a framework for optimizing CART product composition to enhance therapeutic persistence and durability.

ABSTRACT:Relapse after remission remains the primary cause of treatment failure in acute myeloid leukemia (AML), underscoring the need for strategies to eliminate residual leukemic cells. The bone marrow (BM) microenvironment, largely orchestrated by the CXC chemokine receptor 4 (CXCR4)-CXC motif chemokine 12 axis (CXCL12), enables leukemia cell survival and chemoresistance by anchoring blasts in their protective BM niche. Motixafortide, a selective CXCR4 antagonist, mobilizes leukemic cells and disrupts tumor microenvironment interactions in preclinical models. In this randomized, double-blind, placebo-controlled phase 2 trial, 128 patients in first remission received high-dose cytarabine plus motixafortide or placebo. Median relapse-free survival did not substantially differ between groups: 10.3 months (95% confidence interval [CI], 8.0-12.0) for motixafortide and 11.5 months (95% CI, 8.6-24.1) for placebo (log-rank P = .98). But single-cell measurable residual disease (scMRD) analysis, performed before consolidation, demonstrated heterogeneity of CXCR4 inhibition benefit; in the placebo group, higher CXCR4 expression was associated with increased relapse risk (P = .02), whereas in the motixafortide group, higher CXCR4 expression was linked to a reduced relapse rate (P = .047). Exploratory analyses identified scMRD levels at which higher MRD burden was associated with inferior overall survival. Taken together, combining functional MRD profiling with biomarker-driven patient selection, such as CXCR4 expression, may enable more precise and effective postremission interventions in AML. This trial was registered at www.clinicaltrials.gov as NCT02502968 and at EudraCT as 2014-002702-21.

Abstract Background: The CXCL12-CXCR4 axis mediates bone marrow niche mediated protection of leukemic stem cells and contributes to chemoresistance. Motixafortide is a potent CXCR4 antagonist that disrupts leukemic cell retention in the microenvironment. We leveraged single-cell minimal residual disease (scMRD) profiling to quantify CXCR4 expression and clonal persistence after consolidation therapy within the BLAST trial (NCT02502968), aiming to define predictive biomarkers of Motixafortide response. Methods: Among 128 randomized AML patients in first complete remission (CR), 56 underwent bone marrow scMRD analysis using the Tapestri platform (Mission Bio) targeting 40 recurrent mutation hotspots and 19 surface markers including CXCR4. CXCR4 expression was quantified at single-cell resolution and dichotomized by cohort median. Clinical outcomes were correlated with treatment and scMRD metrics. Results: Although relapse-free survival (RFS) was not different between treatment arms in the ITT population (median RFS: 10.3 vs. 11.5 months, p=0.98), scMRD profiling uncovered a CXCR4 dependent treatment effect. In the placebo arm, high CXCR4 expression correlated with increased relapse risk (p=0.02). In contrast, in the Motixafortide arm, high CXCR4 expression was associated with reduced relapse (p=0.047). A multivariable Cox model confirmed a significant treatment by CXCR4 interaction (HRinteraction 0.032, 95% CI 0.004 to 0.269, p=0.0015), supporting CXCR4 as a predictive biomarker for Motixafortide benefit. scMRD revealed persistent clonal diversity, but conventional MRD positivity did not associate with outcome. Conclusion: While the BLAST trial showed no overall difference in RFS, single-cell MRD analysis revealed differential treatment effects according to CXCR4 expression. These findings suggest that CXCR4 may serve as a biomarker for response to Motixafortide and highlight the potential of single-cell profiling to inform patient stratification. Citation Format: Enise Ceran, Sonia Jaramillo-Segura, Anne Kahtrin Merbach, Christian Rohde, Robert Durruthy-Durruthy, Adam Sciambi, Marc Arribas-Layton, Michelle Lotze, Maxi Wass, Kathrin Rieger, Mathias Hänel, Regina Herbst, Christoph Röllig, Edgar Jost, Richard Schlenk, Katharina Götze, Marina Scheller, Marion Subklewe, Simone Kowoll, Jörg Steighardt, Bayram Edemir, Lutz P. Müller, Irit Glicko-Kabir, Abi Vainstein-Haras, Cora Gromann, Ella Sorani, Shaul Kadosh, Andreas Wienke, Claudia Baldus, Uwe Platzbecker, Hubert Serve, Martin Bornhäuser, Carsten Müller-Tidow. Single-cell MRD profiling identifies CXCR4-high AML as a responder population to CXCR4 inhibition with Motixafortide [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2440.

