{"paper_id":"049ca84f-d02e-4fdd-ba35-e11cf2e893e8","body_text":"1 \nTET2 loss promotes premalignant survival and clonal selection in MYC -\ndriven B cell lymphoma \n \nSarah Spoeck 1, Nadine Kinz 1*, Irene Rigato 2*, Katharina Hoppe 1, Julia Heppke 1, Paul Y. \nPetermann1, Johannes G. Weiss 1,3, Miriam Erlacher 4, Michael Schubert5, Joel S. Riley1, \nAndreas Villunger1,6, Francesca Finotello2, Verena Labi1,7 \n \n \n1Institute for Developmental Immunology, Biocenter, Medical University of Innsbruck, \nInnsbruck, Austria \n2Department of Molecular Biology , Digital Science Center  (DiSC), University of Innsbruck, \nInnsbruck, Austria  \n3Department of Paediatrics I, Medical University of Innsbruck, Innsbruck, Austria \n4Department of Pediatrics and Adolescent Medicine, University Medical Center Ulm, Ulm, \nGermany \n5Institute for Bioinformatics, Medical University of Innsbruck, Innsbruck, Austria \n6The Research Center for Molecular Medicine of the Austrian Academy of Sciences, Vienna, \nAustria \n \n7Corresponding author contact:  \nVerena Labi, PhD \nInstitute for Developmental Immunology \nBiocenter, Medical University of Innsbruck \nInnrain 80, A-6020, Innsbruck, AT \nPh: +43-512-9003-70380 \nFax: +43-512-9003-73960 \nEmail: verena.labi@i-med.ac.at \n \n* N. Kinz and I. Rigato contributed equally to this work \n \nRunning Title: TET2 facilitates MYC-driven lymphomagenesis \n \nKeywords: TET2, BCL2 family, MYC-driven B cell lymphoma , EµMyc mouse model, clonal \nselection \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 2 \nAbbreviations:  \nBCL2 - B-cell lymphoma 2 \nBCR - B cell receptor \nBIM - BCL2 interacting mediator of cell death \nDEGs - differentially expressed genes \nEdU - 5-ethynyl-2´-deoxyuridine \nFACS - fluorescence-activated cell sorting \nIgD - Immunoglobulin D \nIghv - Immunoglobulin heavy chain variable regiou \nIgM - Immunoglobulin M \nLOF - loss-of-function \nMCL1 - myeloid cell leukemia 1 \nMFI - mean fluorescence intensity \npAKT - phosphorylated protein kinase B \npH3 - phosphohistone H3  \nTCF3/E2A - transcription factor 3 (E2A immunoglobulin enhancer-binding factors E12/E47) \nTET2 - ten-eleven translocation 2  \nVH - Variable Heavy chain \nγH2AX - gamma-histone variant H2AX \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 3 \nABSTRACT \nThe DNA demethylase ten-eleven translocation enzyme 2 (TET2) is frequently inactivated in \nhematologic malignancies, yet how its loss shapes oncogene -driven transformation remains \nunclear. Using a mouse model in which MYC is overexpressed  in the B cell lineage , driving \naggressive B cell lymphoma, we show that Tet2 loss increases lymphoma penetrance and \nbiases disease toward an IgM⁺ immunophenotype. Established lymphomas are broadly similar \nacross genotypes, suggesting that Tet2 loss exerts much of its effects before lymphoma onset. \nAccordingly, Tet2 loss expands a  premalignant IgM⁺ B cell subset with reduced apoptotic \nsensitivity and an increased frequency of BCL2 ⁺BIMhi cells. Consistently, Tet2 deficient IgM⁺\t\nB cells persist better in in vitro cultures, show increased clonogenic survival, and exhibit clonal \nskewing. These findings support a model in which Tet2 loss heightens MYC-driven lymphoma \npenetrance by promoting the survival and selection of premalignant B cells under apoptotic \nstress. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 4 \nINTRODUCTION \nHematopoiesis, the lifelong production of blood cells,  depends on continuous epigenetic \nremodeling, in which the ten-eleven translocation enzyme 2 (TET2) plays a central role. TET2 \nis an iron- and α-ketoglutarate (αKG)-dependent dioxygenase that regulates gene expression \nby catalyzing the stepwise oxidation of 5 -methylcytosine (5mC) to 5-hydroxymethylcytosine \n(5hmC) and further  oxidized derivatives, thereby initiating DNA demethylation and restoring \nunmethylated cytosine1-6. In a cell type -specific manner, TET2 is preferentially recruited to \nenhancers, but can also act at promoters and gene bodies, where it helps maintain chromatin \naccessibility and lineage-appropriate gene expression programs7-11. Through these activities, \nTET2 is critical for proper  hematopoietic differentiation, lineage commitment, and immune \nhomeostasis, and its perturbation can have broad consequences12-16. \nTET2 is a well -established tumor suppressor in hematopoietic cells. Tet2-deficient mouse \nmodels show enhanced hematopoietic stem and progenitor cell (HSPC) self-renewal, myeloid \nbias, altered B and T cell homeostasis, and predisposition to myeloid and lymphoid \nmalignancies12-15,17. These phenotypes are commonly more pronounced  when Tet2 loss is \ncombined with Tet3 loss, consistent with partial functional redundancy between the two \nenzymes5,8,18,19. In B cells, TET2 and TET3 help maintain B-lineage gene regulatory programs \ncontrolled by core transcription factors (TFs) such as EBF1 and PAX58,11,20. Accordingly, loss \nof TET activity in mouse models disrupts stage-specific enhancer demethylation, impairs B cell \ndifferentiation and function, including germinal center responses  and antibody output , and \npredisposes to B cell malignancies 8,11,21-24. In myeloid cells, TET deficiency causes \nhypermethylation of lineage -specific enhancers and  of binding sites for key  transcription \nfactors such as PU.1, RUNX1 and CEBPA, thereby promoting myeloid bias, a proinflammatory \nstate, and myeloid transformation in mice19,25-27.  \nIn humans, pathogenic alterations of TET genes most commonly affect TET2, whereas they \nare substantially less frequent in TET35. Rare germline TET2 variants have been associated \nwith immune dysregulation and predisposition to hematologic malignancy, including childhood \nB and T cell lymphoma28-30. Far more commonly, TET2 loss-of-function (LOF) mutations arise \nin hematopoietic cells, often as early events in stem and progenitor compartments. They rank \namong the most frequent genetic alterations in human hematologic malignancies, including \nacute myeloid leukemia (AML), chronic myelomonocytic leukemia (CMML), myelodysplastic \nsyndromes (MDS), myeloproliferative neoplasms (MPN), and T and B cell lymphomas 31-35. \nConsistent with a role as a  tumor suppressor, TET2 LOF mutations are the second most \ncommon lesions in clonal hematopoiesis (CH), a preclinical condition in which mutant HSPC \nclones expand with age . Here, inflammatory cues contribute to the selective expansion of \nTET2-deficient HSPCs, and may facilitate malignant transformation27,36,37.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 5 \nAcross mouse models of myeloid malignancy , Tet2 deficiency recurrently cooperates with \noncogenes such as FLT3-ITD, JAK2 V617F, or RAS  to shorten disease latency or increase \npenetrance31,38-42. Additional c ooperating events include epigenetic modifiers  such as  \nDNMT3A or ASXL1 , as well as loss of the tumor suppressor TP53 43-46. To date, only a few \nstudies have addressed cooperation of Tet2 loss with additional oncogenic lesions in B cells, \nnotably BCL6 and TCL1A, where it accelerates lymphomagenesis16,47. However, whether Tet2 \nloss similarly cooperates with other oncogenic drivers to promote B cell transformation remains \nunknown, despite recurrent TET2 LOF mutations in human B cell malignancies. This gap is \nparticularly notable given the central role of oncogenic MYC deregulation in aggressive human \nB cell lymphomas, including Burkitt lymphoma (BL) and diffuse large B -cell lymphoma \n(DLBCL)48,49. Although physiologically required for hematopoietic progenitor function and early \nB cell development, enforced MYC expression imposes strong proliferative pressure while \nsensitizing cells to apoptosis, a constraint that must be overcome during transformation 50-54. \nThis raises the question of how Tet2 loss shapes premalignant and malignant B  cell states \nduring MYC-driven lymphomagenesis in vivo. \nTo investigate this, we used the EμMyc mouse model, in which enforced MYC expression in \nthe B cell lineage drives aggressive lymphoma55,56. TET2 deficiency increased MYC-driven B \ncell lymphoma penetrance and shifted the tumor spectrum toward IgM ⁺ disease. Although \nestablished lymphomas were broadly similar across genotypes, premalignant EμMyc Tet2−/− \nmice showed a partial block in peripheral B cell maturation, reflected by accumulation of IgM⁺ \nimmature-like B cells that are largely CD21− and CD23−. Within this compartment, Tet2 loss \nenriched a BCL2 ⁺BIMhi subpopulation associated with enhanced in vitro survival, increased \nfunctional BCL2-dependence, greater colony-forming capacity, and increased clonal skewing. \nTogether, these findings identify Tet2 loss as a cooperating lesion that facilitates MYC-driven \nlymphomagenesis by restraining B cell maturation, buffering apoptotic stress, and promoting \nearly clonal skewing. \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 6 \nRESULTS \nTet2 loss increases B cell lymphoma penetrance in EμMyc mice \nTo test how TET2 LOF influences MYC-driven lymphomagenesis, we monitored mouse \ncohorts with germline Tet2 deletion13 with or without the B cell-specific EμMyc transgene55,56. \nIn the absence of MYC overexpression, wildtype, Tet2+/−, and Tet2−/− animals remained free of \novert disease over the observation period of 300 days (Fig. 1A). As expected, EμMyc mice \ndeveloped aggressive B cell lymphoma with a median overall survival of 127 days, and 26.8% \nof animals remained tumor-free until day 300 (Fig. 1A). Tet2 loss reduced overall survival on \nthe EμMyc background, with 122 days for EμMyc Tet2+/− mice and 103 days for EμMyc Tet2−/− \nmice (Fig. 1A). Notably, all EμMyc Tet2−/− mice developed tumors, whereas long-term tumor-\nfree survivors persisted in the EμMyc and EμMyc Tet2 +/− cohorts. Disease latency of the \nanimals that ultimately succumbed was equal  across genotypes ( Fig. 1B; EμMyc 105 days; \nEμMyc Tet2+/− 108 days; EμMyc Tet2−/− 103 days), indicating that Tet2 loss primarily increases \nlymphoma penetrance upon oncogenic MYC expression. \nTo assess tumor burden  at two major sites of disease involvement in the EμMyc model, we \nanalyzed spleens and bone marrow. As expected, spleen to body weight ratio was elevated in \ntumor-bearing EμMyc mice compared with non-transgenic controls,  however, Tet2 loss \nincreased spleen weight further (Fig. 1C), and total splenocyte and bone marrow cell numbers \nshowed a similar trend (Fig. S1A). Flow cytometric profiling of spleen and bone marrow \ncomposition revealed comparable frequencies of major immune lineages and malignant B cells \nbetween EμMyc and EμMyc Tet2−/− mice (Fig. 1D and S1B), suggesting that germline Tet2 \nloss does not induce major shifts in overall immune composition in established disease. \nWe next classified tumors by surface B cell receptor (BCR) expression of the immunoglobulin \nM (IgM) isotype, which is commonly used to stratify EμMyc lymphomas by cell-of-origin55,57. \nIgM⁻ tumors are typically linked to transformation within a  proliferating, developmentally \narrested compartment of IgM⁻ progenitor B cells55,57. IgM⁺ tumor cells are commonly IgD⁻ and \nshow minimal IgV (Immunoglobulin variable region) somatic hypermutation with low activation-\ninduced cytidine deaminase ( AID) expression, while recombinant activating gene ( RAG) \nactivity has been reported, supporting an immature-like, proliferating pre-germinal center state \nas cell-of-origin58,59. A majority of  EμMyc mice presented IgM⁻ tumors, wh ile a smaller \nproportion developed IgM⁺ or IgM⁻/IgM⁺ mixed disease (Fig. 1E and S1C). In contrast, EμMyc \nTet2−/− mice predominantly developed IgM ⁺ tumors, and EμMyc Tet2 +/− animals showed an \nintermediate phenotype (Fig. 1E and S1C). This  pattern is consistent with a gene -dosage \neffect, as also reported in other hematopoietic mouse models5,60. Together, these data identify \nTET2 as a suppressor of MYC -driven transformation that appears particularly relevant within \nthe BCR-expressing compartment. