key: cord-0033721-a1w69sw3
authors: nan
title: Oral Sessions
date: 2014-03-28
journal: Bone Marrow Transplant
DOI: 10.1038/bmt.2014.43
sha: 351c1e0e833e0359779d5eb30cf8369e66bf8159
doc_id: 33721
cord_uid: a1w69sw3

nan

Introduction: There are still many debates whether there is a place for autologous HSCT in the treatment of adult acute lymphoblastic leukemia. Since 2009 the RALL study group is conducting the ALL-2009 trial (ClinicalTrials.gov public site; NCT01193933) where the main principle is non-aggressive but non-interruption treatment with prolonged L-asparaginase (Σ=590.000 IU), and autologous HSCT with BEAM conditioning followed by maintenance in T-cell ALL pts without HLA-identical donors (ASH, 2012 (ASH, , poster 2572 . The autologous HSCT was planned as a late high dose consolidation (on the 20-22 weeks of the protocol). In patients with HLA-identical siblings allogenic BMT was an option. Materials (or patients) and Methods: From Nov, 2008 , till Nov, 2013 centers enrolled 220 patients. In 5-phenotype was unknown. B-cell precursor phenotype was diagnosed in 64.6% (n=139/215), T-cell precursor phenotype -in 34,9% (n=75); biphenotypic AL -in 0,5% (n=1/215). T-cell ALL patients were young -29 y (19-55); male gender prevailed -29f/46m. T-ALL subtypes distribution in our study was as follows: 34 (47.1%) patients had early T-ALL (T-I/II), 31 (42.9%)-thymic (T-III), 10 (10%)-mature (T-IV). For the whole group medians were: Hb-110 g/l (42-180), L-26,3*10/ 9 l (0.5-313), plt -82*10/ 9 (5-943), b/m blasts -74.9% (0-99), LDH -975 IU (131-12 000). Cytogenetics was available in 54.6% of pts (n=41/75), 43,9% of them (n=18/41) had normal (NK), 20 pts -abnormal karyotype, 3 pts (7%) had no mitosis. Mediastinum involvement was registered in 40 (53.3%) and CNS disease -in 9 (12%). Results: The analysis was performed in November, 2013. Induction and follow-up data were available in 72 pts. CR was achieved in 63/72 (87.5%): after prephase=6, after 1 st ind.phase=33, after 2 nd ind.phase=24. Induction death occurred in n=5 (6.9%). Primary resistance and even progression during induction was registered in n=4 (5.6%). Due to few transplantation centers in Russia HSCT procedures were carried out in 2 centers. 18 pts proceeded to autologous HSCT, all of them were successfully harvested at a median time of 18-19 weeks, autologous HSCT was applied at a median 5 months from CR. No treatment related deaths occurred in autologous group. All pts after autologous HSCT receive chemotherapy maintenance according to the protocol though with Introduction: In adult patients with Philadelphia chromosomenegative (Ph -) acute lymphoblastic leukemia (ALL) or in an allogeneic hematopoietic stem cell transplantation (allo-HSCT) setting, a detailed analysis of genetic alterations from diagnosis to relapse remain largely unexplored. Materials (or patients) and methods: We carried out whole-exome sequencing analysis in longitudinal matched samples from diagnosis to relapse after allo-HSCT in 61 adult patients with Ph -B-ALL. Results: (1) Whole-exome sequencing was conducted for germline DNA from 3 relapsed cases at 3 specifi c time points including: diagnosis, complete remission (CR) after induction chemotherapy pre-HSCT, relapse after allo-HSCT (discovery cohort). We identifi ed candidated relapse-associated mutations by comparing variants identifi ed in leukemic samples of diagnosis or relapse against germline variants present in the sample of CR from the same individual. After validation by Sanger sequencing, we ultimately confi rmed 25 candidate relapse-associated somatic mutations involving 23 unique genes. In addition to mutations in genes known to be involved in leukemogenesis (MYC and KRAS), we discovered novel mutated genes, PTPN21 and USP54 etc, which were not previously reported in hematologic malignancies. (2) To further validate our fi ndings, we then performed whole coding-region sequencing for 23 candidated mutated genes in an extended validation cohort including 58 adult Ph -B-ALL patients (extension cohort), where 27 patients experienced relapse at a median time of 6.5 (range 2-33) months after allo-HSCT and 31 patients did not relapse after allo-HSCT at a median follow-up for 34 (range 12-56) months. We discovered novel associations of recurrent mutated genes (CREBBP, KRAS, PTPN21) with the pathogenesis of adult Ph -B-ALL relapse after allo-HSCT, which were mutated in at least two relapsed cases, but were not mutated in non-relapsed patients. (3) We also included additional 9 genes (PAX5, CDKN2A/ B, IKZF1(IKAROS), VPREB1, EBF1, TCF3(E2A), NR3C1 and ETV6) in targeted sequencing analysis, which have been identifi ed to be mutated in relapsed childhood ALL after chemotherapy by recent genome-wide analysis. Five genes (ETV6, PAX5, CDKN2A, TCF3, and NR3C1) also mutated in adult ALL cases. Our data raised the Introduction: Currently, the only curative treatment approach for patients with myelodysplastic syndromes (MDS) is allogeneic hematopoietic stem cell transplantation (HSCT). Unfavorable cytogenetic abnormalities have been shown to serve as predictors for MDS-relapse after HSCT. There is evidence that the novel 5-group cytogenetic classifi cation has a better predictive value for outcome after HSCT than standard IPSS-cytogenetics. The aim of this study was to determine the impact of the new 5-group cytogenetic MDS classifi cation on outcome after HSCT in a large international cohort. Materials (or patients) and Methods: In total, 903 patients with MDS or AML evolving from MDS (secondary AML, sAML) and suffi cient cytogenetic information, reported to the EBMTdatabase, were included into the study. According to the 5-group cytogenetic classifi cation 19 (2.1%) patients had very good risk, 204 (22.6%) normal risk, 438 (48.5%) intermediate risk, 178 (19.7%) poor risk, and 64 (7.1%) very poor risk cytogenetics. The impact of cytogenetic classifi cation was analyzed in uni-and multivariate models regarding overall survival (OS) and relapse free survival (RFS) after HSCT. Predictive performance of the two classifi cations was compared by means of the cross-validated log partial likelihood. Results: Estimated 5-year RFS and OS in all patients were 32% and 36% respectively. 5-group cytogenetic information was found to be strongly associated with OS and RFS in univariate analysis (OS: log-rank test P<.0001, RFS: P<.0001). Further patient and disease characteristics showed a signifi cant impact on impaired OS and RFS: Disease status at HSCT (RA/RARS no pretreatment; RAEB(t)/ sAML in CR; RAEB(t)/sAML not in CR, RAEB(t)/sAML untreated) (OS: P<.0001, RFS: P<.0001) and IPSS cytogenetics (good; intermediate; poor) (OS: P<.0001, RFS: P<.0001). The patient age showed only an impact for RFS (P=.05), but not for OS (P=.096). In multivariate analysis, statistically signifi cant predictors for RFS and OS at HSCT were either 5-group-or IPSS-cytogenetics, disease status and patient's age. Patients with poor risk [(RFS: HR=1.40 (95% CI: 1.15-1.71); HR=1.38 (95% CI: 1.12-1.70)] or very poor risk cytogenetics [(RFS: HR=2.14 (95% CI: 1.6-2.9); OS: HR=2.14 (95% CI: 1.59-2.87)] had worse RFS and OS than patients in the other 3 risk groups. Furthermore, patients with very poor risk had worse RFS and OS compared to patients with poor risk cytogenetics [(RFS: HR=1.53 (95% CI: 1.11-2.11) , OS: HR=1.55 (95% CI: 1.11-2.15) ]. When comparing the predictive performance of a series of 3 models both for OS and for RFS - (1) with only classical risk factors, (2) these extended with IPSS cytogenetics, (3) extended with 5-group classifi cation instead, the model with 5-group cytogenetics performed best. Discussion: In this international, multicentric study we confi rm that poor and very poor risk cytogenetics, according to the new 5-group cytogenetic classifi cation, independently predicted worse patient outcome after HSCT in MDS and sAML patients. Strategies to prevent relapse after HSCT in these patients are of major importance. *CK and GG contributed equally. Disclosure of Interest: None Declared.

A. Bacigalupo 1,* , C. Di Grazia 1 , F. Gualandi 1 , M. T. van Lint 1 , A. Ghiso 1 , A. Signori 2 , A. M. Raiola 1 1 

Introduction: Background. Allogeneic transplants for patients with myelodisplastic syndromes (MDS) have been associated with a signifi cant risk of transplant related mortality (TRM), ranging from 20% to over 50%. This is partly due to advanced age of the patients and advanced disease phase. Aim of the study. We studied 225 MDS patients grafted in our Unit between 1983 and 2013, to assess whether TRM has been reduced in recent years and to what extent. Materials (or patients) and Methods: Patients. We identifi ed in our data base 225 MDS patients: which were divided in 4 groups transplants before year 2000 (n=78) , between 2001 and 2005 (n=58) , 2006-2010 (n=58) , beyond 2010 (n=36). The median age in the four groups was as follows: 40, 48, 51, 58 years (p<0.0001); the proportion of patients with RAEB was comparable. The proportion of alternative donors grafts was 20%, 64%, 60%, 80% (p<0.0001). The conditioning regimen was reduced intensity (RIC) in 61%, 61%, 51%, 41% (p=0.1). One third of the patients were prepared with full dose TBI, with no diff erence in the groups (p=0.5) . The more recent group of patients comprised 40% grafted from haploidentical family donors, wioth high dose post transplant cyclophosphamide (PT-CY). Results: GvHD II-IV was 55%, 44%, 32%, 2/% in the 4 groups respectfully (p=0.01); The cumulative incidence (CI) of TRM in the 4 groups was 54%, 44%, 34%, 5% (p<0.0001). The cumulative incidence of relapse was 30%, 35%, 20%, 25% (p=0.3) . The actuarial 2 year survival was 24%, 33%, 54% and 65% (p<0.0001). In multivariate Cox analysis on survival showed year of transplant >2010, to be the strongest positive predictor, followed by patients female gender. Patients age, and the intensity of the conditioning regimen were not predictive. Discussion: Conclusion. We have seen a very signifi cant reduction of TRM in the recent years, resulting in improved 2 year survival, in a group of elderly patients with MDS. The reason for this appears to be associated with a reduced incidence of acute GvHD , possibly associated with a greater use of post-transplant cyclophosphamide in our Unit. Disclosure of Interest: None Declared.

Introduction: Allogeneic Hematopoietic Stem Cell Transplantation (HSCT) is the best curative option for many patients with myeloid malignancies, mostly due to the antileukemic activity of immune cells contained in the graft. Still, relapse remains an unmet issue. Next-generation sequencing provides the opportunity to track over time mutations and clonal dynamics, to detail disease history and identify mechanisms by which leukemia evade elimination. Materials (or patients) and methods: Serial samples collected during the complex clinical history of one patient with therapy-related myelodysplastic syndrome (tMDS) were analyzed by genomic HLA typing and by high-depth exome sequencing (minimum depth of coverage 70x). Patient fi broblasts and peripheral blood mononucleated cells harvested at remission served as germline reference. Sequencing data were analyzed using the tools developed at the Broad institute to identify mutations and quantitative clonal dynamics. Results: A 54 years old patient with high-risk tMDS underwent T cell-replete HLA-haploidentical HSCT in the presence of active disease, relapsing after one year. HLA typing of the leukemic blasts harvested at relapse demonstrated the selective de novo genomic loss of the mismatched HLA haplotype targeted by donor T cells, a frequent mechanism of leukemia immune evasion after haploidentical HSCT (Vago et al., New Engl J Med, 2009) . The patient was re-transplanted from a diff erent haploidentical donor mismatched for the HLA alleles retained by the mutated variants; accordingly, T cells from this donor were expectedly alloreactive against the relapsed leukemia. Remission was re-obtained and maintained for fi ve years, before the occurrence of a second relapse. Unexpectedly, at this relapse leukemic cells had lost the HLA molecules mismatched with the second donor, and expressed those that had gone amiss at fi rst relapse. Since at both relapses HLA haplotype loss was due to a stable genomic alteration, chromosome 6p acquired uniparental disomy (aUPD), any linear relation between the fi rst and second relapse could be ruled out.

To gain more detailed insights on the genetic relation between the three phases of disease, leukemic samples at diagnosis, fi rst relapse and second relapse were characterized by exome sequencing. Whereas a direct clonal evolution could be demonstrated from diagnosis to fi rst relapse, the large majority of mutations present at second relapse were unique to the sample: only fi ve non-synonymous coding mutations could be identifi ed in all three disease presentations, possibly comprising the `driver mutations' harbored by the original leukemic stem cell clone. Amongst these shared mutations we are currently validating the functional and epidemiological relevance of missense alterations in the TP53 tumor suppressor gene, in the NHEJ1 gene (related to chromosomal instability), and in the BTNL8 gene, encoding the T cell costimulatory receptor B7-H5. Discussion: Our results demonstrate the profound impact played by the graft-versus-tumor eff ect on leukemia clonal evolution through the selection of immuno-privileged subclones. Importantly, we provide new evidence to the long-term persistence of leukemic stem cells by identifying and tracking their putative `driver' mutations. By characterizing these mutations, we might envisage to further our insights on leukemogenesis, and to identify potential new targets for novel leukemia-eradicating therapies. Disclosure of Interest: None Declared.

Introduction: The biological bases of relapse after allogeneic Hematopoietic Stem Cell Transplantation remains to date largely unclear. One possible explanation comes from the "leukemia immunoediting" hypothesis, according to which relapses may be expression of immune-resistant leukemic variants outgrowing upon the selective pressure of the transplanted immune system. We provided a clinically relevant proof-of-principle of this model in transplanted patients (Vago et al., N Engl J Med, 2009) , but ex vivo studies often lack mechanistic insights, due to high interindividual variability and lack of suitable controls. Conversely, "mouse-in-mouse" models of cancer immunoediting can provide precious and reproducible insights into molecular mechanisms, but those results are often diffi cult to translate into clinical practice due to species-specifi city. To overcome these limitations, here we set up a novel mouse-human chimeric model of antileukemic adoptive immunotherapy, to dissect how T cell immune pressure can sculpt leukemia features. Materials (or patients) and Methods: Purifi ed primary human AML blasts were infused into non-irradiated immunodefi cient NOD/ SCID γ-chain null (NSG) mice. Upon leukemia engraftment, mice received serial infusions of human T cells, either autologous or allogeneic (HLA-identical, HLA-haploidentical or HLA-disparate) to the leukemic cells to mimic immune pressure. Absolute counts of human leukemic and T cells were monitored weekly in mice peripheral blood. At sacrifi ce, leukemic cells were FACS-purifi ed, total RNA was extracted and gene expression profi le was analyzed using Illumina microarray. Deregulated genes and signatures were identifi ed by pairwise LIMMA analysis. Gene Ontology (GO) and Gene Set Enrichment Analysis (GSEA) curated databases were interrogated to identify deregulated processes. Results: Infused leukemic cells stably and reproducibly engrafted in NSG mice, could be transferred to secondary and tertiary recipients, and displayed a stable gene expression profi le amongst littermates and upon serial transfer. HLA-disparate and HLAhaploidentical T cells eradicated AML from 10/10 treated mice. Conversely, HLA-identical T cells granted only temporary control in 6/6 mice, while autologous T cells were completely ineffi cacious in 6/6 mice.

