Search

Cancers are composed of genetically distinct subpopulations of malignant cells. DNA-sequencing data can be used to determine the somatic point mutations specific to each population and build clone trees describing the evolutionary relationships between them. These clone trees can reveal critical points in disease development and inform treatment. Pairtree is a new method that constructs more accurate and detailed clone trees than previously possible using variant allele frequency data from one or more bulk cancer samples. It does so by first building a Pairs Tensor that captures the evolutionary relationships between pairs of subpopulations, and then it uses these relations to constrain clone trees and infer violations of the infinite sites assumption. Pairtree can accurately build clone trees using up to 100 samples per cancer that contain 30 or more subclonal populations. On 14 B-progenitor acute lymphoblastic leukemias, Pairtree replicates or improves upon expert-derived clone tree reconstructions. Significance: Clone trees illustrate the evolutionary history of a cancer and can provide insights into how the disease changed through time (e.g., between diagnosis and relapse). Pairtree uses DNA-sequencing data from many samples of the same cancer to build more detailed and accurate clone trees than previously possible. See related commentary by Miller, p. 176. This article is highlighted in the In This Issue feature, p. 171.

Cancers are composed of genetically distinct subpopulations of malignant cells. DNA-sequencing data can be used to determine the somatic point mutations specific to each population and build clone trees describing the evolutionary relationships between them. These clone trees can reveal critical points in disease development and inform treatment. Pairtree is a new method that constructs more accurate and detailed clone trees than previously possible using variant allele frequency data from one or more bulk cancer samples. It does so by first building a Pairs Tensor that captures the evolutionary relationships between pairs of subpopulations, and then it uses these relations to constrain clone trees and infer violations of the infinite sites assumption. Pairtree can accurately build clone trees using up to 100 samples per cancer that contain 30 or more subclonal populations. On 14 B-progenitor acute lymphoblastic leukemias, Pairtree replicates or improves upon expert-derived clone tree reconstructions.Significance:Clone trees illustrate the evolutionary history of a cancer and can provide insights into how the disease changed through time (e.g., between diagnosis and relapse). Pairtree uses DNA-sequencing data from many samples of the same cancer to build more detailed and accurate clone trees than previously possible.See related commentary by Miller, p. 176.This article is highlighted in the In This Issue feature, p. 171.

Supplementary Data from Reconstructing Complex Cancer Evolutionary Histories from Multiple Bulk DNA Samples Using Pairtree

Targeted-sequencing analysis of PDX and PairTree predicted mutational population clusters

dPDX and dRI-PDX leukemia-initiating cell frequencies

Disease recurrence causes significant mortality in B-progenitor acute lymphoblastic leukemia (B-ALL). Genomic analysis of matched diagnosis and relapse samples shows relapse often arising from minor diagnosis subclones. However, why therapy eradicates some subclones while others survive and progress to relapse remains obscure. Elucidation of mechanisms underlying these differing fates requires functional analysis of isolated subclones. Here, large-scale limiting dilution xenografting of diagnosis and relapse samples, combined with targeted sequencing, identified and isolated minor diagnosis subclones that initiate an evolutionary trajectory toward relapse [termed diagnosis Relapse Initiating clones (dRI)]. Compared with other diagnosis subclones, dRIs were drug-tolerant with distinct engraftment and metabolic properties. Transcriptionally, dRIs displayed enrichment for chromatin remodeling, mitochondrial metabolism, proteostasis programs, and an increase in stemness pathways. The isolation and characterization of dRI subclones reveals new avenues for eradicating dRI cells by targeting their distinct metabolic and transcriptional pathways before further evolution renders them fully therapy-resistant.Significance:Isolation and characterization of subclones from diagnosis samples of patients with B-ALL who relapsed showed that relapse-fated subclones had increased drug tolerance and distinct metabolic and survival transcriptional programs compared with other diagnosis subclones. This study provides strategies to identify and target clinically relevant subclones before further evolution toward relapse.See related video: https://vimeo.com/442838617See related article by E. Waanders et al.

Patient characteristics and patient genomic analysis (WES and SNP 6.0)

Leukemia-initiating cell frequencies of paired diagnosis and relapse patient samples

RNA-sequencing data from Patient 9 PDX

RNA-sequencing, pathway enrichment (GSEA) reports and GSVA results (including gene list HSC vs B) of PDX and paired patient samples

PDX targeted-sequencing tissue concordance

The leukemia stem cell (LSC) populations of acute myeloid leukemia (AML) exhibit phenotypic, genetic, and functional heterogeneity that contribute to therapy failure and relapse. Progress toward understanding the mechanistic basis for therapy resistance in LSCs has been hampered by difficulties in isolating cell fractions that enrich for the entire heterogeneous population of LSCs within individual AML samples. We previously reported that CD200 gene expression is upregulated in LSC-containing AML fractions. Here, we show that CD200 is present on a greater proportion of CD45dim blasts compared with more differentiated CD45high cells in AML patient samples. In 75% (49 of 65) of AML cases we examined, CD200 was expressed on ≥10% of CD45dim blasts; of these, CD200 identified LSCs within the blast population in 9 of 10 (90%) samples tested in xenotransplantation assays. CD200+ LSCs could be isolated from CD200+ normal HSCs with the use of additional markers. Notably, CD200 expression captured both CD34– and CD34+ LSCs within individual AML samples. Analysis of highly purified CD200+ LSC-containing fractions from NPM1-mutated AMLs, which are commonly CD34–, exhibited an enrichment of primitive gene expression signatures compared with unfractionated cells. Overall, our findings support CD200 as a novel LSC marker that is able to capture the entire LSC compartment from AML patient samples, including those with NPM1 mutation.

Relapse of acute lymphoblastic leukemia (ALL) remains a leading cause of childhood cancer-related death. Prior studies have shown clonal mutations at relapse often arise from relapse-fated subclones that exist at diagnosis. However, the genomic landscape, evolutionary trajectories, and mutational mechanisms driving relapse are incompletely understood. In an analysis of 92 cases of relapsed childhood ALL incorporating multimodal DNA and RNA sequencing, deep digital mutational tracking, and xenografting to formally define clonal structure, we identified 50 significant targets of mutation with distinct patterns of mutational acquisition or enrichment. CREBBP, NOTCH1, and RAS signaling mutations arose from diagnosis subclones, whereas variants in NCOR2, USH2A, and NT5C2 were exclusively observed at relapse. Evolutionary modeling and xenografting demonstrated that relapse-fated clones were minor (50%), major (27%), or multiclonal (18%) at diagnosis. Putative second leukemias, including those with lineage shift, were shown to most commonly represent relapse from an ancestral clone rather than a truly independent second primary leukemia. A subset of leukemias prone to repeated relapse exhibited hypermutation driven by at least three distinct mutational processes, resulting in heightened neoepitope burden and potential vulnerability to immunotherapy. Finally, relapse-driving sequence mutations were detected prior to relapse using droplet digital PCR at levels comparable with orthogonal approaches to monitor levels of measurable residual disease. These results provide a genomic framework to anticipate and circumvent relapse by earlier detection and targeting of relapse-fated clones. Significance: This study defines the landscape of mutations that preexist and arise after commencement of ALL therapy and shows that relapse may be propagated from ancestral, major, or minor clones at initial diagnosis. A subset of cases exhibits hypermutation that results in expression of neoepitopes that may be substrates for immunotherapeutic intervention. See related video: https://vimeo.com/442838617 See related commentary by Ogawa, p. 21. See related article by S. Dobson et al . This article is highlighted in the In This Issue feature, p. 5