Microglia are parenchymal brain macrophages that are established during embryogenesis and form a self-containing cellular compartment that resists seeding with cells derived from adult definitive hematopoiesis. We report that monocyte-derived macrophages (MoMΦs) accumulate in the brain of aging mice with distinct topologies, including the nigrostriatum and medulla but not the frontal cortex. Parenchymal MoMΦs adopt bona fide microglia morphology and expression profiles. Due to their hematopoietic stem cell (HSC) derivation, monocyte-derived microglia (MoMg) are unlike yolk-sac-derived cells, targets of clonal hematopoiesis (CH). Indeed, using a chimeric transfer model, we show that the hematopoietic expression of DNMT3AR882H, a prominent human CH variant, renders MoMg pathogenic and promotes motor deficits resembling atypical Parkinsonian disorders. Collectively, we establish that MoMg progressively seed the brain of healthy aging mice, accumulate in selected areas, and, when carrying a somatic mutation associated with CH, can cause brain pathology.

Hematopoietic mutations in epigenetic regulators like DNA methyltransferase 3 alpha (DNMT3A), play a pivotal role in driving clonal hematopoiesis of indeterminate potential (CHIP), and are associated with unfavorable outcomes in patients suffering from heart failure (HF). However, the precise interactions between CHIP-mutated cells and other cardiac cell types remain unknown. Here, we identify fibroblasts as potential partners in interactions with CHIP-mutated monocytes. We used combined transcriptomic data derived from peripheral blood mononuclear cells of HF patients, both with and without CHIP, and cardiac tissue. We demonstrate that inactivation of DNMT3A in macrophages intensifies interactions with cardiac fibroblasts and increases cardiac fibrosis. DNMT3A inactivation amplifies the release of heparin-binding epidermal growth factor-like growth factor, thereby facilitating activation of cardiac fibroblasts. These findings identify a potential pathway of DNMT3A CHIP-driver mutations to the initiation and progression of HF and may also provide a compelling basis for the development of innovative anti-fibrotic strategies.