Leukemic blasts subjected to T cell-mediated immune pressure showed a specifi c and reproducible gene signature. GO and GSEA demonstrated the selective deregulation of genes involved in immune processes. Among the top-ranked upregulated genes we identifi ed genes related to response to interferons, comprising proteasome and immunoproteasome subunits, as well as molecules and receptors involved in antigen processing and presentation, comprising classical and non-classical HLA Class I and II molecules (HLA-DMA, HLA-DRA, CD74, B2M, TAPBP, TAPBPL). Discussion: Our fi ndings provide further proof of the leukemia immunoediting hypothesis, demonstrating that leukemic cells modify their expression profi le, and their immunogenicity, in response to T cell-mediated immune pressure, and that the antigen presentation pathway represents a key target in these processes. The model we set up provides a novel and valuable tool to investigate these mechanisms in detail. To fi ne-tune immune pressure, and thus model a phase of equilibrium before leukemia immune escape, experiments with in vivo modulation of T cells reactivity are currently ongoing. Disclosure of Interest: None Declared.

H. Li 1 , D. Zacharaki 1 , R. Ghazanfari 1 , M. Ekblom 1 , S. Méndez-Ferrer 2 , S. Scheding 1, 3, * 1 Stem Cell Center, University of Lund, Lund, Sweden, 2 Stem Cell Biology, Centro Nacional de Investigaciones Cardiovasculares Carlos III, Mardrid, Spain, 3 Hematology, University Hospital Lund, Lund, Sweden Introduction: Human bone marrow (BM) contains a rare population of non-hematopoietic mesenchymal stem cells (BM-MSC) which can diff erentiate toward skeletal lineages such as osteoblasts, adipocytes, chondrocytes and hematopoiesis supporting stromal cells. In vivo, BM-MSC are essential constituents of the hematopoietic stem cell niche. We were fi rst to report that low/ negative expression of PDGFRα on lin neg /CD45 neg /CD271 pos cells allowed identify the candidate primary stromal stem cell population in adult human bone marrow, i.e lin neg /CD45 neg /CD271 pos / PDGFRα low/neg cells were highly enriched in CFU-F (1 in 4) and showed all typical BM-MSC properties in-vitro and in-vivo (Li et al. Blood, 2012, 120:3460) . Based on this fi ndings, the current study aimed to characterize the molecular signature and hematopoiesis-supporting function of primary BM-MSC. Materials (or patients) and Methods: Comparative gene expression profi ling was performed on human BM comparing lin neg /CD45 neg / CD271 pos /PDGFRα low/neg and lin neg /CD45 neg /CD271 pos /PDGFRα pos cells using Illumina Human HT-12 expression v4 BeadChips. Furthermore, phenotype and function of sorted cells primary MSC populations were studied using rt-PCR, co-culture assays with hematopoietic stem cells (HSC), and in-vivo HSC transplantation models using NSG mice. Results: Gene expression analysis comparing lin neg /CD45 neg / CD271 pos /PDGFRα low/neg and lin neg /CD45 neg /CD271 pos /PDGFRα pos cells revealed that 365 genes, including 22 novel surface antigens, such as ITGB5, ANTXR2, CD63, CD230, APLNR and CD74 were signifi cantly higher expressed in the PDGFRα low/neg subset. RT-PCR furthermore demonstrated expression of high levels of CXCL12, VCAM1 and osteopontin in PDGFRα low/neg cells, i.e. genes known to be related to hematopoietic stem cell (HSC) supportive function. Therefore, we tested the hematopoiesis-supporting capacity of primary PDGFRα low/neg BM-MSC. Cord blood CD34 + cells were short-term cultured (7 days) in SCF, TPO, and FL-supplemented serum-free medium±stroma cells. Co-culture with PDGFRα low/neg cells did not only amplify total hematopoietic cells (38.8±14.4 fold) but, importantly, also expanded CD34 + cells very eff ectively (10.2±3.8 fold, compared to 2.4±0.2, 3.6±0.7, and 3.2±0.8 fold for no stroma, CD271 neg /PDGFRα neg , and PDGFRα pos cells, respectively; p<0.05, n=3) . Moreover, the percentage of CD34 + cells was 2-3 fold higher in co-cultures with PDGFRα low/neg cells. Finally, xenotransplantation experiments into NSG mice showed that lin neg /CD45 neg /CD271 pos /PDGFRα low/neg cells eff ectively mediated the ex vivo expansion of transplantable CD34 + long-term engrafting hematopoietic stem cells with potent myeloid and lymphoid diff erentiation capacity. Discussion: Primary lin neg /CD45 neg /CD271 pos /PDGFRα low/neg stroma stem cells show a distinct molecular signature and potent hematopoiesis-supporting function. The study of primary BM-MSC will lead to a better understanding of the nature and the physiological role of these cells in the human bone marrow in situ. Disclosure of Interest: None Declared.

Introduction: The bone/bone marrow microenvironment that maintains the hematopoietic stem cell (HSC) state by controlling stem cell self-renewal and diff erentiation has been defi ned as stem cell niche.

We conceived a novel model in which cartilage pellets in vitro diff erentiated from MSCs generated complete ossicles upon heterotopic implantation in the absence of exogenous scaff olds, reproducing the HSC niche in vivo. Materials (or patients) and Methods: Human bone marrow (BM) and cord blood (CB) derived MSCs were cultured as micromasses for 21 days in the presence of TGFβ1 to diff erentiate into cartilage pellets. The generated pellets were implanted in the subcutaneous tissue of SCID-beige mice and harvested at 8 weeks to assess bone and bone marrow formation through histology and FACS analysis. Results: Heterotopic ossicles, featuring human bone and bone marrow stroma, were regularly observed at 8 weeks. Of note, a near-perfect architecture of a miniature bone organ, including cortical bone, marrow cavity, donor-derived marrow stroma, hostderived sinusoidal circulation, and host-derived hematopoietic tissue developed in a timed fashion. Hematopoietic tissue within the ossicles included erythroid, myeloid and megakaryocytic lineages with comparable relative proportions with the hematopoietic population in the BM of the ossicle-bearing animals. Methylcellulose colony-forming assays revealed the presence of CFC-M, CFC-G, CFC-GM, CFC-GEMM and BFU-E in hematopoiesis derived from BM-and CB-ossicles. BMossicles contained fewer LSK cells (Lin -Sca-1 + c-kit + ) than those of BM (0.096±0.007% vs. 0.201±0.034%, N=3). Instead, the frequency of LSK in CB-ossicles was similar to that of BM (0.128±0.048%, N=3). Immunophenotypically defi ned ST-HSC (LSK + CD34 + Flk2 -) and LT-HSC (LSK + CD34 -Flk2 -) were detectable in the ossicles. ST-HSC and LT-HSC were similar between BM and CB-ossicles (respectively, ST-HSC, 0.054±0.014% vs. 0.064±0.004%, N=2; LT-HSC, 0.037±0.016 vs. 0.030±0.006%, N=2) and slightly decreased in the BM-ossicles (ST-HSC, 0.029±0.004%, N=2; LT-HSC, 0.012±0.004%, N=2). Of note, frequencies of LSK cells, LT-HSC and CFC in the ossicles marrow were highly increased in comparison with peripheral blood of the recipient animal. Finally we investigated whether human hematopoietic cells could stably engraft into the generated human niche. CD34 + cells isolated from human CB were injected into sublethally irradiated mice previously implanted with cartilage pellets. The human cell engraftment resulted in similar proportion in the ossicle marrow generated from both sources and in the BM of the ossicle-Introduction: Acute leukemia remains one of the greatest challenges in oncology. Increasing evidence indicates that the bone marrow microenvironment (BMM) plays a pivotal role in both the pathogenesis and relapse of leukemia. Recent studies confi rmed galectin-3 (gal-3), a multifunctional member of the β-galactosidebinding protein family, is critical in leukemia drug resistance and patient's prognosis. Here we determined the role of gal-3 in the promotion of BMM-induced acute leukemia cells' (ALCs') survival, and the mechanisms involved. Materials (or patients) and Methods: Our study applied the co-culture system with mesenchyml stromal cells (MSCs) to mimic the leukemia BMM in vitro. We validated our hypothesis in diff erent sub-types of ALC lines including Reh, Sup-B15, Jurkat and Kasumi-1. Bone marrow was obtained from healthy adult donors with informed consent and MSCs were cultured and identifi ed as previously reported. Results: CCK-8 test demonstrated MSCs improved viable ALC number with or without cytotoxic agents Idarubicin (IDA) or Etoposide . PI/Annexin V assay confi rmed reduced apoptotic level of MSCs-conditioned ALCs induced by IDA or Vp-16 (P<0.05) ( Figure 1A ). To clarify the underlying mechanisms we performed Western-blot and real-time polymerase chain reaction. We found that compared to ALCs cultured alone, both mRNA and protein expression of gal-3 were signifi cantly up-regulated in MSCs-conditioned ALCs (P<0.05). Increased gal-3 levels correlated with Akt phosphorylation, β-catenin stabilization and higher expression of β-catenin target gene cyclin D1 (P<0.05) ( Figures 1B and 1C ). Then we used gal-3 small interfering RNA (siRNA) to silence its expression in ALCs to determine whether gal-3 modulates the protective eff ects of MSCs in ALCs. Compared with MSCs-conditioned vector-transfected Kasumi-1, we detected an increased apoptotic percentage against IDA and a decrease of β-catenin protein level in gal-3 siRNA-transfected ones ( Figure 1D ). Discussion: Altogether, our fi ndings reveal that gal-3-induced Wnt/β-catenin signaling is involved in MSCs-mediated drug resistance. Silencing gal-3 sensitized the ALCs to chemotherapy and decreased β-catenin stabilization, suggesting gal-3 can be a novel therapeutic target in acute leukemia. Disclosure of Interest: None Declared.

Introduction: Cord blood (CB) has grown substantially as an alternative source of hematopoietic stem cells for unrelated donor trasplantation. When compared with other sources, it is associated with a delayed hematological/immunological recovery. This may lead to high TRM in the early post-SCT period, mostly due to infection. Haplo-cord blood transplantation (haploCBT) has been used to accelerate neutrophil recovery, potentially reducing the incidence of serious neutropenia-related infections and allowing the use of drugs with myelosuppressive side eff ects; however few data are available comparing serious infections in single CBT and haploCBT. Materials (or patients) and Methods: We assessed the occurrence of severe infections, infection-related mortality and long term survival in adult patients diagnosed with high-risk hematologic malignancies receiving myeloablative conditioning and T cell depletion with anti-thymocyte globulin followed by a single CBT (n=74) or haploCBT (n=74). Results: We observed no diff erences in rejection rates between the two groups (9% vs. 5%, p=0.3), but neutrophil recovery was signifi cantly delayed in single CBT recipients (median 23 vs. 15 days, p<0.01). The incidence of grade 2-4 acute GVHD and severe chronic GHVD did not diff er (22% vs. 29%, p=0.2 and 14% vs. 9%, p=0.6, respectively) . HaploCBT recipients had a higher risk of developing any severe viral infection at 4-yr follow-up (60% vs 74%, p=0.05). The rate of CMV reactivation was similar between the two groups (67% vs 77%, p=0.06) while CMV disease occurrence was higher after haploCBT (5% vs. 21%, p<0.01) in CMV seropositive patients. In addition, haploCBT recipients presented higher rates of visceral or disseminated HSV or VZV infection (8% vs. 26%, p<0.01), HHV-6 encephalitis (2% vs. 7%, p=0.05) and BK polyomavirus related hemorrhagic cystitis (4% vs 11%, p=0.01). Single CBT recipients had a higher incidence of EBV reactivation (14% vs 4%, p=0.05), but EBV related PTLD was similar between the two groups (7% vs. 3%, NS). The rate of probable or proven invasive mould infection was also higher in single CBT (20% vs. 8%, p=0.03) , without signifi cant diff erences in the number of cases doccumented in the early post-trasplant period. The 4-yr incidence of NRM was 30% for single CBT and 41% in haploCBT recipients (NS) . Incidences of 4-yr infection related mortality (IRM) and relapse of the underlying malignancy did not diff er between groups (23% vs. 23% and 24 vs. 18%, NS). The 4-yr overall survival was 48 vs. 44%, respectively (NS). In multivariate analysis, the most signifi cant risk factors for IRM were advanced disease status (HR:2.7, 95% CI 1.2-6, p=0.02), CMV seropositive recipient (HR: 3.8, 95% CI 0.99-16, p: 0.06) and low number (<1.4×10e 5 /kg) of CD34+ cells in the CB unit (HR: 2.8, 95% CI 1.3-6.1, p: 0.01). Discussion: These data show that haplo-CBT results in much shorter periods of post-trasplant neutropenia, but the long-term rates of severe infections and IRM were comparable to those of single CBT, since immune-recovery is equally delayed in both grafts. Thus, an ideal future strategy may consist in rapid hematopoietic reconstitution with a haploCBT followed by attempts to improve immune reconstitution. Disclosure of Interest: None Declared.

