ALBERT

Description: Background: C/EBPa is a lineage determining transcription factor, critical for terminal cell differentiation in different tissues including the bone marrow (BM), lung, liver, and adipocytes. Adequate CEBPA levels are needed to maintain the haematopoietic stem cell (HSC) pool and to promote neutrophilic differentiation. This knowledge points towards the importance of CEBPA dosage at different stages of differentiation. Aim: To investigate how CEBPA dosage is regulated at the transcriptional level by specific enhancer(s) and to study their role in haematopoiesis. Results: Chromatin immunoprecipitation for the active mark H3K27ac followed by deep sequencing (ChIP-seq), revealed eight putative regulatory elements within the CEBPA locus. One enhancer showed marked H3K27ac enrichment at +42Kb downstream of CEBPA particularly in CD34+ haematopoietic stem cells (HSCs), implying a role for CEBPA regulation at earlier stages of haematopoiesis. ChIP-Seq experiments revealed binding of RUNX1, ERG, PU.1, FLI1 and GATA2 to this +42Kb enhancer in CD34+ BM cells. Moreover, myeloid cell lines MOLM1 and U937 also showed H3K27ac enrichment at this enhancer, indicating that its activity is maintained upon myeloid differentiation. In contrast, active histone marks were completely devoid at this element in CEBPA expressing lung (A549) and liver (HepG2) cell lines, indicating hematopoietic specificity of this enhancer. Chromatin looping between the +42Kb enhancer and the CEBPA promoter was demonstrated by 4C-Seq, highlighting the specificity of this enhancer in CEBPA regulation. Furthermore, using CRISPR/Cas9 technology we deleted the +42Kb enhancer in the myeloid cell line THP-1 and showed 70% reduction of CEBPA mRNA levels. The +42Kb enhancer is conserved and located at +37Kb near Cebpa in mice. Using CRISPR/Cas9 system, we deleted the +37Kb enhancer in one-cell stage zygotes. Heterozygous mice were inter-crossed and F1 generation mice were born at normal mendelian ratios. Morphological and flow-cytometric analysis of peripheral blood and BM at 8-10 weeks of age, showed 10-20 fold decrease in (MAC1+/GR1+) neutrophil counts in homozygous +37Kb-/- mice, compared to +37Kbwt/wt and +37Kb-/wt controls. In line with the block of neutrophil development, flow-cytometric analysis revealed an increase (2 fold) in CD34+CD16/32low common myeloid progenitors and a decrease (2 fold) in the CD34+CD16/32high granulocyte/ monocyte progenitorsof +37Kb-/- mice. From these findings we hypothesized that the absence of the +37Kb enhancer disturbs the myeloid differentiation program via reduced Cebpa levels. In fact, Cebpa expression levels were reduced by 60-80% in bone marrow of +37Kb-/- mice, but unchanged in other Cebpa -expressing tissues such as lung and liver, indicating tissue specificity of this enhancer. Diminished Cebpa expression levels were accompanied by decreasedexpression of Cebpa target genes, including Csf3r. In line with this, bone marrow progenitor cells from +37Kb-/- mice were completely unresponsive to Csf3 in a colony forming assay. Given the importance of Cebpa in HSCs maintenance, we investigated the HSC population and found that long-term HSCs (CD48- CD150+) and short-term HSCs (CD48- CD150-) were depleted in the bone marrow of the +37Kb-/- mice. HSC depletion was accompanied by an increase in the CD48+/CD150- multipotent progenitors (MPPs). The +37Kb-/- MPPs, unlike controls, were able to serially replate in vitro under IL-3, GM-CSF, IL-6, SCF growth factor conditions with minimal evidence of differentiation, suggesting a leukemogenic potential. Reintroduction of Cebpa cDNA into +37Kb-/- MPPs fully recovered neutrophil development. Conclusion: We conclude that the +37Kb enhancer is tissue-specific and plays a central role in haematopoiesis regulating Cebpa dosage. Our study reveals that the bone marrow maintains its integrity through the activity of the +37Kb enhancer, which (1) prevents HSC exhaustion and (2) preserves neutrophilic development. The in vitro replating capacity of MPPs isolated from +37Kb-/- animals suggests that aberrant control of this enhancer may be a primary leukaemogenic event. In line with this, it is important to note that the conserved enhancer in humans (+42 KB) is a frequent target for oncogenic transcription factors such as AML1-ETO or EVI1, two oncogenes which are found in two distinct subtypes of AMLs with very low C/EBPa expression. Disclosures No relevant conflicts of interest to declare.

