Abstract: Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is an aggressive acute leukemia/lymphoma recently classified as a malignant transformation of plasmacytoid dendritic cells (pDCs) and a subtype of acute myeloid leukemia (AML). BPDCN has no standard treatment and a poor prognosis, with median survival 〈 1 year. A significant roadblock to better understanding BPDCN is a lack of adequate model systems. We generated patient-derived xenografts (PDX) of BPDCN in NOD/Scid/IL2rgnull (NSG) mice. Bone marrow or peripheral blood cells involved by BPDCN blasts (CD45 low, CD123 high, HLA-DR high, CD3 neg) were transplanted into irradiated NSG recipients. Nine of 16 BPDCNs caused lethal leukemia involving blood, spleen, and bone marrow 2-6 months after transplantation. All nine BPDCN PDXs were serially transplantable. Flow characterization of each patient's BPDCN and corresponding xenograft revealed no major differences in BDCA2, BDCA4, FCeR1, ILT7, or cytoplasmic TCL1 staining. All samples maintained high expression of the human interleukin-3 (IL3) receptor (IL3Ralpha/CD123), a hallmark feature of BPDCN. To further characterize BPDCN pathogenesis we performed whole transcriptome sequencing (RNA-seq) on sorted blasts from 11 patients and on normal pDCs isolated from 4 healthy donors. These were compared to RNA-seq in six PDXs. The spectrum of mutations in BPDCN transcriptomes overlapped with that seen in other hematologic malignancies, particularly myeloid disorders, and was similar to reported DNA mutations in BPDCN, including in ASXL1, CTCF, IDH2, NRAS, RUNX1, STAG2, TET2, and TP53. Particularly striking was the presence of a canonical mutation in an RNA splicing factor in 7 of 11 cases (SRSF2 P95H/L/R in four, ZRSR2 R295* and gene locus deletion in two, and SF3B1 K666N in one). Known oncogenic mutations in the original disease were retained in the PDXs, including all splicing factor mutations, with the exception of an IDH2 R140Q that was lost in one PDX. BPDCN PDXs grouped together in unsupervised clustering of expression profiles, distinct from AML and ALL PDXs in an analysis of 134 models from the DFCI Public Repository of Xenografts (http://PRoXe.org). Gene set enrichment analysis (GSEA) of KEGG and REACTOME pathways associated with differentially expressed genes between primary BPDCNs and non-malignant pDCs revealed signatures related to dendritic cell activation, cell cycle, and apoptosis. In addition, 3 of the top 11 sets were genes involved in mRNA processing, mRNA splicing, and processing capped intron-containing pre-mRNAs (all FDR 〈 1e-6). To test the efficacy of BPDCN-targeted therapy using primary human leukemias in vivo, we performed a pre-clinical trial in NSG mice using SL-401, a recombinant biologic consisting of a fusion protein of IL3 and diphtheria toxin. Three independent BPDCN xenografts were injected into 20 NSG mice each, and followed by weekly peripheral blood monitoring for human CD45 and CD123. When leukemia burden reached 〉 0.5% in at least half of the mice in the cohort, animals were randomized to receive SL-401 at 100 ug/kg or vehicle intraperitoneally daily for 5 days. Two mice in each group were sacrificed at day 7 for response assessment, and peripheral blood was followed weekly in the remaining mice for evidence of progression ( 〉 5% human CD45/CD123-positive cells). 7 days after treatment, mice receiving SL-401 had dramatic reductions in BPDCN in the peripheral blood, spleen, and bone marrow (0.31% vs 37.6% in marrow of SL-401 vs vehicle). SL-401 prolonged progression-free survival in all BPDCNs tested (12 vs 48 days, P 〈 0.0001 by log-rank test). At the time of progression after SL-401, relapsing mice were re-randomized to receive a 2nd and in some cases 3rd cycle of SL-401 or vehicle. Repeated treatment in mice that progressed after SL-401 resulted in second and third peripheral blood remissions. All PDXs responded to SL-401 including those with and without splicing factor and TP53 mutations. CD123 expression was maintained at high levels on all SL-401 treated BPDCNs even after repeated cycles. Primary xenografts of BPDCN are faithful models of the human disease, maintain genetic and transcriptomic characteristics of the original tumor, and respond to multiple courses of IL3-DT in vivo, suggesting that they provide a valuable resource to study disease biology and response/resistance to targeted therapy. Disclosures Chen: Stemline Therapeutics, Inc.: Employment. Brooks:Stemline Therapeutics, Inc.: Employment, Equity Ownership, Patents & Royalties. Lane:Stemline Therapeutics, Inc.: Research Funding.