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 7 \nTo assess whether Tet2 loss is associated with persistent transcriptional differences in \nestablished disease, we performed bulk RNA -seq on fluorescence-activated cell sorting  \n(FACS)-purified IgM ⁻ and IgM ⁺ tumor cells. In both immunophenotypic subsets, comparing \nEμMyc Tet2−/− with EμMyc tumors revealed only a small number of significantly differentially \nexpressed genes (DEGs) (Fig. 1F and Table S1). Hallmark gene set analyses likewise showed \nno major genotype -associated changes ( Fig. S1D). These  data argue against widespread \ntranscriptional remodeling in established Tet2-deficient tumors. \nGiven prior work linking Tet2 or combined Tet2/Tet3 loss to altered expression of DNA repair \ngenes and increased DNA damage in hematologic malignancies17,19,61,62, we examined \nproliferative state and genomic integrity in established lymphomas  ex vivo . DNA content \nprofiling revealed similar cell cycle distribution across G 0/G1, S, and G 2/M phases between \nEμMyc and EμMyc Tet2−/− tumors (Fig. 1G and Fig. S1E). Polyploidy was rare but comparable \nbetween genotypes (Fig. 1H and Fig. S1F). We also inferred copy-number profiles from our \nbulk RNA-seq data, which revealed heterogeneous aneuploidy patterns across tumors without \na consistent genotype -associated signature ( Fig. S1G). Accordingly, Hallmark gene set \nanalyses did not indicate genotype-dependent changes in DNA repair programs (Fig. S1D and \nTable S1), and γH2AX staining did not demonstrate a significant increase in DNA double-strand \nbreaks in Tet2-deficient tumor cells compared with controls (Fig. 1I and Fig. S1H). Finally, \nfractions of tumor cells with cleaved Caspase-3 were comparable between genotypes (Fig. 1J \nand Fig. S1 I), indicating  that Tet2 loss does not substantially alter baseline apoptosis in \nestablished lymphomas. \nOverall, Tet2 loss increases EμMyc lymphoma penetrance and raises the fraction of  IgM⁺ \ntumors, while established lymphomas are largely similar across genotypes.  \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 8 \n \nFigure 1: Tet2 loss increases B cell lymphoma penetrance in EμMyc mice. Kaplan-Meier survival analysis \ndisplaying (A) overall survival probability, and (B) median time to death in days for all malignant mice (EµMyc: n=27, \nEµMyc Tet2+/-: n=44, EµMyc Tet2−/−: n=27). (C) Spleen to body weight ratio in non-malignant (280-320 days) mice \n(wildtype: n=7, Tet2+/-: n=13, Tet2−/−: n=6) and malignant mice, excluding animals without overt disease at 300 days \n(EµMyc: n=13, EµMyc Tet2+/-: n=11, EµMyc Tet2−/−: n=23). (D) Splenic immune cell composition assessed by flow \ncytometry: monocytes/macrophages (CD11b+Gr1-), granulocytes (CD11b+Gr1+), erythroid progenitors (nucleated \nTer119+), NK cells (TCRb-NK1.1+), CD4+ T cells (TCRb+CD4+), CD8+ T cells (TCRb+CD8+), and malignant B cells \n(B220+CD19+) (EµMyc: n=10, EµMyc Tet2−/−: n=12). (E) Lymphoma immunophenotype in the spleen. Mixed tumors \nwere defined as tumors where neither IgM- nor IgM+ cells constituted >80% of the total tumor population. (F) Volcano \nplots displaying RNA-seq-derived transcriptional profiles of FACS-sorted IgM- tumor (left; B220+CD19+IgM-IgD-) and \nIgM+ tumor (right; B220+CD19+IgM+IgD-) cells. Comparison between EµMyc Tet2−/− and EµMyc (IgM- tumors: n=4 \nvs. n=5, IgM+ tumors: n=4 vs. n=3) mice was performed separately for each cell type. Significance was defined as \nadjusted p-value<0.05 and absolute log2(fold change)>1. Downregulated genes in EµMyc Tet2−/− tumors are shown \nin blue; upregulated genes are shown in red. (G) DNA content and (H) polyploid cell fraction of splenic lymphoma \ncells, assessed by TO-PRO-3 staining via  flow cytometry ( EµMyc: n=7, EµMyc Tet2−/−: n=8). Flow cytometry \nassessment of (I) DNA double strand breaks by γH2AX staining (EµMyc: n=6, EµMyc Tet2−/−: n=7) and (J) apoptosis \nby cleaved Caspase-3 staining (EµMyc: n=5, EµMyc Tet2−/−: n=5) within splenic lymphoma cells. Bar graphs show \nmedian with interquartile range. Statistical significance was determined using (A) Mantel-Cox test, (B, C) one-way \nANOVA, (D, G) Mann -Whitney test, or (H, I, J) unpaired t -test, depending on normality (Shapiro -Wilk test) with \nHolm-Šidák correction for multiple comparisons. ns = not significant, *p<0.05, **p<0.005. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 9 \nTet2 loss enriches for IgM⁺IgD⁻ immature-like B cells in premalignant EμMyc mice \nTo determine whether the premalignant phase already shows changes that could underlie the \nlater enrichment of IgM ⁺ tumors, we analyzed animals at day 50, when most mice are still \nwithout overt disease. EμMyc Tet2−/− mice exhibited a mild increase in spleen to body weight \nratio and  splenocyte number compared with EμMyc controls (Fig. 2A, B). Flow cytometric \nprofiling of the spleen did not reveal major genotype dependent shifts in the relative abundance \nof major immune cell subsets in the EμMyc context, including premalignant B cells (Fig. 2C). \nIn non -EμMyc littermate controls , immune subset frequencies were likewise largely \nunchanged, aside from a modest increase in monocytes/macrophages in Tet2−/− controls (Fig. \nS2A), consistent with prior reports 12,14. Interestingly, however, specific to the EμMyc context \nTet2 loss skewed peripheral B cell maturation in the spleen  (Fig. 2D). Thus, EμMyc Tet2−/− \nmice showed an increased fraction of IgM⁺IgD⁻ immature-like B cells and a marked reduction \nof IgM⁺IgD⁺ mature B cells, while changes in the IgM⁻ progenitor B cell fraction were modest. \nIn the bone marrow, B cells were dominated by IgM⁻ progenitors, with only minor differences \nin the IgM⁺IgD⁻ immature-like fraction and the expected underrepresentation of IgM⁺IgD⁺ \nmature B cells in the presence of oncogenic MYC (Fig. S2B). This fits normal B cell biology, \nwhere IgM⁺IgD⁻ immature B cells accumulate in the spleen during peripheral maturation, a \nbias that is also evident in EμMyc mice and further enhanced by Tet2 loss. \nTo gain molecular insight into this phenotype, we FACS-sorted IgM ⁻ progenitors and IgM ⁺ \nimmature-like B cells from premalignant EμMyc and EμMyc Tet2−/− spleens and performed bulk \nRNA-seq. In IgM ⁻ progenitors, genotype -associated differences were minimal, with only 3 \nDEGs (Fig. 2E and Table S2). By contrast, Tet2 loss was associated with extensive \ntranscriptional changes in IgM⁺ immature-like B cells, with 985 DEGs (Fig. 2E and Table S2). \nIn the non-EμMyc context, the same analysis revealed only few DEGs in the corresponding B \ncell subsets (Fig. S2C and Table S3). Altogether, the skewed peripheral B cell maturation and \nthe pronounced transcriptional phenotype  in EμMyc Tet2−/− mice depend on aberrant MYC \nexpression and are not a dominant effect of Tet2 loss alone. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 10 \n \nFigure 2: Tet2 loss enriches for IgM⁺IgD⁻ immature-like B cells  in premalignant EμMyc mice. Analysis of \nsplenocytes from premalignant (day 50) EµMyc and EµMyc Tet2−/− mice. (A) Spleen to body weight ratio for each \nmouse and (B) absolute splenocyte counts (*107) (EµMyc: n=35, EµMyc Tet2−/−: n=26). (C) Splenic immune cell \ncomposition assessed by flow cytometry: monocytes/macrophages (CD11b +Gr1-), granulocytes (CD11b +Gr1+), \nerythroid progenitors (nucleated Ter119 +), NK cells (TCR b-NK1.1+), CD4 + T cells (TCR b+CD4+), CD8 + T cells \n(TCRb+CD8+), and B cells (B220 +CD19+) (EµMyc: n=18, EµMyc Tet2−/−: n=13). (D) Splenic B cell subsets were \nassessed via flow cytometry: IgM - progenitors (B220 +CD19+IgM-IgD-), IgM +IgD- immature-(like) \n(B220+CD19+IgM+IgD-), and IgM+IgD+ mature (B220+CD19+IgM+IgD+) B cells. The upper panel summarizes all data \n(wildtype: n=10, Tet2−/−: n=11, EµMyc: n=18, EµMyc Tet2−/−: n=13), the lower panel shows representative dot blots \nfor the IgM/IgD gate. (E) Volcano plots display RNA -seq-derived transcriptional profiles of premalignant FACS-\nsorted splenic IgM- progenitors (left; B220+CD19+IgM-IgD-) and IgM+ immature(-like) (right; B220+CD19+IgM+IgD-) \nB cells. Comparisons between EµMyc Tet2−/− (n=5) and EµMyc (n=6) mice were performed separately for each cell \ntype. Axis ranges were kept identical across volcano plots to allow direct comparison. Significance was defined as \nadjusted p-value<0.05 and absolute log2(fold change)>1. Downregulated genes in EµMyc Tet2−/− subsets are shown \nin blue; upregulated genes are shown in red. Bar plots show median with interquartile range. Statistical significance \nwas determined using (A) unpaired t-test or (B, C) Mann-Whitney test and (D) two-way ANOVA, with Holm-Šidák \ncorrection for multiple comparisons. Normality was assessed using the Shapiro-Wilk test. n.d. = not detected, ns = \nnot significant, *p<0.05, **p<0.005, ***p<0.0005. \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 11 \nTet2 loss skews IgM⁺ immature-like B cell sub-states on the EμMyc background \nGiven the extensive transcriptional changes in premalignant IgM ⁺ immature-like B cells (Fig. \n2E), we next asked which gene  programs are most perturbed by Tet2 loss. GO -term \nenrichment analysis of the DEG set highlighted biological processes  linked to B cell \ndifferentiation/maturation and immune receptor/immunoglobulin programs (Fig. 3A). This is \nconsistent with altered maturation programs within this subset. To infer TF activity from the \nRNA-seq profiles we performed “footprint” analysis using decoupleR63. This revealed a marked \nreduction in E2A/TCF3 (E12/E47) -associated activity in EμMyc Tet2−/− IgM⁺ immature-like B \ncells (Fig. 3B), alongside increased activity scores linked to KLF4, JUNB, NR4A1, and RUNX1. \nTogether, these shifts are consistent with reduced maturation-associated activity (TCF3/E2A) \nand increased stimulus-response/state-regulatory programs (KLF4, JUNB, NR4A1, RUNX1). \nBecause no single widely adopted signature robustly captures the murine IgM ⁺ immature-to-\nmature transition, we curated a focused B -lineage regulatory/maturation gene panel \ncomprising core B-lineage transcription factors, canonical peripheral maturation markers, and \nsignaling/activation nodes. Within this panel, genotype-dependent changes were restricted to \na subset of genes (13 up, 14 down , 76 unchanged) (Fig. 3C ), supporting a selective \nperturbation rather than a global collapse of the B cell maturation program. \nWe next validated cell surface markers identified as DEGs in Fig. 3C via flow cytometry in IgM⁺ \nimmature-like B cells, using the IgM ⁻ progenitor compartment as an internal control. At the \nprotein level, the most prominent changes were  reduced frequencies of  CD21/Cr2⁺ and of \nCD23/Fcer2a⁺ cells within the IgM + immature-like compartment in EμMyc Tet2−/− mice (Fig. \n3D), two markers typically upregulated during splenic peripheral B cell maturation toward the \nIgM+IgD+ mature stage64,65. CD21 and CD23 mean fluorescence intensities (MFIs) within their \nrespective marker-positive IgM⁺ gates were comparable between genotypes (Fig. 3D). Aicda \nand Bcl6 expression was low in EμMyc IgM⁺ immature-like B cells and further decreased upon \nTet2 loss, consistent with a pre -germinal center state (Fig. S3B, C and Table S2). We then \nextended this analysis to additional markers identified as DEGs in Fig. 3C to further resolve \nheterogeneity within the IgM⁺ immature-like compartment beyond CD21 and CD23. Tet2 loss \nprimarily enhanced the frequencies of CD9⁺, CD80⁺ and CD86⁺ subsets (linked to co-\nstimulatory and interaction programs), with modest changes in the CD5⁺, CD38⁺, CD40⁺ and \nCD44⁺ subsets (associated with broader signaling and interaction programs) (Fig. 3E and Fig. \nS3A). Again, MFIs within the respective marker-positive gates were largely comparable \nbetween genotypes  (Fig. 3E and Fig. S3 A), suggesting that Tet2 loss mainly alters  \nsubpopulation frequencies rather than per-cell expression. \nFinally, as apparent shifts along the peripheral maturation axis can be confounded by changes \nin core B  cell identity or BCR surface abundance, we assessed CD19 and IgM in  the IgM⁺ \nimmature-like gate.  