Transposable elements shape stemness in normal and leukemic hematopoiesis Authors: Giacomo Grillo, Bettina Nadorp, Aditi Qamra, Bryce Drylie, Amanda Mitchell, Christopher Arlidge, Ankita Nand, Naoya Takayama, Alex Murison, Seyed Ali Madani Tonekaboni, Komaldeep Kaur Kang, Andrea Arruda, Jean C Y Wang, Mark D Minden, Özgen Deniz, Héléna Boutzen, John E Dick, Mathieu Lupien Published: 2026 ABSTRACT AbstractDespite most acute myeloid leukemia (AML) patients achieving complete remission after induction chemotherapy, two-thirds relapse within 5 years. AML follows a cellular hierarchy sustained by leukemia stem cells (LSCs), which drive tumor progression and relapse. Little is known about the genetic determinants driving LSCs stemness properties. By identifying chromatin variants from accessibility measurements across LSCs, hematopoietic stem cells and downstream progeny, we identified transposable elements (TEs) as genetic determinants of primitive versus mature populations. Accessibility at 121 TE subfamilies distinguished LSCs from mature leukemic cells and stratified AML patients by stemness and survival. Functional assays revealed that these TE subfamilies serve as docking sites for genome topology regulators or lineage-specific transcription factors, including LYL1 in LSCs. Chromatin editing established the necessity of accessibility at LTR12C elements to maintain LSC stemness. Thus, TEs regulate primitive versus mature cell states, with distinct subfamilies underlying stemness in normal versus leukemic stem cells.Subject terms: Cancer stem cells, Epigenomics Main Lifelong blood production originates from self-renewing, multipotent hematopoietic stem cells (HSCs) that give rise to a hierarchy of progenitor and mature hematopoietic populations1. Phenotypic heterogeneity across this hierarchy relies on population-specific genetic architectures, where distinct populations draw on different sequences of a shared genome2–5. Comparing chromatin accessibility across hematopoietic populations identified chromatin variants, defined as DNA elements with population-specific chromatin states6, that revealed how CTCF reshapes the genome topology to govern dormant, low-primed HSCs’ transition into activated, primed states7. By contrast, chromatin variants specific to mature populations serve as cis-regulatory elements by providing binding sites for lineage-specific transcription factors to promote lineage-specific expression profiles7,8. Akin to normal hematopoiesis, acute myeloid leukemia (AML) is hierarchically organized9, with leukemia stem cells (LSCs) at the apex10,11. LSCs self-renew and sustain the long-term generation of malignant cells and represent the source of evolutionary diversity, driving leukemic progression, therapy failure and relapse12–14. Two observations underscore the clinical relevance of LSCs. First, a LSC-specific gene expression signature (LSC17) predicts overall survival and response to induction therapy15,16. Second, mutational tracing in paired diagnosis/relapse AML samples showed genetically diverse, LSC-driven subclones already present at diagnosis, some fated to survive treatment and fuel relapse13. Links between HSCs and LSCs are also supported by clonal hematopoiesis mutations that arise early in HSCs, creating a preleukemic reservoir that can evolve into LSCs and fulminant AML13,17,18. Although nongenetic events have been broadly linked to cancer stem cell properties12,19,20, the specific chromatin-based events that drive HSC-to-LSC transformation remain poorly defined, hindering our ability to explain cancer stemness and devise therapies to selectively target LSCs. To date, searches for chromatin variants underlying phenotypic heterogeneity in normal hematopoiesis and AML have mainly focused on the nonrepetitive genome, excluding transposable elements (TEs)21,22. Yet, TEs have diverse roles in development and pathogenesis23. TEs are classified into subfamilies composed of hundreds to thousands of individual elements traced back to a common ancestral DNA sequence24,25. In AML, as in many other cancer types, specific TE subfamilies are repressed to prevent activation of the viral mimicry response otherwise activated by the double-stranded RNAs they encode26–28. Other TE subfamilies lie in cancer-associated chromatin variants to function as cis-regulatory elements29,30, to potentiate the expression of neighboring genes and/or impact the three-dimensional genome organization by recruiting sequence-specific transcription factors, including CTCF31–33. This results in TE-mediated oncogene expression in AML, as in various other cancer states30,32,34. Despite these roles, whether TEs regulate stemness in LSCs remains unknown. Here, we show how distinct TE subfamilies are enriched in chromatin variants across both normal hematopoietic and LSC-driven leukemic hierarchies, where they modulate stemness properties. Results Unique TEs lie in chromatin variants of HSPC versus mature cells Because chromatin accessibility enables the identification of chromatin variants that define cell-state identity (Extended Data Fig. 1a)6–8,35,36, we systematically studied TE subfamily enrichment within accessible chromatin across normal human hematopoietic populations. For this, we analyzed assay for transposase-accessible chromatin sequencing (ATAC–seq; Methods) data from six highly purified hematopoietic stem and progenitor populations (HSPC), including long-term (LT) and short-term (ST) HSCs, common myeloid progenitors, megakaryocyte-erythroid progenitors, multilymphoid progenitors (MLP) and granulocyte-monocyte progenitors, as well as seven mature myeloid and lymphoid fated populations isolated from human cord blood donors (Extended Data Fig. 1b)7. These populations were isolated using our sorting scheme (Methods), which achieves the highest purity compared to other studies8 and separates HSCs with distinct repopulation and self-renewal potential (~30%) for improved resolution at the apex of the normal hematopoietic hierarchy (Methods). Chromatin accessibility at specific TE subfamilies distinguishes HSPC from mature hematopoietic cells. a) Unsupervised hierarchical clustering of hematopoietic cell states based on the presence/absence of ATAC peaks across repetitive and non-repetitive DNA sequences. b) Schematic representation of hematopoietic population and cell surface markers used for sorting. c) Heatmap of differential accessibility s of TE subfamilies with q < 0.01 across bulk ATAC-seq profiles comparing megakaryocyte-erythroid lineage and all other hematopoietic samples. d) Heatmap displaying differential accessibility Z-scores of all TE subfamilies across bulk ATAC-seq profiles of HSCs, progenitors and mature hematopoietic populations excluding megakaryocyte-erythroid lineage TE subfamilies enriched in accessible chromatin from c. e) Volcano plot showing the median difference in accessibility Z-scores for each TE subfamily between the stem/progenitor and differentiated hematopoietic cell populations versus the −log10 q-value for that difference. f) Examples of non-stem and stem-enriched families of the ERV1, ERV3, SINE1 and SINE2 TE families. Boxplots show differential accessibility Z-scores in the different hematopoietic populations. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5 × interquartile range (IQR). The TE subfamilies shown here were found differentially enriched between HSPC versus mature hematopoietic populations (Benjamini-Hochberg corrected two-sided Wilcoxon signed rank test, q < 0.01). g) ATAC-seq signal profile across all LTR33 elements accessible in HSPC and mature hematopoietic populations. h) Violin plots for the ATAC-seq signal over LTR33 elements accessible in HSPC (n = 18) and mature (n = 16) hematopoietic populations (left). Violin plot for the difference of ATAC-seq signal between HSPC and mature hematopoietic populations over LTR33 elements (each dot corresponds to one LTR33 element—right). Violin plots show the median value of the distributions. P-value results of two-sided paired Wilcoxon test are showcased. i) Number of LTR33 elements accessible in HSPC and mature hematopoietic populations. P-value results of two-sided Student’s t-test are showcased. j) Same as panel g but for LTR50 elements. k) Same as panel h but for LTR50 elements in HSPC (n = 18) and mature (n = 16) hematopoietic populations. Violin plots show the median value of the distributions. p-value results of two-sided paired Wilcoxon test are showcased. l) Same as panel i but for LTR50 elements. P-value results of two-sided Student’s t-test are showcased. m) Number of called peaks in the different hematopoietic cell populations, note that stem (n = 6) and progenitor cells (n = 12) have a higher number of called peaks than mature (n = 16) hematopoietic cells. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). Panel b created in BioRender; Lupien, M. https://biorender.com/3exlppo (2026). We quantified enrichment of all TE subfamilies (n = 971, hg38; Methods) within accessible chromatin using ChromVAR (Methods), substituting TE subfamily elements for transcription factor DNA motif positions (Methods) while controlling for GC content. Unsupervised clustering of TE accessibility z scores revealed that LT-, ST-HSCs and later progenitor populations, collectively referred to as HSPC, share similar TE enrichment profiles and differ from mature populations (Fig. 1a). The megakaryocyte-erythroid lineage is the exception, with megakaryocyte-erythroid progenitors clustered with erythroid precursors (Fig. 1a), consistent with strong lineage-specific enrichment of TE subfamilies. Indeed, removing megakaryocyte-erythroid lineage-specific TE subfamilies (Benjamini–Hochberg corrected P value (q) < 0.0001; Extended Data Fig. 1c) before clustering yielded a clear HSPC versus mature populations separation (Extended Data Fig. 1d). Thus, chromatin accessibility at specific TE subfamilies distinguishes HSPC from mature hematopoietic cells. Distinct TE subfamilies populate the accessible chromatin of primitive and mature hematopoietic cells. a, Heatmap displaying differential accessibility z scores of all TE subfamilies across bulk ATAC–seq profiles of HSC, progenitor and mature hematopoietic populations; note that HSC and progenitors show similar TE subfamily enrichment profiles. b, Association of TE families enriched in accessible chromatin across HSC and progenitor (‘HSPC’) versus mature cells. Y axis lists TE families with enriched TE subfamilies (each dot denotes a different TE subfamily) and x axis presents the median z score difference between HSPC and mature cells (left panel). Number of enriched TE subfamilies in HSPC or mature cells, across TE families (right). TE families are ordered from the most common in HSPC to mature cells. c, Examples of TE subfamilies differentially enriched in accessible chromatin in HSPC and mature cells. Boxplots show differential accessibility z scores in hematopoietic stem, progenitor and mature populations. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. The TE subfamilies shown here were found differentially enriched between HSPC versus mature hematopoietic populations (Benjamini–Hochberg-corrected two-sided Wilcoxon signed-rank test, q < 0.01). d, Genome browser view of ATAC–seq signal from hematopoietic populations for one element of each HSPC-enriched TE subfamily shown in c. e, Transcription factor cistromes enrichment at 2 of 23 TE subfamilies enriched in stem populations (HSCs). Every dot corresponds to a transcription factor cistrome. Red dashed lines correspond to −log10(q) = 1.3 (q = 0.05) and log(OR) = 1 thresholds. The q values correspond to FDR-corrected Fisher’s exact test. f, Individual transcription factors (left), including HOXA9, or three-dimensional genome organization factor (right) cistromes enriched over HSC-specific TE subfamilies (according to e). Boxplots showing enrichment GIGGLE scores of individual cistromes profiled in cell lines derived from a variety of tissue of origin. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. Cistromes showcased here were found significantly enriched over the TE subfamilies of interest (Fisher’s exact two-tailed, P < 0.05). g, Enrichment of HOXA9 and CTCF DNA motifs within TE subfamilies enriched in HSCs. Bars represent −log10(q) for each TE subfamily for transcription factor DNA motifs. The red dashed line corresponds to −log10(q) = 1.3 (q = 0.05) threshold. The q values correspond to Benjamini–Hochberg corrected, one-sided hypergeometric test, P values. CMP, common myeloid progenitors; MEP, megakaryocyte-erythroid progenitors; MLP, myeloid-lymphoid progenitors; GMP, granulocyte-monocyte progenitor; IQR, interquartile range; NS, not significant. To identify TE subfamilies differentially enriched between HSPC and mature populations, we compared TE accessibility z scores using a two-sided Wilcoxon signed-rank test. We found 81 and 128 TE subfamilies belonging to 12 families enriched in HSPC and mature populations, respectively (q < 0.01; Fig. 1b, Extended Data Fig. 1e and Supplementary Tables 1–3). HSPCs were mainly enriched for subfamilies part of ERV1 and ERV3 families (Fig. 1b), including MER61A (q = 1.10 × 10−5), LTR39 (q = 1.87 × 10−6), LTR16E1 (q = 4.61 × 10−7), LTR33 (q = 9.05 × 10−7) and LTR50 (q = 4.22 × 10−5; Fig. 1c and Extended Data Fig. 1f–l), exemplified at unique elements (Fig. 1d). This aligns with previous studies of regulatory activity for LTR16E1 and LTR33 in pluripotent stem cells37. In contrast, mature populations were mainly enriched for SINE1, SINE2 and transposon subfamily members (Fig. 1b), such as AluJB (q = 1.26 × 10−7) and AluSx (q = 1.08 × 10−7), MIR (q = 2.93 × 10−7) and MIR3 (q = 2.93 × 10−7; Fig. 1c and Extended Data Fig. 1f). Although mature populations had similar numbers of accessible DNA elements (peaks; Extended Data Fig. 1m), the enrichment of SINE1, SINE2 and transposon TE subfamilies relied on population-specific accessible elements. In contrast, HSPC shared accessibility at the same ERV1 and ERV3 subfamily elements across stem and progenitor populations. Together, our study reveals distinct TE subfamily accessibility patterns in HSPCs versus mature populations, supporting a role for TE-associated chromatin variants in hematopoietic cell-state identity. Accessible TEs are docking sites for stem factors in HSPC We next assessed if chromatin variants over TE subfamilies could distinguish LT-HSCs (the most primitive HSC subset) from other hematopoietic populations (Extended Data Fig. 1b). One-to-one comparison of TE subfamily accessibility z scores, such as LT-HSCs versus ST-HSCs, did not identify differentially-enriched TE subfamilies (Supplementary Table 4). We next compared stem versus mature populations, revealing 61 TE subfamilies enriched in stem versus 119 in mature populations (q < 0.01; Extended Data Fig. 2a and Supplementary Table 5). Progenitor versus mature populations identified 56 TE subfamilies enriched in progenitor and 89 in mature populations (q < 0.01; Supplementary Table 6). Among TE subfamilies distinguishing HSPC from mature populations (Fig. 1a), 23 were specific to stem populations (Extended Data Fig. 2b,c). Of these, only LTR39-int, LTR85a, MER61D, MLT1E2, HERV4_I-int, and UCON26 were enriched in stem relative to progenitor populations (q < 0.05; Extended Data Fig. 2d and Supplementary Table 7). Notably, MER4B-int, MER57B1, MER57B2, MER57-int, LTR16A2, LTR67B, and LTR85a showed the highest z scores within LT/HSPC and/or Act/HSPC chromatin signatures that discriminate between LT-HSC, ST-HSC and other progenitors7 (Supplementary Table 8). Collectively, chromatin variants over specific TE subfamilies separate HSPC from mature populations, with 23 being stem specific. Stem-specific TE subfamilies provide docking sites for ‘stem’ transcription factors and regulators of genome topology. a) Heatmap of differential accessibility Z-scores of TE subfamilies with q < 0.01 across bulk ATAC-seq profiles comparing stem and all mature hematopoietic samples. b) UpSet plot showing the intersection of TE subfamilies enriched in stem or progenitor hematopoietic populations; please note that 23 TE subfamilies are enriched exclusively in stem populations. c) 23 TE subfamilies enriched exclusively in stem samples (Benjamini-Hochberg corrected two-sided Wilcoxon signed rank test, q < 0.01). Boxplots show differential accessibility Z-scores in the stem (n = 6), progenitor (n = 12) and mature (n = 16) cell state groups. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5 × interquartile range (IQR). d) Heatmap of differential accessibility Z-scores of TE subfamilies with q < 0.05 across bulk ATAC-seq profiles comparing stem and progenitor hematopoietic samples. e) Transcription factor cistromes enriched at MER57B2 (one of the 23 TE subfamilies enriched in stem populations (HSCs)). Every dot corresponds to a transcription factor cistrome taken into account in the analysis. Red dashed lines correspond to -log10(q)=1.3 (q = 0.05) and logOddsRatio(logOR)=1 thresholds. q-values correspond to FDR-corrected Fisher exact test. f) HOXA9 mRNA expression in stem (n = 6), progenitor (n = 12) and mature (n = 16) populations. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5 × interquartile range (IQR). p-value results of two-sided Wilcoxon test are showcased. g) Same as panel e but for MLT1E2. h) RUNX1 mRNA expression in stem (n = 6), progenitor (n = 12) and mature (n = 16) populations. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5 × interquartile range (IQR). P-value results of two-sided Wilcoxon test are showcased. i) Same as panel e but for LTR16A2. j–l) Enrichment GIGGLE score for cistromes enriched over the HSCs specific TE subfamilies MER57B2 (j), MLT1E2 (k) and LTR16A2 (l) identifying HOXA9 (j), RUNX1 (k) or three-dimensional genome organization factors (l). Individual cistromes profiles are from cell lines derived from a variety of tissues of origin. Cistromes showcased here were found significantly enriched over the TE subfamilies of interest (Fisher’s Exact two-tailed p < 0.05). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). Because TE subfamilies harbor DNA motifs for transcription factors30,31,38–41, we used the ReMap atlas of 485 cistromes (Methods) to assess if the 23 stem-specific TE subfamilies could serve as docking sites. HOXA9 binding to the chromatin was enriched over MER57B1 (q = 2.88 × 10−13, log(OR) = 27.69) and MER57B2 (q = 0.0007, log(OR) = 11.95) elements (Fig. 1e and Extended Data Fig. 2e). HOXA9 is a canonical HSC factor42, and more highly expressed in stem than mature populations (Extended Data Fig. 2f). The cistrome of RUNX1, another known HSC transcription factor43, was found enriched over MLT1E2 elements (q = 0.03, log(OR) = 2.12; Extended Data Fig. 2g), despite being expressed at similar levels in stem and mature populations (Extended Data Fig. 2h). While many stem-specific TE subfamilies showed no enrichment for any of the 485 cistromes in ReMap (HERV4_I-int, HERVK9-int, LTR80A, LTR85a, LTR9A1, MamGypLTR1b, MER127, MER4B-int, MER57-int, MER83B-int, MLT2E and PRIMA41-int), CTCF and cohesin complex factors (SMC1A2, SMC3 and RAD21), known regulators of genome topology44, were enriched over MamGypsy2-LTR and LTR16A2 elements (Fig. 1e, Extended Data Fig. 2i and Supplementary Table 9). We independently validated these observations using GIGGLE enrichment score45, confirming HOXA9 enrichment over MER57B1 and MER57B2 (Fig. 1f and Extended Data Fig. 2j), RUNX1 enrichment over MLT1E2 elements (Extended Data Fig. 2k) and CTCF/cohesin enrichment over MamGypsy2-LTR and LTR16A2 elements (Fig. 1f and Extended Data Fig. 2l). Sequence analysis further supported these findings by identifying enrichment of HOXA9 or CTCF DNA motifs within MER57B1 and MamGypsy2-LTR elements, respectively (Fig. 1g). Collectively, these results suggest that TE subfamilies uniquely accessible in stem populations provide docking sites for ‘stem’ transcription factors and regulators of genome topology. LSCs show HSPC-like TE enrichment in accessible chromatin Considering the role of LSCs in AML, we tested whether chromatin variants over TE subfamilies could discriminate LSC+ from LSC− leukemic fractions, and from normal HSPC and mature hematopoietic populations. Uncultured AML patient samples (n = 15) were sorted into fractions based on CD34 and CD38 expression, and fractions were functionally classified as LSC+ (n = 11) or LSC− (n = 24) based on leukemic engraftment upon xenotransplantation into immunodeficient mice15,46. In parallel, we performed ATAC–seq on these fractions and measured TE subfamily enrichment within accessible chromatin. Unsupervised clustering of TE accessibility z scores across AML fractions and normal hematopoietic populations grouped all LSC+ fractions with HSPCs (Fig. 2a), whereas most LSC− fractions (17/24) clustered with mature hematopoietic populations (Fig. 2a). Seven LSC− fractions clustered with LSC+ (Fig. 2a,b), potentially reflecting low LSC frequency within these fractions. Notably, clustering based on genome-wide ATAC–seq peaks, including nonrepetitive sequences, did not reproduce these observations (Extended Data Fig. 3a,b), suggesting that TE-centered chromatin variants provide a nuanced classification of AML biology related to stemness. Leukemia stem cells containing fractions cluster with HSPCs and show distinct and common TE subfamily enrichment. a, Heatmap of accessibility z scores of all TE subfamilies (rows) across bulk ATAC–seq profiles of HSPC, mature hematopoietic populations, LSC+ and LSC− fractions clustered based on correlation. LSC+ fractions clustered with HSPC, while LSC− fractions mostly clustered with mature populations. b, Heatmap of TE subfamilies with differential accessibility z scores (rows, q < 0.01) between ATAC–seq profiles of LSC+ (n = 11) and LSC− fractions (n = 24). c, Enrichment of TE families across LSC+ versus LSC− samples. The y axis shows all TE families; each dot denotes a TE subfamily. SINE1, SINE2 and transposons families are