1 Abstract Cancers are composed of genetically distinct subpopulations of malignant cells. By sequencing DNA from cancer tissue samples, we can characterize the somatic mutations specific to each population and build clone trees describing the evolutionary ancestry of populations relative to one another. These trees reveal critical points in disease development and inform treatment. Pairtree is a new method for constructing clone trees using DNA sequencing data from one or more bulk samples of an individual cancer. It uses Bayesian inference to compute posterior distributions over the evolutionary relationships between every pair of identified subpopulations, then uses these distributions in a Markov Chain Monte Carlo algorithm to perform efficient inference of the posterior distribution over clone trees. Unlike existing methods, Pairtree can perform clone tree reconstructions using as many as 100 samples per cancer that reveal 30 or more cell subpopulations. On simulated data, Pairtree is the only method whose performance reliably improves when provided with additional bulk samples from a cancer. This suggests a shortcoming of existing methods, as more samples provide more information, and should always make clone tree reconstruction easier. On 14 B-progenitor acute lymphoblastic leukemias with up to 90 samples from each cancer, Pairtree was the only method that could reproduce or improve upon expert-derived clone tree reconstructions. By scaling to more challenging problems, Pairtree supports new biomedical research applications that can improve our understanding of the natural history of cancer, as well as better illustrate the interplay between cancer, host, and therapeutic interventions. The Pairtree method, along with an interactive visual interface for exploring the clone tree posterior, is available at https://github.com/morrislab/pairtree.

TO THE EDITOR: BCR-ABL -negative myeloproliferative neoplasms (MPNs) are a group of clonal stem cell disorders characterized by the overproduction of mature myeloid cells. However, MPN patients also have a 2.8- to 3.4-fold increased risk of developing a lymphoproliferative disorder (LPD) compared

Acute myeloid leukemia (AML) is an aggressive hematologic malignancy with poor survival, especially in older patients. Despite high remission rates after chemotherapy, relapse and death are frequent due to persistence of leukemia stem cells (LSCs), which possess properties linked to therapy resistance. Thus, there is an urgent need for a deeper understanding of the unique properties of LSCs. MicroRNAs (miRNAs) are non-coding RNAs that decrease expression of their target mRNAs by post-translational silencing. miRNA profiling of human AML samples fractionated based on LSC activity revealed that miR-125b is expressed at significantly higher levels on cell fractions enriched in LSCs. To evaluate the role of miR-125b in LSCs, expression of miR-125b was enforced in a hierarchical AML model cell line (OCI-AML-8227). miR-125b overexpression (OE) resulted in a significantly lower percentage of CD14+CD15+ differentiated myeloblasts (Figure 1A) and enhanced clonogenic potential in vitro (Figure 1B). Xenotransplantation of four AML patient samples with miR-125b OE revealed a significant increase in the proportion of CD117+ cells, a marker of hematopoietic and leukemic progenitors (Figure 1C). Secondary transplantation of cells harvested from primary engrafted mice at limiting dilution demonstrated a marked increase in LSC frequency with miR-125b OE compared to controls for the two AML samples tested (Figure 1D). Together, these data strongly suggest that miR-125b enhances the self-renewal of LSCs. To investigate the mechanisms by which miR-125b enhances self-renewal, proteomic analysis of miR-125b-OE Ba/F3 cells as well as in silico target prediction were performed and identified PTPN18 as a top putative target for miR-125b. PTPN18 is a tyrosine phosphatase that has been reported to dephosphorylate auto-phosphorylated kinases such as Her2 and Abl to prevent their activation. To evaluate whether PTPN18 OE can rescue the effects miR-125b on LSCs, we carried out transduction of an AML patient sample with control, miR-125b OE, PTPN18 OE, or both miR-125b and PTPN18 OE vectors followed by xenotransplantation. Similar to previous findings, miR-125b OE alone significantly reduced the frequency of CD11b+CD15+ differentiated myeloblasts. Co-transduction of miR-125b/PTPN18 OE vectors resulted in generation of significantly more CD11b+CD15+ cells compared to miR-125b OE alone (Figure 1E), suggesting that suppression of PTPN18 contributes to miR-125b-mediated enhancement of LSC self-renewal. To identify putative phosphotyrosines that might be altered through the miR-125b-PTPN18 signalling axis, we performed immunoprecipitation of phosphotyrosines followed by mass spectrometry in miR-125b-OE Ba/F3 cells and identified increased GSK3 tyrosine phosphorylation as a top target. Additionally, miR-125b OE was confirmed to enhance GSK3 tyrosine phosphorylation, whereas PTPN18 OE reduced it (Figure 1F), together strongly suggesting that miR-125b could enhance tyrosine phosphorylation of GSK3 by silencing PTPN18. GSK3A and GSK3B (GSK3A/B) are paralogous genes that share a high degree of sequence homology and belong to the glycogen synthase kinase 3 (GSK3) family. Tyrosine phosphorylation activates the kinase activity of GSK3, whereas serine phosphorylation inactivates it. We recently identified GSK inhibitors as top candidates targeting LSCs in a stemness-based drug screen using OCI-AML-8227 cells (data not shown). Treatment of OCI-AML-8227 cells with two selective inhibitors of GSK3 selectively reduced the proportion of CD34+ cells while concomitantly increasing expression of myeloid markers CD14 and CD15 (Figure 1G). Overall, our results support an important functional role for PTPN18 and GSK3 in LSC function, and present a potential novel therapeutic target against LSCs. This study highlights the importance of understanding the role of miRNAs and may identify a new druggable vulnerability in LSCs that could lead to the development of new treatment options for AML patients. Figure 1 Disclosures Dick: Bristol-Myers Squibb/Celgene: Research Funding. Wang:Trilium Therapeutics: Patents & Royalties.