Introduction: Reactivation of cytomegalovirus (CMV) after allogeneic stem cell transplantation can cause manifold clinical problems including myelo-and nephrotoxicity, pneumonia, colitis and encephalitis. CMV-seropositive hosts are particularly prone to reactivation when receiving a graft from a seronegative donor. A CMV phosphoprotein 65 (CMVpp65) derived nonamer peptide NLVATVPMV has been characterized as a HLA-A2 restricted T cell epitope peptide. Therefore we designed a vaccine with 300 microgram of the peptide in an oil-in-water emulsion. Materials (or patients) and Methods: Seven CMV-seropositive patients after allogeneic stem cell transplantation for leukemia or lymphoma received four subcutaneous applications of the vaccine at a biweekly interval. Additionally 75 microgram/day GM-CSF was applied subcutaneously on the two days preceding, the day of and the two days following peptide vaccination. Six patients had already experienced several episodes of CMV viremia and got the vaccine in a preemptive manner, one patient underwent prophylactic vaccination.Patients were observed clinically and CMV antigenemia was monitored using immunofl uorescence tests. Blood samples before and after each vaccination were collected and evaluated extensively by multi-color fl ow cytometry including tetramer staining for HLA-A2 and -B7 CD8+ T cells and regulatory CD4+CD25hiFoxP3+ T cells. To further assess CD8+CCR7+CD45RA eff ector T cells, ELISPOT assays were performed for the secretion of interferon gamma and granzyme B. Serological tests to evaluate dynamics in CMV-IgG and -IgM-titers are ongoing. Results: All 28 scheduled vaccinations were administered. As expected, side eff ects were restricted to CTC grade I toxicity of the skin, i.e. rash and induration of the skin at the site of injection. These side eff ects resolved after a maximum of three weeks after the last vaccination. No other toxicities were observed. Five of six patients with CMV antigenemia cleared the CMV after four vaccinations and are free from viremia till present (maximum more than one year) despite cessation of antiviral prophylaxis. The patient receiving a prophylactic vaccination never developed viremia. We could detect CMVpp65-specifi c CD8+ T cells in all patients after vaccination. An increase of secretion of interferon gamma and granzyme B was observed. Discussion: In summary, CMVpp65 peptide vaccination was save and well tolerated in all patients. Patients had a clinically favorable outcome with clearance of the virus after vaccination. Activation and proliferation of CMV65 specifi c T cells was observed. The analysis of humoral responses is still pending and will be reported in the conference. We consider this therapeutic and prophylactic vaccination as an interesting option for patients at risk for CMV reactivation and will extend this clinical trial also to patients expecting kidney transplantation in the near future. Disclosure of Interest: None Declared. Introduction: Neutropenic enterocolitis (NEC) is a life threatening complication of leukemic and solid tumors patients (pts) treated with chemotherapy (CHT) with mortality rate up to 50%. It is a clinical syndrome in neutropenic patients (pts) characterized by abdominal pain (AP), fever (F) and diarrhoea (D). Ultrasound (US) was used to evaluate bowel-wall thickening (BWT), and >4 mm is considered diagnostic of NEC. Perforation occurs in 5%>10% of cases. Early diagnosis is crucial to start conservative medical management (CMM) which appears the optimal strategy for most cases. Authors have proposed objective criteria for immediate surgical treatment. Objective: evaluate prospectively if Bed-side-US(BUS) can detect early signs of NEC and guide a prompt treatment (CMM or surgical) in order to reduce mortality. Materials (or patients) and Methods: in the last 5 years all pts admitted in Our Hematology/BMT Unit wards at University of Pisa (Italy), undergoing chemotherapy (CHT), autologous or allogeneic transplant (AutoTx,AlloTx) were enrolled. Abdominal US was performed, baseline before treatment, and as only one symptom (or a combination) appeared within 12h from onset: F and/or D and/or AP in CHT-related neutropenic pts. Results: out of 1499 neutropenic pts 72 episodes were identifi ed (4.8%). Seven pts had 2 separate episodes of NEC. Thus we diagnosed 72 episodes in 65 pts. Disease diagnosis were HD (N=10), ALL (N=8), AML (N=18), MM (N=9) and NHL (N=27). Treatment received was intensive CHT (N=32), AutoTx (N=36) and AlloTx (N=4). At time of diagnosis symptoms were: F+AP+D in 72.5%, D+P in 15%, F+D in 7.5%, F+P 2.5%, AP in 2.5%. F and D alone were never present at diagnosis of NEC. As control group we considered pts with CHT related mucositis and pts restaged with US during neutropenia in absence of symptoms. None of them had BWT. Fever at diagnosis of NEC was absent in 26% of episodes (N=19/72). Overall 11 pts died (15%). Treatment was CMM in 92% of patients, and was promptly started as BUS diagnosis was made. Mortality in pts treated with CMM was 12% (88% of pts survived). Six pts underwent surgery, guided by US features, during neutropenia, and 50% are alive. Median BWT was 8.6 mm in surviving pts (range 4.2-30 mm) and 11mm in deceased (range 9.3-15 mm). Authors have suggested BWT to be prognostic of outcome; in our study pts with >10 mm had 60% survival. Median time to response from beginning of CMM was 24h and the fi rst sign of was a decrease in AP, while median time to death was 26h (range 10.5-72h). Discussion: BUS allowed to detect early signs of NEC and to start prompt treatment in this life threatening complication, which was CMM in 92%. With BUS pts do not live the isolation room. Early diagnosis and intervention allowed us to reduce mortality. US guided surgical intervention with 50% survival rate. Images of BUS and CT were superimposable. Fever is not a conditio sine qua non for NEC diagnosis. A prompt BUS in neutropenic patients as just one symptom presents allows to make early diagnosis of this life threatening complication and guide prompt treatment (conservative or surgical). Disclosure of Interest: None Declared. S14 Factors infl uencing the risk of BSI and survival after BSI were analysed. Results: There were 159 episodes of BSI diagnosed in 129 patients (median time to the fi rst BSI: 8 days, range: -8; +46); among them 23 patients developed more than one BSI. Overall, 171 pathogens were isolated, with the distribution as follows: 55% (n=94) Gram+ (staphylococci : 94, 20%; enterococci: 39, 23%; viridans streptococci: 18, 11%; Corynebacterium: 2, 1%), 40% (n=69) Gram-(Enterobacteriaceae: 53, 31%, mainly E. coli: n=44; Pseudomonas aeruginosa 11, 6%; other Gram-: 5, 3%), and 5% (n=8) fungi (4 C. krusei, 1 C. glabrata, 1 C. parapsilosis, 2 Fusarium). The Gram+/Gram-ratio did not diff er between primary and secondary episodes of BSI. The risk of developing BSI was signifi cantly increased in case of older patients (median age 48 vs. 43, p=.002), active disease at transplant vs. disease in remission (34% vs. 22%, p=.006), and longer neutropenia (18 vs. 17 days, p=.019). Additionally, there was signifi cantly diff erent risk of BSI according to the type of donor: matched related 8%, matched unrelated or mismatched related (MUD/MMR) 22%, cord blood transplant (CBT) 45% and haploidentical 37%, p<.0001. Sex, type of the underlying disease and conditioning regimen (myeloablative vs. reduced intensity) did not infl uence the risk of BSI. The overall mortality at 60 days after HSCT was higher in patients developing BSI (p<.0001), in case of CBT (p=.014) and in those with an active disease at HSCT (p=.001). The overall mortality at 7 and 30 days after the fi rst BSI was 6% and 16%, respectively. It did not diff er among the 4 types of donors: 0% and 8% for matched related, 13% and 22% for MUD/MMR, 10% and 19% for CBT, 3% and 14% for haploidentical, respectively for 7-day (p=.23) and 30day (p=.69) mortality. Considering only the fi rst BSI episode, two periods of the study were compared: 2008-2010 vs. 2011-2013. The incidence of BSI increased from 22% to 42% (p=.06), a decrease in Gram+/ Gramratio (35/19 vs. 41/39, p=.15 ) was observed, but post-BSI mortality decreased from 12% to 2 % for 7-day mortality (p=.057) and from 29% to 10% for 30-day mortality (p=.004). Discussion: Conclusions Pre-engraftment BSI were particulate frequent in case of CBT or a haploidentical donor, but post-BSI mortality was not increased in these groups. Despite an increased number of BSI during the last 3 years, the mortality was reduced compared to the previous period. Disclosure of Interest: None Declared. Introduction: Human herpesvirus 6 (HHV-6) frequently reactivates after allogeneic hematopoietic cell transplantation (HCT) and is associated with many adverse eff ects, including central nervous system dysfunction (HHV-6 CD). The frequency and clinical significance of HHV-6 DNA detection in cerebrospinal fl uid (CSF) after HCT is not well defi ned. Materials (or patients) and Methods: We identifi ed all patients at our center with HHV-6 DNA detected in CSF using quantitative PCR. Patients with neurologic symptoms and HHV-6 DNA detection in CSF without identifi cation of an alternative etiology were categorized as having HHV-6 CD. Results: Among 3,902 allogeneic HCT recipients from 1998-2012, we found 56 of 124 tested patients with HHV-6 DNA in CSF; 5 patients with pre-HCT or donor-derived chromosomally integrated HHV-6 were excluded. In this cohort of 51 patients, 37 met criteria for HHV-6 CD and 14 (27%) did not. Among patients with HHV-6 DNA in CSF without HHV-6 CD, 8 patients had an alternative diagnosis for their neurologic symptoms (drug toxicity or infection), 3 had spontaneous resolution of their symptoms prior to antiviral therapy, and 3 had no neurologic symptoms (CSF testing prompted by HHV-6 viremia). Diff erences between patients with and without HHV-6 CD were driven by those with an alternative diagnosis. This group had longer time to HHV-6 detection and lower viral load (VL) in CSF (Table 1) . Patients with HHV-6 CD compared to those with no or spontaneous resolution of neurologic symptoms had no signifi cant diff erences. Ganciclovir and/or foscarnet were administered to all patients with and 8 (57%) without HHV-6 CD for at least 1 week. Two patients died due to progressive HHV-6 CD within 1 month. Six patients without HHV-6 CD were not treated and had no morbidity attributable to HHV-6. Kaplan-Meier survival curves after CSF HHV-6 DNA detection demonstrate poor overall survival among all patients (~50% at 180 days). In univariate and multivariable Cox proportional hazards models considering demographic and clinical characteristics (including antiviral treatment, plasma VL, and CSF VL), alternative diagnosis was the only variable [PH-O024] S15 associated with mortality (adjusted HR, 4.4; 95% CI, P<0 .01 at 90 days). Discussion: We identifi ed 14 allogeneic HCT recipients with HHV-6 DNA in CSF who did not meet criteria for HHV-6 CD. Later detection of HHV-6 DNA at lower VL in CSF was more common in patients with an alternative diagnosis, but VL ranges overlapped and plasma VL were similar. All patients had poor survival; whether CSF HHV-6 DNA detection in patients without HHV-6 CD independently contributes to mortality or is a marker for overall poor health due to secondary factors is unclear. In conclusion, patients may have HHV-6 DNA in CSF without HHV-6 CD, and their management warrants further investigation given high mortality rates. Disclosure of Interest: J. Hill: None Declared, M. Boeckh Confl ict with: Chimerix Inc, Confl ict with: Roche/Genentech, Confl ict with: Consulting for Clinigen, K. Jerome: None Declared, R. Sedlak: None Declared, D. Zerr: None Declared. Introduction: Adoptive immunotherapy against viral infections is a promising treatment option for patients after hematopoietic stem cell transplantation. However, the generation of virus-specifi c T-cells is either cost-intensive or time-consuming or requires large volumes of fresh blood which lead to a substantial delay before therapy could be administered. Materials (or patients) and Methods: We have developed the fi rst fully GMP-compliant protocol to generate donor-derived adenovirus (HAdV)-, cytomegalovirus-and Epstein-Barr virus-specifi c T-cell lines (TCLs) within 12 days in a total working time of 15h. For this, PBMCs from only 100ml of peripheral blood were stimulated with overlapping polypeptides derived from diff erent viruses in combination with IL-15. The phenotypic and functional characterization of expanded TCLs were analyzed via MHC-class-I-multimers, an IFN-g ELISpot assay and a fl ow cytometry-based cytotoxicity assay. Results: After 12 days, virus-specifi c TCLs showed highly specifi c activity against the appropriate antigens (mean 662 spot forming cells/10E5 cells) and low or even absent alloreactivity. Furthermore, two patients -after haploidentical HSCT with HAdV viremia displaying rising viral loads despite treatment with cidofovir -received 1x10 4 donor-derived HAdV-specifi c TCLs per kg body weight. In both patients, HAdV-specifi c T-cells could be detected by IFN-γ-ELISpot 30 and 22 days post infusion. The procedure resulted in complete clearance or >1.5 log reduction of viral load within 15 and 18 days, respectively. Although the fi rst patient did not show any side eff ects and is still alive and well, the second patient developed skin GvHD 10 days post therapy that progressed to GvHD grade IV. Discussion: In principle, it is possible that the seHAdV-TCLs contributed to GvHD. However, taking into account the low T-cell dosage infused, the early time interval (10 days following the infusion) and the fact that GvHD emerged 12 days prior to the detection of seHAdV-TCLs, we consider this rather unlikely. In summary, we report the fi rst-in-man use of short-term polypeptide-generated HAdV-specifi c GMP-compliant T-cells, with a clinical outcome comparable to other more cost-and time-intensive protocols. If confi rmed in a larger patient cohort, this approach might pave the way to broad clinical implementation. Disclosure of Interest: None Declared.