Somatic mosaicism in blood cells, referred to as clonal hematopoiesis (CH), arises when hematopoietic stem cells (HSCs) acquire somatic mutations that provide a substrate for clonal selection and subsequent clonal outgrowth over time. This condition becomes common with aging and is associated with adverse clinical outcomes related to mutated progenitor cells and increased inflammation. It has been shown that mutated clones have growth rates ranging from 5% to more than 50% per year and that they are acquired decades before they reach a substantial clone size. In CH, the most frequently mutated gene is DNMT3A encoding a de novo methyltransferase enzyme and the mutations are predominantly heterozygous, scattered throughout the three functional domains. In contrast, more than half of DNMT3A mutations in acute myeloid leukemia are missense alterations within the catalytic domain of the enzyme at residue R882. The high frequency of mutations at this specifc site suggests a gain-of-function activity. Beyond haploinsufficiency, the common heterozygous R882H allele creates an altered protein with dominant negative activity. Despite the increasing knowledge of how CH develops over time, the question regarding which mechanisms lead to clonal selection of mutant clones remains unresolved. Here, we performed single-cell multi-omics profiling of FACS-enriched CD34+ hematopoietic stem and progenitor cells (HSPCs) and CD34- mature blood cells from n=9 DNMT3Amut peripheral blood samples with G-CSF mobilized HSPCs (autologous stem cell grafts) collected from multiple myeloma patients in remission. In total, 140,000 single cells were sequenced. Simultaneous mutation analysis enabled intra-sample comparison between DNMT3Amut and wild-type cells. Differential gene- and surface protein expression analysis (CITEseq) revealed downregulated MHC-II molecules in DNMT3AR882mut compared to wild-type HSCs. This observation was restricted to DNMT3AR882mut comparted to DNMT3ANon-R882mut samples. To validate this phenotype, we established a FACS-sorting strategy using HLA-DR, an MHC-II cell surface receptor involved in presenting peptides to CD4+ T cells, and compared mutant fractions (VAF) in HLA-DRlow vs. HLA-DRhigh FACS-sorted CD34+ HSPCs by digital-droplet PCR, thereby confirming an up to 2.5-fold higher DNMT3AR882mut fraction in HLA-DRlow sorted HSPCs. We have previously shown that HSPCs are capable of activating CD4+ T cells upon presentation of both, endogenous and exogenous antigens via MHC-II and that the presentation of immunogenic antigens via MHC-II by HSPCs mediates bidirectional interactions with antigen-specific CD4+ T cells. Therefore, we hypothesize that the observed downregulation of MHC-II expression in DNMT3AR882mut HSCs can alter the activation of CD4+ T cells, which can functionally impact the development of CH over time. To explore mechanisms linking DNMT3AR882 mutations to the MHC-II phenotype in a model system, we studied interactions between DNMT3AR882Hmut Lin-Sca1+c-Kit+ cells and CD4+ T cells using an (pI:pC)-inducible humanized mouse model for CH that conditionally expresses human DNMT3A cDNA carrying the R882H hotspot mutation (Mx1-Cre+:DNMT3AWT/R882H). We performed multi-parameter flow cytometry to characterize MHC-II expression on HSPCs and progenitor cell populations. Consistent with our human single-cell data, we observed that MHC-II (I-A/I-E) expression on HSPC subpopulations collected from mice with monoallelic DNMT3AR882H expression was downregulated compared to wild-type mice. In contrast, there was no difference at the progenitor cell level. To understand whether the observed downregulation of MHC-II molecules impacts the antigen-specific activation of CD4+ T cells, we co-cultured FACS-sorted HSPCs with CD4+ T cells from OT-II mice that express transgenic T cell receptors specifically recognizing the chicken ovalbumin (OVA329-339 peptide), when presented via MHC-II. Our preliminary analysis revealed that despite lower MHC-II expression, DNMT3AR882H HSPCs are still capable of activating CD4+ T cells in vitro. Together, our data provide evidence in primary human CH samples and in an engineered mouse model for DNMT3AR882Hmut CH that HSCs carrying DNMT3AR882 hotspot mutations in CH down-regulate MHC-II molecules, which potentially leads to altered immunosurveillance mechanisms in DNMT3AR882 mutant CH.

Measurable residual disease (MRD) assessment is important in acute myeloid leukemia (AML), as post-remission MRD indicates higher relapse risk. Current techniques, however, lack precision in individual outcome prediction, complicated by AML's clonal heterogeneity. Novel single cell MRD (scMRD) workflows based on DNA mutations offer a potential breakthrough, simultaneously capturing genotypic and immunophenotypic changes at a single-cell level. The Tapestri® platform (Mission Bio) incorporates a microfluidic droplet technology with cell barcoding beads and gene-specific primers that enables amplification and comprehensive identification of low-frequency subclones that may lead to overt disease or resistance. With this approach, complex heterogeneity and differentiation between leukemic clones and clonal hematopoiesis can be deconstructed. Objective: This study represents the first large-scale scMRD analysis in AML using this novel scMRD technology to deconvolute clonal heterogeneity to identify potential relapse markers and to differentiate between leukemic clones and clonal hematopoiesis. Methods: We analyzed a cohort of 69 AML patients in first complete remission (CR) as assessed by bone marrow morphology from the BLAST trial (NCT02502968). For 11 patients, we analyzed paired samples from both first CR and relapse. We performed single-cell DNA + protein sequencing on a total of 250,000 cells across 23 multiplexed pools. Our methodology included 40 selected mutation hotspot genes and a 19-plex antibody oligonucleotide cocktail for AML-specific surface markers as well as patient-specific hash tag antibodies for enhanced demultiplexing. We used magnetic activated cell sorting (MACS) to enrich CD34/CD117-positive blasts to ensure high-quality samples for subsequent analysis. Results: Our analyses revealed a diverse genetic landscape with 25 different mutated genes. DNMT3A and NPM1:p.W228Cfs mutations were the most common alterations. The genetic distribution included a predominance of preleukemic genes (51.5%), chromatin remodelers (12.1%), splice factors (9.1%), transcription factors (6.1%), cell cycle regulators (6.1%), and other functional categories (15.2%). Lower MRD (and clonal hematopoiesis) levels were consistently associated with improved overall survival (OS) and relapse-free survival (RFS) across different thresholds (<5%, <10%, <20%, <30%, <40% and <50% MRD burden per patient), with patients in the <5% MRD group showing a significant difference in RFS (p = 0.0013). The median overall survival in this group was 18.0 months, while the median relapse-free survival was 11,7 months. OS and RFS remained significantly different in all MRD groups even after exclusion of DNMT3A and TET2 mutations, which most likely represented clonal hematopoiesis. In paired samples, TP53 mutations were notably prominent during follow-up, remaining the dominant clone, indicating a mixed composition, and the persistence of treatment resistant clones. Overall, our study highlighted the dynamic nature of clonal evolution under therapeutic pressure by demonstrating continuous clonal diversity during clinical remission and notable changes in clonal composition during relapse. Conclusion: In this AML cohort, our study provides insights into the genetic landscape and clonal dynamics of AML in remission. Thus, our findings emphasize the importance and the potential of single-cell DNA + protein sequencing in refining MRD assessment compared to current techniques.