Because CD19 and surface IgM MFIs were comparable between \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 12 \ngenotypes (Fig. 3F), we next  asked whether downstream BCR signaling may be altered in \nEμMyc Tet2−/− IgM⁺ immature-like B cells. \n \n \nFigure 3: Tet2 loss skews IgM ⁺ immature-like B-cell sub-states on the EμMyc background. (A) GO-term \nenrichment analysis of RNA -seq data from FACS -sorted IgM + immature-like B cells from the comparison of \npremalignant EµMyc Tet2−/− (n=5) versus EµMyc (n=6) mice. DEGs (adjusted p-value<0.05 and absolute log2(fold \nchange)>1) were subjected to GO -term biological process analysis in RStudio. The bar graph depicts the top \nenriched biological processes ranked by adjusted p -value, with the number of DEGs contributing to each term \nindicated in italics at the end of each bar. (B) DecoupleR analysis on the RNA-seq experiment performed in RStudio \nusing the CollecTRI transcription factor (TF)-target network. Bar plot displays the top 14 TFs with at least 45 targets \nand p-value<0.05. (C) Volcano plot comparing transcriptional profiles, with a specific focus on a self -defined B-\nlineage regulatory/maturation gene panel from the RNA-seq analysis described in (A). Significance was defined as \nadjusted p-value<0.05 and absolute log 2(fold change)>0.5. Downregulated genes in EµMyc Tet2−/− subsets are \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 13 \nshown in blue; upregulated genes are shown in red; B -lineage regulatory/maturation genes, which are not \ndifferentially expressed, are shown in dark grey, and all other genes in bright grey. (D-E) Flow cytometric validation \nof selected B-lineage regulatory and maturation genes in IgM - progenitors (filled symbols; B220+CD19+IgM-IgD-) \nand IgM+ immature-like B cells (empty symbols; B220+CD19+IgM+IgD-). The upper panel displays the percentage \nof cells expressing the indicated markers, while the lower panel shows the geometric MFI within the respective \nmarker-positive gate. (F) MFI of surface IgM and CD19 within IgM- progenitors (filled symbols; B220+CD19+IgM-IgD-\n) and IgM+ immature-like B cells (empty symbols; B220+CD19+IgM+IgD-). Bar plots show median with interquartile \nrange. Statistical significance was assessed using unpaired t-test for %CD21+ and CD21 MFI, %CD23+, %CD80+, \n%CD86+, IgM MFI, CD19 MFI, or Mann-Whitney test for CD23 MFI, %CD9+ and CD9 MFI, CD80 MFI, CD86 MFI, \nwith Holm -Šidák correction for multiple comparisons. Normality was evaluated using the Shapiro -Wilk test.  \nMFI = mean fluorescence intensity, ns = not significant, *p<0.05, **p<0.005, ****p<0.0001. \n \n \nTet2 loss selects an apoptosis-buffered IgM⁺ B cell sub-state on the EμMyc background \nTo profile BCR-associated signaling, we performed intracellular flow cytometry in EμMyc and \nEμMyc Tet2−/− IgM⁺ immature-like splenic B cells at steady state . We observed enhanced \nabundance of the  non-phosphorylated proteins CD79A and SYK in EμMyc Tet2−/− cells  \n(Fig. S4). Overall, these changes were modest and affected proximal BCR-associated proteins \nwithout a corresponding shift across the phosphorylation readouts . Interestingly, we also \nobserved elevated phosphorylated AKT (pAKT) in EμMyc Tet2−/− cells (Fig. S4), a readout that \ncan reflect BCR-PI3K signaling but also signals from other growth and survival pathways66,67. \nThus, we used Hallmark gene set enrichment analysis on the RNA-seq data from the \npremalignant IgM⁺ immature-like B cells to identify transcriptional programs accompanying this \nchange. The top enriched pathway was IL -2/STAT5 signaling, with additional enrichment of  \nIL-6/JAK/STAT3 signaling and TNFα signaling via NF -κB in EμMyc Tet2−/− relative to EμMyc \ncells (Fig. 4A). In this premalignant setting, we interpret these signatures as a MYC -linked \nstress/survival response rather than overt cytokine stimulation. Among the top hits were also \np53 pathway and apoptosis (Fig. 4A), in line with prior reports on MYC-driven proliferative \npressure and mitochondrial apoptotic priming 68,69. BCL2 -family transcript analys es further \nshowed increased expression of pro-apoptotic BH3-only genes (including Bcl2l11 and Bbc3) \nalongside elevated anti-apoptotic factors (including Bcl2 and Bcl2a1 isoforms) in EμMyc Tet2−/− \nIgM⁺ immature-like cells (Fig. 4B). Thus, Tet2 loss is associated with co-induced pro- and anti-\napoptotic BCL2 -family signatures, consistent with enrichment of an apoptosis -primed yet \nbuffered sub-state68,70. \nAt the protein level, intracellular flow cytometry revealed an increased fraction of BCL2 ⁺ IgM⁺ \nimmature-like cells in EμMyc Tet2−/− compared with EμMyc mice, while the BCL2 MFI within \nthe BCL2⁺ gate was comparable between genotypes (Fig. 4C). MCL1 and BCL-XL levels were \nunchanged (Fig. 4D). Pro-apoptotic BIM, the product of the Bcl2l11 gene, was detected across \nessentially all cells, but a distinct BIM hi subpopulation was evident predominantly in Tet2 \ndeficient EμMyc cells (Fig. 4E). Interestingly, these BIMhi cells largely overlapped with the \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 14 \nBCL2+ population (Fig. 4F), suggesting a selective enrichment of a BCL2⁺BIMhi IgM⁺ immature-\nlike sub-state in EμMyc Tet2−/− mice. These data point to  a cell subset with increased BIM-\ndependent death pressure which is buffered by enhanced protection through BCL2. \nBecause expression changes and steady-state protein abundance do not necessarily predict \nsurvival dependence, w e next evaluated spontaneous apoptosis in short -term culture . As \nreported previously , EμMyc IgM⁺ immature-like B cells rapidly undergo spontaneous \napoptosis53,55,56,71,72, however, Tet2 loss provided partial protection from death, most evident \nat 6 hours (Fig. 4G). \nGiven the prominent BCL2⁺BIMhi fraction in EμMyc Tet2−/− IgM⁺ immature-like B cells (Fig. 4F), \nwe next assessed drug-induced changes in cytochrome c release, which reflect mitochondrial \nouter membrane permeabilization and thus commitment to intrinsic apoptosis. Splenic B cells \nfrom premalignant mice were treated for 1 hour with the BCL2 inhibitor ABT-199 (Venetoclax) \nor the MCL1 inhibitor S63845. ABT-199 induced greater cytochrome c release in EμMyc Tet2−/− \nIgM⁺ immature-like B cells compared with EμMyc controls in this short -term in vitro priming \nassay (Fig. 4H), consistent with increased functional BCL2 dependence in this subset. I n \ncontrast, S63845 had little effect on cytochrome c release (Fig. 4I). \nAltogether, these findings show the  selective enrichment of an apoptosis -buffered IgM ⁺ \nimmature-like sub-state upon Tet2 loss in the EμMyc background. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 15 \n \nFigure 4 : Tet2 loss selects an apoptosis -buffered IgM ⁺ B cell sub -state on the EμMyc background.  \n(A) GO-term analysis of RNA -seq data from FACS -sorted IgM+ immature-like B cells from the comparison of \npremalignant EµMyc Tet2−/− (n=5) versus EµMyc (n=6) mice. DEGs (adjusted p-value<0.05 and absolute log2(fold \nchange)>1) were subjected to MSigDB Hallmark 2020 in Enrichr. The bar graph depicts the top enriched pathways \nranked by p-value, with the number of DEGs contributing to each term indicated in italics at the end of each bar. (B) \nVolcano plot comparing transcriptional profiles, with a specific focus on apoptotic/BCL2-family gene panel from the \nRNA-seq analysis described in (A). Significance was defined as adjusted p -value<0.05 and absolute log 2(fold \nchange)>0.5. Downregulated genes in EµMyc Tet2−/− subsets are shown in blue; upregulated genes are shown in \nred; apoptotic/BCL2-family genes, which are not differentially expressed in dark grey, and all other genes in bright \ngrey. (C) Flow cytometric analysis determining the fraction of BCL2+ cells and MFI within the BCL2+ gate in IgM+ \nimmature-like B cells (B220 +CD19+IgM+IgD-) (EµMyc: n=6, EµMyc Tet2−/−: n=4). (D) MFI of BCL -XL and MFI of \nMCL1 in IgM + immature-like B cells, quantified by flow cytometry ( EµMyc: n=8, EµMyc Tet2−/−: n=4). (E) Flow \ncytometric analysis determining the fraction of BIMhi cells and representative histograms for the BIM staining in the \nIgM+ immature-like B cell compartment ( EµMyc: n=8, EµMyc Tet2−/−: n=4). (F) Flow cytometric assessment \ndetermining the fraction of  BCL2+BIMhi cells and representative dot plots of the BCL2 +BIMhi population in IgM + \nimmature-like B cells ( EµMyc: n=3, EµMyc Tet2−/−: n=4). (G) Cell survival kinetics assessed in vitro  for IgM + \nimmature-like B cells (DAPI+B220+CD19+IgM+IgD-) from EµMyc (n=3) and EµMyc Tet2−/− (n=3) mice, at 0, 2, 6, and \n10 hours of culture. Assessment of mitochondrial apoptotic sensitivity via cytochrome c release of IgM+ immature-\nlike B cells (ZombieDye-B220+CD19+IgM+IgD-cytochromec-) treated with (H) 30 µM ABT-199/Venetoclax and (I) 1 \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 16 \nµM S63845. For both treatments, cytochrome c release in DMSO controls and treated samples is shown as a line \nplot (left) and as a bar graph (right), depicting the fold change relative to DMSO (EµMyc: n=3, EµMyc Tet2−/−: n=4). \nBar plots show median with interquartile range. Statistical significance was assessed using unpaired t-test (A-F, H, \nI), or two-way ANOVA (G) with Holm-Šidák correction for multiple comparisons. Normality was evaluated using the \nShapiro-Wilk test. MFI = mean fluorescence intensity, ns = not significant, *p<0.05, **p<0.005. \n \n \nTet2 loss enhances persistence and clonal skewing in premalignant IgM⁺ EμMyc B cells \nGiven the reduced apoptosis sensitivity in premalignant EμMyc Tet2 −/− IgM⁺ immature-like  \nB cells (Figure 4), we asked whether this compartment also contains cells with an improved \ncapacity to persist and divide. Therefore, we tracked cell division of premalignant splenic B \ncells over 48 hours in culture via proliferation dye loss. M ost IgM⁺ immature-like B cells were \nrapidly lost over 48 hours (Fig. 5A), consistent with the pronounced in vitro apoptosis sensitivity \nobserved in Figure 4 G. The  few surviving  cells in EμMyc cultures largely failed to divide, \nwhereas a small fraction of EμMyc Tet2−/− IgM⁺ immature-like B cells persisted and underwent \nmultiple divisions within 48 hours (Fig. 5B). To assess whether the dye dilution phenotype \ncould be explained by increased proliferative capacity, we profiled DNA content ex vivo and \nperformed a  2 hour  in vitro  5-ethynyl-2´-deoxyuridine (EdU) pulse combined with \nphosphohistone H3⁺ (pH3) co-staining in premalignant B cells. DNA content analysis showed \nan increased fraction of IgM⁺ immature-like cells in the S and G2/M cell cycle phases (Fig. S5A, \nB). In line, EdU incorporation was moderately enhanced in EμMyc Tet2−/− IgM⁺ immature-like \nB cells, while the fraction of EdU ⁺pH3⁺\tcells, which reflects progression into mitosis among \nEdU-labeled cells , was comparable between genotypes (Fig. S5C, D) . Of note , IgM⁻ \nprogenitors showed no clear genotype -associated differences across these measures (Fig . \nS5A-D). Together, these data suggest improved persistence and modest redistribution across \ncell cycle phases, which could reflect altered cell cycle kinetics within the IgM⁺ immature-like \ncompartment. \nTo test whether this phenotype associates with enhanced colony forming capacity, we plated \nsplenocytes in methylcellulose with or without the cytokine IL -7, which is required for the \nsurvival of IgM⁻ progenitors but not IgM+ immature B cells65,73. Strikingly, EμMyc Tet2−/− cells \nshowed an increased  capacity to form colonies  without IL-7 (Fig. 5C), indicating  enhanced \nclonogenic potential . Consistent with a fitness advantage emerging in only a subset of \npremalignant IgM⁺ immature-like B cells, we next asked whether Tet2 loss is accompanied by \nearly clonal skewing within this compartment. Using immunoglobulin heavy chain variable \nregion (Ighv) transcript abundance as a proxy for clonal representation, premalignant EμMyc \nTet2−/− IgM⁺ immature-like B cells displayed a more restricted repertoire than EμMyc controls \n(Fig. 5D), whereas overall Variable Heavy chain (VH) gene usage patterns were broadly similar \nbetween genotypes (Fig. 5D). These data support the view