enriched in LSC− samples, whereas ERV1 and ERV3 families are enriched in LSC+ samples (left). Number of enriched TE subfamilies in LSC+ or LSC− samples, groups based on TE family (right). TE families are ordered from the most common in LSC+ versus LSC−. d, Comparison of the proportion of TE subfamilies enriched in HSPC and hematopoietic mature populations in LSC− and LSC+ fractions grouped and color-coded according to their TE family classification. The overall distribution of TE families within the hg38 build of the human genome (Genomic) is also shown as a reference. e, UpSet plot showing the intersection of TE subfamilies enriched in accessible chromatin across normal and leukemic populations, including between LSC+/HSPC (21 subfamilies) and LSC−/mature hematopoietic populations (53 subfamilies). Colors within bar graphs denote the grouping of TE subfamilies according to their TE families classification, as annotated in d. Overall chromatin accessibility alone does not capture LSC stemness as TE-centered chromatin variants. a) Unsupervised hierarchical clustering of hematopoietic cell states and LSC fractions based on the presence/absence of ATAC peaks across repetitive and non-repetitive DNA sequences. b) Same as a but for LSC fractions alone. c) Volcano plot showing the median difference in bias-corrected accessibility deviation Z-scores for each TE subfamily between the LSC+ and LSC- fractions versus the −log10(q-value) for that difference. d) Bar plots showing the number of total or accessible TE elements belonging to the 44 TE subfamilies enriched in LSC+ fractions (left) or in the 77 TE subfamilies enriched in LSC- fractions (right). Direct comparison of TE accessibility z scores between LSC+ and LSC− fractions revealed 44 subfamilies enriched in LSC+ and 77 in LSC− fractions (Fig. 2b,c, Extended Data Fig. 3c,d and Supplementary Tables 10–12). LSC+ fractions were mostly enriched for ERV1 and ERV3 related subfamilies, whereas LSC− fractions showed enrichment for SINE1, SINE2 and transposon related subfamilies (Fig. 2c). Comparing AML fractions to normal hematopoietic populations showed that the proportions of TE subfamilies discriminating HSPC versus mature and LSC+ versus LSC− fractions were skewed relative to genome-wide TE composition (Fig. 2d). Specifically, ERV1 and ERV3 related subfamilies were preferentially accessible in HSPC and LSC+, while accessible SINE1 and transposons related subfamilies were overrepresented in mature and LSC− populations (Fig. 2d). Most accessible TE subfamilies in LSC+ or LSC− fractions were shared with HSPC (21/44) or mature populations (53/77), respectively (Fig. 2e and Extended Data Fig. 4a,b). Shared subfamilies between LSC+ and HSPC were largely from the ERV1 and ERV3 families (14/21 subfamilies; Fig. 2e and Extended Data Fig. 4c,d), whereas subfamilies shared between LSC− fractions and mature populations were mainly SINE1 and transposons related (Fig. 2e and Extended Data Fig. 4c,d). Although LINE1 subfamilies did not distinguish LSC+ from LSC−, those accessible in HSPC versus mature populations were even more enriched in LSC+ than in HSPC (Extended Data Fig. 4d), in line with diverse LINE1 functions in normal tissues and cancers47, including regulating genome topology through L1M3f48, and gene expression regulatory functions through young LINE1 subfamilies (for example, L1PA5 and L1PA7) in colon and breast cancer49–51. Enrichment levels among shared subfamilies were largely concordant between HSPC/LSC+ and mature/LSC− (Extended Data Fig. 4d). Two TE subfamilies accessible in LSC+, namely LTR12C (ERV1 family) and L1PA7 (LINE1 family), showed the largest discordance with HSPC (Extended Data Fig. 4d). Overall, our data suggest that LSCs can be distinguished from committed leukemic cells based on accessible TE enrichment patterns that are largely shared with normal HSPC populations. Similar and distinct TE subfamily enrichment in accessible chromatin in LSC+ and HSPC and in LSC- and mature hematopoietic populations. a) Venn diagram showing unique and shared enriched TE subfamilies over accessible chromatin observed in HSPC hematopoietic populations versus LSC+ fractions (left) and mature hematopoietic populations versus LSC- fractions (right). b) Genome browser views of ATAC-seq signal at individual TE elements in HSPC, LSC+, mature and LSC- samples. c) TE subfamilies are differentially enriched only in LSC+, LSC-, HSPC and mature populations of the hematopoietic system, as shown in Fig. 2e. The Cleveland dot plot shows the median z-scores for the relevant comparisons. d) Shared TE subfamilies between LSC+, LSC-, HSPC and mature populations of the hematopoietic system, as shown in Fig. 2e. The Cleveland dot plot shows the median Z-scores for the relevant comparisons. Accessible TE signature scores stemness and stratify AML risk Stemness is related to survival and relapse in AML and other cancers52,53. We hypothesized that TE subfamilies differentially accessible in LSC+ and LSC− fractions (44 in LSC+ and 77 in LSC−; Fig. 2b,c), hereafter the LSCTE121 signature, could score TE-associated stemness in AML and be related to clinical outcomes. Thus, we performed ATAC–seq on peripheral blood from three independent AML cohorts (cohort1, n = 29; cohort2, n = 60; cohort3, n = 77; Supplementary Table 13). Unsupervised clustering using the LSCTE121 signature grouped subsets of patients with the LSC− fractions (cohort1, 9/29; cohort2, 20/60; cohort3, 36/77) while the remaining patients clustered with LSC+ fractions (cohort1, 20/29; cohort2, 40/60; cohort3, 41/77; Extended Data Fig. 5a–c). This AML patient classification into LSC-like versus non-LSC-like profiles mirrors recent gene expression-based patient stratification54,55. We next computed a single LSCTE121 z score by combining enrichment of LSC+ TE subfamilies and depletion of LSC− TE subfamilies using Stouffer’s method (Methods). This score is intended to measure LSC-ike stemness properties rather than serve as a clinical biomarker. In all three cohorts, the dendrogram branch containing LSC+ fractions showed the highest LSCTE121 z scores, and LSC+ fractions scored higher than LSC− fractions (Extended Data Fig. 5a–d). LSCTE121 z score did not correlate with blast count (cohort1—R = 0.031, P = 0.88; cohort2—R = −0.17, P = 0.21; cohort3—R = 0.025, P = 0.83; Extended Data Fig. 5e–g), suggesting that the score reflected stemness rather than cellularity. In cohort1, Kaplan–Meier analysis comparing the top versus bottom quartile of LSCTE121 z score showed significant differences in disease-free and overall survival (Fig. 3a–c and Extended Data Fig. 5h). We also compared LSCTE121 with LSC17 (ref. 15). Sample classification differed between the two approaches (Fig. 3a,d). LSC17 best predicted overall survival (Extended Data Fig. 5i; log-rank, P = 2 × 10−4) and LSCTE121 showed low correlation with LSC17 (R = 0.27, P = 0.16; Fig. 3d). Mechanistically, TE accessibility near LSC17 genes supported their distinction. Among 1646 TE elements within +/−2 kb of the 17 LSC17 genes, 811 (49%) belonged to TE subfamilies enriched in LSC− fractions, whereas 33 (0.02%) belonged to LSC+ enriched TE subfamilies (Supplementary Table 14). In agreement, ATAC signal over these nearby TE elements was higher in LSC− than LSC+ fractions (Extended Data Fig. 5j), was not elevated relative to TE elements surrounding other genes (Extended Data Fig. 5k), and did not differ from 100 random, size-matched 17 genes sets in LSC+ fractions (Extended Data Fig. 5l). Next, we assessed accessibility at Transcription Start Sites (TSS) in cohort1 patients stratified based on LSC17 (high = 12, low = 17) or LSCTE121 (high = 14, low = 15), and gene set enrichment analysis to identify pathways linked to increased TSS accessibility in LSC17-high or LSCTE121-high patients. We identified 311 pathways with higher TSS accessibility in LSC17-high patients (Extended Data Fig. 5m and Supplementary Tables 15 and 16), with cell cycle-related pathways preferentially detected in LSC17-high versus LSCTE121-high patients (Fig. 3e,f and Extended Data Fig. 5n). In contrast, LSCTE121-high patients were associated with the interleukin 10 signaling (Fig. 3g,h and Extended Data Fig. 5o), a pathway known to promote LSC stemness56. Concordantly, LSCTE121 z scores were highest in fractions with the elevated limiting dilution assay-defined LSC frequency (Fig. 3i and Supplementary Table 17). Disease-free survival differences observed with LSCTE121 were also observed in cohort2 and cohort3 AML patients (Fig. 3j–o and Extended Data Fig. 5p,q). Together, our results suggest that LSCTE121 captures stemness biology distinct from that of LSC17, and that each provides AML risk stratification based on different stemness discriminators. The LSCTE121 signature captures stemness biology distinct from the LSC17 score. a–c) Heatmap showing differential accessibility Z-scores for LSC TEs (q < 0.01 LSC+ vs LSC-) between ATAC-seq profiles of LSC + , LSC- fractions and bulk AML tumors (cohort1 in a, cohort2 in b and cohort3 in c), clustered based on correlation. Note that 20, 40 and 41 samples cluster with LSC+ fractions in cohort1, cohort2 and cohort3 bulk AML patients, respectively. d) Boxplot of LSCTE121 Z-score in LSC fractions (LSC+, n = 11; LSC-, n = 24). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). P-value results of Benjamini-Hochberg corrected two-sided Wilcoxon test are showcased. e) Pearson correlation between LSCTE121 Z-score and blast count in peripheral blood for cohort1 bulk AML patients. The linear regression line is shown in blue and the confidence interval is shown in gray. P-value results of null-hypothesis testing are showcased on the plots. f) Same as panel e but for cohort2 bulk AML patients. The linear regression line is shown in blue and the confidence interval is shown in gray. P-value results of null-hypothesis testing are showcased on the plots. g) Same as panel e but for cohort3 bulk AML patients. The linear regression line is shown in blue and the confidence interval is shown in gray. P-value results of null-hypothesis testing are showcased on the plots. h) Kaplan–Meier estimates using overall survival as a clinical endpoint for LSCTE121 Z-score top vs bottom 25 percentile in cohort1 (n = 8/group). log rank (Mantel–Cox) p-values are shown on the Kaplan–Meier curves. i) Kaplan-Meier estimates using overall survival as a clinical endpoint, patient division based on LSC17-high vs low scores in cohort1 bulk AML patients; LSC17-high: n = 12, LSC17-low: n = 17. log rank (Mantel–Cox) p-values are shown on the Kaplan–Meier curves. j) Violin plots for the ATAC-seq signal over TE elements around (+/-2kb) LSC17 genes in LSC+ and LSC- fractions. P-value results of the Wilcoxon test are showcased. k) Violin plots for the ATAC-seq signal over TE elements around (+/-2kb) LSC17 genes compared to all other genes in the human genome. P-value results of the Wilcoxon test are showcased. l) Violin plots for the ATAC-seq signal over TE elements around (+/-2kb) LSC17 genes and around 100 independent groups of 17 size-matched, randomly selected genes. p-value results of Benjamini-Hochberg corrected pairwise Wilcoxon test are showcased. m) UpSet plot showing the intersection of REACTOME pathways with increased chromatin accessibility in LSC17-high or LSCTE121-high patients (cohort1). n) Volcano plot showing the log2 fold-change chromatin accessibility for each cell cycle gene promoter (TSS) between LSCTE121-high and LSCTE121-low patients (cohort1) versus the −log10, resulted from Wald test, p-value for that fold-change. o) Volcano plot showing the log2 fold-change chromatin accessibility for each IL10 signaling gene promoter (TSS) between LSC17-high and LSC17-low patients (cohort1) versus the −log10 p-value for that fold-change. p) Same as panel h but for cohort2 patients (LSCTE121-high: n = 15; LSCTE121-low: n = 45). q) Same as panel h but for cohort3 patients (LSCTE121-high: n = 35; LSCTE121-low: n = 42). The LSCTE121 signature stratifies AML patients with distinct disease-free intervals. a, Heatmap showing differential accessibility z scores for LSCTE121 (q < 0.01 LSC+ versus LSC−) in bulk cohort1 AML tumors. Please note that bulk tumors do not cluster according to their LSC17 scores with LSC+ or LSC− fractions. b, Boxplot displaying the LSCTE121 z score in cohort1 patients (n = 29). Patients were divided into LSCTE121-high or low based on the heatmaps in a. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. P-value results of the two-sided Wilcoxon test are showcased on the boxplots. c, Kaplan–Meier estimates using disease-free interval as a clinical endpoint for LSCTE121 z score top versus bottom 25 percentile in cohort1 (n = 8 per group). The log-rank (Mantel–Cox) P values are shown on the Kaplan–Meier curves. d, Pearson correlation between LSC17 score and LSCTE121 z score in cohort1 bulk AML patients. The linear regression line is shown in blue and the confidence interval is shown in gray. P-value results of null-hypothesis testing are showcased on the plots. e, GSEA enrichment plots showcasing significantly increased chromatin accessibility at cell cycle gene promoters (TSS) in LSC17-high patients (cohort1). P-value results of two-sided weighted Kolmogorov–Smirnov test are showcased on the plots. f, Volcano plot showing the log2 fold change chromatin accessibility for each cell cycle gene promoter (TSS) between LSC17-high and LSC17-low patients (cohort1) versus the −log10(P), resulting from the Wald test, for that fold change. g, GSEA enrichment plots showcasing significant increased chromatin accessibility at IL10 signaling gene promoters (TSS) in LSCTE121-high patients (cohort1). P values results of weighted Kolmogorov–Smirnov test are showcased on the plots. h, Volcano plot showing the log2 fold change chromatin accessibility for each IL10 signaling gene promoter (TSS) between LSCTE121-high and LSCTE121-low patients (cohort1) versus the −log10, resulting from the Wald test, P values for that fold change. i, Boxplot of LSCTE121 z score in LSC fractions (n = 35; high, n = 1; med, n = 6; low, n = 4; NE, n = 24). Please note that engrafting fractions show a higher LSCTE121 z score than nonengrafting fractions. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. P values results of two-sided paired Wilcoxon test are showcased. j, Same as a, but for cohort2 patients (LSCTE121-high, n = 15; LSCTE121-low, n = 45). k, Same as b, but for cohort2 patients (LSCTE121-high, n = 15; LSCTE121-low, n = 45). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. P-value results of the two-sided Wilcoxon test are shown in the boxplots. l, Same as c, but for cohort2 patients (LSCTE121-high, n = 15; LSCTE121-low, n = 45). The log-rank (Mantel–Cox) P values are shown on the Kaplan–Meier curves. m, Same as a, but for cohort3 patients (LSCTE121-high, n = 35; LSCTE121-low, n = 42). n, Same as b, but for cohort3 patients (LSCTE121-high, n = 35; LSCTE121-low, n = 42). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. P-value results of the two-sided Wilcoxon test are showcased on the boxplots. o, Same as c, but for cohort3 patients (LSCTE121-high, n = 35; LSCTE121-low, n = 42). The log-rank (Mantel–Cox) P values are shown on the Kaplan–Meier curves. GSEA, gene set enrichment analysis. LSC-accessible TEs dock AML essential factors Using the ReMap cistrome data, we asked whether accessible TE subfamilies shared between HSPC and LSC+, or unique to either group (Fig. 2e), could provide docking sites for transcription factors. Some TE subfamilies showed enrichment for many cistromes, while others showed few (Extended Data Fig. 5), aligning with the role of certain TE subfamilies as transcription factor ‘hubs’30,34,57,58. ERG, RUNX1, LMO2 and TRIM28 cistromes were consistently enriched across accessible TEs in HSPC and LSC+ (Fig. 4a-b, Extended Data Fig. 6a-b and Supplementary Tables 18–20). These transcription factors contribute to normal hematopoiesis and/or leukemia43,59–61. Beyond roles in erythroblast differentiation60, TRIM28 also suppresses TE function62, suggesting involvement in HSPC and LSC+ fate commitment. The interplay between ERG and RUNX1 in hematopoiesis and AML63,64 is reflected by their enrichment over >70% of the same accessible TE subfamilies (10/15 HSPC-specific, 8/10 LSC+ specific, 6/8 shared; Extended Data Fig. 6c), including LTR78 (ERG—q = 0.005, log(OR) = 1.74; RUNX1—q = 0.0004, log(OR) = 1.9), LTR67B (ERG—q = 1.12 × 10−8, log(OR) = 2.73; RUNX1—q = 5.77 × 10−17, log(OR) = 4.09) and MLT1E2 (ERG—q = 0.0003, log(OR) = 2.02; RUNX1—q = 0.009, log(OR) = 1.73; Extended Data Fig. 6d and Supplementary Table 18–20). Over half of the TE subfamilies enriched for ERG and/or RUNX1 harbored their DNA motifs (8/25 for ERG motif and 11/32 for RUNX1 motif; Extended Data Fig. 6e,f). ERG, RUNX1, LMO2 and TRIM28 were more expressed in HSPC than mature populations (Extended Data Fig. 6g). Collectively, these results suggest that HSPC and LSC+ share a core TE-based regulatory layer driven by these four transcription factors. TE subfamilies enriched in accessible chromatin in LSC provide binding sites to transcription factors essential for AML. a, Frequency plot of transcription factor cistromes enriched at TE subfamilies in accessible chromatin in LSC+ fractions. The top 5% most frequently enriched transcription factor cistromes are shown. b, UpSet plot showing the intersection of transcription factor cistromes enriched at accessible TE subfamilies in LSC+ and HSPC populations. c, Enrichment GIGGLE score for individual LYL1, NFYA and NFYB cistromes profiled in cell lines derived from various tissue states across TE subfamilies specifically accessible in LSC+. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. Cistromes showcased here were found significantly enriched over the TE subfamilies of interest (Fisher’s exact two-tailed, P < 0.05). d, Enrichment of NFY DNA motifs within TE subfamilies enriched in the accessible chromatin of LSC+. The red dashed line corresponds to −log10(q) = 1.3 (q = 0.05) threshold. The q value corresponds to Benjamini–Hochberg-corrected one-sided hypergeometric test P values. e, Overview of essentiality scores of LSC+ specific transcription factors across all cancer types available in DepMap with more than five cell lines, based on CRISPR data. The distribution of the essentiality scores of each transcription factor calculated in AML cell lines was compared to the distribution of the same transcriptional factor calculated in cell lines from every other cancer type. Rectangles inner color corresponds to the median essentiality score, while border color corresponds to the two-sided pairwise t test, Benjamini–Hochberg-corrected P values (adjusted P value). TE subfamilies enriched in accessible chromatin in both LSC and HSPC provide binding sites to both ERG and RUNX1. a) Frequency plot of transcription factor cistromes enriched at TE subfamilies accessible in HSPC populations. The top 5% most frequently enriched transcription factor cistromes are shown. b) Frequency plot of transcription factor cistromes enriched at TE subfamilies enriched in HSPC and LSC+ fractions. The top 5% most frequently enriched transcription factor cistromes are shown. c) UpSet plot of the intersection of TE subfamilies showing enrichment for ERG and/or RUNX1 cistromes in HSPC and/or LSC+. Please note that most of the TE subfamilies are shared between ERG and RUNX1. d) Example of TE subfamilies enriched in HSPC and/or LSC+ with enrichment of both ERG and RUNX1 cistromes and TE subfamily accessible in HSPC populations with enrichment of ZNF143, CTCF and other looping factor cistromes. Dots correspond to each transcription factor cistrome considered in the analysis. Red dashed lines correspond to -log10(q)=1.3 (q = 0.05) and logOddsRatio(logOR) = 1 thresholds. q-values correspond to FDR-corrected Fisher exact test. e) Enrichment of ERG DNA recognition sequences within TE subfamilies accessible in LSC+ and/or HSPC populations. Bars represent -log10(q). The red dashed line corresponds to -log10(q)=1.3 (q = 0.05) threshold. q-value corresponds to Benjamini-Hochberg corrected p-values. f) Same as panel e but for the RUNX1 DNA recognition sequences. q-value corresponds to Benjamini-Hochberg -corrected p-values. g) Expression levels of the transcription factors ERG, LMO2, RUNX1 and TRIM28 enriched over HSPC TE subfamilies and LSC+ TE subfamilies in HSPC (n = 18) and mature (n = 16) populations. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). p-value results of two-sided Wilcoxon test are showcased on the box plot. Cistrome enrichment over accessible TEs also identified 11 transcription factors specific to HSPC, including HIF1A, SETDB1, GATA1, ZNF143, TEAD4, PBX2, EZH2, CBX3, SMC3, MAX and HDAC2 (Fig. 4b and Extended Data Fig. 6a). Although CBX3 and MAX have not been linked to hematopoiesis, both were more expressed in HSPC than mature populations (Extended Data Fig. 7a). PBX2 is dispensable for normal hematopoiesis or immune response65, TEAD4 favors embryonic-to-blood commitment66, whereas GATA1, HIF1A, SETB1, EZH2 and HDAC2 regulate HSCs self-renewal and/or differentiation67–71. ZNF143 and SMC3 together with CTCF regulate 3D genome organization72–74. In HSPC, the cistrome for these factors, and those of known partners (SMC1A, RAD21 and CHD8)72,75,76, enrich over LTR41B elements (q ≤ 3.22 × 10−34, OR ≤ 9.36; Extended Data Fig. 6d and Supplementary Table 18), aligning with a known role for LTR41B in genome topology77. DNA motif enrichment at some of these TEs (Extended Data Fig. 7b) is consistent with the role of CTCF as a gatekeeper of stemness in hematopoietic populations7. GATA1 and HDAC2 were expressed at higher levels in HSPC than mature populations (Extended Data Fig. 7a), consistent with their role in normal hematopoiesis70,71. Together, these results support the role of HSPC-accessible TE subfamilies as docking sites for HSC regulators and genome topology factors. Distinct transcription factors bind at TE subfamilies enriched in accessible chromatin in LSC and HSPC. a) Expression levels of the 11 transcription factors enriched over HSPC TE subfamilies in HSPC (n = 18) and mature (n = 16) populations. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5 × interquartile range (IQR). p-value results of two-sided Wilcoxon test are showcased on the box plot. b) Enrichment of GATA1 and CTCF DNA recognition sequences within TE subfamilies enriched in HSPC populations. Bars represent -log10(q). The red dashed line corresponds to -log10(q)=1.3 (q = 0.05) threshold. q-value corresponds to Benjamini-Hochberg corrected p-values. c) Transcription factor cistromes enriched over ATAC peaks that do not overlap TE elements in LSC+ fractions. Every dot corresponds to a transcription factor cistrome considered in the analysis. Red dashed lines correspond to −log10(q-value) = 1.3 (q-value = 0.05) and OR = 1 thresholds; q-values correspond to FDR-corrected Fisher exact test. d) Violin plots for essentiality scores of LSC+ specific transcription factors, based on CRISPR data. For each transcription factor, the distribution of the essentiality scores calculated in 23 AML cell lines was compared to the distribution of all the other genes tested. p-values based on two-sided t-tests are reported. The red line corresponds to the essentiality score = 0 threshold. e) Violin plots for essentiality scores of LSC+ specific transcription factors, based on RNAi data. For each transcription factor, the distribution of the essentiality scores calculated in 10 AML cell lines was compared to the distribution of all the other genes tested. p-values based on two-sided t-tests are reported. The red line corresponds to the essentiality score = 0 threshold. f) Overview of essentiality scores of LSC+ specific transcription factors across all cancer types available in DepMap with more than five cell lines, based on RNAi data. The distribution of the essentiality scores of each transcription factor calculated in AML cell lines was compared to the distribution of the same transcriptional factor calculated in cell lines from each other cancer types. Rectangles inner color corresponds to the median essentiality score, while border color corresponds to Pairwise t-test Benjamini-Hochberg corrected p-values (adjusted p-value). Finally, the cistrome of LYL1, NFYA, NFYB and POU5F1 was enriched over LSC-specific accessible TEs (Fig. 4a,b and Supplementary Table 19). GIGGLE analysis showed a biased enrichment towards cistromes generated in blood or bone marrow samples for LYL1, NFYA and NFYB (Fig. 4c). These cistromes were not enriched over ATAC peaks outside repeats in LSC+ (Extended Data Fig. 7c), suggesting that TE accessibility drives their enrichment. NFYA/NFYB DNA motifs were significantly detected within cistrome-enriched TE subfamilies (Fig. 4d), which are known to be prevalent within LTR12 subfamilies34,78,79. To test function relevance, DepMap CRISPR knock-out essentiality screen results (Methods) revealed negative essentiality scores for all four transcription factors in most AML cell lines (12/23; Extended Data Fig. 7d). Across 23 AML cell lines, LYL1, NFYA and NFYB were consistently essential (P < 0.05) whereas POU5F1 was not more essential than other genes (Extended Data Fig. 7d). This essentiality was not detected in DepMap RNAi screen (Extended Data Fig. 7e). Across cancer types, LYL1 was significantly more essential in AML than in other cancers (CRISPR screen—97%, 34/35; RNAi screen—75%, 17/21) (Fig. 4e, Extended Data Fig. 7f and Supplementary Tables 21 and 22), aligning with its pro-oncogenic role in AML80. NFYA and NFYB were broadly essential across cancers (Fig. 4e and Extended Data Fig. 7f). Overall, our results suggest that LSC-specific TE accessibility defines docking sites for transcription factors, including LYL1, NFYA and NFYB, essential in AML. LSC-accessible TEs are essential for stemness We undertook functional studies to assess how TEs contribute to stemness using the OCI-AML22 model, which harbors functionally assessed LSC+ and LSC−, enrichable using FACS sorting (CD34+CD38−, very high LSC frequency; CD34+CD38+ , very low LSC frequency; CD34−CD38+/−, LSC−)46. Scoring for the LSCTE121 signature on ATAC–seq from each OCI-AML22 fraction revealed that the LSC+ fraction of OCI-AML22 cells clustered with LSC+ AML fractions (Fig. 5a), whereas the OCI-AML22 LSC−46 clustered with most LSC− AML fractions (Fig. 5a). Thus, OCI-AML22 recapitulates TE accessibility heterogeneity seen in primary AML. In agreement, the CD34+/CD38− fractions showed a higher LSCTE121 z score than CD34+/CD38+ fractions (Extended Data Fig. 8a), supporting this score as a readout of stemness potential. Accessibility at LTR12C elements is essential for LSC stemness properties. a, Heatmap showing unsupervised clustering from differential accessibility z scores for TE subfamilies accessible across LSC populations (q < 0.01 LSC+ versus LSC−), OCI-AML22 fractions. b, Boxplot displaying the LTR12C accessibility z scores in LSC+ (n = 11) and LSC− (n = 24). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× IQR. The q value corresponding to the Benjamini–Hochberg corrected adjusted two-sided Wilcoxon signed-rank test is shown on the boxplot. c, ATAC–seq signal profile across all LTR12C elements accessible in LSC+ and/or LSC− fractions. d, Violin plots for the ATAC–seq signal over LTR12C elements accessible in LSC+ or LSC−. P values results of two-sided paired Wilcoxon test are showcased. e, Same as panel c but for OCI-AML22 fractions, OCI-AML3 and MOLM13. f, Same as d, but for OCI-AML22 fractions, OCI-AML3 and MOLM13. P-value results of two-sided paired Wilcoxon test are showcased. g, Violin plots showcasing H3K27ac signal distribution in OCI-AML3 stably expressing both CRISPRi and scramble (control in gray) or LTR12C (purple) guide RNA combos, over LTR12C (target) or LTR12D (negative control). Every dot represents an individual TE element. Please note that the loss of H3K27ac signal is specific to LTR12C elements only. P-value results of two-sided Wilcoxon test are showcased. h, Same as g, but for H3K9me3 signal. P-value results of two-sided Wilcoxon test are showcased. i, Direct H3K27ac and H3K9me3 signal comparison over LTR12C elements. Axes represent log2(fold change) H3K27ac signal (x axis) or H3K9me3 signal (y axis) between CRISPRi OCI-AML3 expressing LTR12C guide RNA combo and CRISPRi OCI-AML3 expressing scramble guide RNA combo. Every dot is an LTR12C element. Please note that the majority of LTR12C elements (purple dots) showcase loss of H3K27ac signal (−log2(fold change)) and gain of H3K9me3 signal (+log2(fold change)). j, Growth curves of CRISPRi OCI-AML3 cells expressing either scramble (control) or LTR12C guide RNA combos. The results are shown as means ± s.d. from three independent transductions with the vector promoting the expression of either scramble (control) guide RNAs or guide RNAs targeting LTR12C elements. P-value results of two-sided Student’s t test are showcased on the plot. k, Violin plot for the chromatin accessibility signal distribution in OCI-AML22-LSC+ stably expressing both CRISPRi and scramble (control in gray) or LTR12C (purple) guide RNA combos over LTR12C (target) or LTR12D (negative control) elements. A representative biological replicate of three is shown. Every dot represents an individual TE element. P-value results of two-sided paired samples Wilcoxon test are showcased on the violin plot. l, Percentage of CD34+/CD38−, CD34+/CD38+ , CD34−/CD38+ and CD34−/CD38− cells upon CRISPRi-mediated chromatin editing at LTR12C elements in OCI-AML22-LSC+. The results are shown as means ± s.d. from three independent transductions with the vector promoting the expression of either five scramble (control) guide RNAs or five guide RNAs targeting LTR12C elements. P value was generated by a two-sided t test. CRISPRi reduces chromatin accessibility through gain of H3K9me3 and loss of H3K27ac specifically at LTR12C elements in OCI-AML3. a) Boxplot displaying the LSCTE121 Z-score in LSCs (n = 35), OCI-AML22 fractions (n = 8), OCI-AML3 (n = 3) and MOLM13 (n = 3) cell lines. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). p-values results of the two-sided Wilcoxon test are showcased. b) Violin plot for the difference of ATAC-seq signal between LSC+ and LSC- fractions over LTR12C elements (each dot corresponds to one LTR12C element). c) Boxplot displaying the L1PA7 accessibility Z-scores in LSC+ (n = 11) and LSC− (n = 24). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). q-value corresponding to Benjamini-Hochberg corrected two-sided Wilcoxon signed rank test is showcased. d) ATAC-seq signal profile across all L1PA7 elements accessible in LSC+ and/or LSC- fractions. e) Violin plot for the ATAC-seq signal over L1PA7 elements accessible in LSC+ (n = 11) and/or LSC- (n = 24) fractions (left). Violin plot for the difference of ATAC-seq signal between LSC+ and LSC- fractions over L1PA7 elements (each dot corresponds to one L1PA7 element—right). Violin plots show the median value of the distributions. p-value results of paired samples Wilcoxon test are showcased. f) Number of LTR12C (left) and L1PA7 (right) elements accessible in LSC+ (n = 11) and LSC- (n = 24) fractions. p-value results of two-sided Student’s t-test are showcased. g) Boxplot displaying the LTR12C accessibility Z-scores in OCI-AML22 fractions (n = 8). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). h) Boxplot displaying the L1PA7 accessibility Z-scores in OCI-AML22 fractions (n = 8). Boxplots show the median and upper and lower quartiles; whiskers represent 1.5× interquartile range (IQR). i) Boxplot displaying the LTR12C accessibility Z-scores in OCI-AML3 (n = 3) and MOLM13 (n = 3) cell lines. Boxplots show the median and upper and lower quartiles; whiskers represent 1.5 × interquartile range (IQR). j) Violin plots showcasing the chromatin accessibility signal distribution in OCI-AML3 stably expressing both CRISPRi and scramble (control in gray) or LTR12C (purple) guide RNA combos, over LTR12C or LTR12D. Every dot represents an individual TE element. Note that the loss of chromatin accessibility is specific to LTR12C elements only. p-values results of two-sided paired Wilcoxon test are showcased. k) Violin plots showcasing the chromatin accessibility signal distribution in OCI-AML3 stably expressing both CRISPRi and scramble (control in gray) or LTR12C (purple) guide RNA combos over all LTR12 TE subfamilies, except LTR12C and LTR12D (in I). Every dot represents an individual TE element. Violin plots show the median value of the distributions. p-values results of two-sided paired Wilcoxon test are showcased. l) ATAC signal over LTR12C elements in CRISPRi OCI-AML3. Every dot corresponds to one LTR12C element; the x-axis represents the log2 of the normalized ATAC signal intensity in OCI-AML22-LSC+ stably expressing both CRISPRi and scramble guide RNA combo, while the y-axis represents the log2 normalized ATAC signal intensity in OCI-AML22-LSC+ stably expressing both CRISPRi and LTR12C guide RNA combo. m) Same as panel k but for H3K27ac signal. LTR12C and LTR12D subfamilies are shown in Fig. 5g. Violin plots show the median value of the distributions. P-values results of two-sided paired Wilcoxon test are showcased. n) Same as panel l but for H3K27ac signal. o) Same as panel k but for H3K9me3 signal. LTR12C and LTR12D subfamilies are shown in Fig. 5h. Violin plots show the median value of the distributions. P-values results of two-sided paired Wilcoxon test are showcased. p) Same as panel l but for H3K9me3 signal. To directly test whether TE subfamilies are required for stemness, we used chromatin editing to repress accessibility across elements from an individual TE subfamily. We focused on LTR12C, which, with L1PA7, exhibited strong differential accessibility between LSC+ and HSPC (Extended Data Fig. 4d). LTR12C and L1PA7 were more accessible in LSC+ than LSC− fractions in terms of ATAC–seq signal and absolute number of accessible elements (Fig. 5b–d and Extended Data Fig. 8b–f). Genes within 10 kb of the most accessible LTR12C and L1PA7 elements in LSC+ included known stemness-associated genes, such as TRIM71 (ref. 81) and PAK3 (ref. 82; Extended Data Fig. 8b,e and Supplementary Table 23). In OCI-AML22, LTR12C accessibility was highest in CD34+/CD38− LSC+ fractions and decreased as cells lost stemness across the cellular hierarchy, whereas L1PA7 did not show the same pattern (Fig. 5e,f and Extended Data Fig. 8g,h). Although commonly used AML cell lines (OCI-AML3 and MOLM13) showed low overall LTR12C accessibility, a subset of LTR12C elements was accessible (Fig. 5e,f and Extended Data Fig. 8i). We therefore first tested LTR12C function in OCI-AML3 by repressing accessibility at LTR12C elements using CRISPR/dCas9-KRAB (CRISPRi). Using Repguide83, we designed pool guide RNAs targeting the 2765 LTR12C elements (predicted to hit 2400 copies, 87%), and a scrambled guide control pool (Supplementary Table 24). In OCI-AML3 CRISPRi cells34, ATAC–seq and CUT&RUN for H3K27ac and H3K9me3 revealed a strong and specific reduction of chromatin accessibility over LTR12C elements (P = 9.8 × 10−16) with limited off-target effect on related TE subfamilies (for example, LTR12D; Extended Data Fig. 8j–l), accompanied by loss of H3K27ac and gain of H3K9me3 at LTR12C elements (Fig. 5g–i and Extended Data Fig. 8m–p). Repressing LTR12C decreased OCI-AML3 growth (Fig. 5j). To assess whether LTR12C accessibility supports LSC+ properties, we stably expressed CRISPRi6,84,85 in OCI-AML22 and introduced either scrambled or LTR12C guides (Extended Data Fig. 9a-c; Methods). Across three replicates, ATAC–seq confirmed reduced accessibility at LTR12C elements (P ≤ 8 × 10−4) with limited off-target effects (Fig. 5k and Extended Data Fig. 9d–j). Functionally, LTR12C repression reduced the LSC+ CD34+/CD38− fraction (62.5–44.3%) and increased the more committed CD34+/CD38+ fraction (27.2–45.4%; Fig. 5l). Collectively, these results establish that LTR12C accessibility is a determinant of LSC stemness. CRISPRi reduces chromatin accessibility specifically at LTR12C elements in OCI-AML22-LSC+. a) Schematic overview of the CRISPRi chromatin editing approach in OCI-AML22. Panel a created in BioRender; Provitera, L. https://biorender.com/8weuoey (2026). b) dCas9-KRAB expression levels in OCI-AML22 cells. ProteinSimple (western blot) was performed on whole extracts, showcasing dCas9-KRAB expression in positive control30. The asterisk indicates the dCas9-KRAB band in CRISPRi OCI-AML22. c) Quantification of the chemiluminescence signal. Signal peaks corresponding to GAPDH (loading control) and dCas9-KRAB (CRISPRi) are indicated in the figure. d) Genome browser views of ATAC-seq signal reduction at individual LTR12C elements in OCI-AML22-LSC+ expressing both CRISPRi and LTR12C guide RNA combo (purple) compared to scramble gRNA combo (control in gray). e) Violin plots showcasing the chromatin accessibility signal distribution in replicate #1 over all LTR12 TE subfamilies, except LTR12C and LTR12D (in Fig. 5k). Every dot represents an individual TE element. Violin plots show the median value of the distributions. P-values results of two-sided paired samples Wilcoxon test are showcased. f) ATAC signal over LTR12C elements in replicate #1. Every dot corresponds to one LTR12C element; the x-axis represents the log2 of the normalized ATAC signal intensity in OCI-AML22-LSC+ stably expressing both CRISPRi and scramble guide RNA combo, while the y-axis represents the log2 normalized ATAC signal intensity in OCI-AML22-LSC+ stably expressing both CRISPRi and LTR12C guide RNA combo. g) Violin plots showing the chromatin accessibility signal distribution in replicate #2 OCI-AML22-LSC+ stably expressing both CRISPRi and scramble (control in gray) or LTR12C (purple) guide RNA combos, over LTR12C or LTR12D (left) and all other LTR12 TE subfamily elements (right). Every dot represents an individual TE element. Plesae note that the loss of chromatin accessibility is specific to LTR12C elements only. Violin plots show the median value of the distributions. P-values results of two-sided paired samples Wilcoxon test are showcased. h) Same as panel g but for replicate #3 OCI-AML22-LSC+ stably expressing both CRISPRi and scramble or LTR12C guide RNA combos. Violin plots show the median value of the distributions. P-value results of two-sided paired samples Wilcoxon test are showcased. i) Same as panel f but for replicate #2 OCI-AML22-LSC+ stably expressing both CRISPRi and scramble or LTR12C guide RNA combos. j) Same as panel f but for replicate #3 OCI-AML22-LSC+ stably expressing both CRISPRi and scramble or LTR12C guide RNA combos. Discussion Our study establishes that enrichment of specific TE subfamilies within chromatin variants is a hallmark of stemness in normal hematopoiesis and AML, with clinical implications for AML patients. Many accessible TE subfamilies in LSC and HSPC are shared with those active in embryonic stem versus mature cells30,86, including ERV families that serve as docking sites for stem transcription factors (for example, OCT4, NANOG)30,38,39. Expanding the role of TEs as cis-regulatory elements29,30,38–41,86, we find that HSPC-accessible TEs (for example, LTR41, LTR41B, LTR55) permit binding of genome topology factors, including ZNF143, CTCF, and cohesin. Given that CTCF-mediated topology changes regulate quiescence exit and long-term self-renewal in primitive hematopoietic populations7, our identification of additional factors, namely ERG, RUNX1, LMO2, and TRIM28, capable of binding to TEs accessible in LSC/HSPC (for example, LTR78, LTR67B, MLT1E2), reinforces their established roles in hematopoiesis and AML43,59–62. Together, our results suggest a model in which distinct TE subfamilies shape genome topology and transcriptional control, thereby defining stemness and the hierarchical organization of normal and leukemic cells. By revealing that TE subfamilies enrichment in accessible chromatin distinguishes LSCs from mature leukemic cells, we provide a chromatin-based approach to interrogate repeat DNA and classify AML samples by leukemia-initiating potential. Applied to patient samples, our LSCTE121 signature stratifies patients by TE accessibility associated with LSC+ biology, identifying those with higher relapse risk and shorter survival. The lack of correlation between LSCTE121 and the RNA-based LSC17 score15 argues that the two are complementary measures of LSC biology. Because large AML transcriptomic resources exist but chromatin state-based datasets are limited, our findings highlight the need to study the repetitive genome in its chromatin context to understand stemness in leukemia and other cancers. In cancer, chromatin variants frequently occur at TEs87 and can promote oncogene activation32. However, most TE studies have emphasized expression, such as ERV1 expression linked to worse survival in kidney cancer patients88, and a 14 TE-based signature predicting survival in AML89. We instead focus on TEs as regulatory elements. Among LSC-accessible TE subfamilies, LTR12C comprises potent regulatory elements, controlled in part by ZNF676 KRAB zinc finger-mediated repression32,34,90–96, that drive oncogene expression across cancers32,95, work as enhancers in AML34 and can act as alternative promoters in hepatocellular carcinoma96. Our work extends these regulatory functions to LSC stemness. We further show that accessible TEs provide docking sites for oncogenic transcription factors in LSCs. Using this framework, we identified LYL1 as an AML driver, in agreement with previous reports80, and showed that it is preferentially essential in AML cell lines. Thus, the repetitive genome directly contributes to oncogenic transcriptional control in AML. Finally, building on prior CRISPRi studies targeting TE in AML and prostate cancer30,34, we demonstrate that TE subfamilies accessible in LSC are required to maintain stemness. Repressing hundreds of LTR12C elements at the chromatin level, without altering DNA sequence, reduced the LSC+ fraction and increased the proportion of more committed cell types in OCI-AML22. These results reveal LSC-specific TE dependencies and suggest vulnerabilities that could be targeted while sparing HSPC. Altogether, this work provides a framework to capture cancer’s primitive nature by defining TE subfamilies as genetic determinants of stemness in normal and leukemic stem cell hierarchies. Methods Patient samples All biological samples were collected with written informed consent according to procedures approved by the Research Ethics Board of the University Health Network (UHN; REB 01-0573-C) and viably frozen in the Leukemia Tissue Bank at Princess Margaret Cancer Centre/University Health Network. Patient details are provided in Supplementary Table 13. Cell sorting and xenotransplantation assay Hematopoietic populations were selected using the highest purity sorting scheme available97,98. Peripheral blood AML samples were sorted into fractions based on the expression of CD34 and CD38. LSC activity in each fraction was assessed by xenotransplantation into NOD.Prkdcscid.Il2rgnull (NSG) mice as previously described15. Antibodies used are shown in Supplementary Table 25. Bulk ATAC–seq ATAC–seq was used to profile the accessible chromatin landscape of 11 LSC+, 24 LSC− fractions, three independent cohorts of peripheral blood AML bulk samples (cohort1—n = 29, 12 LSC17-high and 17 LSC17-low; cohort2—n = 60, unknown LSC17 score; cohort3—n = 77, unknown LSC17 score), sorted subpopulations of OCI-AML22 based on CD34 and CD38 status (CD34+/CD38−, CD34+/CD38+, CD34−/CD38+ and CD34−/CD38−), CRISPRi OCI-AML22 expressing either scramble or LTR12C guide RNA combos, untreated MOLM13 and OCI-AML3, and CRISPRi OCI-AML3 expressing either scramble or LTR12C guide RNA combos. 