Disease recurrence causes significant mortality in B-progenitor acute lymphoblastic leukemia (B-ALL). Genomic analysis of matched diagnosis and relapse samples shows relapse often arising from minor diagnosis subclones. However, why therapy eradicates some subclones while others survive and progress to relapse remains obscure. Elucidation of mechanisms underlying these differing fates requires functional analysis of isolated subclones. Here, large-scale limiting dilution xenografting of diagnosis and relapse samples, combined with targeted sequencing, identified and isolated minor diagnosis subclones that initiate an evolutionary trajectory toward relapse [termed diagnosis Relapse Initiating clones (dRI)]. Compared with other diagnosis subclones, dRIs were drug-tolerant with distinct engraftment and metabolic properties. Transcriptionally, dRIs displayed enrichment for chromatin remodeling, mitochondrial metabolism, proteostasis programs, and an increase in stemness pathways. The isolation and characterization of dRI subclones reveals new avenues for eradicating dRI cells by targeting their distinct metabolic and transcriptional pathways before further evolution renders them fully therapy-resistant. Significance: Isolation and characterization of subclones from diagnosis samples of patients with B-ALL who relapsed showed that relapse-fated subclones had increased drug tolerance and distinct metabolic and survival transcriptional programs compared with other diagnosis subclones. This study provides strategies to identify and target clinically relevant subclones before further evolution toward relapse. See related video: https://vimeo.com/442838617 See related article by E. Waanders et al .

Umbilical cord blood (CB) is a convenient and broadly used source of hematopoietic stem cells (HSCs) for allogeneic stem cell transplantation. However, limiting numbers of HSCs remain a major constraint for its clinical application. Although one feasible option would be to expand HSCs to improve therapeutic outcome, available protocols and the molecular mechanisms governing the selfrenewal of HSCs are unclear. Here, we show that ectopic expression of a single microRNA (miRNA), miR-125a, in purified murine and human multipotent progenitors (MPPs) resulted in increased self-renewal and robust long-term multi-lineage repopulation in transplanted recipient mice. Using quantitative proteomics and western blot analysis, we identified a restricted set of miR-125a targets involved in conferring long-term repopulating capacity to MPPs in humans and mice. Our findings offer the innovative potential to use MPPs with enhanced self-renewal activity to augment limited sources of HSCs to improve clinical protocols.

Chromosomal rearrangements are a hallmark of acute lymphoblastic leukemia (ALL) and are important ALL initiating events. We describe four different rearrangements of the erythropoietin receptor gene EPOR in Philadelphia chromosome-like (Ph-like) ALL. All of these rearrangements result in truncation of the cytoplasmic tail of EPOR at residues similar to those mutated in primary familial congenital polycythemia, with preservation of the proximal tyrosine essential for receptor activation and loss of distal regulatory residues. This resulted in deregulated EPOR expression, hypersensitivity to erythropoietin stimulation, and heightened JAK-STAT activation. Expression of truncated EPOR in mouse B cell progenitors induced ALL invivo. Human leukemic cells with EPOR rearrangements were sensitive to JAK-STAT inhibition, suggesting a therapeutic option in high-risk ALL.

Therapy resistance and relapse in acute myeloid leukemia (AML) are driven by leukemia stem cells (LSCs). Recent evidence highlighting functional and genetic heterogeneity among LSC subclones underscores the importance of capturing the entire LSC compartment in studies of LSC biology. Although LSCs are often enriched in the CD34+CD38- cell fraction, they are frequently detected in other phenotypic fractions, and in some cases are restricted to the CD34- population. In order to discover novel LSC markers, we examined genes differentially expressed between functionally-validated LSC+ and LSC- cell fractions obtained from primary AML samples, and identified CD200 as a candidate cell surface marker for LSCs. CD200 expression in 57 primary AML samples was analyzed by flow cytometry using anti-human CD200 clone 1B9(kindly provided by Trillium Therapeutics Inc.). CD200 was present on a greater proportion of CD45dim blasts compared to more differentiated CD45high non-blast populations (54.4% versus 21.7%, p To validate CD200 as a LSC marker, leukemic blasts were sorted from 15 primary AML samples based on CD45 and CD200 expression and transplanted into NSG mice. Samples were selected based on either the presence of both CD200+ and CD200- blasts, or CD200 expression on 5% on blasts. Overall, these results indicate that CD200 expression can be used to segregate LSCs from bulk leukemia cells. CD200 expression may be a particularly useful LSC marker in cytogenetically normal AMLs with NPM1 mutation (CN-AMLNPM1c), which have low or negative CD34 expression and commonly possess CD34- LSCs. Among 20 CN-AMLNPM1c samples, the proportion of CD200+ blasts was higher than that of CD34+ blasts irrespective of FLT3-ITD status, although there was a trend towards higher CD200 expression in FLT3-ITD+ samples. In xenotransplantation assays, 7 of 8 CN-AMLNPM1c samples tested contained CD200+ LSCs while the remaining sample contained both CD200+ and CD200- LSCs. Principal component analysis of gene expression profiles demonstrated that functionally-validated CD200+ LSC-containing fractions from CN-AMLNPM1c patients clustered separately from LSC fractions from NPM1wt or cytogenetically-abnormal cases, and were enriched for stem cell genes by gene set enrichment analysis. Furthermore, ATAC-Seq analysis demonstrated greater chromatin accessibility in CD200+ LSC-containing versus CD200‒ LSC-depleted fractions from CN-AMLNPM1c patients, with unique enrichment of HOX motifs. These data validate CD200 as an LSC marker in CN-AMLNPM1c cases. In summary, CD200 is a valuable tool for capturing heterogeneous LSC populations including both CD34+ and CD34- LSCs in many primary AML samples. It will be particularly useful for future studies of LSCs in CN-AMLNPM1c where CD34 expression does not identify LSCs. Disclosures No relevant conflicts of interest to declare.