Introduction: Viral infections remain important problems after allogeneic hematopoietic stem cell transplantation (HSCT) from unrelated donors. There is little information about the diff erence of the risk of viral infections between cord blood transplantation (CBT) recipients and bone marrow transplantation (BMT) recipients. Materials (or patients) and Methods: To evaluate incidences, risk factors, and outcomes of viral infections after unrelated CBT by comparison with those after unrelated BMT, we retrospectively analyzed the records of 6,400 Japanese adult patients who underwent unrelated HSCT for the fi rst time between 2006 and 2010, using the data collected by the Transplant Registry Unifi ed Management Program of Japan Society for Hematopoietic Cell Transplantation. Results: Overall, 1,488 patients had at least one viral infectious episode with a cumulative incidence of 23.9%, at a median of 45.5 (1-1400) days after HSCT. The incidence was signifi cantly greater (27.9% vs. 21.5%, P<0.001) and the median onset was signifi cantly earlier (39 days vs. 50 days, P<0.001) in CBT versus BMT recipients. One-hundred twenty-fi ve patients (8.4%) had 2 or more infectious episodes caused by diff erent viruses, and a total of 1,617 episodes were documented. Causative agents for hemorrhagic cystitis (HC) including adenovirus (ADV) and BK virus (BKV) were the most common (25.2%), followed by cytomegalovirus (CMV; 24.9%), human herpes virus-6 (HHV-6; 20.0%) and varicella-zoster virus (VZV; 19.5%). The major clinical manifestations caused by CMV were gastrointestinal diseases (58.2%) and pneumonia (17.4%). HHV-6 most frequently caused encephalitis (52.1%), about 80% of which was defi nitely diagnosed by the virus detection in cerebrospinal fl uid. VZV predominantly caused localized zoster (95.6%). Among the all HSCT recipients, CBT was identifi ed as the only signifi cant pre-transplant risk factor (HR 1.38, P<0.001), whereas grade II-IV acute GVHD and chronic GVHD were identifi ed as timedependent post-transplant risk factors (HR 1.34, P<0.001 and HR 1.30, P=0.013) . In the CBT recipients, GVHD prophylaxis with methotrexate signifi cantly decreased the risk of viral infections compared to the prophylaxis with mycophenolate mofetil or calcineurin inhibitors alone (HR 0.75, P=0.006), whereas the impact of grade II-IV acute GVHD and chronic GVHD on the risk were not apparent. In contrast, HLA mismatch, grade II-IV acute GVHD and chronic GVHD were signifi cant independent risk factors for viral infections after BMT (HR 1.23, P=0.024, HR 1.57, P<0.001 and HR 1.51, P=0.004) . HHV-6 encephalitis and VZV reactivation occurred signifi cantly more frequent after CBT compared to BMT (3.4% vs. 1.3%, P<0.001 and 6.4% vs. 4.8%, P=0.012), while the incidences of HC caused by ADV/BKV and CMV diseases were not signifi cantly diff erent between the two groups. Viral infections signifi cantly Introduction: During the last decade substantial progress has been made in the fi eld of allogeneic stem cell transplantation (SCT). Allogeneic SCT has therefore become an established treatment option for patients with high-risk leukemias, but the therapeutic success is still limited by patients' relapse. Evidence was presented, that minimal residual disease (MRD) status in principle could be treated by ``pre-emptive'' immunotherapy including donor lymphocyte infusion (DLI). However, DLI is only effi cacious for defi ned diseases and the high T cell doses required during overt relapse raise the risk for severe graft versus host disease (GvHD). Allogeneic cytokine-induced killer (CIK) cells demonstrated potent cytotoxicity against hematological malignancies, but displayed negligible alloreactivity and caused minimal GvHD in vitro and in vivo in pre-clinical mouse models. Materials (or patients) and Methods: Between August 11, 2011 and December 12, 2013 CIK cell infusions approved by the regulatory authorities (Regierungspräsidium Darmstadt, Germany) were given repetitively on compassionate use basis to 9 patients (<18 years of age, n=8, median age 9, range 1-16 years; >18 years of age, n=1, age 69 years) with haematological malignancies (AML, n=6; Interfant ALL, n=1; T-ALL, n=1; CML, n=1) and evidence of relapse in the absence of acute GvHD >grade I after allogeneic SCT. CIK cells were generated in vitro within 10 days from peripheral blood mononuclear cells of original stem cell donors under good manufacturing practice (GMP)-conditions in the presence of interferon, anti-CD3 antibody, interleukin-2 and interleukin-15. Results: Altogether 43 CIK cell infusions (median number 5, range 1-9 per patient) from matched unrelated (n=5) or haploidentical (n=4) stem cell donors were off ered at a minimum of three weeks after allogeneic SCT and an interval of 4-6 weeks between infusions (median follow up after 1 st infusion 8, range 1-15 months). Patients with hematological relapse underwent cytoreductive chemotherapy before infusions. Based on our cumulative experience starting doses of CIK cell infusions were 5×10 6 CD3 + CD56 -CIK cells/kg in pediatric and 1×10 6 CD3 + CD56 -CIK cells/kg in adult patients. Regardless of donor type, dose escalation continued until a maximal dose of 100×10 6 CD3 + CD56 -CIK cells/kg was reached (median dose 10x10 6 , range 0.1-100×10 6 CD3 + CD56 -CIK cells/kg). Acute GvHD grade I occurred in two patients after infusions of 1×10 6 CD3 + CD56 -CIK cells/kg. In 3/9 patients with hematological relapse CIK cell infusions provided transient remission, but ultimately all three patients relapsed and succumbed to their diseases. Six of nine patients with impending relapse indicated by mixed chimerism or MRD status remained in complete remission after serial CIK cell infusions. Discussion: Altogether, allogeneic CIK cell infusions seemed to be very promising and may improve cell-based immunotherapy approaches for patients with impending relapse after allogeneic SCT in the future. Disclosure of Interest: None Declared. Introduction: Cytomegalovirus, EBV and adenovirus are particularly problematic in patients after HSCT and cord blood transplant (CBT) and are associated with signifi cant morbidity and mortality. Defi ciencies in conventional therapeutics have increased interest in an immunotherapeutic approach to viral disorders. Materials (or patients) and Methods: We have developed two strategies to grow multivirus-specifi c donor-derived T-cells from peripheral blood (PB) and naive cord blood (CB). Using a chimeric adenovirus-vector to modify monocytes, DC and EBV-LCL we generated a single culture of PB or CB mononuclear cells that gave rise to multivirus-specifi c cytotoxic T-cells (CTL). Results: We have infused 26 patients with PB-derived CTL (9 were recipients of haploidentical donor grafts) and 9 patients with Cord Blood-derived CTL. Patients received CTL infusions from 35 to 164 days (median 84d) post HSCT or CBT at a median of 5×10 e7 cells/m 2 . None developed >grade II GvHD or other toxicities over 3 months of safety monitoring after infusion. We observed up to a 5-fold increase in CMV-and EBV-specifi c T-cells by 4 weeks post-CTL as measured by IFN-g ELISPOT assay. 26 viral reactivations were observed in patients before or immediately after CTL infusion. In the absence of conventional therapy, 7 of the 11 patients with CMV infection became negative for CMV in the blood within 7 days of CTL infusion, with a corresponding rise in CMV-specifi c CTL in peripheral blood. Each of 8 patients with high EBV loads cleared their virus, as did 6 of 7 patients with adenoviral infections/disease. Discussion: This study demonstrates that multivirus-specifi c CTL derived from the peripheral blood of seropositive donors as well as the cord blood of virus naive donors expand in vivo and are active against multiple viruses. Furthermore, by restoring immunity to multiple viruses simultaneously, the need for continued prophylaxis with pharmacotherapy is eliminated, thus, improving the effi ciency and cost eff ectiveness of protecting HSCT and CBT recipients from these potentially lethal viruses. Disclosure of Interest: None Declared. S17 effi cacy of anticancer monoclonal antibodies (mAb), of which central tumoricidal activity is an antibody-dependent cellular cytotoxicity (ADCC) exerted by FcγRIIIa (CD16) expressing eff ector cells. To circumvent this drawback, we examined in vivo the feasibility of T cells gene-modifi ed to express a newly generated chimeric CD16-CD3ζ receptor as a transferable alternative eff ector for cancer mAb therapy. Materials (or patients) and Methods: A novel affi nity-matured chimeric CD16 with a 158V/V-CD3ζ (cCD16ζ ) gene construct was lentivirally introduced into CD3 + T cells from both healthy individuals (n=3) and patients with B-cell lymphoma (n=3) (cCD16ζT cells). In the context of ADCC activity, functional properties of cCD16ζ-T cells were extensively examined both in vitro and in vivo, compared with those of NK cells (n=3). Finally, we examined the feasibility of double gene-modifi ed CD8 + T cells to express WT1specifi c TCR and cCD16ζ receptor. Results: cCD16ζ-T cells were readily expandable in ex vivo culture, and successfully displayed ADCC-mediated tumoricidal activity against opsonized cancer cells with mAbs in vitro. During ADCC, ligation of opsonized cancer cells to cCD16ζ receptor stimulated cCD16ζ-T cells to produce proinfl ammatory cytokines and release toxic granules through activation of the NFAT pathway after phosphorylation of the CD3ζ chain. In parallel, these stimulated cCD16ζ-T cells transiently proliferated and diff erentiated into eff ector memory T cells. On the other hand, although activated NK cells using rhIL-2 displayed similar ADCC activity, they failed to proliferate. Human cCD16ζ-T cells infused concomitantly with rituximab synergistically inhibited the growth of disseminated Raji cells, a CD20 + lymphoma cell line, in immunodefi cient mice, whereas similarly infused rhIL-2-treated NK cells survived for a shorter time and displayed less eff ective tumor suppression. Finally, the double-gene modifi ed CD8 + T cells could successfully recognize the target cells through both the rituximabopsonized CD20 and WT1 epitope/HLA-A24 complex on the cell surface. Discussion: Our fi ndings strongly suggest the clinical feasibility of cCD16ζ-T cells as adoptively transferable ADCC eff ector cells that could potentially enhance the clinical responses mediated by currently available anticancer mAbs. Disclosure of Interest: None Declared. Introduction: Studies using T-cell receptor (TCR) or chimeric antigen receptor (CAR) transduced T-cells have shown eff ectiveness of adoptive immunotherapy to treat diff erent malignancies. The effi cacy and safety of such interventions greatly depends on good target selection to prevent on-target toxicity. Furthermore, the broad application of TCR-based adoptive immunotherapy is hampered by a lack of an eff ective immune response against self-antigens. Through self-tolerance, T-cells carrying high-affi nity TCRs reactive to self-antigens are deleted during thymic selection. An attractive strategy is to exploit the immunogenicity of foreign human leukocyte antigen (HLA) molecules to generate an eff ective immune response against these antigens. Here, we describe a protocol to effi ciently isolate high-avidity alloHLA-restricted T-cells targeting the B-cell compartment. Materials (or patients) and Methods: From a B cell peptide elution library 15 peptides derived from genes exhibiting B-cell restricted expression patterns were identifi ed and peptide-MHC multimers (pMHC) of HLA-A*02:01 were generated. Via MACSorting and FACSorting a plethora of pMHC-binding T-cell clones from HLA-A*02:01 negative individuals were isolated. Generated T-cell clones were selected based on peptide-specifi city and avidity for further characterization.

Results: We successfully isolated two distinct T-cell clones carrying high-affi nity TCRs specifi c for a CD20 peptide presented in HLA-A*02:01. CD20 dependent recognition could be demonstrated by genetically engineering of CD20 negative K562-A2 cells to express CD20. Furthermore, the isolated T-cell clones effi ciently recognized CD20-expressing HLA-A*02:01 primary chronic lymphocytic leukemia (CLL), acute lymphoblastic leukemia (ALL) and mantle cell lymphoma (MCL), while recognition of CD20-negative hematopoietic and non-hematopoietic cell-subsets was absent. Furthermore, the CD20 specifi c T-cell clones were able to more effi ciently recognize ALL cell-lines than CD20 specifi c antibodies. We demonstrated that on target cells with only very low CD20 surface expression, the CD20 specifi c T cell clones could still efficiently recognize endogenously processed CD20 derived peptides in the context of HLA-A2. In addition, we observed that for the isolated T-cell clones pMHC-multimer binding is a poor indicator of functional activity. T-cell clones with similar fl uorescence intensity when stained with their cognate pMHC-multimer exhibited a varying degree of peptide affi nity spanning several logs. Discussion: In summary, we developed a protocol for the identifi cation of high-affi nity TCRs of therapeutic relevance targeted to self-antigens. CD20-specifi c TCRs could broaden the application of immunotherapies targeted to CD20 in those cases where CD20cell surface expression is low. Disclosure of Interest: None Declared.