Abstract: Blastic plasmacytoid dendritic cell neoplasm is a clonal disease derived from precursors of plasmacytoid dendritic cells (pDC). It is a rare neoplasm involving the skin which may or may not be associated from the outset with a leukemic component. The disease invariably progresses to aggressive leukemic dissemination, leading to a differential diagnosis with acute leukemia. In 2004, we set up a French network to recruit biological data at diagnosis. Diagnosis was according to recommendations (Swerdlow et al, 2008), with, in addition, a mandatory panel of pDC markers (Garnache-Ottou et al, 2009) detected by flow cytometry or by immunohistochemistry on infiltrated blood, bone marrow or cutaneous lesions. In total, 109 cases of BPDCN were included in 35 hospitals (2000-2013). BPDCN is more prevalent in men (sex ratio 4.4/1) and in elderly subjects (median age: 63 years; 7 patients were 〈 20 yo).S kin lesions are very prevalent (85%) with variable lesion types. Blood cell counts show variable leukocytosis (figure 1A) with presence of blasts in 65% of cases. Anemia and thrombopenia were present in 59% and 76% of cases respectively. Bone marrow aspiration showed blastic infiltration in 94% of cases. Indeed, in 7 cases, there was isolated cutaneous involvement at diagnosis, with neither blood nor bone marrow infiltration. Morphologies of blast cells were heterogeneous. Typical morphologies were the most frequent, including medium-sized cells with a blastic round or irregular nucleus, cytoplasm displayed faint and irregular basophilia and no granulation. In a contingent of the blastic populations, we observed small vacuoles in a peculiar arrangement under the cytoplasmic membrane (42%) or the presence of large pseudopodia (28%) or both (17%) (Figure 1B, C, D). Some cases showed a more immature morphology with larger cells, higher nucleo-cytoplasmic ratio, very visible nucleoli and reinforced basophilia (28%) or a pseudomonoblastic morphology (5%) (Figure 1E). Rare cases presented a pseudolymphocytic form (7%, figure 1F) or large granulations in the cytoplasm (2%). Peroxidase and esterase were negative in all cases. Dysplasia of hematopoietic lineages was observed in 29% (figure 1G). For 8% of patients, myelodysplastic syndrome was diagnosed before the diagnosis of BPDCN. Immunophenotype showed that HLA-DR and CD4 were expressed in all cases, but 4 cases did not express CD56 (confirmed using 3 different antibodies). Expression of markers of others hematopoietic lineages was frequent. Among myeloid markers, the most frequent was CD33 (46%), followed by CD117 (23%), whereas CD13, CD11c, CD15 and CD65 were rarely expressed. Monocyte markers (CD14, CD64, CD11b) and myeloperoxidase were never expressed. For the T lineage, CD2 and CD7 were the most frequent (62% and 58% respectively) whereas CD5 was rare (7%). No cytoplasmic or surface CD3 were detected. For the B lineage, CD22 was expressed in 16%, and low levels of cCD79a in 5%. Both were never expressed together, and no CD19, CD20 and immunoglobulins were found. Generally, we observed one of these antigens (Ags) per case, but in 44% of cases, there was a combination of 2 or 3 Ags from 2 or 3 different lineages. Immature Ags such as CD34 and CD133 were never found, and Tdt was found in 14% of cases. Cytogenetic analysis revealed abnormal caryotype in 65% of the 78 caryotypes evaluated, with 20 cases having a complex caryotype. The frequency of the chromosomal abnormalities involved are shown in Figure 1H. In conclusion, we describe the largest series of BPDCN to date in the literature. Detailed clinical and biological data at presentation allow improved recognition of this rare form of acute and aggressive leukemia, enabling early initiation of appropriate management. Figure 1. A: Blood cell count in 109 BPDCN patients at diagnosis. Bars represent the median. B: Typical BPDCN morphology. C: In this case, the nuclei were peripheral, cytoplasm presented heterogenous basophilia, vacuoles were rare but large pseudopodia are frequent. D: Typical morphology with frequent microvacuoles under the cytoplasmic membrane. E: Immature morphology. F: Pseudolymphocytic morphology. G: Presence of dysplasia in myeloid cells with Auer Rods in the granulocytes. The morphology of the Blastic