Disease recurrence remains a significant cause of mortality in B-cell acute lymphoblastic leukemia (B-ALL). Genomic analysis of matched diagnosis and relapse samples have demonstrated that relapse arises from a minor subclone already present at diagnosis and not the dominant clone in the majority of patients. However, the reasons why only some clones survive therapy and generate relapse are obscure and elucidation of the mechanisms that underlie these differing fates may be revealed by functional analysis of isolated subclones. Previous work has shown that the subclonal diversity in B-ALL exists at the level of the leukemia-initiating cells capable of generating patient derived xenografts (Notta et al., Nature, 2011). In order to investigate the functional consequences of genetic clonal evolution during disease progression, we performed in-depth genomic and functional analysis of 14 paired diagnosis/relapse samples from adult and pediatric B-ALL patients with varying cytogenetic abnormalities. Diagnosis-specific, relapse-specific, and shared clonal and subclonal variants were identified by whole exome sequencing of the patient samples. Targeted sequencing of these variants in 372 xenografts generated by transplantation of CD19+ cells in a limiting cell dilution assay uncovered clonal variation. This analysis provided for the unequivocal identification of minor subclones ancestral to the relapse, termed diagnosis Relapse-Initiating (dRI) clones, in the diagnostic sample. Our xenografting approach enabled the physical isolation of dRI clones providing a unique opportunity to interrogate their epigenetic and transcriptional landscapes in order to unravel their relapse initiating capacity. To this end, representative diagnosis, dRI and relapse clones from 5 of the 14 patients were subjected to RNAseq and ATACseq (assay for transposase-accessible chromatin using sequencing) analysis. Despite the differences in transcriptional and chromatin openness between patients, principal component analysis of subclones from individual patients positioned the dRI clones as evolutionary intermediates between the diagnosis and relapse clones. Hierarchical clustering of the most significantly differentially expressed genes and open chromatin regions demonstrated that dRI clones shared gene expression and chromatin accessibility signatures with both the dominant diagnosis clone as well as the dominant relapse clone. To gain mechanistic insight into the data we used gene set enrichment analysis (GSEA) and identified common molecular pathways present in all patients that were enriched in dRI clones and persisted at the time of relapse in comparison to the dominant diagnosis clone. dRI and relapse clones converged in the activation of genes involved in cellular functions such as endocytosis, autophagy and innate immune response. In addition, cell surface proteins like ABC transporters and ephrins were also upregulated in dRI and relapse clones. Remarkably, functional interrogation of dRI clones in secondary xenografts, in comparison to more representative diagnosis clones, displayed increased tolerance to standard chemotherapeutic agents (dexamethasone, L-asparaginase and vincristine). Investigation of the molecular pathways and cellular receptors/transporters identified by gene expression analysis are being assessed in vitro and in vivo as potential targets for novel therapeutic approaches and disease monitoring. Overall, we have shown evidence that minor subclones at diagnosis, ancestral to the relapse clone, possess functional advantages and unique properties over other diagnostic subclones prior to treatment exposure. In depth analysis of pathways identified in these dRI subclones will shed light on potential new therapeutic approaches for abrogating and reducing disease recurrence in B-ALL. Disclosures Mullighan: Amgen: Honoraria, Speakers Bureau; Cancer Prevention and Research Institute of Texas: Consultancy; Abbvie: Research Funding; Pfizer: Honoraria, Research Funding, Speakers Bureau; Loxo Oncology: Research Funding.

Without prophylactic therapy, B-cell Acute Lymphoblastic Leukemia (B-ALL) spreads to the leptomeninges of the central nervous system (CNS) in up to 70% of patients. CNS involvement is more common in certain high risk B-ALL subgroups, including patients with KMT2A (MLL)-translocations, and disease relapse in the CNS carries a poor prognosis. The genetic determinants and biology of B-ALL dissemination to the CNS are poorly defined and therefore therapies targeting the drivers of CNS disease are lacking. Whereas B-ALL exhibits significant subclonal diversity that contributes to functional heterogeneity and disease relapse, recent reports suggest similar clonal composition of bone marrow (BM) and CNS disease, with the potential for CNS dissemination being a universal property of B-ALL cells (Williams et al. 2016, Bartram et al. 2018). Furthermore, functional studies of leptomeningeal disease have focused on the invasion of B-ALL cells into the CNS but limited studies have addressed the selection of genetic clones with the ability to grow within the subarachnoid space. To better define the evolutionary history and biology of leptomeningeal B-ALL we performed targeted DNA, SNP copy number, RNA sequencing, and functional analysis on cells isolated from matched BM and CNS tissue of patient derived xenografts (PDX) generated from a cohort of paired diagnosis and relapse samples from 14 pediatric and adult B-ALL patients of varying cytogenetics. The majority of primary patient samples yielded CNS disease 20 weeks after intrafemoral injection into NSG mice. CNS disease burden was higher in PDXs derived from relapsed B-ALL samples. Human B-ALL cells isolated from the CNS of PDXs retained competence to repopulate disease in the BM, spleen, and CNS upon serial transplantation. Targeted DNA sequencing results analyzed using a Bayesian clustering method revealed different genetic clonal composition between matched BM and CNS cells in approximately half of the xenografts. PDXs from relapse samples were more likely to show genetic discordance between the BM and CNS. Copy number analysis also confirmed frequent genetic discordance between cells isolated from the BM and CNS from individual PDXs. Interestingly, in one patient all PDXs generated from the relapse sample displayed chromosome 6p and 17p hemi-deletions that were unique to the CNS. In total, PDXs from four patients showed recurrent enrichment of specific lesions in CNS-engrafting cells, suggesting that transit to and/or survival within the subarachnoid space can be the product of selection for genetic clones with increased CNS tropism. RNA-seq of matched BM and CNS cells derived from 45 of the primary PDXs demonstrated that CNS-isolated cells were transcriptionally distinct from their matched BM. These differences were most pronounced in samples from patients with MLL-AF4 translocations, whose CNS isolated cells grouped together in multi-dimensional scaling. Using GSEA, the most highly CNS-enriched gene sets in MLL samples were related to mRNA translation initiation and polypeptide elongation. Translation-related gene sets are similarly enriched in the blasts of MLL B-ALL patients with CNS disease in the COG 9906 study. CNS-isolated cells from PDXs of MLL patients exhibited altered rates of protein synthesis compared to matched BM-isolated cells. Treatment of PDXs with the clinically-approved translational inhibitor omacetaxine mepesuccinate (OMA) effectively decreased rates of translation in CNS-engrafting cells. Moreover, OMA reduced leukemia burden nearly 4-fold in PDXs bearing established CNS infiltration generated from two MLL patients. Our data represent an advance in the understanding of B-ALL CNS disease. We present a rich resource of genomic and transcriptomic data from xenografts spanning multiple B-ALL subgroups across diagnosis and relapse and have identified selection for genetically and biologically distinct clones in the CNS, contrary to the current model. Furthermore, we demonstrate that in MLL patients, dysregulation of protein synthesis occurs at CNS dissemination and targeting this process is a novel therapeutic paradigm that may benefit patients with CNS disease. Disclosures Mullighan: Cancer Prevention and Research Institute of Texas: Consultancy; Pfizer: Honoraria, Research Funding, Speakers Bureau; Loxo Oncology: Research Funding; Amgen: Honoraria, Speakers Bureau; Abbvie: Research Funding.