Introduction: Adoptive T-cell therapy with chimeric antigen receptors (CARs) has showed its curative potential of B-cell malignancies. However, translating this experience to other disease indications requires the development of other CAR specifi cities. To note, off -tumor expression of the target antigen raises safety concerns about newly designed CARs. CD44v6 is expressed on acute myeloid leukemia (AML) and multiple myeloma (MM) cells and is required for their engraftment in immunodefi cient mice. The aim of this project is to develop a new CAR-based strategy for the safe and eff ective treatment of AML and MM. Materials (or patients) and Methods: We developed a CAR specifi c for CD44v6 by cloning the scFv of a humanized mAb in a CD28/ CD3 zeta chain backbone. CD44v6-targeted T cells were tested for their reactivity against primary AML and MM cells both in vitro and in vivo. CD44v6 expression was analyzed by quantitative RT-PCR, fl ow cytometry and immunhistochemistry. Suicide genes were co-expressed with the CD44v6.CAR through a bidirectional LV vector. Results: CD44v6-targeted T cells completely eliminated primary leukemic blasts and malignant plasma cells in vitro, even in the presence of immunosuppressive stroma. When infused into NSG mice, they persisted long-term and exerted extraordinary antitumor eff ects against autologous leukemic blasts, even with a high tumor burden. Interestingly, antitumor effi cacy was dependent on transduction with CD3/CD28-beads and culture with IL-7/IL-15. Among normal non-hematopoietic tissues we found CD44v6 expression only on the skin, albeit at lower levels than primary leukemic blasts. Strikingly, at the E:T ratios allowing the potent antitumor eff ects of CD44v6-targeted T cells, primary keratinocytes were spared and there was no cytokine production. Interest-ingly, keratinocytes expressed lower levels of ICAM-1 and higher levels of PD-L1 than leukemic blasts. Of hematopoietic cells, only circulating CD14+ monocytes expressed CD44v6 and were killed by CD44v6-targeted T cells. Interestingly, neither bone-marrow monocytes nor tissue-resident macrophages expressed CD44v6. Importantly, hematopoietic stem cells (HSCs) and progenitors were CD44v6-negative. Accordingly, CD44v6-targeted T cells did not interfere with their clonogenic potential and where able to discriminate between tumor cells and HSCs and progenitors in bone marrow samples from MM patients. Finally, the infusion of CD44v6-targeted T cells in mice reconstituted with human CD34+ cells resulted in the selective elimination of monocytes, while other cell subsets were spared, indicating the preservation of the stem-cell pool. For enabling rapid and conditional ablation of CD44v6-targeted T cells at will, we fi nally co-expressed the CD44v6-CAR with TK or the inducible caspase-9 and validated the suicide gene approach in hyperacute xenogeneic GVHD surrogating maximal toxicity. Discussion: We demonstrated that suicidal T cells targeting the tumor-promoting antigen CD44v6 have the potential to eradicate AML and MM with acceptable and/or reversible toxicities. Moreover, we showed that target expression on normal tissues do not automatically predict the susceptibility to CAR-mediated T-cell killing. We believe that the extensive preclinical validation of suicidal CD44v6-targeted T cells warrant their investigation in a fi rst-in-man trial to be conducted in the near future. Disclosure of Interest: None Declared. Introduction: Allogeneic hematopoietic stem cell transplantation (alloSCT) followed by donor lymphocyte infusion (DLI) can be curative for hematological malignancies. Benefi cial Graft-versus-Leukemia reactivity (GvL), however, is often accompanied with undesired Graft-versus-Host Disease (GvHD). In HLA-matched alloSCT, donor T-cells can mediate GvL reactivity by recognition of minor histocompatibility antigens (MiHA) on malignant cells, but also GvHD when MiHA are recognized on non-hematopoietic tissues. Development of broadly applicable T-cell therapies to selectively induce GvL after HLA-matched alloSCT without GvHD requires identifi cation of multiple hematopoietic-restricted MiHA. Materials (or patients) and Methods: Peripheral blood samples and skin biopsies for fi broblast (FB) culture were obtained from two patients with acute/chronic myeloid leukemia who entered into complete remission following DLI after HLA-matched alloSCT and suff ered from GvHD. Patient-specifi c CD8 + T-cells were isolated based on expression of CD137 after 48 hours of stimulation with patient cells. T-cell recognition of (peptide-loaded) EBV-B cells, primary leukemic cells and (cytokine treated) fi broblasts was measured by IFN-γ ELISA. MiHA were identifi ed by Whole Genome Association scanning (WGAs). For microarray gene expression analysis, various (malignant) hematopoietic cells and non-hematopoietic cells from GvHD target tissues were collected. Non-hematopoietic cell types were cultured in the absence or presence of IFN-γ to mimic an infl ammatory environment. Results: 170 MiHA-specifi c CD8 + T-cell clones were isolated from two AML/CML patients with combined GvL and GvHD. All clones were analyzed for reactivity against a mix of known MiHA peptides. T-cell clones specifi c for LB-GEMIN4-1V, ZAPHIR, LB-ADIR-IF and LRH-1 were isolated, of which LRH-1 has therapeutic relevance based on hematopoietic-restricted expression. All clones were then tested against patient-derived FB and the majority showed reactivity against FB after IFN-γ pre-treatment. We identifi ed 6 novel MiHA by WGAs, which were all encoded by ubiquitously expressed genes as demonstrated by microarray expression analysis (CLYBL, PFAS, GLE1, PNO1, YIPF1, FBXO7) . Three of these MiHA showed associations with exon SNPs and could be confi rmed by peptide recognition, while the other 3 MiHA were identifi ed by associations with intronic SNPs, and T-cell epitopes have not (yet) been identifi ed. Furthermore, 3 clones lacked reactivity against (cytokine-treated) FB. For 2 clones, MiHA could not be identifi ed by WGAs due to low population frequencies. The remaining Tcell clone recognized a novel MiHA in HLA-B*15:01 and showed association with an intron SNP in the ITGB2 gene. This gene is hematopoietic-restricted as shown by microarray gene expression analysis, and the T-cell clone for this MiHA strongly recognized primary leukemic cells as measured by IFN-γ ELISA, which together with the lack of FB reactivity indicates its therapeutic signifi cance. Discussion: We isolated T-cell clones for 4 known MiHA, and identifi ed 7 novel MiHA by WGAs in two patients with combined GvL and GvHD. One of the novel MiHA is presented by HLA-B*15:01 on primary leukemic cells and was shown to be encoded by a hematopoietic-restricted gene. Our data show that WGAs followed by microarray gene expression analysis allows rapid identifi cation of therapeutically relevant MiHA. Disclosure of Interest: None Declared. Introduction: In humans, the antitumor effi cacy of genetically targeted T cells often associates with toxicity. In particular, off -tumor expression of the target antigen raises concerns when using T cells modifi ed with chimeric antigen receptors (CARs). CD19-targeted T cells, for example, have caused profound B-cell aplasia. We have recently developed a CD44v6-specifi c CAR mediating potent antitumor in vivo eff ects against leukemia and myeloma (Casucci et al, Blood 2013). Since CD44v6 is expressed at some stages of hematopoietic development, i.e. circulating monocytes, preclinically evaluating the potential hematological toxicities of CD44v6-targeted T cells becomes mandatory. Unfortunately, currently available immunocompromised mouse models do not fully recapitulate human hematopoiesis and are therefore unsuited for this purpose. Aim: To develop a humanized mouse model for predicting the potential hematological toxicities of genetically targeted T cells and, in particular, of T cells modifi ed with a CD44-specifi c CAR. Materials (or patients) and Methods: NOD/scid/gamma-chain double knock-out (NSG) mice completely lack T, B and NK cells, and are consequently severely immunocompromised. Adult NSG mice transplanted with human cord blood-derived CD34+ hematopoietic stem cells (HSCs) engrafted, but mainly reconstituted CD19 + B cells, with few or no myeloid cells and T cells. For improving human hematopoietic reconstitution, we explored a recently described NSG mouse model transgenic for human IL-3, c-kit ligand and GM-CSF (NSG-3GS). Concerning the safety profi ling in vivo, we used cord blood derived CD34-selected cells and autologous T cells, which were activated using beads conjugated to mAbs to CD3/CD28. Results: Compared with NSG mice, HSC-transplanted NSG-3GS mice had a signifi cantly better myeloid reconstitution, including CD44v6-expressing monocytes. Interestingly, newborn NSG-3GS mice transplanted intra-liver did not only reconstitute monocytes, but also circulating single-positive CD4 or CD8 T cells and CD4 + / CD25 + /Foxp3 + T regulatory cells. Reconstituting T cells could be expanded ex vivo following clinical-grade protocols for CAR modifi cation and be re-infused in secondary recipients without causing xenogeneic GVHD, suggesting full functionality and xenotolerance. The infusion of CD19-or CD44v6-targeted T cells in reconstituting NSG-3GS mice resulted in the selective elimination of B cells and monocytes, respectively, and in the preservation of all other cell populations. This was accompanied by transient malaise resembling cytokine-release syndromes observed in humans. After in vivo exhaustion of targeted T cells, selectively eliminated populations re-appeared, indicating full preservation of the HSC pool. For a more rapid ablation of CD44v6-targeted T cells in vivo, we co-expressed either TK or the inducible caspase suicide genes 9 and, by administering the respective prodrugs, were capable of rescuing mice deliberately induced to graft-versus-host disease for surrogating maximal, unexpected toxicity. Discussion: We have developed a humanized mouse model that fully recapitulates human hematopoiesis and is suited for preclinically evaluating the whole toxicity spectrum of genetically targeted T cells, and the clinical value of suicide-gene rescue. Disclosure of Interest: None Declared. Introduction: Ruxolitinib (INCB018424) is the fi rst JAK1/JAK2 inhibitor approved for treatment of patients with myelofi brosis (MF). Although ruxolitinib shows limited anti-clonal activity, a profound improvement of QoL and splenomegaly in MF patients is observed and linked to a substantial reduction of MF-associated circulating pro-infl ammatory cytokines. JAK/STAT-signalling is known to be involved in the regulation of various immune cells including CD4 + T cells, which critically orchestrate infl ammatory responses. To better understand how ruxolitinib is modulating CD4 + T cell response, we here provide an in depth analysis of CD4 + T cell function upon ruxolitinib exposure. Materials (or patients) and Methods: Highly purifi ed CD4 + T cells from healthy human PBMC were stimulated for 4 days in the presence of increasing concentrations of ruxolitinib (0.1μM -10μM). Phenotype and function were analyzed by fl ow cytometry. Cytokine production was quantifi ed either by intracellular staining or by fl ow-based bead assays. Proliferation was detected by CFSE dilution analysis. CD4 + CD62L + T cells obtained from C57BL/6 mice were isolated and subsequently diff erentiated into Th1, Th2, Th9, Th17 and iTreg. Polarization into the diff erent CD4 + T cell subsets was induced by cytokine/antibody cocktails (Th1: IL-12, anti-IL4; Th2: IL-4, anti-IL12; Th9: IL-4, TGF-β, anti-IFNγ; iTreg: IL-2, TGFβ; Th17: IL-6, TGFβ, IL-1b, anti-IFNγ,anti-IL4) together with anti-CD3 and anti-CD28. For analysis of apoptosis/necrosis induction, annexin/propidium iodide staining was applied. Signalling events were analyzed by phospho-fl ow technology (pS6, pSTAT1, pSTAT3, pSTAT5, pERK, pAKT, pP38, pFos, pJun and pZAP70). Results: CD4 + T cell proliferation is signifi cantly and dose-dependently suppressed by ruxolitinib. Of note, we could not detect any changes in the viability of ruxolitinib-exposed CD4 + T cells. In line with previous studies, production of pro-infl ammatory cytokines such as IL-1β, IL-5, IL-6 and TNF-α were dose-dependently inhibited in ruxolitinib-exposed CD4 + T cells, although expression of the pro-infl ammatory IL-8 was increased in a dose-dependent manner. Interestingly, we also observed an increase in IL-2 and IFNγ particularly at the lower ruxolitinib concentrations followed by a dose dependent reduction at higher dose-levels. After shortterm activation of ruxolitinib-exposed CD4 + T cells, proximal TCR signaling events were not aff ected, whereas a clear down-regulation of IL-2 induced STAT5 phosphorylation could be detected. After wash-out the ruxolitinib-induced inhibitory eff ects on CD4 + T cell function were fully reversible, as shown by induction of the T cell activation markers CD25 and CD69. Finally, we diff erentiated murine CD4 + naïve T cells into the various T Helper cell subsets and could provide clear evidence that the diff erentiation capacity of naïve CD4 + T cells into Th1, Th9, Th17 and iTreg was markedly reduced, whereas inhibition of Th2 diff erentiation was only marginally aff ected. Discussion: We could show that ruxolitinib potently aff ects CD4 + T cell biology. These data provide a rationale for testing JAK inhibitors in diseases triggered by hyperactive CD4 + T cells, such as autoimmune diseases. However, they also provide an explanation for the increased infection rates (i.e. viral reactivation and urinary tract infection) seen in ruxolitinib-treated patients. Disclosure of Interest: None Declared. Introduction: Despite major improvements in the last decade in the fi eld of HSCT, acute graft versus host disease (aGVHD) remains a life-threatening complication and substantially reduces effi cacy of HSCT. In particular, the outcome of patients with severe steroidresistant aGVHD continues to be poor. Therefore, the search for new therapeutic strategies for the treatment of aGVHD remains of vital importance. Materials (or patients) and Methods: We expanded MSC using platelet lysate (PL) and infused them in patients with steroidrefractory GVHD. Furthermore immunological changes after infusion of MSC were characterized in vitro. Anti-viral and antileukemia responses of reactive T-cells were tested and phenotypical changes in immune cells were followed up as were cytokines implicated in GVHD. MSC from bone marrow of third party healthy volunteers were isolated via plastic adherence using the CellStack system of Macopharma, expanded with PL for up to three passages (P3), harvested using TripLE and stored. In an open-label, non-randomized prospective phase I/II study patients with steroid-refractory GVHD grade II to IV were treated with hPPL-MSC. 50 patients were included and received up to four infusions. Response rates, TRM and other adverse events were assessed for up to 12 months. In addition, a comprehensive phenotypical and functional analysis was performed with PBMCs and serum isolated from all patients before, during, and after infusion of MSC. Results: The production of MSC using PL in CellStacks takes±22 days to expand from bone marrow to P3, resulting in a mean number of 59x10 5 MSC per double layer CellStack. All batches fulfi lled the release criteria. Between January 2009 and July 2012, 48 out of the 50 patients included were eligible for analysis, 7 children and 41 adults. Mean age was 44,9 years (range 1, 9) . Organs involved in aGVHD were the skin (52%), the GI tract (88%) and the liver (35%). Overall GVHD grade was II for 12 (25%), III for 33 (69%), and IV for 3 (6%) patients. Mean number of infusions were 3 (1) (2) (3) (4) . No severe side eff ects were observed upon infusions. Median follow up was 5,0 months (range 0, 5) . Complete overall response of aGVHD was observed in 24 patients (50%) after a median of 53 days (range 3-116 days). Overall survival was signifi cantly improved in responders when compared to nonresponders (p<0.001). Patients who relapsed with GVHD of the gut were again sensitive to steroids, except one patient who then responded well to a second cycle of MSC. Immunological monitoring shows that anti-viral and anti-leukemia reactive T-cells are well preserved in all patients who responded to MSC treatment. In addition, we identifi ed a combination of biomarkers that already 2 weeks after initiation of treatment predicts a complete resolution of GVHD, whilst this usually only became clinically apparent after months. Discussion: Generation and infusion of MSC in patients suff ering from steroid-resistant aGVHD grade II-IV is feasible, safe and appears to be eff ective. Infusion of MSC did not impair anti-virus or anti-leukemia reactive T-cells. Identifi ed biomarkers predict early a usually late clinical resolution of GVHD and thus might be useful to early guide clinical decision making. Disclosure of Interest: None Declared. Introduction: Innate lymphoid cells (ILC) residing in the lamina propria of the gut are activated by small molecule polycyclic ligands that bind the Aryl hydrocarbon receptor (AHR). ILC are critical to the development and regeneration of gut epithelium through local production of IL22 that stimulates proliferation of epithelial stem cells. Commensal bacteria in the gut metabolize dietary tryptophan to Indole 3 carboxaldehyde (ICA) that diff uses across the epithelial barrier and activates ILC through the AHR, generating IL22 that decreases ingress of pathogenic microbes (Behnsen Immunity 2013). We hypothesized that enteral administration of ICA would decrease the severity of gastro-intestinal GvHD in a MHC mis-marched mouse BMT model. Materials (or patients) and Methods: Lethally irradiated B10.BR (H2 k ) mice were transplanted with 3×10E6 T-cell depleted (TCD) BM alone or the combination of TCD BM plus 2×10 6 purifi ed T cells from B6 (H2 b ) donor mice. Mice received daily gavage with 100 mg ICA or vehicle through day 38 post-transplant. Survival and clinical manifestations of GvHD were monitored through day 38. Histopathology of the gut, serum cytokines, intracellular cytokine staining of donor T-cells and bacterial counts in mesenteric lymph nodes were measured in a separate cohort of experimental mice euthanized on day 20. Mass spectrometry of gut luminal contents measured indole metabolites. Results: Daily ICA treatment did not have a negative impact on survival of transplant recipients ( Figure 1A ). ICA-treated recipients of donor T-cells plus TCD BM had better survival than vehicle-treated mice and less weight loss ( Figure 1B ) than vehicle-treated control mice. Bacteria in lymph nodes obtained at planned necropsy on day 20 were only present in mice that received donor T-cells and developed GvHD (p<0.01), and ICA-treatment was associated with less bacteria compared to controls ( Figure 1C ). ICA treatment lead to lower levels of serum G-CSF and increased levels of IL10 and IL12 on day 20 ( Figure 1D ), supporting the observation that ICA enhanced integrity of the gut epithelium, and decreased microbial ingress. ICA treatment decreased frequencies of splenic Th17 CD4+ T-cells (11% vs 5%, p=0.02, Figure 1E ) and activated CD69+ CD8+ T-cells (39% vs 23%, p=0.02, not shown) on day 20 posttransplant, consistent with the observed reduction in GvHD. Discussion: Enteral administration of indole derivatives is a novel approach to reduce gut GvHD through activation of the AHR in ILC. ICA treatment preserves the gut epithelial barrier, reducing trans-epithelial migration of enteric bacteria and reducing numbers of activated Th17+ T cells. Disclosure of Interest: None Declared. Introduction: Acute GVHD predisposes to chronic GVHD with autoimmune manifestations. It is currently unclear how autoimmunity is linked to antecedent alloimmunity but the thymus may play a role in this process. Using murine models of allogeneic hematopoietic stem cell transplantation (alloHSCT) we previously demonstrated that acute GVHD impairs the capacity of medullary thymic epithelial cells (mTEC) to ectopically express a diverse array of tissue-restricted antigens (TRA), an essential mechanism [PH-O036] S21 for clonal deletion of self-reactive T cells. These data led us to propose that purging of individual self-antigens from the total TRA repertoire leads to a failure of central tolerance induction in acute GVHD. Materials (or patients) and Methods: In the present work we tested this hypothesis in a transgenic mouse model. Since the specifi cities of the autoreactive eff ector T cells in chronic GVHD remain unidentifi ed we used membrane-bound ovalbumin (mOVA) as a model neo-TRA. The OT-II→RIP-mOVA transplantation model was suitable to address this question because (a) negative selection of mOVA-specifi c OT-II TCR recapitulates physiological tolerance induction to TRA in the thymus medulla, and (b) a small reduction in mOVA mRNA in mTEC suffi ces to prevent deletion of OT-II cells. Results: The transfer of Balb/c T cells into allogeneic RIP-mOVA recipients (H-2 d → H-2 b ) reduced mTEC compartment size by 10fold (total numbers decreased from 2000 to ≤200 cells/thymus) at four weeks after alloHSCT. In addition, global gene expression levels of OVA in total residual mTEC cell pools were decreased by 50%. These data indicated that mTEC were deleted in the course of acute GVHD, including the population which expressed mOVA. To test whether acute GVHD aff ected negative selection of OVAspecifi c T cells, RIP-mOVA recipients with or without acute GVHD were lethally re-irradiated 5 weeks after alloHSCT and re-transplanted in a 2 nd syngeneic HSCT with T-cell depleted bone marrow from OT-II mice. Sustained intrathymic expression of OVA in RIP-mOVA recipients without acute GVHD effi ciently purged OVAspecifi c TCR from the T-cell repertoire. The acute GVHD-induced loss of OVA expression, however, allowed for the de novo generation of TCRVα2 + Vβ5 + CD4 + thymocytes and was hence suffi cient to prevent eff ective negative selection of OT-II TCR. OT-II cells were exported from the thymus and emerged in spleens and lymph nodes. Frequencies of OT-II in the spleen increased from 1% to 7% after development of acute GVHD whereas in lymph nodes the frequencies rose from 3% to 14%. Discussion: Our data indicate the presence of an etiologic link between acute GVHD and autoimmunity during subsequent chronic GVHD. As pathomechanism, acute GVHD compromises the ability of the recipient's mTEC to produce individual TRA at physiological levels and hence the ability to sustain a diverse repertoire of TRA required for eff ective negative thymic selection. Escape from central deletion in the mTEC compartment suffi ces for the de novo generation and export of TRA-specifi c CD4 + T cells in mice. Disclosure of Interest: None Declared. Introduction: Donor lymphocyte infusion (DLI) after allogeneic stem cell transplantation can mediate both Graft-versus-Leukemia (GVL) reactivity and GVHD. Since expression of HLA-class II under non infl ammatory conditions is predominantly restricted to hematopoietic cells, donor CD4 + T-cells recognizing minor histocompatibility antigens (MiHAs) even from broadly expressed genes, have been hypothesized to selectively target recipient hematopoietic cells resulting in GVL without GVHD. Based on the hypothesis that development of GVHD due to CD4 + T-cells is caused by induced expression of HLA-class II on non-hematopoietic tissues, GVHD can be expected under infl ammatory conditions due to upregulation of HLA-class II expression. Materials (or patients) and Methods: In a clinical trial we treat patients after T-cell depleted transplantation from HLA-identical sibling donors with 10 6 /kg purifi ed donor CD4 + T-cells. To characterize immune responses, in vivo activated T-cells were clonally isolated using fl owcytometric cell sorting at the time of clinical response, expanded and tested for alloreactivity on patient and donor derived target cells using IFNγ-ELISA. We characterized the immune response in two patients who converted from markedly mixed to full donor chimerism after CD4 + DLI. The fi rst patient did not develop GVHD, the second developed limited skin GVHD which was successfully treated with a short period of local steroids. Results: Following cell sorting and expansion, many alloreactive CD4 + T-cell clones were isolated from both patients. Only one alloreactive CD8 + T-cell clone was isolated from the second patient with minimal GVHD. Using whole genome association scanning, we identifi ed 5 new HLA-class II restricted MiHAs from the fi rst patient and one HLA-class I as well as one HLA-class II restricted MiHA from the second patient. All MiHAs were encoded by broadly expressed genes. To investigate whether these MiHA specifi c T-cells recognized fi broblasts, we tested their reactivity with patient skin derived fi broblasts with and without pretreatment with 500 IE/ml IFNγ. All MiHA specifi c CD4 + T-cells from both patients recognized patient derived EBV-LCL, but not patient derived fi broblasts. Even if fi broblasts were pretreated with IFNγ resulting in good expression of HLA-class II, there was no recognition by the MiHA specifi c CD4 + T-cells. Third party control HLA-DR specifi c CD4 + T-cells could recognize the pretreated fi broblasts. The MiHA specifi c CD8 + T-cells isolated from the second patient recognized both IFNγ pretreated as well as non-pretreated fi broblasts, explaining the limited skin GVHD. Discussion: Here, we demonstrate that not only restricted expression of HLA-class II on hematopoietic cells defi nes specifi city of immune responses after CD4 + DLI in patients with HLA identical sibling donors. While infl ammation can cause recognition of HLA-class II expressing fi broblasts by allo-HLA-class II specifi c CD4 + T-cells, the upregulation of HLA-class II did not lead to recognition by CD4 + T-cells specifi c for MiHAs encoded by broadly expressed genes. These fi ndings broaden the therapeutic window of CD4 + DLI in patients with HLA identical sibling donors because even under infl ammatory conditions GVL may be separated from GVHD. Disclosure of Interest: None Declared.