Mutations in DNA methyltransferase 3 alpha (DNMT3A) are the most frequent driver of clonal hematopoiesis of indeterminate potential (CHIP), and associated with higher risk of cardiovascular disease and pro-inflammatory activation of immune cells. Here, we investigated the mechanisms underlying DNMT3A CHIP-associated inflammatory phenotypes in macrophages. We show that monocytes of DNMT3A CHIP-driver mutation carriers are associated with DNA hypomethylation of succinate dehydrogenase A (SDHA) and an altered tricarboxylic acid cycle metabolite profile. Silencing of DNMT3A in monocytes increased SDHA and elevated mitochondria complex II activity. The secreted complex II product, malate, further increased inflammatory activation in wild type monocytes to further augment inflammation in a paracrine manner. Pharmacological inhibition of SDHA (using dimethyl malonate) in mice harboring DNMT3A mutations in hematopoietic stem cells ameliorated the inflammatory response and improved cardiac function after myocardial infarction. Thus, interfering with the altered metabolic state may provide a new therapeutic option to dampen inflammatory activation in DNMT3A CHIP carrying patients. ### Competing Interest Statement The authors have declared no competing interest.

Both fatty bone marrow (FBM) and somatic mutations in hematopoietic stem cells (HSCs), also termed clonal hematopoiesis (CH) accumulate with human aging. However it remains unclear whether FBM can modify the evolution of CH. To address this question, we herein present the interaction between CH and FBM in two preclinical male mouse models: after sub-lethal irradiation or after castration. An adipogenesis inhibitor (PPARγ inhibitor) is used in both models as a control. A significant increase in self-renewal can be detected in both human and rodent DNMT3A Mut -HSCs when exposed to FBM. DNMT3A Mut -HSCs derived from older mice interacting with FBM have even higher self-renewal in comparison to DNMT3A Mut -HSCs derived from younger mice. Single cell RNA-sequencing on rodent HSCs after exposing them to FBM reveal a 6-10 fold increase in DNMT3A Mut -HSCs and an activated inflammatory signaling. Cytokine analysis of BM fluid and BM derived adipocytes grown in vitro demonstrates an increased IL-6 levels under FBM conditions. Anti-IL-6 neutralizing antibodies significantly reduce the selective advantage of DNMT3A Mut -HSCs exposed to FBM. Overall, paracrine FBM inflammatory signals promote DNMT3A -driven clonal hematopoiesis, which can be inhibited by blocking the IL-6 pathway.

Microglia are established in embryogenesis forming a self-containing cellular compartment resisting seeding with cells derived from adult definitive hematopoiesis. We report that monocyte-derived macrophages (MoMΦ) accumulate in the brain of aging mice with distinct topology, including the nigrostriatum and medulla, but not the frontal cortex. Parenchymal MoMΦ adopt bona fide microglia expression profiles. Unlike microglia, these monocyte-derived microglia (MoMg) are due to their hematopoietic origin targets of clonal hematopoiesis (CH). Using a chimeric transfer model, we show that hematopoietic expression of DNMT3A R822H , a prominent mutation in human CH, renders MoMg pathogenic promoting motor deficits resembling atypical Parkinsonian disorders. Collectively, these data establish in a mouse model that MoMg progressively seed the brains of aging healthy mice, accumulate in selected areas, and, when carrying a somatic mutation associated with CH, can contribute to brain pathology.