cells is typical. H. Chromosomal abnormalities in 78 caryotypes evaluated: The histogram represents the number of cases in which each chromosome was involved (deletion, gain, translocations). Figure 1. A: Blood cell count in 109 BPDCN patients at diagnosis. Bars represent the median. B: Typical BPDCN morphology. C: In this case, the nuclei were peripheral, cytoplasm presented heterogenous basophilia, vacuoles were rare but large pseudopodia are frequent. D: Typical morphology with frequent microvacuoles under the cytoplasmic membrane. E: Immature morphology. F: Pseudolymphocytic morphology. G: Presence of dysplasia in myeloid cells with Auer Rods in the granulocytes. The morphology of the Blastic cells is typical. H. Chromosomal abnormalities in 78 caryotypes evaluated: The histogram represents the number of cases in which each chromosome was involved (deletion, gain, translocations). Disclosures Bardet: Celgene: Research Funding. Deconinck:CHUGAI: Other: Travel for international congress; PFIZER: Research Funding; ROCHE: Research Funding; NOVARTIS: Other: Travel for international congress; ALEXION: Other: Travel for international congress; JANSSEN: Other: Travel for international congress; LFB loboratory: Consultancy.

Abstract: A new entity of acute leukemia co-expressing CD4 and CD56 markers without any other lineage-specific markers has recently been identified as arising from plasmacytoid dendritic cells (pDC). We proposed to call it plasmacytoid dendritic cell leukemia (pDCL) (Chaperot et al., 2001; Feuillard et al., 2002). We previously reported that pDCL may express the CD33 myeloid associated marker (Garnache-Ottou et al., 2005). This expression of CD33 on CD4+CD56+ lineage negative cells should not exclude the diagnosis of pDCL and underlines that pDCL specific markers should be identified. In order to better characterize pDCL, we organized a French pDCL network to collect cells and data on this rare leukemia. The aim of this study was to identify specific markers for the routine diagnosis. We focused on previously described pDC specific markers: CD123, BDCA-2, BDCA-4 (neuropilin-1). Fourteen pDCL and 69 other acute leukemias (10 B-ALL, 3 T-ALL, 3 BAL, 2 unclassified AML, 4 AML M0, 12 AML M1, 11 AML M2, 1 AML M3, 11 AML M4, 9 AML M5 according to FAB classification and 3 blastic transformations of myeloproliferative or myelodysplastic syndromes) were analyzed by flow cytometry. Leukemic cells were identified on their low expression of CD45 and using lineage specific markers. pDCL have been functionally characterized by their capacities to activate naïve cord blood CD4+ T cells, to induce a Th2 polarization and to produce IFN-α in response to viral supernatant. BDCA-4 was not specific for pDCL since 10% of acute leukemias (7/69) expressed this marker at the same levels as pDCL. This was not surprising since BDCA-4 has been reported to be expressed by myeloid cells as well as by some B and T cell subsets. Unlike BDCA-4, BDCA-2 was never expressed in the 69 tested acute leukemias. However, BDCA-2 expression was not systemically detected in all the pDCL tested (4 negative cases out of the 14 tested). CD123, also know as IL-3 receptor α-chain, was expressed on nearly all the acute leukemias tested. However, the levels of expression discriminated pDCL from other acute leukemias. These latter acute leukemias expressed lower levels of CD123 (fluorescence intensity ratio: 22 [mean] +/- 8 [SEM] ; range: 2–71) when compared to pDCL (fluorescence intensity ratio: 171 [mean] +/− 36 [SEM] ; range: 71–353). Overall, these results show that BDCA-4 is not a useful marker to identify pDCL. When BDCA-2 is expressed on leukemic cells, this is a strong argument in favor of pDCL. Finally, the level of CD123 expression is an interesting tool to characterize pDCL. We plan to analyze other acute leukemias, and in particular, acute leukemias aberrantly expressing CD4 and/or CD56 markers. This will allow us to confirm that pDCL can be identified on the basis of BDCA-2 and CD123 expression. Since the prognosis of pDCL was poor, an early diagnosis is needed to treat patients with specific therapeutic options (e.g., allogeneic hematopoietic cell transplantation). This can be achieved in the future by using anti-BDCA-2 and CD123 antibodies in addition to classical lineage specific markers. supported by the Goelam and the GEIL