Background Philadelphia-chromosome positive acute lymphoblastic leukemia (Ph+ ALL), the most common form of ALL in adults, is a highly aggressive blood malignancy defined by the BCR-ABL1 fusion. Although inhibitors targeting the BCR-ABL1 oncoprotein, such as imatinib, have significantly improved clinical response rates for this disease, a subset of patients are refractory to therapy or respond initially but relapse soon after. ABL1 kinase domain mutations partly explain differential responses in patients; however, for the majority of cases, a molecular basis that can reconcile this clinical observation is lacking. Methods Flow-sorted blasts from 53 primary samples, 49 de novo Ph+ ALL and 4 lymphoid blast crisis CML, were subjected to RNA sequencing (RNA-seq) and whole genome sequencing (WGS). Response rates were tracked using BCR-ABL1 transcript levels from patient blood. Results Hierarchical clustering of transcriptome data produced two molecular subgroups of Ph+ ALL. One subgroup, which we termed 'Core-B', upregulated key regulators of B-lymphoid differentiation including IL7R and MS4A1 (CD20). By contrast, the second subgroup upregulated an expression program related to hematopoietic stem cell (HSC) and myeloid differentiation, with upregulation of KIT, CD34, MPO, CSF3R, and GATA3. We termed this subgroup 'Aberrant-Stem-Myeloid' (ASM). These subgroups displayed a striking disparity in response rates to intensive chemotherapy with imatinib. Whereas 'Core-B' patients showed highly durable responses often lasting many years, 'ASM' patients frequently relapsed (4% vs 43% relapse; p=0.007). We used WGS analysis to investigate the genetic basis of these molecular subtypes. 'Core-B' Ph+ ALLs were enriched for deletions in PAX5, a B-cell specification gene, and CDKN2A/B, tumor suppressors. The 'ASM' subtype lacked these genetic alterations; instead, these leukemias were enriched for deletions in EBF1, an early B-cell lineage factor that represses T-lymphoid and myeloid lineages and is expressed before PAX5 in B-cell lineage differentiation. Accordingly, blasts from 'ASM' leukemias with EBF1 deletions showed decreased CD19 antigen expression and upregulation of myeloid antigens by clinical flow cytometry. Rare cases with concurrent EBF1 and PAX5 deletions showed expression features of both 'ASM' and 'Core-B' leukemias. Mutations observed in myeloid leukemias (TET2, RUNX1) were only present in the 'ASM' subtype. Loss of IKZF1, found in 77% of cases, also displayed distinct patterns between the two subgroups; deletions leading to the dominant negative isoform (Ik6) were enriched in the 'Core-B' subgroup (45% vs. 14%; p=0.019) while monosomy 7 and large deletions encompassing IKZF1 were enriched in the 'ASM' leukemias (41% vs. 10%; p=0.017). In 1 of 4 diagnosis/relapse patients analyzed, a molecular switch from 'Core-B' at diagnosis to 'ASM' at relapse was observed. The diagnostic 'Core-B' clone from this patient harbored a PAX5 mutation that was lost at relapse, whereas the relapsed 'ASM' clone harbored trisomy 21 and a RUNX1 mutation. Altogether, our data suggest that the 'ASM' leukemias emerge through dysregulation of genes earlier in lympho-myeloid specification compared to 'Core-B' leukemias. These findings led us to investigate if the 'ASM' subtype originates from an HSC and the 'Core-B' subtype originates from a B-cell progenitor. We first looked at the distribution of the long (p210) and short (p190) isoforms of BCR-ABL1 in the two subtypes. The p210 isoform, also the hallmark of CML, is speculated to arise in an HSC, and the p190 is thought to arise in a B-cell progenitor. Neither the p190 or p210 BCR-ABL1 isoform was enriched in either subgroup. We resolved highly purified HSC and progenitor subsets from CD34+CD19- cells, functionally evaluated by methylcellulose assays, and subjected them to a sensitive nested-PCR strategy. Cases from both the 'ASM' and 'Core-B' subtypes showed HSC/myeloid progenitor involvement regardless of the BCR-ABL1 isoform. This data suggest that the cell-of-origin does not play a role in establishing the molecular subtype of leukemia blasts. Conclusion There are two distinct molecular subtypes of Ph+ ALL that demonstrate differential responses to treatment and emerge from independent mutational routes. Moreover, the key genetic determinants that form the molecular subtype are secondary driver alterations that lie downstream of BCR-ABL1. Disclosures No relevant conflicts of interest to declare.

B-cell acute lymphoblastic leukemia (B-ALL) readily disseminates to the leptomeninges of the central nervous system (CNS). CNS involvement is more frequent in certain poor prognosis subgroups including patients with MLL-AF4 translocations, and late CNS relapse is often lethal. The biology and clonal history of CNS leukemia are poorly defined and consequently therapies exploiting drivers of metastasis are lacking. To characterize leptomeningeal leukemia we performed targeted DNA sequencing, SNP copy number analysis, RNA sequencing, and functional analysis on cells isolated from the bone marrow (BM) and CNS of xenografts generated from a cohort of paired diagnosis and relapse samples from 14 B-ALL patients. The majority of patient samples disseminated to the CNS following intrafemoral injection into irradiated NSG mice, with greater CNS involvement in xenografts derived from relapse patient samples. Secondary transplantation of both BM- and CNS-purified cells demonstrated their capacity to re-engraft BM, CNS, and spleen. Targeted-sequencing results were analyzed using a Bayesian clustering method to determine the clonal composition of matched BM and CNS, demonstrating discordance in subclonal prevalence in nearly half the xenografts tested. Xenografts derived from two patient samples demonstrated recurrent enrichment of a particular subclone in the CNS versus BM. Similarly, copy number analysis identified frequent discordance between BM and CNS tissues within individual mice. All xenografts from one patient exhibited chromosome 6p and 17p hemi-deletions that were exclusive to CNS cells. While these data suggest that individual B-ALLs harbor subclones with CNS tropism, there were no recurrently enriched single nucleotide mutations or copy number alterations across all patients. RNA-sequencing of 45 BM and CNS pairs from primary xenografts demonstrated that CNS-isolated cells were consistently distinct from their matched BM. GSEA analysis of xenografts generated from patients with MLL-AF4 translocations (MLL) (n=2 patients, 26 mice), identified CNS cell enrichment of gene sets related to mRNA translation and nascent peptide elongation compared to BM. MLL-CNS cells exhibited altered rates of protein synthesis compared to BM cells from the same mouse. The clinically-approved translation inhibitor omacetaxine mepesuccinate effectively diminished protein translation rates of CNS isolated cells and reduced CNS engraftment by four fold in xenografts derived from two MLL-AF4 patients. These data demonstrate that the CNS microenvironment selects for the outgrowth of B-ALL cells with genetically and/or biologically distinct properties. Moreover, we demonstrate that in MLL-AF4 patients, altered protein synthesis occurs in CNS dissemination and that targeting this process may clinically benefit patients with CNS disease. Citation Format: Robert J. Vanner, Stephanie M. Dobson, Ildiko Grandal, Olga Gan, Jessica McLeod, James Kennedy, Veroniqu Voisin, Abdellatif Daghrach, Erwin M. Schoof, Cynthia Guidos, Jayne Danska, Esme Waanders, Mark Minden, Charles G. Mullighan, John E. Dick. Genetic profiling of central nervous system dissemination of B-acute lymphoblastic leukemia reveals clonal selection and therapeutic vulnerability [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5173.