Introduction: Dendritic cells (DCs) are professional antigen-presenting cells, which display a unique capacity to induce and expand proinfl ammatory CD8 + CTLs and CD4 + T cells. Preclinical studies have suggested that dendritic cells (DCs) play a critical role in T-cell-mediated infl ammation in acute graft-versus-host disease (aGVHD) following allogeneic stem cell transplantation (HSCT). However, little is known about the involvement of human DCs in the pathogenesis of aGVHD. Previously, we have described that human 6-sulfo LacNAc + (slan) DCs display pronounced proinfl ammatory properties and are detecable in aff ected tissues of patients with various infl ammatory diseases. We have also observed a signifi cantly reduced frequency of slanDCs in the blood of patients with severe aGVHD that could be explained by an increased DC migration into aff ected tissues (Tuve et al., Transplantation, 2012) . Following these fi ndings, we explored the tissue infi ltration and cytokine expression of slanDC in colorectal biopsies of patients suff ering from intestinal aGVHD. Materials (or patients) and Methods: The present study includes 65 patients who underwent HSCT at the University Hospital of Dresden. Patients underwent diagnostic colonoscopy whenever symptoms of lower GI-GvHD occurred at a median time of 32 days (range 10-132 days) after HSCT. Histopathological grading was performed as described by Lerner et al. None of the patients had evidence of viral or bacterial infection of the intestinal tract. Most patients (n=58) received steroid treatment (1-2 mg/kg prednisolone) at the time of colonoscopy. The median time of steroid therapy until biopsies was 4 days (range 1-13 days). Patients with clinical improvement of digestive symptoms upon therapy were categorized as steroid-sensitive, whereas patients with no clinical improvement or exacerbation of symptoms upon treatment for more than 1 week were categorized as steroid-refractory. Two colorectal biopsies per patient (total n=124) were formalin-fi xed, paraffi n-embedded, sectioned and stained for the slan molecule, tumor necrosis factor-alpha and interleukin-23. For quantifi cation of slanDCs in tissues, positively stained cells were counted in three diff erent highpower fi elds (HPF) of a section with an Olympus BH-2 microscope (Olympus, Hamburg, Germany) and the mean value was determined. The signifi cance of the results was determined by using the statistical software R Version 2.15.1 and the package 'nlme' Version 3.1-108. Values of *p<0.05 were considered as signifi cant. Results: SlanDCs were found in 119 of 124 tissues at varying frequencies and were preferentially located in the stroma. Remarkably, slanDCs locally expressed the proinfl ammatory cytokines TNF-alpha and IL-23, which are implicated in the pathogenesis of aGVHD.Signifi cantly higher numbers of slanDCs were detected in histologically confi rmed aGVHD grade 1-4 tissues compared to grade 0 tissues. The density of slanDCs was comparable between the diff erent aGVHD grades. Steroid treatment neither infl uenced the number of slanDCs in aGVHD tissues nor their potential to express TNF-alpha and IL-23. The density of slanDCs in aff ected tissues was comparable between steroid-sensitive and steroidrefractory patients. Discussion: The accumulation of slanDCs expressing proinfl ammatory cytokines in aGVHD tissues indicate that slanDCs may contribute to the immunopathogenesis of this disease. Disclosure of Interest: None Declared. Introduction: Infl ammatory diseases, such as GVHD, are characterized by recruitment of infl ammatory cells and by neovascularization. However, the interplay between infl ammation and neovascularization has not been characterized in detail. In particular the initial stimulus for the initiation of infl ammation in target organs remains obscure. To investigate the role of neovascularization during the initiation phase of acute GHVD, we established a chemotherapy-based minor mismatch mouse model and studied neovascularization, leukocyte infi ltration and target organ damage. Materials (or patients) and Methods: We performed allo-BMTs in the LP/J → C57BL/6 model after busulfan-cyclophosphamide conditioning. Colon, liver and skin of syn and allo-BMT recipients were collected on day +2, +5, +7 and +15. Leukocyte infi ltration was assessed by FACS analysis of infl ammatory cells and CD4/CD8 immune staining on tissue sections. FACS analysis of endothelial cells and CD31 immune staining was performed to quantify blood vessels. Quantifi cation was performed with Image J analysis of microscopy image data. The severity of GVHD was monitored by daily scoring of clinical GVHD and by histopathological analysis of tissue sections. Results: We found that neovascularization is a very early event during GVHD. The vascular density, as determined by CD31+ area in immune staining, was already increased at day +2 after allo-BMT (N=10; colon allo-BMT 3,5 % vs 2,1 % syn-BMT; p=0,0008, liver allo-BMT 0,6 % vs 0,4 % syn-BMT; p=0,057, skin allo-BMT 2,9 % vs 1,9 % syn-BMT; p=0,001) (Figure 1 ). At early time points (day +2 and day +5) we found that there were no signifi cant diff erences in lymphocyte infi ltration (Table 1 ) between allo-BMT recipients and syn-BMT recipients. At day +7 we found signifi cantly increased lymphocyte infi ltration (Table 1) in allo-BMT recipients vs. syn-BMT recipients. At day +15 we found extensive lymphocyte infi ltration (Table 1 ) as well as the beginning of clinical GVHD (GVHD-Score N=10; allo-BMT 5,2 vs 1,6 syn-BMT; p=2,2E-07, Histopathology-Score N=10; colon allo-BMT 2 vs 0,75 syn-BMT; p=0,02; liver allo-BMT 3 vs 1 syn-BMT; p=4,3E-12; skin allo-BMT 2,5 vs 0 syn-BMT; p=0,0008). Discussion: Current results from our group, and from others, show that the endothelium is a potential therapeutic target after allo-HSCT. Here we demonstrate that lymphocyte infi ltration and target organ damage are secondary events to neovascularization during GHVD. Our results amend the knowledge on the interplay between the vasculature and infl ammation during GVHD and may facilitate the development of immune-modulatory therapies aiming at the endothelium. Disclosure of Interest: None Declared. Introduction: Adoptive immunotherapy with naturally occurring T regulatory cells (nTregs) followed by conventional T cells (Tcons) prevented acute and chronic GvHD (GvHD) and improved immune reconstitution in HLA-haploidentical stem cell transplantation (Di Ianni et al., Blood 2011) . One major concern is the potential suppression of immune mediated graft-versus-leukemia (GvL) eff ect since FoxP3 + Tregs can also suppress immune response against tumour. However, we showed in immunodefi cient NSG mice engrafted with human myeloid and lymphoblastic acute leukemia cells (7×10 6 iv), Treg (3×10 6 )+Tcon (3×10 6 ) infusion rescued mice from leukemia death in the absence of GvHD. Although the functions of exhausted Mh cells could be rescued by in vivo anti-PD-L1 antibody, we found that this was very ineffi cient if CD8 T cells were transferred without helper CD4 T cells. Evaluation of the surface phenotype of helpless, exhausted Mh cells showed that they retained expression of the TNFR family receptors, CD27 and OX40. To test whether increased co-stimulation via these receptors would enhance anti-PD-L1-mediated rescue of helpless Mh cells, we treated mice following at 5 weeks following adoptive transfer with agonistic anti-CD27 and anti-OX40 antibodies, either alone or in combination with anti-PD-L1. While neither anti-CD27 nor anti-OX40 had any eff ect when applied singly, both acted synergistically with PD-L1 in enhancing both proliferation and eff ector functions of helpless Mh cells. In contrast to anti-OX40, co-stimulation via anti-CD27 in combination with anti-PDL1 led to the rapid onset of GVHD. Transcriptional profi ling of isolated Mh cells from each of the treatment groups showed that anti-CD27 in combination with anti-PDL1 led to marked down-regulation of multiple genes related to tolerance, whereas anti-OX40/anti-PD-L1 co-treatment led to altered expression only of a subset of these genes. Discussion: In conclusion, combined therapy with agonistic co-stimulation and co-inhibitory blockade can reverse exhaustion in unhelped, alloreactive CD8 T cells and this can occur with or without GVHD depending upon the TNFR family receptor targeted. Disclosure of Interest: None Declared. Introduction: Bone marrow transplantation (BMT) is a potential curative approach for the treatment of haematological malignancies. Before transplantation, a chemotherapeutic conditioning regimen ablates the immune system almost completely. In this scenario, the T-cell immune-reconstitution post BMT is initially dependent on the thymic-independent peripheral expansion of donor and chemotherapy-resistant recipient T-cells. This process requires several months, during which the patients are highly susceptible to infections and disease relapse. Recent data suggest that T-cells at the early stages of diff erentiation, and, above all, the self-renewing and multipotent T Stem Cell Memory (T SCM ) discovered by our group, are endowed with superior T-cell immune-reconstitution capacity in preclinical models. Whether T SCM are involved in T-cell recovery following T cell ablation in humans still needs to be defi ned. Materials (or patients) and Methods: Thirty patients underwent non-myeloablative conditioning regimen with fl udarabine, cyclophosphamide (Cy) and TBI and received unmanipulated bone marrow from a haploidentical donor. High-dose Cy was given at days 3 and 4 post-transplant to control graft-versus-host disease. All patients were treated with G-CSF from day 5 until the white blood cell count reached a value ≥ of 1000/mm, and received mycophenolate mofetil and tacrolimus as immunosuppressive therapy after transplant. Peripheral blood was drawn from donors and recipients before transplantation and from recipients every week until day 98, then every month until 1 year after transplantation. Eighteen color polychromatic fl ow cytometry immunophenotyping was used to track T cell dynamics over time.