Abstract: Abstract 2588 Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare and aggressive neoplasm derived from plasmacytoid dendritic cell precursors that involves the skin and invariably leads to aggressive leukemic dissemination (Garnache Ottou et al., 2007). The incidence is more frequent in men and elderly subjects, but BPDCN may occur in young adults and even in children. Current treatment options are not codified, and although conventional chemotherapy often induces an initial response, relapse is frequent and rapid. The prognosis is poor, with a median overall survival of 9–13 months whatever the initial presentation was. Allogeneic hematopoietic transplantation is currently the only potentially curative treatment, but this is not an option for most elderly patients. Therefore, there is an urgent need to develop novel therapies that can specifically target this tumor type. Since BPDCN cells express high levels of the interleukin-3 receptor (IL-3R) α-chain (CD123), we tested SL-401, a novel biologic conjugate that targets IL-3Rα (Konopleva et al. 2010 and Frolova et al. 2010), against 2 BPDCN cell lines (GEN2.2 and CAL-1) and 6 primary BPDCN cells isolated from 5 patients (P#1-5). Cells from P#2 were tested at diagnosis (#2d) and at relapse (#2r). A CD123+ cell line (TF-1/H-ras) sensitive to SL-401 and a CD123− cell line (Daudi) were used as controls. Cytotoxicity was assessed by MTT assay as well as flow cytometry (FC) after Annexin-V (AV) and 7-AAD staining. Primary BPDCN cells were cultured with IL-3 alone (to maintain survival) or IL-3 in combination with SL-401 for 24 h (FC) or 48 h (MTT) (Figure 1). All BPDCN cell lines and primary cells were found to be sensitive to SL-401 by MTT assay (Figure 1A). These results were confirmed by FC with a dose-dependent decrease in cell viability observed for the GEN2.2 and CAL-1 lines, (47 ± 10 to 2 ± 2%, n = 4, and 90 ± 7 to 11 ± 5%, n = 4, respectively) when treated with SL-401 compared to untreated cells (48 ± 9% and 88 ± 5%, for GEN2.2 and CAL-1 respectively). Cell viability decreased for all the 6 primary cells tested by FC (Figure 1B). The Daudi negative control cell line was resistant to SL-401 (Figure 1A), which confirmed SL-401 specificity.Figure 1.In vitro sensitivity of BPDCN cell lines (CAL-1 and GEN2.2) and primary cells from 5 BPDCN patients to SL-401 assessed by FC and MTT assays.Figure 1. In vitro sensitivity of BPDCN cell lines (CAL-1 and GEN2.2) and primary cells from 5 BPDCN patients to SL-401 assessed by FC and MTT assays. This is the first study evaluating the in vitro sensitivity of BPDCN using the IL-3R targeted drug candidate, SL-401, which is currently being evaluated in clinical trials of patients with acute myeloid leukemia (AML), myelodysplastic syndrome, and chronic myeloid leukemia. Importantly, the highest concentration of SL-401 evaluated in these in vitro studies was ten times lower than the peak plasma concentration achieved in AML patients (Frankel et al, 2008). Since all BPDCN patients evaluated to date express high levels of CD123 (Garnache Ottou et al, 2009), these results suggest that BPDCN patients may clinically benefit from SL-401 therapy. This new strategy should be evaluated in a clinical trial. A. Percentage of viable cells from BPDCN patients (#1, #4, and #5, n = 1) or of viable CAL-1 and GEN 2.2 cells (n = 3) assessed by MTT assay after incubation with different concentrations of SL-401 or without drug (cells only). Untreated cells were considered as 100% viable cells. The Daudi cell line (CD123−) was used as a negative control (n = 3). B. Percentage of viable cells (AV and 7-AAD negative, as measured by FC) after incubation with different concentrations of SL-401 or without drug (cells only) for 24 h. The histogram represents a mean of 3 (P#1), 4 (P#2d), and 6 (P#2r) independent experiments. Samples (P#3- 4, and- 5) were each tested once. Disclosures: Frankel: Stemline Therapeutics: Patents & Royalties, Research Funding. Jacobson:Stemline Therapeutics, Inc: Employment, stock Options. Cirrito:Stemline Therapeutics, Inc: Employment, Equity Ownership, Patents & Royalties. Brooks:Stemline Therapeutics, Inc: Employment, equity options.