Adjusting hematopoietic hierarchy In adults, more than 300 billion blood cells are replenished daily. This output arises from a cellular hierarchy where stem cells differentiate into a series of multilineage progenitors, culminating in unilineage progenitors that generate over 10 different mature blood cell types. Notta et al. mapped the lineage potential of nearly 3000 single cells from 33 different cell populations of stem and progenitor cells from fetal liver, cord blood, and adult bone marrow (see the Perspective by Cabezas-Wallscheid and Trumpp). Prenatally, stem cell and progenitor populations were multilineage with few unilineage progenitors. In adults, multilineage cell potential was only seen in stem cell populations. Science , this issue p. 10.1126/science.aab2116 ; see also p. 126

Acute myeloid leukemia (AML) is a hierarchical disease in which the bulk blast population is sustained by a minority population of leukemia stem cells (LSC). Evidence of functional heterogeneity in the LSC compartment, including variable responses to chemotherapy, underscores the importance of examining the entire stem cell compartment in studies of LSC biology. However, there are currently no phenotypic markers that can consistently segregate LSCs within the leukemic blast population. Although LSC activity is most often enriched in the CD34+CD38- cell fraction, LSCs are also frequently detected in other phenotypic fractions, and in some cases are absent from CD34+ fractions. Thus, LSC studies that focus only on CD34+CD38- blasts may miss biologically important clones that are present in other phenotypic cell fractions. To identify novel markers that will enable better enrichment of LSC activity, we examined gene expression data obtained from functionally validated LSC+ and LSC- cell fractions sorted from primary AML samples, and identified CD200 as a candidate cell surface marker for LSCs. In normal adult bone marrow and cord blood samples, CD200 is expressed on >95% of CD34+CD38- cells, and expression decreases on CD34+38+ cells, suggesting that CD200, similar to CD34, is a stem cell marker. Flow cytometric analysis of AML patient samples (n=57) demonstrated that CD200 expression was present on a greater proportion of CD45dim blasts compared to CD45high non-blast populations (69.2% vs 4.5%, p To test whether CD200 can segregate LSC activity within the CD45dim blast population, we sorted 14 primary AML samples into cell fractions based on CD45 and CD200 expression followed by transplantation into cohorts of NSG mice. AML samples were prescreened for leukemic engraftment ability and were selected for sorting if: 1) CD45dim blasts comprised both CD200+ and CD200- cells or 2) the CD200+ fraction was High CD200 expression was significantly associated with shorter overall survival in univariate analysis in multiple independent AML cohorts, but was not significant in multivariate analysis due to association with NPM1 mutation. NPM1-mutated samples exhibited low CD200 expression on the bulk cells. Interestingly, 7 of the 9 AML patients with CD200+ LSCs had normal karyotype and NPM1 mutation. Our results demonstrate that a CD200-based sorting strategy can successfully enrich and/or segregate LSC activity, and separate LSCs from preL-HSCs in primary AML patient samples. This will now enable direct functional studies of the biological properties of these related but distinct stem cell populations. Furthermore, CD200 will be a valuable tool for the study of LSCs in the subset of NPM1-mutated AML. Disclosures No relevant conflicts of interest to declare.

Current cancer therapies are directed at molecular markers or dominant pathways present in the bulk of neoplastic cells. However, recent studies have identified many genetically distinct subclones co-existing within a single neoplasm. In over 50% of patients with relapsed acute lymphoblastic leukemia (ALL), the genetic clones present at relapse are not the dominant clone present at diagnosis, but have evolved from a minor or ancestral clone (Mullighan et al., Science, 2008). Previous work has shown that this subclonal diversity in B-ALL exists at the level of the leukemia-initiating cells (L-IC) capable of generating patient derived xenografts (Notta et al., Nature, 2011). In order to investigate the functional consequences of genetic clonal evolution during disease progression, we performed in-depth genomic and functional analysis of 14 paired diagnosis/relapse samples from adult and pediatric B-ALL patients of varying cytogenetics. Patient samples were subjected to whole exome sequencing (WES), SNP analysis and RNA sequencing. Diagnosis-specific, relapse-specific, and shared variants at both clonal and subclonal frequencies were identified. Limiting dilution analysis by transplantation of CD19+ leukemic blasts into 870 immune-deficient mice (xenografts) identified no significant trend in enrichment in L-IC frequency between paired patient samples with a median frequency of 1 in 2691. Despite similar frequencies of L-IC, functional differences within identically sourced patient xenografts were observed, including increased leukemic dissemination of relapse cells to distal sites such as the central nervous system (CNS), differences in engraftment levels and differences in immunophenotypes. Targeted-sequencing and copy number analysis of the xenografts, in comparison to the patient sample from which they were derived, has uncovered clonal variation and the unequivocal identification of minor subclones ancestral to the relapse in xenografts transplanted with the diagnostic sample from 8 patients. Some of these subclones are rare and were not captured through standard WES analysis of the patient samples, highlighting the value of xenografting to functionally identify and viably isolate subclones for further study. Interrogation of the therapeutic responses of the ‘relapse-like’ diagnosis subclones in secondary xenografts displayed differential resistance to standard chemotherapeutic agents (vincristine and L-asparaginase) pre-existing in the patient diagnosis samples prior to treatment. Furthermore, investigation of different sites of leukemic infiltration in the xenografts provided evidence of distinct clonal selection in the CNS, a known site of disease relapse, in comparison to the bone marrow. Using this data we can begin to draw the evolutionary paths to relapse. We have shown evidence that minor subclones at diagnosis, ancestral to the relapsing clone, possess functional advantages over other diagnostic clones. Overall, this work provides a substantial advance in connecting genetic diversity to functional consequences, thereby furthering our understanding of the heterogeneity identified in B-ALL and its contributions to therapy failure and disease recurrence. Citation Format: Stephanie M. Dobson, Robert Vanner, Esmé Waanders, Olga I. Gan, Jessica McLeod, Ildiko Grandal, Debbie Payne-Turner, Michael Edmonson, Zhaohui Gu, Xioatu Ma, Yiping Fan, Pankaj Gupta, Sagi Abelson, Michael Rusch, Ying Shao, Scott Olsen, Geoffrey Neale, John Easton, Cynthia J. Guidos, Jayne S. Danska, Jinghui Zhang, Mark D. Minden, Charles G. Mullighan, John E. Dick. Evolving functional heterogeneity in B-acute lymphoblastic leukemia. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr LB-341.