Results: Here we show that multiple subsets of naïve and memory T cells are infused with the donor graft but only naïve T cells are responsible for subsequent immune reconstitution. All patients reached ~100% complete donor chimerism by 30 days post transplant. Following transfer in the lymphopenic host, donor memory T cells underwent rapid proliferation, as assessed by Ki-67 expression, and were thus highly suceptible to Cy treatment. Conversely, due to their delayed activation kinetics, naïve T cells survived Cy and rapidly diff erentiated to CD45RO-CD45RA+CD27+CCR7+CD57-CD95+ T SCM . In vitro treatment of FACS-sorted T cell subsets with Cy confi rmed these in vivo fi ndings. Naïve -derived T SCM dominated the T cell compartment early after transplantation (~80% of all donor T cells) while central memory and eff ector memory T cells appeared at later time points, suggesting the diff erentiation of T SCM into more diff erentiated subsets. Importantly, memory T cells specifi c for multipe antigens that were transferred with the graft could not be found in the recipients, indicating the poor role of memory cells in immune reconstitution in the post transplant setting. Discussion: Overall, these fi ndings indicate that naïve -derived T SCM are superior to other memory T cell subsets in mediating immune reconstitution in lymphodepeted hosts and provide further support to the linear model of memory T cell diff erentiation in humans. Our results have important clinical implications as they indicate that adoptive transfer of memory cells is necessary to provide adaptive T cell immunity in patients receiving haploidentical BMT and post-transplant Cy. Disclosure of Interest: None Declared.

Introduction: Allogeneic myeloablative hematopoietic stem cell transplantation (HSCT) is challenged by long-lasting immunodefi ency and acute Graft-versus-Host Disease (aGVHD), both contributing to treatment-related mortality (TRM). Interleukin-7 (IL-7) is a cytokine essential for de novo T-cell generation in thymus and peripheral T cell homeostasis. Accordingly, recombinant IL-7 has been suggested as a potential therapeutic agent in patients with lymphopenia. Studies from our group indicate that polymorphisms in the IL-7 receptor α-chain are associated with the risk of aGVHD and treatment-related mortality (TRM) in HSCT.

Our aim was to investigate associations between plasma IL-7 levels in the early post-transplant phase and lymphocyte reconstitution, aGVHD and mortality. Materials (or patients) and Methods: 81 paediatric and adult patients undergoing HSCT at Rigshospitalet, Denmark, from 2010-2013 were included. Diagnoses included malignant (n=63) and benign diseases (n=18). Donors were either SIB (n=18), MUD (n=55) (BMSC or PBSC grafts) or unrelated UBSC grafts (n=8). Conditioning regimens included TBI (n=46), busulphan plus cyclophosphamide (n=17), fl udarabine-based conditioning (n=12) or other (n=6). Plasma IL-7 was measured by ELISA (R&D Systems) before transplantation, at the day of transplantation and at day +7, +14, +21, +28, +60 and +90. T, B and NK cells were counted using fl ow cytometry at day +30, +60, +90 and +180. Results: Plasma IL-7 levels increased from baseline levels (2.4 pg/ ml), peaking at day +7 (22.1 pg/ml, p<0.0001) and then gradually declined to pre-transplant levels. High IL-7 levels at day +7 were associated with the use of anti-thymocyte globulin (ATG) in multivariate analysis (p=0.027). Although CRP increased signifi cantly during the same period, IL-7 was not correlated with CRP. IL-7 level day +7 was signifi cantly negatively associated with T cell subset counts day +30 to +180 in multivariate analysis (at day +60: CD3+:β=-9.6×10 6 cells/L, p=0.035; CD8+:β=-7.6×10 6 cells/L, p=0.049; CD4+:β=-1.9×10 6 cells/L, p=0.10). In contrast, a positive association was found with CD45+CD19+ B cells at day +60 (β=10×10 6 cells/L, p=0.0019). 33 patients (40.7%) developed aGVHD grade 2-4, of those 70% adults. For the total study cohort IL-7 levels above median (18.2 pg/ml) at day +14 tended to be associated with acute GVHD in multivariate analysis (p=0.071, OR=4.2). This was confi rmed in adults when stratifying for age (p=0.038, OR=12). Overall survival, TRM and relapse were not associated with IL-7 levels in a multivariate model that included recipient age, TBI and treatment with ATG as signifi cant factors. Discussion: IL-7 plasma levels are elevated during the period of lymphopenia following HSCT and most pronounced in patients treated with ATG. IL-7 levels are negatively correlated with the rate of T cell recovery and positively correlated with B cell recovery, and high levels of IL-7 are associated with grade 2-4 aGvHD. Importantly, the data suggest that IL-7 regulation is independent of the general acute phase response in the early post-transplant phase. In contrast, IL-7 appears to be determined by feed-back regulation based on the size of the total T cell population. High IL-7 levels early after HSCT predict inferior T cell reconstitution up to 6 months later, possibly refl ecting lower total T cell numbers early after HSCT or alternatively aGVHD mediated suppression of haematopoiesis. Disclosure of Interest: None Declared. Introduction: Changes in chimeric level are often key indications for the need to modify a patient's therapeutic regimen. It is not widely recognized that in Stable Mixed Chimerism (SMCh) spontaneous, predictable patterns of variability regularly occur even over years. This is the fi rst report of a comparison of these longterm chimerism (Chm) trends in diff erent disease processes, both malignant and non-malignant. Materials (or patients) and Methods: We examined >500 sequential samples from 48 disease free patients at 0 In several cases an inverse relation between NK and TCells was noted, approximately in parallel with magnitude stratifi cation. Within each magnitude stratum, one of two patterns of inter-sample variability were characteristically observed for each patient: STEADY (5-10%) or FLUCTUATING (10-25%, ?aperiodic cycling); such fl uctuations occurred between single or groups of samples. Most cases showed a STEADY pattern, which was typical of βTM. Fluctuating patterns of SMCh were common in SCID, WAS, OP. SCID cases fl uctuated widely in an interval of weeks; WA and OP had a "concave" pattern -often with values <10% at the nadir -extending over months or years. Malignancy, with MRD-hematopoiesis, showed a high stable pattern in 5/6 cases. T and myeloid cells had the same patterns. Discussion: Conclusions. (1) Routine longitudinal Chm tracking commonly detects SMCh in non-malignant disease, and occasionally in malignancy.

(2) Four spontaneous patterns of SMCh were noted due to magnitude stratifi cation and inter-sample variability over time. (3) Once established, a pattern persists in a patient for the duration of follow-up. (4) Patterns of inter-sample variability are often characteristic, but not diagnostic of a specifi c disease. IMPLICATIONS: (a) The fi ndings suggest that tolerance may entail interactive cellular regulation between host and graft, perhaps involving a reciprocal NK vs TC relationship. (b) The limited repertoire of temporal and longitudinal patterning suggests that animal models, to explore the dynamics of engraftment and tolerance in man, are feasible. (c) Despite the variability in Chm levels, these patterns of SMCh defi ne settings that do not require therapeutic intervention, at least in non-malignant disease. Disclosure of Interest: None Declared. Introduction: Allogeneic stem cell transplantation (SCT) is a curative therapy for patients (pts) with AML/MDS. Reduced toxicity conditioning with fl udarabine and treosulfan (FT) is a dose intensive regimen with enhanced anti leukemia eff ect and acceptable toxicity. Relapse after SCT remains the main obstacle to cure. Natural killer (NK) cell alloreactivity has documented role in reducing relapse after haplotype mismatched SCT in AML but its role in

[PH-O045]

HLA matched SCT is more controversial. The missing ligand theory suggests that missing KIR ligands in the recipient may drive donor NK alloreactivity in the absence of HLA mismatch. Materials (or patients) and Methods: In this analysis we defi ned the prognostic factors for relapse and overall survival (OS) in 203 pts with AML (n=129, 29 of them secondary) and MDS (n=74) given FT conditioning. In particular we tested the role of recipient HLA ligands for NK KIRs in controlling relapse. Results: The median age was 58 years (21-76). The donor was an HLA-matched sibling (n=97) or matched unrelated (n=106). Disease status was CR1 (n=65), CR2/later CR (n=24), no CR (n=44) or previously untreated MDS (n=70). 80 pts (39%) had poor-risk cytogenetic or molecular markers. HLA typing revealed that 67% expressed at least one Bw4 antigen, 81% expressed group 1 C alleles and 66% group 2 alleles. With a median follow-up of 48 months (6-108), 86 pts are alive, 66 died of relapse and 51 of non-relapse causes (NRM). 5-year OS and leukemia-free survival (LFS) rates were 39% (95%CI, 32-47) and 36% (95%CI, 28-43), respectively. 5-year cumulative incidence of relapse and NRM was 38% and 27%, respectively. The most signifi cant predictor of relapse was disease status at SCT. 5year relapse rates were 33% in CR/untreated and 55% in refractory disease, respectively (p=0.001). Pts expressing HLA C group 1 alleles had a relapse rate of 45% compared to 26% in pts with missing HLA C group 1 ligands (p=0.03). Missing HLA C group 2 and/or Bw4 ligands had no eff ect on relapse. Multivariate analysis identifi ed no CR at SCT (HR 3.6, p=0.001), missing HLA C group 1 ligand (HR 2.6, p=0.03), sibling donor (HR 1.8, p=0.04), poor cytogenetics (HR 1.7, p=0.05) and female donor to male recipient (HR 0.5, p=0.06) as independent factors predicting relapse. The reduced relapse rate associated with missing HLA C group 1 ligand was more pronounced in pts with MDS/secondary AML, 39% Vs 7%, respectively (p=0.02) and in pts in untreated disease or no CR, 46% Vs. 8%, respectively (p=0.02). Missing HLA ligands were not associated with GVHD or NRM. LFS was 46% in pts with missing HLA C group1 ligand compared to 30% in pts expressing the ligand (p=0.07). Multivariate analysis identifi ed SCT not in CR (HR 2.8, p=0.0007), comorbidity score >2 (HR 1.5, p=0.06) and missing HLA C group 1 ligand (HR 1.9, p=0.02) as independent predicting factors for LFS. Discussion: Missing HLA C group 1 ligand in SCT recipients with AML and MDS may be associated with reduced relapse risk, similar NRM and improved LFS possibly due to enhanced NK alloreactivity according to the missing ligand theory. This eff ect is more pronounced in pts with MDS and secondary AML and in pts with advanced disease. These observations merit further study in larger cohorts and in pts given other conditioning regimens. Disclosure of Interest: None Declared. Introduction: Regulatory Т cells with inhibitory function (Treg) are critical for the maintenance of immune tolerance. The CD39 pos Treg subset which increases after prolonged stimulation and inhibits Tcell responses has not been extensively studied in the settings of allogeneic hematopoietic stem cell transplantation (HSCT). Materials (or patients) and Methods: Samples from 27 patients (pts, age 9-39y) were studied at 1 st , 2 nd , 3 rd , 6 th , and 12 th month (mo) after allogeneic related (N=16) and unrelated (N=11) HSCT. Percentage and absolute counts (AC) of T, B, NK, CD4 and CD8 T cells, as well as the subsets of naïve (N, CD45RA pos CCR7 pos ), central memory (CM, CD45RA neg CCR7 pos ), eff ector memory (EM, CD45RA neg ССR7 neg ), eff ector (E CD45RA pos ССR7 neg ), and terminally diff erentiated (TE, CD28 neg CD57 pos ) cytotoxic CD8posT cells in parallel with CD25 hi CD127 low CD4 Treg and their CD39 pos Treg subset were determined by multicolor fl ow cytometry (FACSCanto II, BD).

Chronic T cell activation was measured by CD38 expression. Cytokine profi les (IL-2, IFNg, TNFa, IL-4, IL-5, IL-10) were determined in PHA-stimulated whole blood samples (BD CBA Human Th1/Th2 kit.). All pts received prophylaxis of graft-versus-host-disease (GVHD). Six pts developed clinical signs of mild to moderate acute (grades I-II) and 11 -of limited chronic GVHD. All pts with GVHD (N=17) were treated with corticosteroids at some time point of this complication. Results: One year after HSCT, the T cell population was characterized by prevalence of eff ector and terminally diff erentiated CD8 T cells, and a predominant Тh2 cytokine profi le. The level of Treg for all pts reached a maximum at 3 rd mo (average 15%) and decreased to reference values thereafter (8% at 12 mo vs. 6% in healthy controls, p>0.05). The proportion of CD39 pos Treg increased after 3 rd mo post HSCT, reaching an average proportion of 50% within Treg population at 1 year (vs. 27%, in healthy controls, p<0.01). This proportion correlated with the percentage of CD38 pos T cells (R=0.42, P<0.01), CM CD4 (R=0.45, P<0.01) and, inversely with TE CD57 + CD8 T cells (R=-0.51, P<0.01) and CD4 T AC (R=-0.61, P<0.01). Further on, the percentage of CD39 + Treg and the level of CD39 expression (MFI) were signifi cantly increased in the group of pts with clinical signs of chronic GVHD (N=13) at 6 th and 12 th mo time points (MW p>0.01) regardless of treatment with immunosuppressive drugs. Discussion: Other studies have shown that low levels of Treg after HSCT are associated with increased risk of acute GVHD while pts with chronic GVHD have high Treg levels. Treg prevent the terminal diff erentiation of eff ectors and maintain the memory cell pool after immune stimulation but subsets with possible adverse eff ects are selected in case of extreme/prolonged stimulation such a chronic GVHD. According to our results, total Treg were not as sensitive marker of immune activation as their CD39 pos Treg subpopulation which increases signifi cantly in the settings of chronic GVHD. In our study, increased CD39 pos Treg levels were associated with poor eff ector cell diff erentiation and lower CD4 pos AC, that might be critical in conditions of immune suppression by increasing the risk of infectious complications. Our results draw the attention to the CD39+Treg subset as a putative biomarker of poor immune reconstitution in post-allogeneic HSCT patients. Disclosure of Interest: None Declared.

than younger patients. This leads to an increased transplantationassociated mortality (TRM) in older patients. As a consequence, reduced intensity conditioning regimens were developed. In recent years this led to a signifi cantly increasing proportion of patients older than 55 years among all transplanted patients. We investigated the HLA-associated risk in elderly patients undergoing unrelated (HSCT) by comparing their risk to that of younger patients. Prior to introduction of 5 th JACIE Standards, the same doctor consented both the RD and their recipient in 34/173 (20%) donations, compared to 0/43 donations after these standards were introduced (p=0.002). On retrospective assessment of RD records using AN acceptance criteria, 81/210 (39%) would have been deferred.

Of the 39% who failed, the most common reasons were age<16 (36%) or >60 (33%, with 21% due to age>60 alone), uncontrolled hypertension BP>150/95 (8%), and diabetes on treatment (2%). 7 RDs (3%) experienced SAEs, and a further 7 were reviewed for severe side eff ects of GCSF but required no, or minimal, intervention. RDs aged >60 were more likely to develop either SAEs or severe side eff ects (p=0.026) and adults who would not be accepted as UDs due to increased risk to donor health, experienced signifi cantly more SAEs (p=0.049).