Description: Introduction: Acute myeloid leukemia (AML) with inv(3)(q21q26) or t(3;3)(q21;q26) [inv(3)/t(3;3)] is associated with aberrant expression of the stem cell regulator EVI1 and dismal prognosis. Recently, we and others (Gröschel et al, Cell, 2014; Yamazaki et al, Cancer Cell, 2014) have shown, as a consequence of the inv(3)/t(3;3) rearrangements, that the proto-oncogene EVI1 is activated upon the structural repositioning of a distal GATA2 enhancer from 3q21 to EVI1, coinciding with loss of GATA2 expression from the rearranged allele. Notably, GATA2 deficiency has been shown to impair hematopoietic stem cell frequency and function (Lim et al, J. Clin. Invest., 2012) and Evi1 activation in inv(3) murine models is followed by leukemia onset after a long latency of 6 months (Yamazaki et al). We therefore hypothesize that additional cooperating genetic lesions, other than EVI1 activation and GATA2 deregulation, are required for full leukemic transformation. We sought to extend the molecular characterization of inv(3)/t(3;3) myeloid malignancies through next-generation sequencing. Methods: We selected 32 AML (including cell lines MUTZ-3 and UCSD-AML1), 4 CML-BC (including cell lines HNT-34 and MOLM-1), and 5 myelodysplastic syndrome (MDS) cases with inv(3)/t(3;3) rearrangements and performed whole transcriptome sequencing (n=40) as well as whole exome sequencing (WES, n=10) on the Illumina HiSeq 2500. For WES germline control we used DNA from cultured CD3+ T-cells from diagnostic bone marrow specimens. Variants were examined when recurrently detected in more than two patients or previously associated with leukemogenesis or cancer pathogenesis. Gene expression profiles (GEP) were constructed for 24 cases for subsequent analyses. Results: All examined variants, validated by Sanger sequencing, were assigned to mutational categories to discern patterns of mutations (Figure 1A). We observed that all 41 cases acquired at least one mutation belonging to the mutational categories reported as relevant for leukemia pathogenesis. Importantly, all AML and CML-BC patients, as well as 4 out of 5 MDS patients contained mutations in genes activating RAS/RTK signaling, amounting to 98% of all inv(3)/t(3;3) myeloid malignancies. An aggregate of 74% of all patients harbor mutually exclusive mutations in RAS-pathway associated genes; NRAS(27%), PTPN11(20%), KRAS(11%), and NF1(9%). Additionally, we found mutations in genes affecting signaling pathways involving RAS; FLT3(13%), BCR-ABL(11%), and KIT(2%) (Figure 1B). NPM1, CEBPA, and IDH1/2 mutations were absent in all cases. Furthermore, frequent mutations were found in transcription factor (32%), splice factor (29%), epigenetic modifier (29%), and tumor suppressor (10%) encoding genes (Figure 1A). GATA2 is the most commonly mutated transcription factor in inv(3)/t(3;3) myeloid malignancies (15%). As these mutations were found on the non-rearranged GATA2 allele, only mutant GATA2 is expressed due to GATA2 silencing mediated by the 3q21q26 rearrangement. Novel somatic truncating mutations and copy number loss were detected in the transcription factor FOXP1(10%). Additionally, we observed frequent mutations in the splice factor gene SF3B1(27%) in AML and MDS cases. Finally, the predominant monosomal karyotype within inv(3)/t(3;3) myeloid malignancies is contrasted by the low incidence of TP53 mutations (5%). No mutational pattern alluded to the high coincidence of monosomy 7 (68%), but inv(3)/t(3;3)/-7 cases were characterized by a strong down regulation of the transcription factor CUX1. In an effort to discriminate MDS and AML with inv(3)/t(3;3) we performed cluster and principle component analyses, revealing no cluster formation. Additionally, differential expression analysis was unable to differentiate AML from MDS, indicating strong homogeneity based on GEPs. Conclusion: Myeloid malignancies with inv(3)/t(3;3) harbor a common set of molecular alterations, i.e. EVI1 activation and GATA2 deregulation, conjugated by mutations activating signaling pathways. Mutational patterns and GEPs precluded the discrimination of MDS and AML with inv(3)/t(3;3), supporting the notion that these myeloid disorders should be categorized as AML irrespective of blast count. Exploiting signaling pathways therapeutically may prove a valuable adjunct to the scarce armamentarium of drugs effective in this malignancy. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