Current cancer therapies are directed at molecular markers or dominant pathways present in the bulk of neoplastic cells. However, recent studies have identified many genetically distinct subclones co-existing within a single neoplasm. In over 50% of patients with relapsed acute lymphoblastic leukemia (ALL), the genetic clones present at relapse are not the dominant clone present at diagnosis, but have evolved from a minor or ancestral clone (Mullighan et al., Science, 2008). Previous work has shown that this subclonal diversity in B-ALL exists at the level of the leukemia-initiating cells (L-IC) capable of generating patient derived xenografts (Notta et al., Nature, 2011). In order to investigate the functional consequences of genetic clonal evolution during disease progression, we performed in-depth genomic and functional analysis of 14 paired diagnosis/relapse samples from adult and pediatric B-ALL patients of varying cytogenetics. Patient samples were subjected to whole exome sequencing (WES), SNP analysis and RNA sequencing. Diagnosis-specific, relapse-specific, and shared variants at both clonal and subclonal frequencies were identified. Limiting dilution analysis by transplantation of CD19+ leukemic blasts into 870 immune-deficient mice (xenografts) identified no significant trend in enrichment in L-IC frequency between paired patient samples with a median frequency of 1 in 2691. Despite similar frequencies of L-IC, functional differences within identically sourced patient xenografts were observed, including increased leukemic dissemination of relapse cells to distal sites such as the central nervous system (CNS), differences in engraftment levels and differences in immunophenotypes. Targeted-sequencing and copy number analysis of the xenografts, in comparison to the patient sample from which they were derived, has uncovered clonal variation and the unequivocal identification of minor subclones ancestral to the relapse in xenografts transplanted with the diagnostic sample from 8 patients. Some of these subclones are rare and were not captured through standard WES analysis of the patient samples, highlighting the value of xenografting to functionally identify and viably isolate subclones for further study. Interrogation of the therapeutic responses of the ‘relapse-like’ diagnosis subclones in secondary xenografts displayed differential resistance to standard chemotherapeutic agents (vincristine and L-asparaginase) pre-existing in the patient diagnosis samples prior to treatment. Furthermore, investigation of different sites of leukemic infiltration in the xenografts provided evidence of distinct clonal selection in the CNS, a known site of disease relapse, in comparison to the bone marrow. Using this data we can begin to draw the evolutionary paths to relapse. We have shown evidence that minor subclones at diagnosis, ancestral to the relapsing clone, possess functional advantages over other diagnostic clones. Overall, this work provides a substantial advance in connecting genetic diversity to functional consequences, thereby furthering our understanding of the heterogeneity identified in B-ALL and its contributions to therapy failure and disease recurrence. Citation Format: Stephanie M. Dobson, Robert Vanner, Esme Waanders, Olga I. Gan, Jessica McLeod, Ildiko Grandal, Debbie Payne-Turner, Michael Edmonson, Zhaohui Gu, Xioatu Ma, Yiping Fan, Pankaj Gupta, Sagi Abelson, Michael Rusch, Ying Shao, Scott Olsen, Geoffrey Neale, John Easton, Cynthia J. Guidos, Jayne S. Danska, Jinghui Zhang, Mark D. Minden, Charles G. Mullighan, John E. Dick. Evolving functional heterogeneity in B-acute lymphoblastic leukemia. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr LB-341.

Background: The hematopoietic roadmap is a compilation of the various lineage differentiation routes that a stem cell takes to make blood. On several occasions over the last six decades, the murine roadmap has been reconceived due to new information overturning old dogmas. The human roadmap, which describes the extraordinary throughput of more than three hundred billion cells daily has changed little, with the classical model of hematopoiesis still prevailing. In this model, blood differentiation initiated at the level of stem cells must pass through a series of increasingly lineage-restricted oligopotent and then unipotent progenitor intermediates. As cellular differentiation flows from stem cells to unipotent cells, the oligopotent progenitors are the quintessential step where lineage choices are consolidated. One such critical oligopotent intermediate is the common myeloid progenitor (CMP), believed to be the origin of all myeloid (My), erythroid (Er) and megakaryocyte (Mk) cells. While murine studies are beginning to challenge the existence of oligopotent progenitors, a comprehensive analysis of human myelo-erythroid differentiation is lacking. Moreover, whether the pool of oligopotent intermediates is fixed across human development is unknown. The classical roadmap is the only guide that describes human hematopoiesis and consequently our view of blood homeostasis and disease is intimately tied to this conception. Rationale: The differentiation roadmap taken by human hematopoietic stem cells (HSCs) is fundamental to our understanding of blood homeostasis, hematopoietic malignancies and regenerative medicine. Results: We mapped the cellular origins of My, Er and Mk lineages across three timepoints in human blood development: fetal liver (FL), neonatal cord blood (CB) and adult bone marrow (BM). Using a new cell-sorting scheme based on markers linked to Er and Mk lineage specification (CD71 and BAH1), we found that previously described populations of multipotent progenitors (MPPs), CMPs, and megakaryocyte-erythroid progenitors (MEPs) were considerably heterogeneous and could be further purified beyond current definitions. Nearly 3000 single cells from 11 cellular subsets from the CD34+ compartment of FL, CB and BM (33 subsets in total) were evaluated for their My, Er and Mk lineage potential using an optimized single-cell assay. In FL, the ratio of cells with multilineage versus unilineage potential remained constant in both the stem cell (CD34+CD38-) and progenitor cell (CD34+CD38+) enriched compartments. By contrast in BM, nearly all multipotent cells were restricted to the stem cell compartment, whereas unilineage progenitors dominated the progenitor cell compartment. Oligopotent progenitors were only a negligible component of the human blood hierarchy of adult BM leading to the conclusion that multipotent cells differentiate into unipotent cells directly. Mk-Er activity predominately arose in the stem cell compartment at all developmental timepoints. In CB and BM, most Mks emerged as part of mixed clones from HSCs/MPPs, indicating that Mks directly branch from a multipotent cell and not from oligopotent progenitors like CMP. In FL, a striking 80% of single-cell clones with Mk activity were derived from a novel progenitor in the stem cell compartment, although less potent Mk progenitors were also present in the progenitor compartment. Interestingly, in a hematological condition of HSC loss (aplastic anemia), Mk-Er but not My progenitors were more severely depleted, pinpointing a close physiological connection between HSC and the Mk-Er lineage. Conclusion: Our data indicate that there are distinct roadmaps of blood differentiation across human development. Prenatally, Mk-Er lineage branching occurs throughout the cellular hierarchy. By adulthood, both Mk-Er activity and multipotency are restricted to the stem cell compartment, whereas the progenitor compartment is composed of unilineage progenitors forming a 'two-tier' system, with few intervening oligopotent intermediates. Disclosures No relevant conflicts of interest to declare.