of increased clonal skewing rather \nthan a genotype -specific shift in VH preference. Consistent with repertoire compression , \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 17 \ndifferential expression analysis in the same bulk RNA -seq dataset showed broad \nunderrepresentation of immunoglobulin variable-region transcripts in Tet2-deficient IgM⁺ cells, \nsupporting repertoire compression rather than uniform regulation of Ig gene expression (Fig. \nS5E). \nTogether, these data suggest that Tet2 loss enables a small subset of premalignant IgM ⁺ \nimmature-like EμMyc B cells to persist and expand. This early fitness and clonal skewing likely \ncontribute to the later increased representation of IgM ⁺ lymphomas and the fully penetrant \ndisease observed in EμMyc Tet2−/− mice. \n \n \nFigure 5: Tet2 loss enhances persistence and clonal skewing in premalignant IgM⁺ EμMyc B cells. Analysis \nof proliferative capacity in splenic IgM + immature-like B cells from EµMyc (n=9) and EµMyc Tet2−/− (n=10) mice. \nTotal splenocytes were labeled in vitro with a proliferation dye, and flow cytometric analysis was performed at 0 and \n48 hours. (A) Representative dot plots showing an overlay of total cells (grey) and live IgM+ immature-like B cells \n(red) at 0 and 48 hours. Two replicates of each genotype are depicted. (B) The fraction of dividing IgM+ immature-\nlike B cells was quantified based on proliferation dye dilution. Representative flow cytometry histograms of \nproliferation dye intensity in IgM + immature-like B cells from EµMyc (left) and EµMyc Tet2−/− (right) mice. (C) \nSplenocytes from premalignant EµMyc (n=3) and EµMyc Tet2−/− (n=3) were plated in methylcellulose with or without \nIL-7, and colonies were counted after 7 days. (D) Immunoglobulin heavy chain variable region (Ighv) gene usage \nin IgM+ immature-like B cells was determined by RNA-seq. Each bar represents an individual mouse, with Ighv \ngenes comprising >4% of total Ighv transcripts displayed in distinct colors with corresponding gene names. Bar \nplots show median with interquartile range. Statistical significance was determined using (B) unpaired t -test. \nNormality was assessed using the Shapiro-Wilk test. *p<0.05 \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 18 \nDISCUSSION \nThe gene encoding the DNA demethylase TET2 is recurrently affected by LOF mutations \nacross a broad range of hematologic malignancies. However, how Tet2 loss contributes to \ntransformation remains incompletely understood. Here, we used the EμMyc mouse model to \nassess whether and how Tet2 loss affects MYC-driven B cell lymphomagenesis. We show that \nTET2 deficiency increases lymphoma penetrance and shifts the tumor spectrum toward IgM ⁺ \ndisease, while once established, lymphomas appeared broadly similar when comparing EμMyc \nand EμMyc Tet2-/- mice. These findings suggest that Tet2 loss acts mainly during premalignant \nstages of MYC-driven lymphomagenesis. \nIn mouse models, Tet2 loss per se is only weakly tumorigenic and associates with long disease \nlatency, suggesting a role as tumor facilitator rather than autonomous driver12-15. In B cells, \ndirect evidence that Tet2 loss cooperates with defined oncogenic drivers remains limited to a \nfew studies, most notably TCL1A - and BCL6 -driven malignancies 16,47. This contrasts with \nmouse models of myeloid malignancies such as MDS, MPN-like disease, CMML, and AML, \nwhere the facilitator role is firmly established. Here, TET2 deficiency cooperates with lesions \naffecting signaling pathways and genome surveillance, including JAK2 V617F, oncogenic KIT, \nFLT3-ITD, oncogenic RAS, and Tp53 loss14,38,42,46,60. Notably, several of these alterations are \nthemselves associated with MYC activation or MYC -associated transcriptional programs74-76. \nIn the context of oncogenic KIT, TET2 deficiency has been linked to a hyperactive MYC \nsignature downstream of PI3K signaling 60. Consistent with this, Myc/MYC and Tet2/TET2 \ntranscripts are inversely related in MYC-driven mouse T-cell acute lymphoblastic leukemia (T-\nALL) and in the human Burkitt cell line  P493-6, where  MYC also  binds the TET2 locus77. \nBeyond expression control, MYC can engage TET2 both at chromatin and through \nmetabolism: in U2OS cells via SNIP1-dependent recruitment to chromatin, and in human and \nmurine B lymphoma models through αKG/2 -hydroxyglutarate-dependent modulation of TET \nactivity and 5hmC states 78-80. Against this background, our EμMyc model provides the first \ndirect in vivo evidence that Tet2 loss cooperates with oncogenic MYC in tumorigenesis. This \nis particularly relevant in B cells, where MYC alterations are prominent drivers of lymphoma \ndevelopment81,82. \nOne of the most striking phenotypes in EμMyc Tet2-/- mice is the shift toward IgM⁺ tumors. In \npremalignant mice, this is mirrored by developmental skewing at the transition from immature \nto mature B cells, making IgM ⁺ immature-like B cells the key population. This compartment \nwas less homogeneous than expected, and our data suggest that Tet2 loss enriches distinct \nsubstates within it. The strongly reduced CD21 ⁺ and CD23⁺ cell frequencies in premalignant \nEμMyc Tet2 -/- mice place the enriched substates toward the less mature end of the IgM ⁺ \nimmature-like compartment. Consistent with this, we do not detect appreciable Aicda or Bcl6 \nexpression, in line with prior work indicating that these  cells retain an immature state 58,59. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 19 \nExcluding immunoglobulin genes, which likely reflect clonal skewing, GO-term enrichment for \nlymphocyte differentiation, negative regulation of cell development, and regulation of \nhemopoiesis was evident in EμMyc Tet2-/- IgM⁺ immature-like B cells. Together with reduced \ninferred HNF1A and TCF3/E2A activity by decouple R, this is more consistent with impaired \nprogression toward mature B cells rather than with a strong developmental block83,84. Related \nperturbations of  differentiation have likewise been described upon Tet2 loss in other \nhematopoietic contexts, including HSPCs13-15 and in germinal center (GC) B cells24.  \nApart from the CD21/CD23 phenotype, increased CD9 ⁺, CD80 ⁺, and CD86 ⁺ cell fractions \nindicate substates with enhanced interaction and activation potential. This is supported by  \nGO-term enrichment for negative regulation of cell activation, regulation of T cell activation, \nand immune response -regulating signaling, together with increased inferred NR4A1, KLF4, \nJUNB, and RUNX1 activity by decouple R. CD86 and CD9 are best supported by prior \nliterature. In mice, B cell-specific TET loss caused Cd86 de-repression via reduced HDAC1/2 \nrecruitment and altered chromatin at the Cd86 locus22. In murine STAT5-driven leukemic stem-\ncell models, CD9 marks a high -fitness, self -renewing subpopulation linked to JAK/STAT -\nassociated persistence under oncogenic stress85. In our data, enrichment of survival-promoting \nJAK/STAT, TNF/NF-κB, and related signaling programs likely reflects a broader logic also seen \nin myeloid  models, in which Tet2 loss repeatedly shifts signaling responsiveness, stress \nhandling, and competitive fitness toward persistence under oncogenic pressure14,38,42,46,60. This \nsuggests that, under oncogenic MYC,  premalignant B cells use Tet2 loss in a manner that  \nparallels what has been observed in oncogene-driven myeloid cells. \nUnder strong MYC-imposed apoptotic pressure, such signaling states are well positioned to \nfavor survival of cells that can better  buffer mitochondrial death signaling.  Notably,  \nEμMyc Tet2-/- mice selectively enriched a BCL2 ⁺BIMhi IgM⁺ immature-like subpopulation that \nrepresents only a minor fraction in EμMyc controls. The EμMyc model is characterized by \nMYC-induced BIM and reduced BCL2 in IgM ⁺\tB cells, and by preferential expansion of these \ncells upon BIM loss56. Such a BCL2⁺BIMhi population fits a scenario in which BIM-associated \napoptotic priming is buffered by BCL2, thereby promoting cell survival 86-92. Accordingly,  \nEμMyc Tet2-/- IgM⁺ immature-like cells showed a modest but reproducible survival advantage \nduring spontaneous apoptosis in vitro, reflecting increased fitness under conditions of growth \nfactor withdrawal and loss of microenvironmental interactions. The BCL2 -inhibitor ABT-199, \nbut not the MCL1-inhibitor S63845, increased cytochrome c release in these cells, consistent \nwith enrichment of a functionally BCL2-dependent BCL2⁺BIMhi substate.  \nBeyond our system, evidence linking Tet2 loss to BCL2-family dependencies and cell survival \nremains limited. In B cells, the clearest survival -related evidence comes from a murine CLL \nmodel, in which Tet2 loss was associated with stronger BCR signaling dependency16. In EμMyc \nTet2-/- IgM⁺ immature-like B cells, however, the phospho -readouts do not support enhanced \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 20 \ntonic or proximal BCR signaling. Although pAKT was increased, this readout is not specific for \nBCR activity and is equally compatible with broader survival -associated signaling. In HSPC \nand myeloid contexts, Tet2 loss is most consistently linked to enhanced self -renewal, \ncompetitive fitness, and resistance to inflammatory stress rather than to broad proliferative \nactivation5,42,93. Along this line, premalignant EμMyc Tet2 -/- IgM⁺ immature-like B cells, \naberrantly driven into cycle by MYC, showed only a modest shift toward S and G 2/M and no \nconsistent transcriptomic evidence of enhanced proliferation. Yet after 48 hours in culture, the \nfew surviving EμMyc Tet2 -/- IgM⁺ immature-like B cells had proliferated, unlike their EμMyc \ncounterparts. Further, EμMyc Tet2 -/- IgM⁺ immature-like B cells showed increased colony -\nforming capacity in methylcellulose and evidence of repertoire compression and clonal \nskewing based on Ighv transcript abundance. Overall, our data suggest that  Tet2 loss \npromotes persistence and selective outgrowth rather than broad proliferative activation, \nextending to premalignant B cells a principle previously established in TET2 -deficient HSPC \nand myeloid settings.  \nSeveral aspects of the present study also define the limits within which this model should be \ninterpreted. Here, the EμMyc system is used as a mechanistic model of MYC -driven \nlymphomagenesis rather than a preclinical surrogate for a single human lymphoma entity. In \naddition, our data suggest that heterogeneity within the premalignant IgM⁺ compartment has \nbeen underappreciated in EμMyc mice. While our bulk analyses do not fully resolve this \nheterogeneity, they still provide consistent transcriptomic, phenotypic, and functional evidence \nthat Tet2 loss has biologically meaningful effects in this compartment.  