50,000 cells from each sample were processed as previously described99. Libraries were sequenced with 50-bp single-end (LSC fractions) or 50-bp paired-end (AML bulk samples) reads and mapped to hg38. Reads were filtered to remove duplicates, unmapped or poor quality (Q < 30) reads, mitochondrial reads, chrY reads and those overlapping the ENCODE blacklist. Following alignment (unique mapping), accessible chromatin regions/peaks were called using MACS2. Default parameters were used except for the following: --keep-dup all -B --nomodel --SPMR -q 0.05. ATAC–seq profiles of hematopoietic cell populations were downloaded from GSE125064 and processed similarly. We included only LSC fractions from patients without NPM1 mutations and set a threshold of at least 10 million reads and more than 15,000 called peaks. ATAC–seq data of LSC fractions and bulk AML samples can be accessed at EGAS00001004893 (LSC fractions), EGAS00001004896 (cohort1), EGAS00001007191 (cohort2) and EGAS50000001147 (cohort3). Narrowpeak files are submitted to Gene Expression Omnibus under ID GSE234237. Our approach (detailed below) for mapping ATAC peaks to repetitive sequences sometimes relies on single-ended reads, thereby limiting our ability to call peaks restricted to highly repetitive sequences, such as evolutionarily young repetitive sequences. TE families and families Bedfiles of repeats classified as TE families (excluding simple repeats, satellites, snRNA/rRNA and repeats of unknown families) and with more than 100 individual genomic elements (n = 971) in the hg38 human genome build were downloaded from the UCSC Genome Browser. Subfamilies were classified into families according to Repbase (https://www.girinst.org/repbase/)100,101. In such classification, LTR subfamilies that did not belong to ERV1, ERV2 or ERV3 were classified as ‘solo-LTRs’. TE enrichment in accessible chromatin analysis TE enrichment analysis was performed using ChromVAR102. The binary matrix for ChromVAR was generated by identifying the presence/absence of accessible chromatin peaks in each sample of interest. For the same set of peaks, a binary annotation matrix was generated, with rows corresponding to peaks and columns to each individual TE subfamily. TE subfamily-to-peak assignment was generated by checking the overlap of TEs for each subfamily in the accessible chromatin peak set. This operation was performed for each TE subfamily independently of the others. As a consequence, multiple TE elements belonging to different TE subfamilies could be assigned to the same peak. ChromVAR was run with default parameters on the two input matrices, generating a bias-corrected deviation z score. Deviation z scores were then compared between the samples of interest using a nonparametric two-sided Wilcoxon signed-rank test. For plotting, heatmap colors for z scores were truncated to minimum and maximum values of −10 and 10, respectively, and clustering was based on Euclidean distance unless otherwise stated. For the LSCTE121 z score, we combined all LSC+/LSC− enriched TEs in two separate.bed files and calculated the z score over all LSC+/LSC− enriched TEs separately. This operation will generate a ‘LSC+ z score’ and a ‘LSC− z score’. To combine both z scores, we used Stouffer’s method with equal weights of the LSC+ z score and the inverse LSC− z score: ATAC signal was quantified over LTR33 and LTR50 elements accessible in HSPC and/or mature hematopoietic populations and over LTR12C and L1PA7 elements accessible in LSC+ and/or LSC− using deeptools103. Given the different number of samples for LSC+ and LSC− fractions (11 and 24, respectively), we reported ATAC signal of top and bottom 25% of samples based on LSCTE121 score. The top 25% corresponds to LSC+, while the bottom 25% is composed uniquely of LSC− and was therefore named LSC−. Transcription factor cistromes and motifs enrichment analysis The enrichment of cistromes and motifs was assessed at the TE subfamilies enriched specifically in LT-HSCs, HSPC populations, LSC+ and commonly enriched in HSPC populations and LSC+ (23, 60, 21 and 23 TE families respectively; Extended Data Fig. 2c and Fig. 2b). Each TE subfamily was analyzed individually, restricting our analysis to the elements overlapping with a given consensus set of ATAC peaks (for example for a subfamily enriched in HSPC populations, we inquired transcription factor cistromes and motifs enriched at the elements overlapping with the ATAC peaks catalog of HSPC populations). Enrichment of transcription factor cistromes was performed using the Bioconductor package LOLA (https://bioconductor.org/packages/release/bioc/html/LOLA.html)104, version 1.20.0. 485 transcription factor cistromes were obtained from ReMap (http://remap.univ-amu.fr/)105. Enriched transcription factor cistromes at individual TE subfamilies were called against all elements from the 971 TE subfamilies overlapping any of the HSPC/mature/LSC+/LSC−z score ATAC peaks (q < 0.05). The top 5% most frequently enriched are shown in Fig. 4a and Extended Data Fig. 5a,b. The HOMER motif discovery tool (v.4.7; findMotifsGenome.pl) was used to compute motif enrichment. Enriched motifs within individual TE subfamilies (as explained above) were called against all elements from the 971 TE subfamilies overlapping any of the HSPC/mature/LSC+ /LSC− ATAC peaks (Benjamini–Hochberg, q < 0.05). GIGGLE enrichment scores were calculated as described in ref. 30 with the following parameters: species—human hg38; data type in cistrome—transcription factor, chromatin regulator; peak number of cistrome sample to use—all peaks in each sample. Essentiality scores Essentiality scores were obtained from the Broad Institute project, Achilles (DepMap)106. CRISPR (Avana 21Q1) and combined RNAi genetic dependencies data were downloaded from the ‘Download’ page at https://depmap.org/portal/achilles/. For CRISPR data, we classified cell lines into cancer types based on the following sample information available on the DepMap website. In particular, the following 23 cell lines were classified as AML: HEL, HEL 92.1.7, MV411, OCI-AML2, THP1, NOMO-1, EOL-1, KASUMI-1, NB4, OCI-AML3, MOLM13, TF1, U937, F36P, KO52, AML-193, M-07e, P31/FUJ, MONO-MAC-1, MOLM14, MUTZ-8, OCI-M2 and SHI-1. For RNAi data, our analysis was restricted to the following AML cell lines analyzed by the project Achilles: HL60, MOLM13, MOLM16, MV411, NB4, OCI-AML2, OCI-AML3, OCIAML5, THP1 and U937. LSC17 scores The LSC17 scores for all AML bulk samples were obtained using the clinical LSC17 assay. In general, the LSC17 score can be calculated as previously reported15. In short, log2-normalized values for each of the 17 genes were multiplied by their respective coefficients and then summed. For genes with multiple probes, the probe with the highest overall signal was selected. LSC17 scores versus LSCTE121 scores: chromatin accessibility at TSS To investigate differences between LSC17 and LSCTE121 scores, we focused on AML patients belonging to cohort1, which is the only cohort with both scores because of the availability of chromatin accessibility (LSCTE121) and gene expression (LSC17) data. We quantified chromatin accessibility (ATAC–seq signal) in TSS regions of all genes in the human genome (gencode v38). The TSS was defined as between 2 kb upstream and 1 kb downstream of the annotated TSS. DESeq2 was then used to normalize the signal across patients and to identify differentially accessible TSSs between LSC17-high and LSC17-low patients, and between LSCTE121-high and LSCTE121-low patients (within the same cohort). As in RNA-seq analysis, we performed gene set enrichment analysis to identify pathways with significant differences in chromatin accessibility. This analysis showed that the LSC17 score correlates with actively cycling cells, whereas LSCTE121 correlates with stem features not captured by LSC17. Survival analysis The disease-free interval was defined as the time from remission to the first new tumor progression event (locoregional recurrence, distant metastasis, development of a new primary tumor or death). Survival analysis was performed using the Kaplan–Meier estimate method. P values comparing Kaplan–Meier survival curves were calculated using the log-rank (Mantel–Cox) test. OCI-AML22 and OCI-AML3 culturing OCI-AML22 cells were expanded in culture in the following medium, with final concentrations as indicated: X-VIVO 10 (Lonza, BE04-380Q) supplemented with 20% BIT 9500 Serum Substitute (STEMCELL Technologies, 09500), 1× GlutaMAX Supplement (Thermo Fisher Scientific, 35050061), Primocin (0.1 mg ml−1; InvivoGen), SCF (200 ng ml−1; Miltenyi Biotech, 130-096-696), IL3 (20 ng ml−1; Miltenyi Biotech, 130-095-069), TPO (20 ng ml−1; Peprotech, 300-18), FLT3L (40 ng ml−1; Peprotech, 300-19 B), IL6 (10 ng ml−1; Miltenyi Biotech, 130-093-934) and G-CSF (10 ng ml−1; Miltenyi Biotech, 130-093-861). Cells were maintained at a density of 0.8 × 106 cells per ml and passaged every 3–5 days in a 96-well flat-bottom plate. The CD34+CD38− fraction or CD34+ fraction was regularly sorted to serially expand the cells. OCI-AML3 cells were routinely cultured in RPMI 1640 (Gibco, 21875-034). Media supplemented with 1× antimicrobial antimycobacterial (Gibco, 15240062), 2 mM GlutaMAX (Gibco, 35050061) and 10% fetal bovine serum (Life Technologies) at 37 °C in 5% carbon dioxide. Cells were maintained and split every 2–3 days at a density of 0.4-0.8 × 106 cells per ml. Lentiviral production VSV-G pseudotyped lentiviral vector particles were produced by polyethyleneimine-based cotransfection107 of 10.5 μg of pVSVg, 20.5 μg of pCMVR8.74 (both from Addgene) and 38 μg of transfer vector (pHR-SFFV-KRAB-dCas9-P2A-mCherry, from Addgene, plasmid 60954) per HEK293FT cells T175 flask. Viral particles collected after 44 and 70 h were pooled and filtered (0.22-µm filter), concentrated 100× using Lenti-X concentrator protocol (Takaka, 631232) optimized with Lenti-X concentrator beads at a ratio of 1:8 (beads:virus supernatant). Concentrated virus was resuspended in X-VIVO 10 (Lonza) and stored at –80 °C until use. Lentivirus titration Viral titers were assessed in MOLM13 cells. For titration, MOLM13 cells were plated at 25,000 cells per well in 96-well round-bottom plates, with 100 μl per well. Viral supernatant was added in three serial dilutions in duplicate. After mixing by pipetting, cultures were incubated O/N. The next day, 60–80 μl of supernatant was carefully removed, and 150 μl fresh medium was added. After three additional days, transduction efficiency (%BFP+) was determined with a flow cytometer (Symphony, BD). Titers were calculated by using the mean of all six transductions of duplicate serial dilutions, considering the Poisson distribution: TU/ml = ln(%untransduced cells/100) × 25,000 cells/(vector input, for example, 0.5 μl/1,000). CRISPRi experiments in OCI-AML3 For dCas9-KRAB-mediated repression of LTR12C, four gRNAs targeting the LTR12C consensus sequence were designed using Benchling biology software (v.2022). Four sgRNAs (Sigma-Aldrich) targeting the LTR12C family were cloned into the lentiviral expression vector pKLV-U6gRNA(BbsI)-PGKpuro2ABFP (Addgene, 50946). Lentivirus was produced in HEK293T cells by triple