The limbus of the eye is the location of the corneal epithelial stem cell niche. These cells are necessary for continuous renewal of the corneal epithelium. In the case of limbal stem cell deficiency, these cells are damaged, and the whole cornea becomes opaque. It is important to be able to identify stem cells that could be applied for new therapeutic strategies. There are various known markers to characterize these cells, including p63, Nanog, oct4 and FGFR2. However, none of these markers are exclusively expressed in these stem cells (they are also expressed in transient amplified cells). It seems likely that a combination of stem cell markers will be necessary for corneal stem cell identification. The aim of this study was to detect IRF8 in limbal epithelial stem cells and to determine its function. In a mouse model, IRF8 could be detected in limbal and basal epithelial cells of the cornea by histological and immunohistological staining of wild-type mouse eyes. Furthermore, the limbus of the eye was significantly smaller in IRF8-knockout mice than in wild-type mice, and the expression of Nanog was lower in IRF8-knockout mice. This suggests that IRF8 has an influence on the maintenance of stem cell properties in the limbus, possibly by affecting the expression of Nanog. Furthermore, IRF8 has an impact on E-cadherin and N-cadherin expression in the mouse eye.

Background: Leukemia-cell-stroma interactions are an attractive therapy target in AML. The CXC chemokine receptor 4 (CXCR4) and its ligand CXCL12 (SDF-1) are expressed and secreted by myeloid leukemia blasts and play an essential role in the migration, homing, differentiation, proliferation, and retention of myeloid blasts within the protective bone marrow (BM) niche. The high-affinity CXCR4 antagonist BL-8040 is directly toxic for AML blasts and interrupts the interaction between blasts and the protective BM microenvironment. Objective: The aim of this clinical trial was to evaluate the impact of BL-8040 in combination with standard consolidation therapy on relapse free survival (RFS) in patients with acute myeloid leukemia (AML) in first complete remission (CR, CRp, CRi). Methods: Adult patients ≥ 18 years with documented 1st CR/CRi/CRp, after induction therapy with cytarabine and an anthracycline, and not scheduled for allogeneic stem cell transplantation were randomized in a 1:1 ratio to receive two cycles (≥ 60 years) or three cycles (<60 years) of high-dose cytarabine plus BL-8040 or placebo as consolidation therapy. The dosage of cytarabine was 1g/m² per dose for subjects older than 60 years and 3 g/m² per dose for subjects younger than 60 years. Results: Between November 2015 and November 2019, 134 patients of the originally planned number of 202 were recruited at 29 trial sites, 6 patients violated inclusion and/or exclusion criteria and were excluded. Considering the results of the interim analysis, the data monitoring committee recommended to stop recruitment. Thus, no further patients were enrolled. One hundred and twenty-eight patients were randomized (63 placebo; 65 BL-8040). Baseline characteristics were balanced between treatment arms: median age 61 vs 63 years; 2010 ELN favorable or intermediate-risk 95% vs 98%; CXCR4 expression 38% vs 30% in the BL-8040 and standard arm, respectively. Overall median follow-up time was 25.4 months. Median RFS was similar between arms 10.3 months [95% CI 8.0,12.0] for the BL-8040 arm and 11.5 months [95% CI 8.6,24.1] for the placebo arm (two-sided p=0.98 by log-rank test).The median OS was not reached at the time point of the analysis. Intention to treat (ITT) analyses revealed no difference between placebo and BL-8040, neither for the total cohort, nor for the subset of patients with expression of CXCR4 with respect to RFS (Hazard ratio (HR) 1.0, 95% CI 0.7,1.6, p =0.9, and HR 0.9, 95% CI 0.1,2.9, p =0.9,respectively) or overall survival (OS; HR 1.1, 95% CI 0.6,2.0, p = 0.8, and HR 2.5, 95% CI 0.4,13.8, p = 0.3, respectively). Regarding safety endpoints, overall, 739 and 576 adverse events were documented in the BL-8040 and placebo arm, respectively. Serious adverse events were similar between therapies. Conclusion: The addition of BL-8040 to intensive consolidation therapy did not improve RFS or OS. Although more side effects were noted, no differences in serious adverse events were observed.