Abstract: Microvesicles (MVs) with procoagulant properties may favor liver parenchymal extinction, then cirrhosis-related complications and mortality. In a longitudinal cohort of cirrhotic patients, we measured plasma levels of platelet-derived MVs (PMVs), endothelial-derived MVs, and red blood cell–derived MVs, expressing phosphatidylserine (annexin V–positive [AV + ]) or not, and evaluated their impact on Model for End-Stage Liver Disease (MELD) score and transplant-free survival. METHODS: MVs were quantified using flow cytometry in plasma from 90 noninfected cirrhotic patients and 10 healthy volunteers matched for age and sex. Impact of plasma microvesicle levels on 6-month transplant-free survival was assessed using log-rank tests and logistic regression. RESULTS: Microvesicle levels, mostly platelet-derived, were 2.5-fold higher in healthy volunteers compared with cirrhotic patients. Circulating small AV + PMV levels were lower in cirrhotic patients ( P = 0.014) and inversely correlated with MELD scores ( R = −0.28; P = 0.0065). During 1-year follow-up, 8 patients died and 7 underwent liver transplantation. In the remaining patients, circulating microvesicle levels did not change significantly. Six-month transplant-free survival was lower in patients with low baseline small AV + PMV levels (72.6% vs 96.2%; P = 0.0007). In multivariate analyses adjusted for age, ascites, esophageal varices, encephalopathy, clinical decompensation, total platelet counts, MELD score, and/or Child-Pugh C stage, patients with lower small AV + PMV levels had a significant 5- to 8-fold higher risk of 6-month death or liver transplant. Other PMV levels did not impact on survival. DISCUSSION: Decreased circulating small AV + PMV levels are associated with significantly lower transplant-free survival in cirrhotic patients independently of MELD score and platelet counts.

Abstract: Chimeric Antigen Receptor (CAR) therapy has led to great successes in patients with leukemia and lymphoma. Umbilical Cord Blood (UCB), stored in UCB banks, is an attractive source of T cells for CAR-T production. We used a third generation CD123 CAR-T (CD28/4-1BB), which was previously developed using an adult’s Peripheral Blood (PB), to test the ability of obtaining CD123 CAR-T from fresh or cryopreserved UCB. We obtained a cell product with a high and stable transduction efficacy, and a poorly differentiated phenotype of CAR-T cells, while retaining high cytotoxic functions in vitro and in vivo. Moreover, CAR-T produced from cryopreserved UCB are as functional as CAR-T produced from fresh UCB. Overall, these data pave the way for the clinical development of UCB-derived CAR-T. UCB CAR-T could be transferred in an autologous manner (after an UCB transplant) to reduce post-transplant relapses, or in an allogeneic setting, thanks to fewer HLA restrictions which ease the requirements for a match between the donor and recipient.

Abstract: Chimeric antigen receptor (CAR) T cells express an extracellular domain consisting of a single-chain fragment variable (scFv) targeting a surface tumor-associated antigen. scFv selection should involve safety profiling with evaluation of the efficacy/toxicity balance, especially when the target antigen also is expressed on healthy cells. Here, to assess differences in terms of efficacy and on-target/off-tumor effects, we generated five different CARs targeting CD123 by substituting only the scFv. In in vitro models, T cells engineered to express three of these five CD123 CARs were effectively cytotoxic on leukemic cells without increasing lysis of monocytes or endothelial cells. Using the IncuCyte system, we confirmed the low cytotoxicity of CD123 CAR T cells on endothelial cells. Hematotoxicity evaluation using progenitor culture and CD34 cell lysis showed that two of the five CD123 CAR T cells were less cytotoxic on hematopoietic stem cells. Using a humanized mouse model, we confirmed that CD123− cells were not eliminated by the CD123 CAR T cells. Two CD123 CAR T cells reduced tumor infiltration and increased the overall survival of mice in three in vivo models of blastic plasmacytoid dendritic cell neoplasm. In an aggressive version of this model, bulk RNA sequencing analysis showed that these CD123 CAR T cells upregulated genes associated with cytotoxicity and activation/exhaustion a few days after the injection. Together, these results emphasize the importance of screening different scFvs for the development of CAR constructs to support selection of cells with the optimal risk–benefit ratio for clinical development.