Description: Changes in regulation of transcription factor (TF) networks are a hallmark of acute myeloid leukaemia (AML). Deregulation in expression of the leucine zipper transcription factor C/EBPa, is common in AML with recurrent oncogenic events, including 3q-rearrangementsEVI1+ or translocation t(8;21)RUNX1-ETO. We aimed to study how the chromatin organization of the CEBPA locus control transcription in myeloid cells and how AML-related oncoproteins functionally interfere with these enhancers to deregulate CEBPAexpression. Human CEBPA is located on the short arm of chromosome 19 in a sub-topological associated domain that spans a genomic region of approximately 170kb. CEBPA is flanked by CEBPG and SLC7A10, at the 5’ and 3’, respectively. We hypothesized that upon myeloid commitment the 3D chromatin configuration of the locus facilitates the transcriptional activation of CEBPA by bringing enhancers in proximity to the CEBPA promoter and therefore enhances its expression. To identify potential enhancers that interact with the CEBPA promoter, we applied circularized chromatin conformation capture-sequencing (4C-seq) and compared the chromatin configuration of myeloidCEBPA+ (n=4) with that of lymphoidCEBPA- (n=2) cell lines. Using the CEBPA promoter as a viewpoint, all cell lines investigated showed a similar profile of interactions of multiple chromatin sites, localized within the locus. Non-haematopoietic cell lines also exhibited the same profile of interactions, suggesting no restricted chromatin configuration specific for haematopoietic origin. Our 4C-seq data show that the chromatin organization of the locus is highly conserved across different tissues, independent of CEBPA expression. However, the major differences between myeloidCEBPA+ cells and lymphoidCEBPA- are shown by the abundance of active enhancer marks, H3K27ac and p300 in chromatin immunoprecipitation-sequencing (ChIP-seq) analysis. These enhancers are located at the 3’ end of CEBPA and concentrated in a genomic cluster containing at least 6 potential enhancers. Non-haematopoieticCEBPA+ cells also showed multiple active enhance marks in ChIP-seq analysis at the 3’ end of CEBPA, but exhibited a different pattern than myeloidCEBPA+ cells, suggesting tissue specificity. At least two potential enhancers were exclusive for myeloid cells. These enhancers were investigated for transactivation potential in a luciferase based reporter system where only in myeloidCEBPA+ cells, as compared to non-haematopoieticCEBPA+cell lines, showed high transactivation. Combining the complete 4C-seq and ChIP-seq set of data, we found a potential myeloid-specific enhancer as a strong candidate, located at +42kb downstream from CEBPA. A combination of in-house and public available ChIP-seq data show selective binding of haematopoietic stem cell and myeloid related TFs to this enhancer, such as LMO2, SCL, RUNX1, ERG, GATA2, and PU.1, in human CD34+ bone marrow cells and myeloid cell lines. In AML transformed cell lines and in human AML samples, EVI1 and AML1-ETO show strong binding to this enhancer by ChIP-seq, which may provide an explanation for the strongly reduced CEBPA expression levels in corresponding leukaemias. To study the role of this enhancer and CEBPA expression, we conducted CRISPR/Cas9 genome editing and deleted this enhancer (800bp) in a myeloidCEBPA+ cell line, THP-1. CEBPA expression was reduced in all targeted clones by 5 fold when compared to control clones. In summary, we showed that the 3D chromatin organization of the human CEBPA locus is pre-configured at an early stage of development and thus, it is maintained across different cell types independent of CEBPA expression. Combining functional genomics data, we identified a novel human CEBPA enhancer important for CEBPA expression in myeloid cells. Oncogenic usage of this enhancer implies involvement in AML and can, partially, explain CEBPA deregulation in human AML subtypes. Disclosures No relevant conflicts of interest to declare.

Description: Key Points The DNA glycosylase MBD4 acts as a safeguard against damage from 5mC deamination. Germ line MBD4 deficiency stimulates clonal hematopoiesis and guides the development of leukemia via recurrent mutations in DNMT3A.