transfection with packaging plasmids (psPAX2 and pMD.G) and transfer vector pHR-SFFV-KRAB-dCas9-P2A-mCherry (Addgene, 60954) and sgRNA vector using XtremeGENE reagent (Merck, 6366244001). The negative control used in this experiment was the empty lentiviral expression vector pKLV-U6gRNA(BbsI)-PGKpuro2ABFP (Addgene, 50946). The next day, the media was changed for fresh RPMI media. Viral particles were collected at 48 and 72 h, filtered with a 0.45-μM filter, and stored at −80 °C until use. In total, 1 × 106 OCI-AML3 cells were co-transduced with dCas9-KRAB and lentiviral sgRNA expression vector supplemented with 4 µg ml−1 polybrene. Cells expressing mCherry and BFP were sorted on a FACSAria II (Extended Data Fig. 10). For the cell proliferation assay, 25,000 live cells were seeded in triplicate on a 12-well plate in 50 µl of medium. Cells were counted every 48 or 72 h over a period of 8 days using a haematocytometer and trypan blue. Before counting, cells were spun down and resuspended in fresh RPMI 1640 for continued culture. FACS gating and sorting strategies to collect leukemia stem cells from model systems. a) Sorting strategy used in OCI-AML22. b) Gating strategy used in OCI-AML22. c) Gating strategy used in OCI-AML3. Cleavage under targets and release using nuclease (CUT&RUN) CUT&RUN was performed as described in refs. 108,109. In total, 500 × 103 OCI-AML3 cells per reaction were pelleted, resuspended in wash buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 0.5 mM spermidine, plus PIC) and incubated with 10 µl concanavalin A bead for 10 min. The cells were incubated with antibodies against H3K27ac (Active Motif, 39034), H3K9me3 (Diagenode, C15410193) or IgG control (Abin101961) at 0.5 µ per reaction. The cells were washed three times and incubated with 1:300 pAG-MNase (gift from the Hurd laboratory) for 10 min at RT. The digestion was carried out by adding 2 μl of 100 mM CaCl2 to 100 μl of the sample for 1 h at 4 °C. After the reaction stopped, DNA was extracted, and library preparation was then carried out per the NEBNext Ultra II DNA Library Prep Kit for Illumina. Libraries were quality assessed using D1000 DNA ScreenTapes on an Agilent 4200 TapeStation. Generation of dCas9-KRAB (CRISPRi) OCI-AML22 and transduction with LTR12C or control vectors After expansion, the OCI-AML22 CD34+CD38− fraction was sorted (Extended Data Fig. 10; MoFlo Astrio, BD), washed and resuspended in its own media. Cells were mixed with concentrated virus produced as previously described, at a ratio of 1:10, supplemented with LentiBOOST research grade, Mayflower Biosciences, ref SB-P-LV-101-03 (1/100). Cells were then centrifuged at 1,200 rpm for 10 min before being left overnight in the incubator. The following day, the medium was replaced with OCI-AML22 culturing medium. To obtain sufficient cells, OCI-AML22-transduced cells were expanded over 1.5 months before being sorted for the mCherry+CD34+ population. After two additional months of cell expansion, mCherry+CD34+ were sorted again (MoFlo Astrio, BD) to obtain 82 million cells. In total, 1.79 million OCI-AML22 CD34+mCherry+ cells were collected and used for the next steps and are referred to as the dcas9-KRAB (CRISPRi) OCI-AML22. The mCherry+CD34+ sorted cells were transduced the day following sorting, with three independently produced viruses (four wells per condition) in a 96-well plate. Each well received 1.6 × 105 OCI-AML22 mCherry+CD34+ cells; lentiviral vector (LTR12C or control), concentrated as described in the virus section, was added at a 1:10 dilution together with LentiBOOST research grade (Mayflower Biosciences, SB-P-LV-101-03; 1:100). Cells were then centrifuged at 1,200 rpm for 10 min before being left overnight in the incubator. The following day, the medium containing virus was washed and replaced with OCI-AML22 culturing medium. Five days after transduction with LTR12C or control vector guides lentivirus, cell surface marker, mCherry, viability and GFP expression were assessed on the Symphony (BD) using CD34-APC (BD, clone 8G12; ref 340441, 1/100), CD38-PeCy7 (BD, clone Hb7; ref 335790, 1/100). Seven days post-transduction, the mCherry+GFP+ population was sorted (MoFlo Astrio, BD) for each replicate condition. Enough cells were collected to perform ATAC–seq on each replicate. CRISPRi ATAC–seq and CUT&RUN analysis ATAC–seq signal (for both OCI-AML22 and OCI-AML3) and H3K27ac and H3K9me3 CUT&RUN signal (for OCI-AML3) were normalized and quantified over LTR12, LTR12_, LTR12B, LTR12C, LTR12D, LTR12E and LTR12F individual elements using DESeq2 (ref. 110). This operation helped to identify on-target effects (over LTR12C elements) and off-target effects (on other LTR12 subfamilies) in our assay. For OCI-AML22, each biological replicate was compared with the corresponding control, whereas for OCI-AML3, ATAC–seq and CUT&RUN analyses used three biological replicates, which were grouped and analyzed together. Differences in chromatin accessibility or histone mark abundance were compared between cells coexpressing CRISPRi and scramble guide RNA combo and CRISPRi and LTR12C guide RNA combo. Signal over individual LTR12C elements was plotted using a scatter dotplot. Tornado plots and relative quantification plots were generated using deepTools103. Design of guide RNAs targeting TEs Guide RNAs targeting LTR12C elements were designed using Repguide (https://tanaylab.github.io/repguide/; ref. 111), using default parameters. We selected the combination of five guide RNAs targeting the highest number of LTR12C elements with limited targeting of other LTR12 families (LTR12, LTR12_, LTR12B, LTR12D, LTR12E and LTR12F). Five negative control guide RNAs83 were used as negative control guide RNA combination. To perform our experiments, guide RNAs were cloned into a lentiviral vector (https://en.vectorbuilder.com/), which also carries the gene encoding GFP. Sequences of guide RNAs used in this study are provided in Supplementary Table 24. Protein extraction and western blot analysis OCI-AML22 cells stably expressing dCas9-KRAB (CRISPRi) were lysed in modified RIPA (10 mmol l−1 Tris–HCl, pH 8.0; 1 mmol l−1 EDTA; 140 mmol l−1 NaCl; 1% Triton X-100; 0.1% SDS; 0.1% sodium deoxycholate) containing protease inhibitor (Complete, EDTA-free Protease Inhibitor combination, Roche). Lysates were sonicated for 5 min (30-s on, 30-s off) using a Diagenode Bioruptor 300. Cell debris was removed by centrifugation (10 min, 15,000 rpm at 4 °C) followed by protein quantification using the BCA Protein Assay Kit (Thermo Fisher Scientific). Western blot was performed using the ProteinSimple Automated system (Jess, Bio-Techne) following the manufacturer’s instructions. We used an anti-Cas9 antibody (Diagenode, C15200203; 1:5,000) and an anti-GAPDH antibody (Cell Signaling Technology, GAPDH (14C10) rabbit mAb 2118; 1:300). Statistics and reproducibility No statistical method was used to predetermine sample size. No data were excluded from the analyses. The experiments were not randomized. The Investigators were not blinded to allocation during experiments and outcome assessment. The most appropriate parametric or nonparametric statistical test was chosen for individual analyses in the article. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Online content Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at 10.1038/s41588-026-02585-z. Supplementary information Reporting Summary Supplementary Tables 1–25. Acknowledgements The authors acknowledge the Princess Margaret Genomics and Bioinformatics group for providing the infrastructure required to conduct analyses included in this study. This study was funded by the Princess Margaret Cancer Foundation, Ontario Institute for Cancer Research, with funding from the Province of Ontario, Canadian Institutes for Health Research (CIHR), CEEHRC team grant from the CIHR, Medicine by Design, the Canadian Cancer Society Research Institute (all grants to M.L. and J.D.); by the Terry Fox Research Institute Program Project and a Canada Research Chair (to J.D.); by JSPS KAKENHI (grant JP17K0989 9to N.T.). M.L. holds an investigator award from the Ontario Institute for Cancer Research, a Bernard and Francine Dorval Award from the Canadian Cancer Society and the Joey and Toby Tanenbaum Chair. Extended data Author contributions G.G., B.N., A.Q. and M.L. designed the studies. G.G. and B.N. performed the computational analysis with help from A.Q., A.N. and L.S.R. A. Mitchell performed the functional xenograft transplantation, while C.A., K.K.K. and N.T. prepared ATAC–seq libraries. A. Murison processed part of ATAC–seq data and S.A.M.T gave computational input. Ö.D., B.D. and H.B. performed experimental validations. A.A. provided clinical annotation for AML samples. M.D.M. provided AML samples. G.G., B.N., A.Q., Ö.D., H.B., J.C.Y.W., D.T., M.D.M., J.E.D. and M.L. interpreted the data; M.L. and J.E.D. supervised the study. G.G., B.N. and M.L. wrote the manuscript. Peer review Peer review information Nature Genetics thanks Dixie Mager, Nakul Shah and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Data availability The data generated in this study are publicly available in Gene Expression Omnibus (GEO) under accessions GSE234237 and GSE288771 (OCI-AML3 validations). Fastq files are submitted to the European Genome-Phenome Archive (EGA) under the following accessions: EGAS00001004893 (LSC fractions), EGAS50000001141 (OCI-AML22 subpopulations), EGAS00001004896 (bulk AML samples, cohort1), EGAS00001007191 (bulk AML samples, cohort2) and EGAS50000001147 (bulk AML samples, cohort3). Code availability Code for data processing, statistical analysis, and plotting can be found on CodeOcean (https://codeocean.com/capsule/4918318/tree). Competing interests G.G., B.N. and A.Q. report that at the time of publication, they were employees of T-One Therapeutics, Bristol Myers Squibb and Hoffman-La Roche Canada, respectively. The remaining authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Giacomo Grillo, Bettina Nadorp. Contributor Information John E. Dick, Email: john.dick@uhn.ca Mathieu Lupien, Email: mathieu.lupien@uhn.ca. Extended data is available for this paper at 10.1038/s41588-026-02585-z. Supplementary information The online version contains supplementary material available at 10.1038/s41588-026-02585-z. References Associated Data Supplementary Materials Reporting Summary Supplementary Tables 1–25. Data Availability Statement The data generated in this study are publicly available in Gene Expression Omnibus (GEO) under accessions GSE234237 and GSE288771 (OCI-AML3 validations). Fastq files are submitted to the European Genome-Phenome Archive (EGA) under the following accessions: EGAS00001004893 (LSC fractions), EGAS50000001141 (OCI-AML22 subpopulations), EGAS00001004896 (bulk AML samples, cohort1), EGAS00001007191 (bulk AML samples, cohort2) and EGAS50000001147 (bulk AML samples, cohort3). Code for data processing, statistical analysis, and plotting can be found on CodeOcean (https://codeocean.com/capsule/4918318/tree).