Somatic mutations in DNA methyltransferase 3A (DNMT3A) are among the most frequent alterations in clonal hematopoiesis (CH) and acute myeloid leukemia (AML), with a hotspot in exon 23 at arginine 882 (DNMT3AR882). Here, we demonstrate that DNMT3AR882H-dependent CH and AML cells are specifically susceptible to the hypomethylating agent azacytidine (AZA). Addition of AZA to chemotherapy prolonged AML survival solely in individuals with DNMT3AR882 mutations, suggesting its potential as a predictive marker for AZA response. AML and CH mouse models confirmed AZA susceptibility specifically in DNMT3AR882H-expressing cells. Hematopoietic stem cells (HSCs) and progenitor cells expressing DNMT3AR882H exhibited cell autonomous viral mimicry response as a result of focal DNA hypomethylation at retrotransposon sequences. Administration of AZA boosted hypomethylation of retrotransposons specifically in DNMT3AR882H-expressing cells and maintained elevated levels of canonical interferon-stimulated genes (ISGs), thus leading to suppressed protein translation and increased apoptosis.

With aging, humans accumulate preleukemic mutations (pLMs) in hematopoietic stem and progenitor cells (HSPCs) which was termed age-related clonal hematopoiesis (ARCH). To gain a better insight changes in the HSPCs-environment crosstalk upon aging that might contribute to ARCH, it is critical to develop a multilayer perspective that integrates information on mutations, epigenetics the cellular context and the bone marrow (BM) microenvironment, since all these layers are changing during ageing. Accordingly, a key question in the field is how the ageing BM microenvironment influences clonal expansion of HSPCs. Fatty bone marrow (FBM) is one of the environmental factors that may influence clonal hematopoiesis (CH) with age. As we age, our bone marrow shifts from red to adipocyte-enriched yellow BM. We hypothesize that age related BMF accumulation may provide a selective advantage to specific pre-leukemic stem and progenitor cells (preL-HSPCs) carrying pLM. To support this hypothesis, we established a FBM model in NSG mice to enable the study of both human and rodent preL-HSPCs. Transplantation of primary human preL-HSPCs from AML patients (DNMT3A, NPM1 mutations ) into FBM resulted in enhanced engraftment compared to control mice without FBM. We further demonstrate that DNMT3A-R882H+/- mice derived BM HSPCs, engrafted significantly higher in NSG mice with FBM compared to controls. Interestingly, when DNMT3A-R882H+/- derived BM cells from middle-aged mice (12-month old) were injected into FBM mice, engraftment increased tenfold. Secondary engraftment of aged DNMT3A-R882H +/-BM derived cells resulted in an increase in engraftment upon transplantation into to FBM, suggesting enhanced in vivo self-renewal capacity of HSPCs in FBM. To study the underlying molecular mechanisms provided by the FBM to preL-HSPCs carrying DNMT3A-R882H +/-, we used a multiplex cytokine assay. In this approach we analyzed 17 common cytokines in BM following transplantation of young, two-month old, or middle-aged, 12-month old, DNMT3A-R882H +/-or control - BM derived cells into FBM. Our results show that transplanting two months old, middle-aged DNMT3A-R882H +/-or control BM derived cells to FBM resulted in a significant increase in BM mIL-6 secretion when compared to transplants into control, non-FBM mice. mIL-6 was secreted by adipocytes following irradiation regardless of which cells are transplanted. We then transplanted middle-aged DNMT3A-R882H+/- BM derived cells to FBM mice that had been treated intraperitoneally with a neutralizing IL-6 Ab. The administration of neutralizing IL-6 Ab resulted in a significant decrease in engraftment of DNMT3A-R882H+/- BM derived cells, confirming that IL6 contributes to the expansion of the DNMT3a-R882H+/- cells in FBM. In summary, these results demonstrate for the first time that the FBM provides a selective advantage to pre-leukemic cells carrying DNMT3A-R882H. Importantly, we show that IL-6 is a one of the major players in the molecular mechanism that confers the FBM advantage specifically to preL-HSPCs carrying R882H both in vitro and in vivo. Mueller-Tidow: Janssen Cilag: Consultancy, Research Funding; Bioline: Consultancy, Research Funding; Pfizer: Consultancy, Research Funding.