Finally, because Tet2 \nloss was analyzed in a germline setting, B  cell-intrinsic and non -cell-autonomous effects \ncannot be fully separated, although this may also reflect biologically relevant aspects of early \nhematopoietic Tet2 loss, including clonal hematopoiesis. More broadly, it will be important to \ndetermine whether similar premalignant survival states also arise in more human -relevant \nhematopoietic disease settings. Even within these constraints, the main conclusions remain \nunchanged. \nUsing the EμMyc model, we provide a mechanistic framework for how Tet2 loss facilitates \nMYC-driven B cell lymphomagenesis through apoptosis buffering and selective clonal \noutgrowth. Together, these findings support a model in which TET2 deficiency promotes \nlymphoma development by enhancing survival under MYC -imposed stress a nd thereby \nincreasing the likelihood of premalignant clonal persistence, selection and outgrowth. \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 21 \nMATERIALS & METHODS \n \nAnimal models \nAll animals were backcrossed and maintained on a C57BL/6N background for at least 10 \ngenerations and bred at the central laboratory animal facility of the Medical University of \nInnsbruck. Animal experiments were approved by the Austrian Federal Ministry of Education, \nScience and Research (BMWF: 66.011/0008 -V/3b/2019) and conducted under standard \nhousing conditions consisting of a 12-hour (h) light/dark cycle, relative humidity of 55-65% and \ntemperature 22 ± 2°C. \nEµMyc (B6.Tg(IghMyc)22Bri) and Tet2-/- (B6(Cg)-Tet2tm1.2Rao) mouse lines were generated and \ngenotyped as previously described13,55. Both male and female mice were used indiscriminately \nand were monitored until either 46 - 55 days of age, 280 - 320 days of age, or until meeting \npredetermined euthanasia criteria. \n \nPreparation of single-cell suspensions \nAll centrifugation steps were performed at 1500 rpm for 5 minutes (min) at 4°C (VWR \nCentrifuge Megastar 1.6R). \nBone marrow cells were isolated by flushing femurs and tibiae with FACS -B buffer (PBS with \n2% FBS (Gibco, 10270106) and 10 µg/ml Gentamicin (Gibco, 15750037)) using a syringe and \na 23G needle. The cell suspension was centrifuged and filtered through a 50 µ m filter (BD \nBioscience, 340632). \nSpleens were dissociated by gently pressing the tissue through 70 µm cell strainers (Corning, \n352350). The resulting cell suspension was centrifuged and resuspended in 1 ml of ice -cold \nred blood cell lysis buffer (155 mM NH4Cl, 10 mM KHCO3, 0.1 mM EDTA; pH  7.5) and \nincubated on ice for 3 min. Lysis was stopped by adding 5 ml of FACS -B, centrifuged and \nfiltered through a 50 µm filter. \nCell numbers in suspension were determined using hemocytometer and trypan blue exclusion.  \n \nCell surface staining for flow cytometry \nSplenocytes and bone marrow cells were stained for flow cytometric analysis, with a minimum \nacquisition threshold of 3*105 cells per sample. All centrifugation steps were performed at 2000 \nrpm for 2 min at 4°C (VWR Centrifuge Megastar 1.6R), and all incubations were conducted at \n4°C in 96 -U-well plates. Cells were incubated for 10 min with 20 µl of αCD16/32 Fc-Block \n(1:200 in FACS-B; Biolegend, 101310), followed by incubation for 15 min with 30 µl of a mixture \nof the following fluorochrome -conjugated anti -mouse antibodies diluted in FACS -B \nsupplemented with Brilliant Stain Buffer (1:4; BD Biosciences, 566349): αCD9-FITC (1:400, \nBiolegend, 124808), αCD40-FITC (1:200, eBioscience, 11 -0402-81), αCD8-PE (1:300, \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 22 \nBiolegend, 100708), αNK1.1-PeCy7 (1:100, Biolegend, 108713), αCD86-PeCy7 (1:100, \nBiolegend, 105013), αIgM-PeCy7 (1:200, Biolegend, 406514), αIgD-PerCP/Cy5.5 (1:200, \nBiolegend, 405710), αCD38-APC (1:200, Biolegend, 102712), αIgM-APC (1:1000, Jackson \nImmunoResearch, 115 -607-020), αCD4-A700 (1:400, Biolegend, 116022), αCD21-A700 \n(1:200, Biolegend, 123431), αCD23-APC/Cy7 (1:200, Biolegend 101629), αGr1-BV421 \n(1:200, Biolegend, 108433), αCD80-BV421 (1:100, Biolegend, 104725), αCD5-BV421 (1:200, \nBiolegend, 100617), αTer119-BV510 (1:100, Biolegend, 116237), αCD44-BV605 (1:200, \nBiolegend, 103047), αTCRβ-BV605 (1:200, Biolegend, 109241), αCD11b-BV650 (1:1000, \nBiolegend, 101259), αCD19-BV711 (1:400, Biolegend, 115555) and αB220-BV785 (1 :400, \nBiolegend, 103246). Finally, cells were washed with FACS-B and immediately acquired on an \nLSR II-Fortessa flow cytometer (BD Bioscience). Data were analyzed using FlowJo software \n(version 10.10.0). \n \nIntracellular staining for flow cytometry \nAll centrifugation steps were performed at 2400 rpm for 2 min at 4°C (VWR Centrifuge \nMegastar 1.6R), and all incubation steps were carried out at 4°C in 96-U-well plates. Following \ncell surface staining (see section Cell surface staining), cells were fixed and permeabilized by \nincubation with 100 µl FixPerm buffer (containing 4% methanol -free formaldehyde (Thermo \nFisher, 28908) and 0.1% Saponin (Sigma-Aldrich, 47036)) for 20 min and then washed three \ntimes with 150 µl PermWash buffer (PBS with 1% BSA, 0.1% Saponin, 0.0025% Natrium Azid). \n \nFor DNA content analysis, in combination with mitotic and DNA damage markers, cells were \nincubated with 30 µl αCD16/32 Fc-Block (1:100 in PermWash Buffer) for 15 min, followed by \nincubation with 30 µl of a mixture of the following fluorochrome -conjugated anti -mouse \nantibodies for at least 30 min: αpH3-PE (1:100, Biolegend, 650807) or cleaved Caspase-3-PE \n(1:100, BD Bioscience, 570183) and αγH2AX-PerCPCy5.5 (1:100, eBioscience, 46-9865-42). \nCells were washed with 100 µl PermWash Buffer and incubated in 100 µl of PBS containing \n250 µg/ml RNase A (Sigma-Aldrich, R5500) at 37° for 20 min, after which 50 µl of TO-PRO-3 \nIodide (1:333 in PBS, Thermo Fisher, T3605) was added and samples were immediately \nacquired. \n \nFor analysis of BCR signaling components, cells were incubated with 30 µl αCD16/32 Fc-Block \n(1:100 in PermWash Buffer) for 15 min, followed by incubation with 30 µl of following primary \nantibodies for at least 30 min: B cell Signaling Antibody Sampler (1:100, Cell Signaling, 9768), \nαpAKT (1:100, Cell Signaling, 4060S), αAKT (1:100, Cell Signaling, 4691S), αSYK (1:100, \nSanta Cruz, sc1077). After a washing step with 100 µl PermWash Buffer, cells were incubated \nwith a goat anti -rabbit IgG (H+L) Alexa Fluor TM 647-conjugated secondary antibody (1:1000, \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 23 \nInvitrogen, A21245) for 15 min at 4°C. Finally, cells were washed with 100 µl FACS -B and \nacquired immediately. \n \nTo assess BCL2 family members, cells were incubated with 30 µl αCD16/32 Fc-Block (1:20 in \nPermWash Buffer) for 15 min, followed by incubation with 30 µl of following antibodies: αBIM \n(1:100, Abcam, ab32158), αBCL2-PE (1:100, Biolegend, 633507),  αMCL1 (1:100, Cell \nSignaling, 5453T) or αBCL-XL (1:100, Cell Signaling, 2764S) for at least 30 min. After a \nwashing step with 100 µl PermWash Buffer, a goat anti -rabbit IgG (H+L) Alexa Fluor TM 647-\nconjugated antibody (1:1000, Invitrogen, A21245) was applied for 15 min. A final washing step \nwith 100 µl PermWash Buffer was performed.  \n \nSamples were immediately acquired on an LSR II -Fortessa flow cytometer (BD Bioscience), \nand data were analyzed using FlowJo software (version 10.10.0). \n \nB cell viability assay \nB cells were enriched from splenic cell suspensions using MagniSortTM Streptavidin Negative \nSelection Beads (ThermoFisher, MSNB-6002-74), according to the manufacturer´s instruction. \nFor depletion of non -B cells, 300 µl of a biotinylated antibody mix (diluted 1:100 in FACS -B) \ncontaining αCD4 (Biolegend, 100404), αCD8 (Biolegend, 100704), αNK1.1 (Biolegend, \n108704), αCD11b (Milteny Biotec, 130 -113-242), αGr1 (Biolegend, 108404) and αTer119 \n(Biolegend, 116204) was used.  \nB cells (5*105) were then cultured in 50 µl FACS-B (untreated) in 96-U-well plates for 0 h, 2 h, \n6 h and 10 h at 37°C. To avoid loss of dead cells, 25 µl of the following antibody mix (diluted \nin FACS-B) was added directly to each well: αCD16/32 Fc-Block (1:67; Biolegend, 101310), \nαCD19-PE (1:67; eBioscience, 12 -0191-83), αB220-PeCy7 (1:67; Biolegend, 103222) ,  \nαIgD-PerCP/Cy5.5 (1:200, Biolegend, 405710), αIgM-APC (1:1000, Jackson \nImmunoResearch, 115-607-020) and DAPI (1:16600; Sigma -Aldrich, D9542). Samples were \nacquired on a LSR II-Fortessa flow cytometer (BD Bioscience), and data were analyzed using \nFlowJo software (version 10.10.0). \n \nDrug-induced cytochrome c release \nApoptotic sensitivity was assessed based on a previously published protocol88,89,94.  \nBriefly, 30 µM ABT -199/Venetoclax, (MedChemExpress, HY -15531) 1 µM S63845 \n(MedChemExpress, HY-100741), 20 µM Alamethicin (MedChemExpress, HY-N6708) and 1% \ndimethyl sulfoxide (DMSO; Merck, D5879) were diluted to 2X the desired final concentration \nin 25 µl of 0.002% digitonin (Merck, D141) prepared in MEB buffer (Mannitol Experimental \nBuffer: 10 mM HEPES (Merck, H0887) pH 7.5, 150 mM mannitol (Merck, M9647), 15 0 mM \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 24 \nKCl, 1 mM EGTA (Merck, E3889), 1 mM EDTA (Merck, ED4S), 0.1% BSA (Sigma -Aldrich, \n12659), 5 mM succinate (Merck, S3674)). These compound plates were pre -arrayed in  \n96-U-well plates and stored at -80°C until use. \nPer treatment,  2*105 splenocytes were stained (see section Cell surface staining ) in low -\nbinding tubes (Sarstedt, 72706600) with 100 µl of the following antibodies diluted in FACS -B \nsupplemented with Brilliant Stain Buffer (1:4; BD Biosciences, 566349): αCD16/32 (1:200, \nBiolegend, 101310), αIgM-PeCy7 (1:200, Biolegend, 406514), αIgD-PerCP/Cy5.5 (1:200, \nBiolegend, 405710), αCD19-BV711 (1:400, Biolegend, 115555) and αB220-BV785 (1:400, \nBiolegend, 103246). Cells were subsequently washed with FACS-B and further incubated with \n300 µl Zombie Green fixable viability dye (1:500 in FACS-B; Biolegend 423111) for 10 min at \n4°C. Per well 130 µl FACS-B was added and centrifuged at 2000 rpm for 2 min at 4°C (VWR \nCentrifuge Megastar 1.6R). For each 96-U-well, 2*105 stained cells (in 25 µl MEB buffer) were \nadded to 25 µl compound solution (pre-arrayed compound plates were thawed 1 hour \nbeforehand), and incubated for 1 h at RT in the dark. Subsequently, 16.5 µl of 4% methanol -\nfree formaldehyde (Thermo Fisher, 28908) was added per well and incubated for 10 min at \nRT, followed by the addition of 16.5 µl per well N2 buffer (1.7 M Tris base (Carl Roth, 5429.4), \n1.25 M Glycin (Fisher Scientific, 10773644); pH 9.1) and incubation for 5 min at RT. Finally, \n10 µl per well anti-cytochrome c antibody (1:400, Biolegend, 612310) diluted in 10x CytoStain \nBuffer (PBS with 2% Tween 20 (Carl Roth, 9127.1) and 10% BSA) was added and incubated \nfor 12 h at 4°C in the dark before flow cytometric analysis on an LSR II-Fortessa flow cytometer \n(BD Bioscience). Data were analyzed using FlowJo Software (version 10.10.0). \n \nColony formation assay (MethoCult) \nColony-forming unit assays using MethoCult were performed according to the manufacturer´s \ninstructions. Briefly, splenocytes (1*105) were resuspended in 100 µl IMDM (Gibco, 21056023), \nsupplemented with 100 units/ml penicillin and 100 µg/ml streptomycin (Sigma-Aldrich, P0781), \nand plated in 35-mm culture dishes (TC Dish35, Suspension, Sarstedt, 83.3900.500) in 1 ml \nmethylcellulose with out IL -7 (MethoCult M3231, STEMCELL Technologies) or with IL -7 \n(MethoCult M3630, STEMCELL Technologies). Dishes were incubated at 37°C, and colonies \nwere counted after 7 days.  \n \nIn vitro 5-ethynyl-2-deoxyuridine (EdU) assay \nSplenocytes (4*10 5) were incubated with 10 µM EdU using the Click -iTTM Plus EdU Flow \nCytometry Assay Kit (Invitrogen, C10632) for 2 h according to the manufacturer´s instructions. \nFollowing EdU labeling, cells were harvested and washed twice with 200 µl FACS-B.  \nCell surface staining was subsequently performed (see section Cell surface staining) using the \nfollowing fluorochrome-conjugated anti -mouse antibodies diluted in FACS -B supplemented \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 25 \nwith Brilliant Stain Buffer (1:4; BD Biosciences, 566349): αCD16/32 (1:200, Biolegend, \n101310), αIgM-FITC (1:200, Biolegend, 406505), αIgD-PerCP/Cy5.5 (1:200, Biolegend, \n405710), αCD19-BV711 (1:400, Biolegend, 115555) , and αB220-BV785 (1:400, Biolegend, \n103246). Following surface staining, cells were fixed, permeabilized, and stained as described \nin section Intracellular staining for flow cytometry  with αpH3-PE (1:200, Biolegend, 650807). \nEdU detection was then performed according to the manufacturer´s inst ructions. Samples \nwere acquired on an LSR II-Fortessa flow cytometer (BD Bioscience) and data were analyzed \nusing FlowJo software (version 10.10.0). \n \nProliferation assay \nSplenocytes (2*106) were labeled with 10 µM Cell Proliferation Dye eFlour TM 450 (Thermo \nFisher, 65-0842-90) according to the manufacturer´s instructions. Briefly, cells were harvested \nand washed twice with 200 µl pre -warmed PBS (centrifugation: 2000 rpm for 2 min at 24°C; \nEppendorf centrifuge 5424R) to remove serum. A 20 µM dye s olution was prepared in pre -\nwarmed PBS and mixed 1:1 with the cell suspension while vortexing. The mixture was \nincubated for 10 min at 37°C in the dark. Labeling was stopped by adding 4 volumes of cold \ncomplete B cell medium (DMEM medium (Sigma-Aldrich, D6429) supplemented with 10% FBS \n(Gibco, 10270106), 2 mM L -glutamine (Sigma -Aldrich, G7513), 50 µM ß -mercaptoethanol \n(Sigma-Aldrich, M3148), 10 mM HEPES (Sigma -Aldrich, H0887), 1 mM sodium pyruvate \n(Gibco, 11360039), 1X nonessential amino acids (Gibco, 11140 035), 100 units/ml penicillin \nand 100 µg/ml streptomycin (Sigma-Aldrich, P0781)), followed by incubation on ice for 5 min. \nAfter three washing steps with 1 ml complete B cell medium (centrifugation: 2000 rpm for  \n2 min at 24°C; Eppendorf centrifuge 5424R),  labeled cells were cultured in 500 µl B cell \nmedium per well in 24 -well plates. Following 48 h of culture, cells were gently resuspended, \ntransferred to a 96 -U-well plate, and stained (see section  Cell surface staining) with 30 µl of \nthe following fluorochrome -conjugated anti -mouse antibodies diluted in FACS -B:  \nαIgD-PerCP/Cy5.5 (1:200, Biolegend, 405710), αIgM-APC (1:1000, Jackson \nImmunoResearch, 115 -607-020), αCD19-BV711 (1:400, Biolegend, 115555) , and  \nαB220-BV785 (1:400, Biolegend, 103246). After cell surface staining, cells were incubated in \n100 µl PBS containing fixable viability dye (1:2000, eBioscience 65-0866-14) for 10 min at 4°C \nand washed with 130 µl FACS-B. Samples were resuspended in FACS-B buffer and acquired \non an LSR II -Fortessa flow cytome ter (BD Bioscience) . Data were analyzed using FlowJo \nSoftware (version 10.10.0). \n \nCell sorting \nSplenocytes were incubated for 15 min at 4°C in 300 µl of the following fluorochrome -\nconjugated anti -mouse antibodies: αB220-FITC (1:200, eBioscience, 553088), αCD25-PE \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 26 \n(1:200, Biolegend, 101904), αIgD-PerCP/Cy5.5 (1:200, Biolegend, 405710), αIgM-APC \n(1:1000, Jackson ImmunoResearch, 115 -607-020), αCD117(c-kit)-BV421 (1:200, Biolegend, \n105828), and αCD19-BV605 (1:200, Biolegend, 115540). Subsequently, 1 ml of FACS-B was \nadded, cells were centrifuged (1500 rpm for 5 min at 4°C; VWR Centrifuge Megastar 1.6R), \nresuspended in 0.5-3 ml FACS-B, and filtered through a 50 µm cell strainer (BD Bioscience, \n340632).  \nCell sorting was performed at 4°C on a FACS Aria III (70 µm nozzle; BD Bioscience). Doublets \nwere excluded based on FSC-H/FSC-W and SSC-H/SSC-W gating. Sorted cell subsets were \ndefined as follows: large pre-B cells (B220+CD19+IgM-IgD-cKit-CD25+FSC-Ahi), IgM- progenitor \nB cells (B220+CD19+IgD-IgM-), IgM+ (immature) B cells (B220+CD19+IgD-IgM+), and tumor cells \n(B220+CD19+IgD-IgM- or B220+CD19+IgD-IgM+). Cells were collected into low -binding tubes \n(Sarstedt, 72706600) containing 300 µl FACS-B. Sorted cells were centrifuged (2200 rpm for \n2.5 min at 4°C; Eppendorf centrifuge 5424R), resuspended in 500 µl PBS, and centrifuged \nagain. Resulting cell pellets wer e either snap -frozen directly or resuspended in 100 µl of \nDNA/RNA Shield (Zymo Research, R1200-25) and stored at -80°C. \n \nRNA extraction, sequencing, and data analysis \nTotal RNA was extracted from frozen samples (pellets or in DNA/RNA Shield) using the Quick-\nRNA Micro Prep Kit (Zymo Research, R1050) with DNase digestion according to the \nmanufacturer´s instructions. Library preparation and bulk RNA sequencing were performed \nusing poly(A) selection, strand -specific library preparation and sequencing on an Illumina® \nNovaSeqTM platform with 2x150 bp paired -end reads, targeting 20 million reads per sample. \nData quality was guaranteed to be ≥85% bases with Q30 or higher. \nRaw FASTQ reads were processed using the nf -core RNA -seq pipeline 95 with the STAR -\nsalmon96 option and aligned to the mouse reference genome GRCm39 (version M30). Genes \nwith names. Starting with “Gm” or “Rp” or ending. With “rik” were not. Considered for the \ndownstream analysis. In addition, genes were required to have at least 5 raw counts in at least \n“M” samples, where “M” was the size of the smallest group considered in the comparison. DEG \nanalysis was performed using the DESeq2 97 (version 1.44.0) package in RStudio (version \n2024.12.0+467). Unless otherwise stated, genes were considered significantly differentially \nexpressed (DEG) if they met the criteria of p-adjusted<0.05 and absolute log2(fold change)>1. \nFor visualization of the B-lineage regulatory/maturation (Fig. 3C) and BCL2 family gene panel \n(Fig. 4B), volcano plots were generated using relaxed significance thresholds (p-adjusted<0.05 \nand absolute log 2(fold change)>0.5) to highlight biologically relevant genes with modest \nexpression changes. Volcano plots displayed all detected genes, except for the tumor cell \ncomparison (Fig. 1F), in which immunoglobulin ( Ig) genes were excluded to improve \nvisualization clarity. \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 27 \nThe B -lineage regulatory/maturation gene panel  comprised 103 genes, including 13 \nupregulated genes ( Blk, Cd5, Cd9, Cd38, Cd44, Cd80, Cd83, Lef1, Myb, Pim2, Rbpj, Syk, \nTcf4), 14 downregulated genes (Aicda, Bach2, Bcl6, Cd40, Cd93, Cebpa, Cr2, Ebf1, Fcer2a, \nIrf8, Lmo2, Slamf6, Slamf1, Zap70) and 76 genes not significantly deregulated ( Actb, Bank1, \nBcl11a, Blnk, Btk, Ccnd3, Ccr7, Cd14, Cd19, Cd22, Cd24a, Cd27, Cd28, Cd48, Cd53, Cd69, \nCd72, Cd74, Cd79a, Cd79b, Cd81, Cd86, Cd200, Cd300a, Cxcr4, E2f1, Ets1, Fcer1g, Foxo1, \nFoxp1, Foxp4, Ikzf1, Ikzf2, Ikzf3, I l7r, Irf4, Itga4, Itgam, Itgb2, Lcp2, Ly9, Lyn, Mafb, Mcl1, \nMs4a1, Myc, Notch2, Pax5, Paxip1, Plcg2, Ppp3cb, Prdm1, Ptpn6, Rag1, Rag2, Rela, Runx1, \nSmad3, Spi1, Spib, Stat5a, Stat6, Tcf3, Tnfrsf13b, Tnfrsf13c, Tnfrsf17, Vav1, Vav2, Vpreb1, \nXbp1, Zbtb7a, Zbtb7b, Zfp386, Zfp423, Zfp521, Zyx).  \nThe BCL2 family gen e panel consisted of 17 genes, including 6 upregulated genes ( Bbc3, \nBcl2, Bcl2l11, Bcl2a1b, Bcl2a1d, Blk ), 1 downregulated gene ( Bmf), and 10 genes not \nsignificantly deregulated (Bcl2l1, Bcl2l2, Mcl1, Bad, Bid, Bak1, Bax, Pmaip1, Hrk, Bik). \nGO-term enrichment analysis for biological processes was performed on DEGs using the \nclusterProfiler package 98 (version 4.12.6 ) in RStudio (version 2024.12.0+467). Enrichment \nanalysis for MSigDB Hallmark genes was conducted using Enrichr 99-101. Transcription factor \nactivity was inferred using the decoupleR package 63 (version 2.9.7) based on CollecTRI \nregulons102 in RStudio (version 2024.12.0+467) with a minimum threshold of 45 target genes \nper transcription factor and a significance cutoff of p-value<0.05. \nImmunoglobulin heavy chain variable (Ighv) transcript usage is based on transcript per million \n(tpm) reads. The relative proportion of each Ighv gene was calculated by normalizing to the \ntotal Ighv transcript abundance (set to 100%). Individual Ighv transcripts exhibiting >4% were \nhighlighted by individual color and gene name in bar plots, with each bar representing a single \nmouse.  \nThe aneuploidy score shown in Figure S1G was  calculated as follows: c hromosome-wide \ncopy-number profiles were inferred from bulk RNA-seq (premalignant and malignant) using the \ngene-dosage signal in differential expression with the R package DESeq297 (version 1.50.2). \nFor each sample/contrast, gene-level log2(fold change) estimates and standard errors (lfcSE) \nwere mapped to genomic coordinates using GRCm38 Ensembl mouse gene annotations \n(chromosomes 1–19 and X). Genes with high uncertainty were removed (lfcSE ≥ 5). To place \nsamples on a comparable baseline, gene-level log2(fold changes) were centered to an inferred \neuploid reference by computing a weighted mean log 2(fold change) per chromosome with \nweights = 1/lfcSE, estimating the mode of the chromoso me-mean distribution by kernel \ndensity, and subtracting this mode from all gene -level values within the sample. Smooth \nchromosome profiles were then obtained by first computing the running mean of weighted fold \nchanges (window size 100 , including partial sizes at both ends) and then fitting weighted \ngeneralized additive models separately for each chromosome and sample using the R \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 28 \npackage mgcv 103 (version 1.9 -3). Model predictions were evaluated on an evenly spaced \ngenomic grid (~10 Mb spacing) to generate genomic tiles, and predicted log2 fold -changes \nwere converted to inferred absolute copy number compared to the diploid baseline. \n \nAI-based literature discovery \nThe Elicit web application (Pro subscription; https://elicit.com; accessed 2026) was used to \nsupport literature discovery by generating topic-focused research reports.  A ll cited sources \nused for the manuscript were manually screened and verified by the authors . All prompts \nrelevant to this manuscript are reported below in full. \n06.02.2026: “Which mouse hematologic malignancy models show that Tet2 loss -of-function \nfacilitates oncogene - or tumor -suppressor–driven disease (accelerates latency and/or \nincreases penetrance) in myeloid and lymphoid lineages (B cell and T cell)? Primary outcomes \n(vs driver-only controls): incidence/penetrance (%) and latency/median survival (time -to-\ndisease). Secondary outcomes (if reported): disease severity/progression (e.g., \ntransformation, grade, leukocytosis/splenomegaly), and transcriptomic changes from RNA-seq \n(e.g., gene set enrichment / pathway signatures / MYC target modules, not just “DE genes”).” \n06.02.2026: “In mouse or human hematopoietic cells and hematologic malignancy models, \nwhat direct functional evidence shows that TET2 loss -of-function impairs tumor -suppressor \nmechanisms in (1) DNA damage/repair, (2) apoptosis, (3) proliferation/cell -cycle control, (4)  \np53 signaling, (5) JAK–STAT signaling, (6) IL-6 signaling, (7) IFN signaling, (8) TNF signaling, \nor (9) NF -κB signaling? Include myeloid, B -cell, and T -cell contexts. Prioritize perturbation -\nbased studies (TET2 loss ± rescue and/or pathway pe rturbation) with functional readouts. \nPrimary outcomes (functional assays): DDR/repair/genomic instability: γH2AX/53BP1 foci, \ncomet, micronuclei, karyotype/WGS/instability metrics Apoptosis/BCL2 -family: BH3 profiling, \nAnnexin V, caspases, mitochondrial pri ming/cytochrome c assays Proliferation/cell -cycle: \nEdU/BrdU, Ki -67, cell -cycle profiling p53 signaling: p53 activation/targets, checkpoint \nresponses, genetic or pharmacologic modulation JAK –STAT / IL -6: pSTAT readouts, IL -6 \ndependence, inhibitor/rescue experiments IFN: ISG induction, IFN stimulation/blockade with \nfunctional consequences TNF / NF -κB: TNF stimulation/blockade, NF-κB activation/readouts \nwith functional consequences” \n04.02.2026: “In the EµMyc mouse model, what peer -reviewed evidence shows that  \nproliferating/tumor-associated IgM ⁺ IgD⁻ B cells are immature B cells (e.g.,  \ntransitional/immature stages) rather than mature activated or memory B cells? Please extract \nthe immunophenotyping panels and gating (e.g., B220,  CD93/AA4.1, CD23, CD21, CD24, \nCD38, GL7, Fas, CD73), tissue location  (BM vs spleen/LN), functional assays \n(transplantation/outgrowth/clonality), and molecular evidence (IgV SHM, AID signatures, BCR \nrearrangements) supporting the authors’ conclusion.” \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 29 \n04.02.2026: “In hematologic cells and tumors (mouse models and human systems), what  \nevidence shows functional or genetic synergy between TET2 loss (KO, LOF  mutation, \nknockdown) and BCL2 -family anti -apoptotic proteins (e.g., BCL2,  BCL-XL/BCL2L1, MCL1, \nBCL-W, A1/BCL2A1)? Focus on B -cell lymphomas, with a special emphasis on the EµMyc \nmouse model. Please extract the  model/system, perturbations, readouts (tumor \nlatency/burden, survival, apoptosis/BH3 profiling, drug sensitivity to BH3 mimetics), and \nwhether the interaction is additive vs synergistic.” \n09.03.2026: “In human and mouse hematologic malignancies, especially MYC -driven B-cell \nmalignancies, which primary studies describe viable BIM -high, apoptosis-primed states that \npersist through buffering by BCL2 family proteins? Focus on premalignant or developmentally \ndefined cell popul ations and on studies using functional mitochondrial apoptosis assays. If \navailable, include evidence from TET2-mutant or TET2-deficient settings.” \n09.03.2026: “In human and mouse hematologic malignancies, especially B-cell malignancies, \nwhich primary studies show that high BCL2 expression or functional BCL2 dependence is \nassociated with increased sensitivity to selective BCL2 inhibition (for example venetoclax/ABT-\n199), particularly in apoptosis-primed cells with elevated BIM or other BH3-only proteins, and \nwhat evidence exists for this relationship in TET2-mutant or TET2-deficient settings?” \n \nQuantification and statistical analysis \nData are presented as median values with interquartile range. Normality of distributions was \nassessed using the Shapiro -Wilk test ( α=0.05). For normally distributed data, statistical \nsignificance was evaluated using one-way ANOVA, two-way ANOVA, or an unpaired t-test, as \nappropriate. For non-normally distributed data, the Mann-Whitney test was applied. The Holm-\nŠidák method was used to correct for multiple comparisons where applicable. Data \nnormalization procedures, when applied, are de scribed in the corresponding Material and \nMethods subsections or in the figure legends. Significance thresholds were defined as *p<0.05, \n**p<0.005, ***p<0.0005, and ****p<0.0001. The specific statistical test and the number of \nbiological replicates (n) for each analysis are indicated in the corresponding figure legends. \nGraphs and statistical analysis were performed using GraphPad Prism (version 10.6.0) and \nMicrosoft Excel (version 16.78), and figures were created using Affinity Designer (version \n1.10.8) and RStudio (version 2024.12.0+467). \n \nACKNOWLEDGEMENTS \nWe thank Emmanuel Derudder, William Olson, Sebastian Herzog and Felix Eichin for insightful \ndiscussions. We are grateful to B. Unterberger, C. Soratroi and I. Gaggl for expert technical \nassistance and M. Saurwein for animal care. The AI-assisted web app tool Elicit was used for \nliterature discovery, the details are provided in the Methods section. This research was funded \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 30 \nin whole by the Austrian Science Fund (FWF) (Grant DOI 10.55776/FG25 , “BCL2 Network \nAdaptations in B Cell Transformation”, to V.L., A.V., J. S.R. and F.F.). M.E. received funding \nfrom Deutsche Forschungsgemeinschaft (DFG) (Grant DOI TRR 353/1 - 471011418 – \n“Regulation of cell death decisions). F.F. was supported by the Austrian Science Fund (FWF) \n(Grant DOI 10.55776/PAT5895324). For open access purposes, the author has applied a CC \nBY public copyright license to any author accepted manuscript version arising from this \nsubmission. The computational results presented here have been achieved in part using the \nLEO HPC infrastructure of the University of Innsbruck. \n \nAUTHOR CONTRIBUTIONS \nConceptualization: V.L., S.S., and A.V. Methodology: S.S., N.K., I.R., J.H., M.S., M.E., J.S.R., \nF.F., and V.L. Investigation: S.S., N.K., I.R., P.Y.P., J.G.W., J.H., K.H . Visualization: S.S. \nSupervision: V.L., J.S.R., A.V., and F.F. Writing – original draft: S.S. and V.L. \n \nCOMPETING INTERESTS \nAll authors declare that they have no competing interests. \n \nDATA AND MATERIALS AVAILABILITY \nAll data needed to evaluate the conclusions in the paper are present in the paper and/or the \nSupplementary Materials. Raw RNA-sequencing data will be deposited in the ArrayExpress \ncollection in BioStudies under the accession number [to be provided upon publication]. All \ncodes used in this study have been previously published and are referenced in the manuscript. \nAdditional data related to this paper are available from the corresponding author upon request.  \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 31 \nLITERATURE \n1 Tahiliani, M.  et al.  Conversion of 5 -methylcytosine to 5 -hydroxymethylcytosine in \nmammalian DNA by MLL partner TET1. Science 324, 930 -935, \ndoi:10.1126/science.1170116 (2009). \n2 Ko, M. et al. Impaired hydroxylation of 5-methylcytosine in myeloid cancers with mutant \nTET2. Nature 468, 839-843, doi:10.1038/nature09586 (2010). \n3 Ito, S.  et al.  Tet proteins can convert 5 -methylcytosine to 5 -formylcytosine and 5 -\ncarboxylcytosine. Science 333, 1300-1303, doi:10.1126/science.1210597 (2011). \n4 Ito, S. et al. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and \ninner cell mass specification. Nature 466, 1129-1133, doi:10.1038/nature09303 (2010). \n5 López-Moyado, I. F., Ko, M., Hogan, P. G. & Rao, A. TET Enzymes in the Immune \nSystem: From DNA Demethylation to Immunotherapy, Inflammation, and Cancer. Annu \nRev Immunol 42, 455-488, doi:10.1146/annurev-immunol-080223-044610 (2024). \n6 Loenarz, C. & Schofield, C. J. Physiological and biochemical aspects of hydroxylations \nand demethylations catalyzed by human 2-oxoglutarate oxygenases. Trends Biochem \nSci 36, 7-18, doi:10.1016/j.tibs.2010.07.002 (2011). \n7 Rasmussen, K. D.  et al.  TET2 binding to enhancers facilitates transcription factor \nrecruitment in hematopoietic cells. Genome Res  29, 564 -575, \ndoi:10.1101/gr.239277.118 (2019). \n8 Lio, C. W. et al. Tet2 and Tet3 cooperate with B-lineage transcription factors to regulate \nDNA modification and chromatin accessibility. Elife 5, doi:10.7554/eLife.18290 (2016). \n9 Hon, G. C.  et al. 5mC oxidation by Tet2 modulates enhancer activity and timing of \ntranscriptome reprogramming during differentiation. Mol Cell  56, 286 -297, \ndoi:10.1016/j.molcel.2014.08.026 (2014). \n10 Sardina, J. L. et al. Transcription Factors Drive Tet2-Mediated Enhancer Demethylation \nto Reprogram Cell Fate. Cell Stem Cell  23, 727 -741.e729, \ndoi:10.1016/j.stem.2018.08.016 (2018). \n11 Orlanski, S.  et al.  Tissue-specific DNA demethylation is required for proper B -cell \ndifferentiation and function. Proc Natl Acad Sci U S A  113, 5018 -5023, \ndoi:10.1073/pnas.1604365113 (2016). \n12 Moran-Crusio, K.  et al.  Tet2 loss leads to increased hematopoietic stem cell self -\nrenewal and myeloid transformation. Cancer Cell  20, 11 -24, \ndoi:10.1016/j.ccr.2011.06.001 (2011). \n13 Ko, M. et al. Ten-Eleven-Translocation 2 (TET2) negatively regulates homeostasis and \ndifferentiation of hematopoietic stem cells in mice. Proc Natl Acad Sci U S A  108, \n14566-14571, doi:10.1073/pnas.1112317108 (2011). \n14 Li, Z. et al. Deletion of Tet2 in mice leads to dysregulated hematopoietic stem cells and \nsubsequent development of myeloid malignancies. Blood 118, 4509 -4518, \ndoi:10.1182/blood-2010-12-325241 (2011). \n15 Quivoron, C. et al. TET2 inactivation results in pleiotropic hematopoietic abnormalities \nin mouse and is a recurrent event during human lymphomagenesis. Cancer Cell 20, \n25-38, doi:10.1016/j.ccr.2011.06.003 (2011). \n16 Mouly, E. et al. B-cell tumor development in Tet2-deficient mice. Blood Adv 2, 703-714, \ndoi:10.1182/bloodadvances.2017014118 (2018). \n17 Pan, F. et al. Tet2 loss leads to hypermutagenicity in haematopoietic stem/progenitor \ncells. Nat Commun 8, 15102, doi:10.1038/ncomms15102 (2017). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 32 \n18 Lio, C. J. et al. TET enzymes augment activation-induced deaminase (AID) expression \nvia 5-hydroxymethylcytosine modifications at the Aicda superenhancer. Sci Immunol 4, \ndoi:10.1126/sciimmunol.aau7523 (2019). \n19 An, J. et al. Acute loss of TET function results in aggressive myeloid cancer in mice. \nNat Commun 6, 10071, doi:10.1038/ncomms10071 (2015). \n20 Guilhamon, P.  et al.  Meta-analysis of IDH -mutant cancers identifies EBF1 as an \ninteraction partner for TET2. Nat Commun 4, 2166, doi:10.1038/ncomms3166 (2013). \n21 Schoeler, K. et al. TET enzymes control antibody production and shape the mutational \nlandscape in germinal centre B cells. Febs j 286, 3566-3581, doi:10.1111/febs.14934 \n(2019). \n22 Tanaka, S. et al. Tet2 and Tet3 in B cells are required to repress CD86 and prevent \nautoimmunity. Nat Immunol 21, 950-961, doi:10.1038/s41590-020-0700-y (2020). \n23 Shukla, V. et al. TET deficiency perturbs mature B cell homeostasis and promotes \noncogenesis associated with accumulation of G-quadruplex and R-loop structures. Nat \nImmunol 23, 99-108, doi:10.1038/s41590-021-01087-w (2022). \n24 Rosikiewicz, W.  et al.  TET2 deficiency reprograms the germinal center B cell \nepigenome and silences genes linked to lymphomagenesis. Sci Adv  6, eaay5872, \ndoi:10.1126/sciadv.aay5872 (2020). \n25 Romanova, E. I. et al. RUNX1/CEBPA Mutation in Acute Myeloid Leukemia Promotes \nHypermethylation and Indicates for Demethylation Therapy. Int J Mol Sci  23, \ndoi:10.3390/ijms231911413 (2022). \n26 Rasmussen, K. D.  et al.  Loss of TET2 in hematopoietic cells leads to DNA \nhypermethylation of active enhancers and induction of leukemogenesis. Genes Dev \n29, 910-922, doi:10.1101/gad.260174.115 (2015). \n27 Huerga Encabo, H.  et al. Human TET2-Mutant Clonal Hematopoiesis Expansion Is \nDriven by Distinct Inflammatory Signaling Responses in Stem Cells Versus Myeloid \nProgeny. Blood Cancer Discov  7, 287 -305, doi:10.1158/2643 -3230.Bcd-25-0070 \n(2026). \n28 Stremenova Spegarova, J.  et al.  Germline TET2 loss of function causes childhood \nimmunodeficiency and lymphoma. Blood 136, 1055 -1066, \ndoi:10.1182/blood.2020005844 (2020). \n29 López-Nevado, M. et al. Novel Germline TET2 Mutations in Two Unrelated Patients \nwith Autoimmune Lymphoproliferative Syndrome -Like Phenotype and Hematologic \nMalignancy. J Clin Immunol 43, 165-180, doi:10.1007/s10875-022-01361-y (2023). \n30 Duployez, N.  et al.  Familial myeloid malignancies with germline TET2 mutation. \nLeukemia 34, 1450-1453, doi:10.1038/s41375-019-0675-6 (2020). \n31 Delhommeau, F. et al. Mutation in TET2 in myeloid cancers. N Engl J Med 360, 2289-\n2301, doi:10.1056/NEJMoa0810069 (2009). \n32 Jankowska, A. M. et al. Loss of heterozygosity 4q24 and TET2 mutations associated \nwith myelodysplastic/myeloproliferative neoplasms. Blood 113, 6403 -6410, \ndoi:10.1182/blood-2009-02-205690 (2009). \n33 Bowman, R. L. & Levine, R. L. TET2 in Normal and Malignant Hematopoiesis. Cold \nSpring Harb Perspect Med 7, doi:10.1101/cshperspect.a026518 (2017). \n34 Lemonnier, F. et al. Recurrent TET2 mutations in peripheral T-cell lymphomas correlate \nwith TFH -like features and adverse clinical parameters. Blood 120, 1466 -1469, \ndoi:10.1182/blood-2012-02-408542 (2012). \n35 Kubuki, Y. et al. TET2 mutation in diffuse large B-cell lymphoma. J Clin Exp Hematop \n56, 145-149, doi:10.3960/jslrt.56.145 (2017). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 33 \n36 Busque, L. et al. Recurrent somatic TET2 mutations in normal elderly individuals with \nclonal hematopoiesis. Nat Genet 44, 1179-1181, doi:10.1038/ng.2413 (2012). \n37 Buscarlet, M.  et al.  DNMT3A and TET2 dominate clonal hematopoiesis and \ndemonstrate benign phenotypes and different genetic predispositions. Blood 130, 753-\n762, doi:10.1182/blood-2017-04-777029 (2017). \n38 Shih, A. H.  et al. Mutational cooperativity linked to combinatorial epigenetic gain of \nfunction in acute myeloid leukemia. Cancer Cell  27, 502 -515, \ndoi:10.1016/j.ccell.2015.03.009 (2015). \n39 Ha, J. S., Jeon, D. S., Kim, J. R., Ryoo, N. H. & Suh, J. S. Analysis of the Ten -Eleven \nTranslocation 2 (TET2) gene mutation in myeloproliferative neoplasms. Ann Clin Lab \nSci 44, 173-179 (2014). \n40 Kunimoto, H.  et al.  Cooperative Epigenetic Remodeling by TET2 Loss and NRAS \nMutation Drives Myeloid Transformation and MEK Inhibitor Sensitivity. Cancer Cell 33, \n44-59.e48, doi:10.1016/j.ccell.2017.11.012 (2018). \n41 Jin, X. et al. Oncogenic N-Ras and Tet2 haploinsufficiency collaborate to dysregulate \nhematopoietic stem and progenitor cells. Blood Adv  2, 1259 -1271, \ndoi:10.1182/bloodadvances.2018017400 (2018). \n42 Chen, E. et al. Distinct effects of concomitant Jak2V617F expression and Tet2 loss in \nmice promote disease progression in myeloproliferative neoplasms. Blood 125, 327-\n335, doi:10.1182/blood-2014-04-567024 (2015). \n43 Patnaik, M. M.  et al. Prognostic interaction between ASXL1 and TET2 mutations in \nchronic myelomonocytic leukemia. Blood Cancer J 6, e385, doi:10.1038/bcj.2015.113 \n(2016). \n44 Kim, M.  et al.  