Introduction: Despite risk stratification according to presenting clinical and genetic features, 10-25% of children with acute lymphoblastic leukemia (ALL) relapse, which is associated with a poor prognosis. Here, we sought to provide a comprehensive overview of the genetic alterations associated with relapse in ALL. Methods: We studied 93 children (27 female, 43 male) diagnosed with ALL (62 B-progenitor, 25 T-lineage) between 1987 and 2008 and treated on total therapy studies XI-XVI who experienced relapse and/or a second tumor. Age at diagnosis ranged from 3 months to 18 years. Median time to relapse was 3 years (range 3 months to 10 years). Seventy patients had a single relapse, 15 cases had 2 relapses, and 8 cases developed a second tumor of different lineage (B-cell lymphoma, chronic myeloid leukemia (n=1 each) and acute myeloid leukemia (n=6)). Diagnosis, relapse and matched normal samples (n=299) were studied using Affymetrix SNP 6.0 microarrays and whole genome or whole exome sequencing. Results: We found 2692 copy number aberrations (CNAs) with a median of 9 (range 0-109) in the diagnosis samples (n=91) compared to a median of 10 (range 0-112) in the relapse samples (n=89) and 12 (range 0-70) in subsequent samples (n=20). The number of CNAs did not differ significantly between diagnosis, relapse or subsequent samples. We identified a 7286 non-silent single nucleotide variants (SNVs) and small insertions or deletions (indels) in 5002 genes, 1392 of which were recurrent. The median number of variants was 12 (range 0-70) at diagnosis (n=91), 21 (range 0-858) at relapse (n=91; P=0.0029 v. diagnosis) and 60 (range 10-650) in subsequent samples (n=20; P The most frequently mutated genes were NOTCH1 (n=33), NRAS (n=24), CREBBP (n=20) and KRAS (n=16). Of the recurrently altered genes, only 87 genes were known to be affected in cancer (Cancer Gene Census, COSMIC database), of which 59 were affected in leukemia and lymphoma tissues, indicating that we have identified 1306 novel recurrently affected genes, most commonly C13orf40 and MKI67. Mutations in epigenetic regulators were particularly frequent, with genes mutated in at least 3 cases altered in over 60% of the cohort (e.g. CREBBP, EP300, MLL2, MLL3, KDM6A/B, CTCF, SETD2, TET2/3, and EZH2). Clonal evolution analyses showed multiple patterns of evolution, with relapses sharing either few or many variants with the diagnosis sample in a frequency that reflects both predominant clones and minor subclones propagating relapse. Variants in NOTCH1, NRAS, and CREBBP were preserved from a major clone at diagnosis in 4, 6, and 5 cases respectively, but acquired at relapse or grown out from a minor subclone at diagnosis in 3, 5, and 8 cases respectively. In contrast, variants in USH2A (n=4), FOXA1 (n=3), and purine/pyrimidine synthesis pathway genes NT5C2 (n=3), PRPS1 (n=3) and NT5C1B (n=1) were exclusively found in relapse samples. Notably, the NT5C2 mutations, which are thought to confer resistance to thiopurines, were subclonal at relapse in the majority of cases. We identified 13 cases (10 B-lineage, 3 T-lineage) in which the diagnosis and relapse were fully discordant for all CNAs and sequence mutations, only 4 of which showed a prolonged remission time (>5 years). This suggests that these patients developed a second primary malignancy and may be predisposed to leukemia development. Indeed, one case revealed focal amplifications on chromosome 1q21.1 encompassing the neuroblastoma breakpoint family genes, which are implicated in cancer development. Comprehensive germline analyses are underway. Conclusion: This study has provided detailed insight into the genetic basis of relapse, implicating multiple new genes and pathways involved in treatment resistance, demonstrating multiple patterns of clonal evolution, and revealing an unexpectedly high frequency of genetically discordant second malignancy in relapse in ALL. Disclosures Evans: Prometheus Labs: Patents & Royalties: Royalties from licensing TPMT genotyping. Mullighan:Amgen: Honoraria, Speakers Bureau; Cancer Science Institute: Membership on an entity's Board of Directors or advisory committees; Incyte: Consultancy, Honoraria; Loxo Oncology: Research Funding.

Recent studies have shown that several miRNA are differentially expressed in hematopoietic stem cells (HSC) and involved in regulating self-renewal, pointing to a new axis of epigenetic control of HSC function. Murine studies have documented a role for miR-125a in regulating HSC as miR-125a enforced expression augments self-renewal. We examined whether these attributes are evolutionarily conserved within human hematopoiesis. Lentiviral vectors over-expressing miR-125a (miR-125OE) were developed and HSC function was investigated using xenotransplantation of CD34+ CD38- human umbilical cord blood (CB) hematopoietic stem and progenitor cells (HSPCs). miR-125OE resulted in significantly increased human bone marrow (BM) chimerism at 12 and 24 weeks post-transplantation and splenomegaly. Within enlarged spleens, there were significantly increased proportions of CD34+CD19+CD10+CD20-B lymphoid cells suggesting a partial B cell differentiation block at the pro-B cell stage. In the BM, CD41+ megakaryocytes, GlyA+ erythroid and CD3+ T cell populations were significantly expanded. Within the primitive compartment, multi-lymphoid progenitors (MLP) were massively expanded by 12 weeks, followed by a combined reduction of immuno-phenotypic HSC and multi-potent progenitors (MPP) by 24 weeks. Given this loss of immuno-phenotypic HSC, we wondered whether stem cell function was compromised in vivo. Secondary transplantation with limiting dilution (LDA) revealed that stem cell frequencies were increased by 4.5 fold in miR-125OE recipients. Using lentivirus sponge-mediated inhibition of miR-125 (miR-125KD) in CD34+CD38-human CB, we were able to directly link these effects to miR-125: B cells increased at the expense of T cells; immuno-phenotypic HSC increased with a concomitant loss of MLP; and functional HSC were decreased by 2.5 fold using secondary LDA assays. Together, these data strongly suggest that miR-125a expression levels regulate human HSC self-renewal and lineage commitment. Since HSC frequency increased so substantially upon miR-125OE, we asked whether more committed cell populations might also be endowed with enhanced self-renewal. Highly purified populations of HSC, MPP and MLP and CD34+CD38+ committed progenitors were transduced and transplanted cells into xenografts. Unexpectedly, miR-125OE transduced CD34+CD38+ progenitors produced a substantial graft after 12 weeks. Control transduced CD34+CD38+ cells did not engraft and only control transduced HSC generated a disseminating graft in recipient mice. miR-125OE transduced HSC and MPP generated robust engraftment, while MLP did not. In all cases, xenografts generated by CD34+CD38+ and MPP transduced with miR-125OE showed multi-lineage repopulation. Moreover, the miR-125OE grafts from CD34+CD38+ and MPP recipients were durable as secondary transplantation generated multi-lineage grafts for at least 20 weeks in 5/7 and 6/10 recipients, respectively; no control transduced groups generated secondary grafts. Thus, the enhancement of self-renewal by enforced expression of miR-125a occurs not only in HSC, but also in MPP and to an as yet unidentified subpopulation within the CD34+38+ committed progenitor compartment. Using protein mass spectrometry, we identified and validated a miR-125a target network in CD34+ CB that normally functions to restrain self-renewal in more committed progenitors. Together, our data suggest that increased miR-125a expression can endow an HSC-like program upon a selected set of non-self-renewing hematopoietic progenitors. Our findings offer the innovative potential to use MPP with enhanced self-renewal to augment limited sources of HSC to improve clinical outcomes. Disclosures No relevant conflicts of interest to declare.