Description: Shwachman-Diamond Syndrome (SDS) is a congenital bone marrow failure disorder characterized by neutropenia and predisposition to leukemia. SDS is associated with loss-of-function mutations in the SBDSgene, involved in ribosome biogenesis, but the cellular and molecular events driving neutropenia in SDS remain poorly defined, largely due to a lack of mammalian disease models recapitulating the hematopoietic features of SDS. To achieve deletion of Sbds in early hematopoietic/myeloid progenitors, we generated Cebpacre/+Sbdsf/f R26 EYFP/+ mice (Sbds Δ/Δ), which were non-viable. Analysis of E14.5 embryos demonstrated global conservation of the hematopoietic hierarchy in Sbds Δ/Δ embryos compared to Cebpacre/+ R26 EYFP/+ controls (Sbds +/+). Functional relevance of Sbds deletion in Cebpa+progenitors was tested by transplanting fetal liver cells from E14.5 Sbds Δ/Δ or Sbds +/+ embryos into lethally irradiated B6.SJL mice. Deficiency of Sbds in different EYFP+ hematopoietic subsets was confirmed by qPCR (log2 fold change: LKS -2.25, CMP -9.42, GMP -9.1 and neutrophils -7.29). Mice transplanted with Sbds Δ/Δ cells developed profound neutropenia (log2FC: -2.56; p=0.002, n=7), which was stable during the time of analysis (16 weeks). FACS and morphological studies of bone marrow EYFP+ cells demonstrated increased frequencies of early progenitor populations, recapitulating the left-shifted hematopoiesis observed in human SDS patients (Dror et al., Ann N Y Acad Sci 2011), with a pronounced accumulation of cKitint Gr1low EYFP+ myelocytes-metamyelocytes (MC-MMs) (frequency of EYFP+ cells: Sbds +/+ 15.3±1.7 %, Sbds Δ/Δ 39.4±1.1 %; p= 8.7x10-8) and a marked decrease in Gr1+ Mac1+ cells (Sbds +/+: 54.5±7.5 %, Sbds Δ/Δ: 22.3±2.8 %; p= 1.6x10-4), indicating that neutropenia was caused by disrupted lineage progression from MC-MMs to neutrophils. In line with this, whole transcriptome analysis of prospectively isolated EYFP+ MC-MMs (RNA-seq, n=4) revealed enrichment for hematopoietic stem and progenitor cell signatures in Sbds Δ/Δ recipients, while myeloid signatures were enriched in Sbds +/+ mice (GSEA). Transcript analysis further showed reduced expression of granule components produced at the MC-MM stage, such as Ltf, Mmp8, Mmp9. As expected, reduced expression of Sbds (FDR=6.6x10-9, log2FC=-2.67) and dysregulation of ribosome proteins (RP) production were observed, with increased expression of over 60 RP genes in SbdsΔ/Δ MC-MMs. To investigate the molecular events underlying impaired lineage progression, we focused on p53 activation, a common mechanism of tissue failure in ribosomopathies. P53 protein was significantly and selectively accumulated in EYFP+ MC-MMs of Sbds Δ/Δ recipients as measured by FACS (Mean Fluorescence Intensity FC: 1.8; p=0.01), with increased expression of p53 targets such as Cdkn1a (p21), Bbc3 (PUMA) and Bax. This was accompanied by increased apoptosis (annexin V FACS analysis) in both MC-MMs and mature neutrophils (FC: 1.36; p=0.02 and FC=2.44; p=1.6x10-4, respectively). However, pharmacological inhibition of the p53 pathway by pifithrin-α (2 months, i.p.) failed to rescue neutropenia, suggesting that alternative mechanisms, such as disrupted expression of transcription factors, may contribute to neutropenia in SDS. Transcript analysis of Sbds-deficient MC-MMs revealed downregulation of Rara (Retinoic acid receptor α) and its target Cdkn1b (p27), implicated in cell cycle arrest and terminal granulopoiesis (Walkley et al., Blood 2004). Congruent with this finding, Ki67 staining showed that end-stage granulocytes from Sbds Δ/Δ recipients fail to exit cell cycle. In conclusion, we established a mouse model of neutropenia caused by Sbds deficiency in the hematopoietic system, providing experimental support for a direct causative link between Sbds deficiency and neutropenia in mammals. The data reveal a previously unanticipated, selective dependency of late myeloid cells on this ubiquitous protein, while its deficiency spares the function of rapid cycling hematopoietic progenitors. Mechanistically, disrupted expression transcription factors governing terminal myeloid differentiation may be implicated. We anticipate the mouse model to be a valuable tool in further dissecting the molecular pathogenesis of SDS and enabling preclinical studies for disease modulation. Disclosures No relevant conflicts of interest to declare.