Mutation in TET2 or TP53 predicts poor survival in patients with \nmyelodysplastic syndrome receiving hypomethylating treatment or stem cell \ntransplantation. Bone Marrow Transplant  50, 1132 -1134, doi:10.1038/bmt.2015.110 \n(2015). \n45 Ramdas, B. et al. Combined heterozygosity of FLT3 ITD, TET2, and DNMT3A results \nin aggressive leukemia. JCI Insight 7, doi:10.1172/jci.insight.162016 (2022). \n46 Zhang, P.  et al.  TP53 mutations and TET2 deficiency cooperate to drive \nleukemogenesis and establish an immunosuppressive environment. J Clin Invest 135, \ndoi:10.1172/jci184021 (2025). \n47 Dominguez, P. M. et al. TET2 Deficiency Causes Germinal Center Hyperplasia, Impairs \nPlasma Cell Differentiation, and Promotes B-cell Lymphomagenesis. Cancer Discov 8, \n1632-1653, doi:10.1158/2159-8290.Cd-18-0657 (2018). \n48 Dalla-Favera, R. et al. Human c-myc onc gene is located on the region of chromosome \n8 that is translocated in Burkitt lymphoma cells. Proc Natl Acad Sci U S A  79, 7824-\n7827, doi:10.1073/pnas.79.24.7824 (1982). \n49 Ladanyi, M., Offit, K., Jhanwar, S. C., Filippa, D. A. & Chaganti, R. S. MYC \nrearrangement and translocations involving band 8q24 in diffuse large cell lymphomas. \nBlood 77, 1057-1063 (1991). \n50 Laurenti, E. et al. Hematopoietic stem cell function and survival depend on c-Myc and \nN-Myc activity. Cell Stem Cell 3, 611-624, doi:10.1016/j.stem.2008.09.005 (2008). \n51 Habib, T.  et al.  Myc stimulates B lymphocyte differentiation and amplifies calcium \nsignaling. J Cell Biol 179, 717-731, doi:10.1083/jcb.200704173 (2007). \n52 Duffy, M. J., Tang, M. & Crown, J. MYC as a Target for Cancer Treatment: from \nUndruggable to Druggable? Target Oncol  20, 791 -801, doi:10.1007/s11523 -025-\n01169-x (2025). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 34 \n53 McMahon, S. B. MYC and the control of apoptosis. Cold Spring Harb Perspect Med 4, \na014407, doi:10.1101/cshperspect.a014407 (2014). \n54 Sander, S.  et al.  Synergy between PI3K signaling and MYC in Burkitt \nlymphomagenesis. Cancer Cell 22, 167-179, doi:10.1016/j.ccr.2012.06.012 (2012). \n55 Adams, J. M. et al. The c-myc oncogene driven by immunoglobulin enhancers induces \nlymphoid malignancy in transgenic mice. Nature 318, 533-538, doi:10.1038/318533a0 \n(1985). \n56 Egle, A., Harris, A. W., Bouillet, P. & Cory, S. Bim is a suppressor of Myc -induced \nmouse B cell leukemia. Proc Natl Acad Sci U S A  101, 6164 -6169, \ndoi:10.1073/pnas.0401471101 (2004). \n57 Lefebure, M. et al. Genomic characterisation of Eμ-Myc mouse lymphomas identifies \nBcor as a Myc co -operative tumour -suppressor gene. Nat Commun  8, 14581, \ndoi:10.1038/ncomms14581 (2017). \n58 Ma, A. et al. Surface IgM mediated regulation of RAG gene expression in E mu-N-myc \nB cell lines. Embo j 11, 2727-2734, doi:10.1002/j.1460-2075.1992.tb05338.x (1992). \n59 Saintamand, A., Garot, A., Saad, F., Moulinas, R. & Denizot, Y. Pre -germinal center \norigin for mature mouse B cell lymphomas: a major discrepancy with human mature \nlymphomas. Cell Cycle 14, 3656-3658, doi:10.1080/15384101.2015.1093708 (2015). \n60 Palam, L. R.  et al.  Loss of epigenetic regulator TET2 and oncogenic KIT regulate \nmyeloid cell transformation via PI3K pathway. JCI Insight  3, \ndoi:10.1172/jci.insight.94679 (2018). \n61 Kafer, G. R. et al. 5-Hydroxymethylcytosine Marks Sites of DNA Damage and Promotes \nGenome Stability. Cell Rep 14, 1283-1292, doi:10.1016/j.celrep.2016.01.035 (2016). \n62 Maifrede, S.  et al.  TET2 and DNMT3A Mutations Exert Divergent Effects on DNA \nRepair and Sensitivity of Leukemia Cells to PARP Inhibitors. Cancer Res 81, 5089-\n5101, doi:10.1158/0008-5472.Can-20-3761 (2021). \n63 Badia, I. M. P. et al. decoupleR: ensemble of computational methods to infer biological \nactivities from omics data. Bioinform Adv  2, vbac016, doi:10.1093/bioadv/vbac016 \n(2022). \n64 Loder, F. et al. B cell development in the spleen takes place in discrete steps and is \ndetermined by the quality of B cell receptor -derived signals. J Exp Med  190, 75-89, \ndoi:10.1084/jem.190.1.75 (1999). \n65 Melamed, D., Kench, J. A., Grabstein, K., Rolink, A. & Nemazee, D. A functional B cell \nreceptor transgene allows efficient IL-7-independent maturation of B cell precursors. J \nImmunol 159, 1233-1239 (1997). \n66 Srinivasan, L. et al. PI3 kinase signals BCR-dependent mature B cell survival. Cell 139, \n573-586, doi:10.1016/j.cell.2009.08.041 (2009). \n67 Brunet, A. et al. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead \ntranscription factor. Cell 96, 857-868, doi:10.1016/s0092-8674(00)80595-4 (1999). \n68 Maclean, K. H., Keller, U. B., Rodriguez -Galindo, C., Nilsson, J. A. & Cleveland, J. L. \nc-Myc augments gamma irradiation-induced apoptosis by suppressing Bcl-XL. Mol Cell \nBiol 23, 7256-7270, doi:10.1128/mcb.23.20.7256-7270.2003 (2003). \n69 Sarosiek, K. A. et al. Developmental Regulation of Mitochondrial Apoptosis by c -Myc \nGoverns Age- and Tissue-Specific Sensitivity to Cancer Therapeutics. Cancer Cell 31, \n142-156, doi:10.1016/j.ccell.2016.11.011 (2017). \n70 Kelly, P. N., Puthalakath, H., Adams, J. M. & Strasser, A. Endogenous bcl -2 is not \nrequired for the development of Emu-myc-induced B-cell lymphoma. Blood 109, 4907-\n4913, doi:10.1182/blood-2006-10-051847 (2007). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 35 \n71 Sidman, C. L., Shaffer, D. J., Jacobsen, K., Vargas, S. R. & Osmond, D. G. Cell \npopulations during tumorigenesis in Eu -myc transgenic mice. Leukemia 7, 887 -895 \n(1993). \n72 Prasad, V. S. et al. Upregulation of endogenous p53 and induction of in vivo apoptosis \nin B -lineage lymphomas of E(mu) -myc transgenic mice by deregulated c -myc \ntransgene. Mol Carcinog 18, 66-77, doi:10.1002/(sici)1098-2744(199702)18:2<66::aid-\nmc2>3.0.co;2-o (1997). \n73 Kikuchi, K., Lai, A. Y., Hsu, C. L. & Kondo, M. IL -7 receptor signaling is necessary for \nstage transition in adult B cell development through up -regulation of EBF. J Exp Med \n201, 1197-1203, doi:10.1084/jem.20050158 (2005). \n74 Funakoshi-Tago, M., Sumi, K., Kasahara, T. & Tago, K. Critical roles of Myc-ODC axis \nin the cellular transformation induced by myeloproliferative neoplasm-associated JAK2 \nV617F mutant. PLoS One 8, e52844, doi:10.1371/journal.pone.0052844 (2013). \n75 Puissant, A. et al. SYK is a critical regulator of FLT3 in acute myeloid leukemia. Cancer \nCell 25, 226-242, doi:10.1016/j.ccr.2014.01.022 (2014). \n76 Zhang, P. et al. Chromatin regulator Asxl1 loss and Nf1 haploinsufficiency cooperate \nto accelerate myeloid malignancy. J Clin Invest 128, 5383-5398, doi:10.1172/jci121366 \n(2018). \n77 Poole, C. J., Lodh, A., Choi, J. H. & van Riggelen, J. MYC deregulates TET1 and TET2 \nexpression to control global DNA (hydroxy)methylation and gene expression to \nmaintain a neoplastic phenotype in T -ALL. Epigenetics Chromatin  12, 41, \ndoi:10.1186/s13072-019-0278-5 (2019). \n78 Chen, L. L. et al. SNIP1 Recruits TET2 to Regulate c-MYC Target Genes and Cellular \nDNA Damage Response. Cell Rep  25, 1485 -1500.e1484, \ndoi:10.1016/j.celrep.2018.10.028 (2018). \n79 Qiu, Z. et al. MYC Regulation of D2HGDH and L2HGDH Influences the Epigenome \nand Epitranscriptome. Cell Chem Biol  27, 538 -550.e537, \ndoi:10.1016/j.chembiol.2020.02.002 (2020). \n80 Lin, A. P.  et al.  MYC, mitochondrial metabolism and O -GlcNAcylation converge to \nmodulate the activity and subcellular localization of DNA and RNA demethylases. \nLeukemia 36, 1150-1159, doi:10.1038/s41375-021-01489-7 (2022). \n81 de Barrios, O., Meler, A. & Parra, M. MYC's Fine Line Between B Cell Development \nand Malignancy. Cells 9, doi:10.3390/cells9020523 (2020). \n82 Nguyen, L., Papenhausen, P. & Shao, H. The Role of c -MYC in B-Cell Lymphomas: \nDiagnostic and Molecular Aspects. Genes (Basel)  8, doi:10.3390/genes8040116 \n(2017). \n83 von Wnuck Lipinski, K.  et al.  Hepatocyte Nuclear Factor 1A Is a Cell -Intrinsic \nTranscription Factor Required for B Cell Differentiation and Development in Mice. J \nImmunol 196, 1655-1665, doi:10.4049/jimmunol.1500897 (2016). \n84 Fischer, U. et al. Cell Fate Decisions: The Role of Transcription Factors in Early B-cell \nDevelopment and Leukemia. Blood Cancer Discov  1, 224 -233, doi:10.1158/2643 -\n3230.Bcd-20-0011 (2020). \n85 Kollmann, S. et al. A STAT5B-CD9 axis determines self-renewal in hematopoietic and \nleukemic stem cells. Blood 138, 2347-2359, doi:10.1182/blood.2021010980 (2021). \n86 Certo, M. et al. Mitochondria primed by death signals determine cellular addiction to \nantiapoptotic BCL -2 family members. Cancer Cell  9, 351 -365, \ndoi:10.1016/j.ccr.2006.03.027 (2006). \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint \n\n 36 \n87 Ni Chonghaile, T.  et al.  Pretreatment mitochondrial priming correlates with clinical \nresponse to cytotoxic chemotherapy. Science 334, 1129 -1133, \ndoi:10.1126/science.1206727 (2011). \n88 Sarosiek, K. A., Ni Chonghaile, T. & Letai, A. Mitochondria: gatekeepers of response \nto chemotherapy. Trends Cell Biol 23, 612-619, doi:10.1016/j.tcb.2013.08.003 (2013). \n89 Montero, J.  et al.  Drug-induced death signaling strategy rapidly predicts cancer \nresponse to chemotherapy. Cell 160, 977-989, doi:10.1016/j.cell.2015.01.042 (2015). \n90 Labi, V., Grespi, F., Baumgartner, F. & Villunger, A. Targeting the Bcl -2-regulated \napoptosis pathway by BH3 mimetics: a breakthrough in anticancer therapy? Cell Death \nDiffer 15, 977-987, doi:10.1038/cdd.2008.37 (2008). \n91 Brinkmann, K., Ng, A. P., de Graaf, C. A. & Strasser, A. What can we learn from mice \nlacking pro-survival BCL-2 proteins to advance BH3 mimetic drugs for cancer therapy? \nCell Death Differ 29, 1079-1093, doi:10.1038/s41418-022-00987-0 (2022). \n92 Vogler, M. et al. The BCL2 family: from apoptosis mechanisms to new advances in \ntargeted therapy. Signal Transduct Target Ther  10, 91, doi:10.1038/s41392 -025-\n02176-0 (2025). \n93 Abegunde, S. O., Buckstein, R., Wells, R. A. & Rauh, M. J. An inflammatory \nenvironment containing TNFα favors Tet2-mutant clonal hematopoiesis. Exp Hematol \n59, 60-65, doi:10.1016/j.exphem.2017.11.002 (2018). \n94 Fraser, C., Ryan, J. & Sarosiek, K. BH3 Profiling: A Functional Assay to Measure \nApoptotic Priming and Dependencies. Methods Mol Biol 1877, 61-76, doi:10.1007/978-\n1-4939-8861-7_4 (2019). \n95 Ewels, P. A.  et al.  The nf -core framework for community -curated bioinformatics \npipelines. Nat Biotechnol 38, 276-278, doi:10.1038/s41587-020-0439-x (2020). \n96 Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast \nand bias -aware quantification of transcript expression. Nat Methods  14, 417 -419, \ndoi:10.1038/nmeth.4197 (2017). \n97 Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion \nfor RNA-seq data with DESeq2. Genome Biol 15, 550, doi:10.1186/s13059-014-0550-\n8 (2014). \n98 Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing \nbiological themes among gene clusters. Omics 16, 284 -287, \ndoi:10.1089/omi.2011.0118 (2012). \n99 Chen, E. Y.  et al. Enrichr: interactive and collaborative HTML5 gene list enrichment \nanalysis tool. BMC Bioinformatics 14, 128, doi:10.1186/1471-2105-14-128 (2013). \n100 Kuleshov, M. V.  et al. Enrichr: a comprehensive gene set enrichment analysis web \nserver 2016 update. Nucleic Acids Res 44, W90-97, doi:10.1093/nar/gkw377 (2016). \n101 Xie, Z.  et al.  Gene Set Knowledge Discovery with Enrichr. Curr Protoc  1, e90, \ndoi:10.1002/cpz1.90 (2021). \n102 Müller-Dott, S. et al. Expanding the coverage of regulons from high -confidence prior \nknowledge for accurate estimation of transcription factor activities. Nucleic Acids Res \n51, 10934-10949, doi:10.1093/nar/gkad841 (2023). \n103 Wood, S. N., Pya, N. & Säfken, B. Smoothing Parameter and Model Selection for \nGeneral Smooth Models. Journal of the American Statistical Association  111, 1548-\n1563, doi:10.1080/01621459.2016.1180986 (2016). \n \n \n \n \n.CC-BY 4.0 International licenseperpetuity. It is made available under a \npreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in \nThe copyright holder for thisthis version posted March 21, 2026. ; https://doi.org/10.64898/2026.03.20.712678doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}