Description: Severe congenital neutropenia (SCN) is characterized by a maturation arrest at the promyelocyte stage and consequently a severe reduction of peripheral neutrophils. Administration of colony stimulating factor 3 (CSF3) restores neutrophil levels in over 90% of SCN patients, leading to an improved survival rate. SCN patients have an increased risk to develop secondary MDS or AML. Leukemic progression of SCN frequently involves the acquisition of a mutation in the gene encoding CSF3 receptor (CSF3R) in the neutropenic phase, followed by a mutation in runt-related transcription factor 1 (RUNX1) prior to transformation to MDS/AML (Skokowa et al, Blood, 2014). CSF3R mutations in SCN truncate the CSF3R C-terminus. RUNX1 mutations in SCN/AML are predominantly single nucleotide variations in the Runt-homology domain (RHD). Frameshift/nonsense mutations in the transactivation domain (TAD) also occur, albeit less frequently. To investigate how the combination of CSF3R and RUNX1 mutations affects myelopoiesis, we isolated hematopoietic stem and progenitor cells (HSPCs) from mice expressing either wild-type (WT) or a truncated (d715) CSF3R, retrovirally transduced them with TAD (S291fsX9) or RHD (D171N) mutants of RUNX1, or an empty vector control (EV), and cultured the cells in the presence or absence of CSF3. FACS analysis was performed to assess numbers of hematopoietic stem cells (LSKs, Lin- Sca1+ Kit+), early myeloid progenitors (LKs, Lin- Sca1- Kit+), common myeloid progenitors (CMPs, LK CD34+ CD16/32low), granulocyte-monocyte progenitors (GMPs, LK CD34+ CD16/32hi), immature (CD11b+, Gr-1lo) and mature (CD11b+, Gr-1hi) neutrophils. Whereas a normal differentiation pattern was observed in the CSF3R-WT background, activation of CSF3R-d715 led to defective neutrophil differentiation. Introduction of the RUNX1 mutants further aggravated the differentiation block and resulted in a selective expansion of HSPC subsets. While the EV and CSF3R-WT controls showed a relatively equal distribution of LK and LSK cells over time, the combination of CSF3R-d715 and RUNX1-TAD led to a 10-fold increase in absolute LSK numbers relative to EV (LSKs day 9: TAD: 1.23x107, EV: 1.27x106, RHD: 4.86x105, n=3, p=0.05) and a 7:1 ratio of LSK over LK. In contrast, the combination of CSF3R-d715 and RUNX1-RHD predominantly expanded LKs (LKs day 9: RHD: 8.77x106, EV: 2.77x106, TAD: 1.69x106, n=3, p=0.027) resulting in a 18:1 ratio of LK over LSK. Further analysis showed that these LKs were blocked in differentiation at the CMP to GMP transition stage (CMP to GMP ratio: RHD: 111.5, EV: 9.3, TAD: 1.8). To interrogate which mechanisms are responsible for the distinct differentiation defects caused by the RUNX1-RHD and -TAD mutants, we performed RNA-Seq on FACS purified LSK and LK populations. CMP to GMP transition is controlled by CCAAT/enhancer-binding proteins (C/EBPs), mainly C/EBPα and C/EBPβ. The combination of CSF3R-d715 and RUNX1-RHD, blocking the transition from CMP to GMP, induced the expression of C/EBPγ, a strong antagonist of C/EBP transcriptional activity (FPKM: day 2: 107.2, day 5: 424.3, day 9: 801.3). A major function of CEBPs in driving CMP to GMP transition is to suppress E2F- and Myc-driven transcription of genes involved in cell cycling. Consistent with this, single sample gene set enrichment analysis with output of hallmark pathways (Broad Institute) showed de-repression of E2F (p=0.002) and Myc (p=0.036) pathways in CSF3R-d715/RUNX1-RHD LKs relative to CSF3R-d715/EV. Interestingly, the RUNX1-TAD mutant that selectively expanded LSKs in combination with CSF3R-d715 did not alter C/EBP expression and function but affected genes involved in ribosomal biogenesis. We conclude that RUNX1-TAD and RHD mutants differentially cooperate with the CSF3R truncating mutations that are frequently acquired in the neutropenic phase of SCN and take alternative routes for LSK versus LK expansion. How this affects the leukemogenic nature of the RUNX1 mutations in combination with CSF3R truncation and how we can functionally interfere with these mechanisms is currently under investigation in in vivo transplantation settings. Disclosures No relevant conflicts of interest to declare.