A median of 4.83X10 6 CD34/kg were harvested from PBSC donors, with no signifi cant diff erence in cell dose or number of apheresis procedures between older and younger donors or between sexes. Despite this, slower engraftment was seen in recipients of PBSC donors >60 years, who were more likely to take 18 days or longer to engraft (p=0.02). Bone marrow donors harvested a median of 2.66x10 6 CD34/kg with an increased probability of not reaching the target cell dose for paediatric RDs donating to an older sibling (p=0.031). From 2010, RDs were followed up with medical questionnaires, achieving a total of 95 donor follow up years. Overall, 30/54 (55.6%) questionnaires were completed and returned, with RDs whose recipient had died less likely to return follow-up questionnaires than those whose recipient was alive (31% vs 66% (p=0.020)). Discussion: HPC donation is less safe in RDs who do not meet UD criteria for medical eligibility. If these criteria (excluding age) were routinely employed, 17% of RDs in our cohort would have been deferred. Identifying specifi c heath issues that defi ne donors with increased risks for SAEs may allow us to be more permissive with respect to other selection criteria. Although changes to JACIE Standards have improved practice, follow-up and AE data must be collected at a national level if we are to further understand the risks of donation for RDs and create specifi c RD deferral criteria. Our fi ndings of slower engraftment in recipients of older donors warrant further investigation in larger studies. Disclosure of Interest: None Declared. To study the impact of HLA-mismatch in more detail we analysed only those 696 patients with high resolution HLA-typing and distinguished a 10/10 (n=363), a 9/10 (n=204), and a ≤ 8/10 (n=129) matched group. 9/10 (2%) and ≤ 8/10 (2.4%) HLA-matched transplants had more graft failure than 10/10 (0%), p=0.02. Severe grade III/IV acute GvHD was more frequently observed in the ≤ 8/10 (23%) vs. 9/10 (14%) vs. 10/10 (14%, p=0.05) groups while chronic GvHD did not diff er between the groups (p=0.5).

In a multivariate analysis for OS 9/10 HLA-mismatch did not differ signifi cantly from 10/10 (HR: 1.118, p=0.4), while ≤ 8/10 had a signifi cantly worse survival (HR: 1.385, p=0.02) . If the role of mismatch is analysed separately between MAC (n=298) and RIC (n=368) the 5-year overall survival after MAC did not diff er signifi cantly between 10/10 vs. 9/10 vs. ≤ 8/10 (62% vs. 54% vs. 54%, p=0.19) while after RIC the 5-year OS for 10/10 vs. 9/10 vs. ≤8/10 was 51% vs. 48% vs. 33% (p=0.002). Discussion: Results of 10/10 and 9/10 HLA-matched unrelated donors are similar after ATG-containing conditioning regimen.

Higher mismatches (≤ 8/10) are associated with poorer outcome, particularly after reduced-intensity conditioning. Disclosure of Interest: None Declared. Introduction: The HLA loci are sometimes classifi ed into high expression loci (HEL) HLA-A, B, C and DRB1 and low expression loci (LEL) HLA-DRB3/4/5, DQ and DP based on their surface expression or impact of mismatches on transplantation outcome. Mismatches in HLA-HEL are strongly associated with adverse transplantation outcome. In contrast, HLA-LEL mismatches are frequently not taken into account by donor selection. However, we and others have illustrated that immune responses against mismatched HLA-DQ and HLA-DP can contribute to the development of both GVHD and Graft-versus-Leukemia reactivity. Here, we show that a mismatch in HLA-DRB3, not considered to be relevant during donor selection, can result in the development of a strong polyclonal immune response associated with severe GVHD after infusion of purifi ed CD4 + donor lymphocytes after T-cell depleted stem cell transplantation with a fully HLA 10/10 matched unrelated donor (MUD). Materials (or patients) and Methods: We are conducting a clinical trial in which we treat patients 3 months after T-cell depleted transplantation from HLA 10/10 MUDs with 0.25×10 6 /kg purifi ed donor CD4 + T-cells to promote immune reconstitution. One of the patients with an HLA-DRB3 mismatch (patient HLA-DRB3*02:02:01, donor HLA-DRB3*01:01:02) in addition to an HLA-DPB1 mismatch, developed severe liver GVHD after CD4 + donor lymphocyte infusion. To characterize the immune response in this patient, in vivo activated HLA-DR expressing T-cells were clonally isolated using fl owcytometric cell sorting at the time of the occurrence of GVHD, expanded and tested for alloreactivity on patient and donor derived target cells using IFNγ-ELISA.

Results: Following cell sorting and expansion of in vivo activated T-cells, a total of 18 alloreactive CD4 + T-cell clones of polyclonal origin were identifi ed, whereas none of the expanding CD8 + T-cell clones showed alloreactivity. To determine the HLA restriction of the CD4 + T-cell recognition, we performed blocking experiments with antibodies against HLA-class II, DP, DQ or DR. Surprisingly, the recognition of only 3 clones was blocked by a HLA-DP antibody. Recognition of 9 clones was blocked by a pan HLA-class II antibody whereas recognition of 6 clones could not be blocked using the standard HLA-class II blocking antibodies. To investigate whether the mismatched HLA-DRB3 could be the target of recognition of the 15 clones from both groups, we retrovirally transduced donor derived EBV-LCL with the patient variant of HLA-DRB3, being HLA-DRB3*02:02. After transduction strong recognition of HLA-DRB3*02:02 expressing target cells by alloreactive CD4 + T-cell clones was found. These experiments illustrate the development of a dominant polyclonal T-cell response against the mismatched HLA-DRB3 molecule, associated with GVHD in this patient. Discussion: In contrast to the assumption that mismatches in HLA-DRB3/4/5 are not of immunogenic signifi cance, we showed that following an HLA 10/10 matched transplantation, mismatches in HLA-DRB3 can have immunogenic consequences and may induce a polyclonal allo immune response associated with severe GVHD. Disclosure of Interest: None Declared. .001) -a potential cost advantage. Discussion: Transplant centres using an expert donor identifi cation and advisory service are able to expedite unrelated donor transplantation through rapid initiation of CT requests and shorter donor selection times, despite having a greater proportion of patients with rare HLA phenotypes. In addition, our experience shows that this service increases the utilisation of national donors. Use of younger donors over male donors refl ects the donor selection preferences of participating GIAS centres. Disclosure of Interest: None Declared. Introduction: Haploidentical allogeneic hematopoietic stem cell transplantation (allo-HSCT) represents a solution for patients who require allo-HSCT but lack a conventional donor. However, haploidentical HSCT may be associated with slow immune reconstitution and a high rate of nonrelapse mortality. We developed an approach of T-cell-replete haploidentical HSCT with lowdose anti-T-lymphocyte globulin and prospectively compared outcomes of consecutive patients undergoing T-cell-replete haploidentical allo-HCT performed at our center with all contemporaneous T-cell-replete allo-HCT using matched sibling donors (MSDs) and unrelated donors (URDs). Materials (or patients) and Methods: Outcomes of 234 consecutive patients undergoing T-cell-replete fi rst allo-HSCT for hematologic malignancies performed contemporaneously at a single center (68 using haploidentical related donors (HRDs); 68, MSDs; 98, URDs) were compared. For patients receiving MSD-HSCT or URD-HSCT, the main myeloablative conditioning regimen used involved busulfan (Bu; 3.2 mg/kg/d IV on days -7 to -4), cyclophosphamide (Cy; 60 mg/kg/d IV on days -3 to -2). Rabbit anti-thymocyte globulin (rATG; thymoglobulin; Genzyme) was administered to patients receiving URD-HSCT (4.5-6 mg/kg total dose). For patients receiving HRD-HSCT, the conditioning regimen consisted of cytarabine (4 g/m 2 /d IV on days −10 to −9), Bu (3.2 mg/kg/d IV on days -8 to -6), Cy (60 mg/kg/d IV on days -5 to -4), and anti-T-lymphocyte globulin-Fresenius (2.5 mg/kg/d IV on days −5 to −2). Results: (1) There were no signifi cant diff erences in terms of underlying disease, risk classifi cation, time from diagnosis to HSCT, disease status at HSCT between the MSD-, URD-, and HRD-HSCT cohorts.

(2) Grades II-IV and severe aGVHD were all signifi cantly more frequent in patients undergoing HRD-HSCT compared with those undergoing MSD-HSCT (II-IV: 42.6% vs 19.1%, P = 0.0015; severe aGVHD: 17.65% vs 5.88%, P = 0.03). However, the incidences of II-IV and severe aGVHD were comparable in patients receiving transplants from HRDs to those from URDs (II-IV: 42.6% vs 40.8%, P = 0.89; severe aGVHD: 17.65% vs 13.27%, P = 0.48). The incidence of cGVHD was not signifi cantly aff ected by donor types. (3) The 4year incidence of relapse was not signifi cantly aff ected by donor types according to all patients (24.2% in the MSD cohort, 22.8% in the URD cohort, 11.9% in the HRD cohort, P>0.05). However, after controlling for high-risk patients, a superior graft-versus-leukemia eff ect was observed in patients undergoing HRD-HSCT compared to MSD-HSCT or URD-HSCT. In high-risk patients receiving MSD, 36.8% experienced relapse, as did 33.6% in the URD cohort, but the incidence decreased to11.1% in the HRD cohort (MSD vs HRD, P = 0.015; URD vs HRD, P = 0 .028). (4) HRD-HSCT yielded comparable rates of 4-year overall survival (OS) and disease-free survival (DFS) to MSD-HSCT or URD-HSCT were receiving cyclosporine or sirolimus, 6 had active uncontrolled infection. All but one had severe kidney injury related to TMA, and dialysis was required for 3 of them. Neurologic dysfunction was present in 4 pts. ADAMTS 13 activity and complement regulatory proteins plasma levels were within normal levels at TMA diagnosis. All pts were treated according to aHUS dosing regimen. Before eculizumab therapy, TMA treatment consisted in: cyclosporine/ sirolimus withdrawal (9/9 pts), plasma exchange in 6 pts (median duration, 14 days), rituximab in 2 pts. Eculizumab was initiated following failure of previous therapy in 10 pts or as fi rst line therapy in 2 pts, at a median of 31 days (3-154 days) after diagnosis of TMA. Five pts did not receive maintenance treatment due to failure in 3, response but progressive aGVHD and/or relapse of primary disease in 2. Seven pts received induction and maintenance eculizumab therapy (median number of infusions, 7 (5-31)). Four pts had a hematological response and 3 pts had no response to eculizumab therapy. Introduction: Myelofi brosis (MF) is a clonal hematological disorder classifi ed as a myeloproliferative neoplasm (MPN). Currently allogeneic stem cell transplantation (HSCT) is the only curative therapy but it is associated with a high morbidity and mortality. Infectious complications, acute and chronic GvHD are major complications which infl uence quality of life after transplantation. Materials (or patients) and Methods: In this cross-sectional study we examined 77 patients (male=37, female=40) with a median age of 61 years (range, 40 to 79 years) who received an allogeneic HSCT for myelofi brosis with a busulfan-based reduced-intensity conditioning from related (n=20) or unrelated donor (n=54) in the period from 1999 to 2011. At time of evaluation, 72 patients were in complete remission, 5 patients were not in complete remission, 25 patients suff ered from GvHD (limited=14, extensive=11). The patient population was divided into four groups according the time between transplantation and evaluation of QoL. First group= 6-36 months (n=24); second group= 36-60 months (n=16); third group= 60-85 months (n=22); fourth group 85-146 months after transplantation (n=15). QoL was measured by using the two validated questionnaires FACT-BMT and MPN-SAF. By means of statistical analysis (SPSS; ANCOVA) it was examined whether there are signifi cant diff erences between the groups. Results: According to the MPN-SAF Total Symptom score (the lower the score the better is the result: possible range= 0-100, the mean scores of the groups one to four are as followed: group1: 18.9 (SD =11.3), group2: 21.1 (SD=17.9) ,group 3:19.3 (SD=15.5) and group 4: 18.9 (SD=12.4). There were no signifi cant diff erences between the groups (p=0,962) . Unfortunately no MPN-SAF Total Symptoms score was available before transplantation, but according the mean value of MPN described in the literature(mean=2 5.3;SD=17.2) the mean score after transplantation (mean=19.5; SD=14.0) of our study population was lower. The mean scores of the FACT-BMT from group one to four were 115.3 (SD=16.6); 120.9 (SD=17.5); 112.4 (SD=15.5) and 115.3 (SD=20.9) There were also no signifi cant diff erences between the 4 groups (p=0.541). We investigated whether there are signifi cant diff erences of the results in both questionnaires regarding gender (male/female), donor (unrelated/related), GvHD (yes/no), HLA (matched/mismatched), primary or secondary MF (post-PV/ post-ET/PMF), remission (complete/ not complete), but none of the variables did infl uence QoL signifi cantly. The subscales BMTS, FACT-G, and its subscales Physical Well-Being (PWB), Functional Well-Being (FWB), Emotional Well-Being (EWB), Social Well-Being (SWB), of the FACT-BMT were calculated and analyzed by ANCOVA. No signifi cant diff erences were found among the groups.

Comparing the mean value of the FACT-G (mean=85.7; SD=13,1; range= 60-108)and its subscales PWB ( mean=22.7; SD= 4.5), SWB (mean=22.4; SD=4.7), EWB (mean 19.9; SD=3.1), FWB (mean=20.6; SD=4.7) of our entire patient population with the mean value of a normal population FACT-G (mean=80. (34), lymphoid (36) malignancies and aplastic anemia (2). The pts received grafts from major (57) and bidirectional (17) ABO incompatible donors, siblings in 45, not related (MUD) in 27 and double-cord blood units in 2 pts, following a myeloablative (MA, 57) and non-myeloablative (NMA, 17) conditioning regimen.

Results: A total of 17 (22.9%) pts were diagnosed with either delayed RBC engraftment (4/17) or PRCA (13/17) with a median of 2 (0-13) % of erythroid precursors in bone marrow after a median time of 31 (20-95) days post alloHCT. The complication was not correlated with granulocyte (p=0.6) or platelet (p=0.1) engraftment delay, the type of donor (sibling vs MUD, p= 0.3) or the intensity of the conditioning regimen (MA vs NMA, p=0.5). All pts were initially treated with erythropoietin and 10 of them responded. The rest of them (7) were given additionally steroids but this was successful in only one pt. Finally, 6 pts with refractory PRCA were treated with plasmapheresis and all of them responded post 6 (5-11) sessions. None of them experienced a PRCA relapse until last follow-up. With a median follow-up of 33 (1-117) months, the transplant outcome was not found to be signifi cantly diff erent between the delayed red cell engraftment/PRCA pts group and the pts not experiencing red blood cell line delay reconstitution (control group The most frequent technique used for irradiation is "patient in one fi eld" using two fi elds per fraction and two patient positions per fraction (n=36, 64%), however, altogether 11 modalities were described with regard to the technique, number of fi elds and positions per fraction. Source to surface distance is 2 to 5 m (most frequently 4 m; n=10, 18%). In 23 (41%) centers patients are immobilized during TBI, using 9 diff erent types of device. Fifty-two centers (93%) use in vivo dosimetry applying 5 types of detectors (mostt frequently semiconductors; n=37, 66%). Accepted discrepancy between planned and measured dose ranges from 1.5 to 10%. In 47 (84%) centers lungs are shielded during irradiation and lung density is considered for treatment planning. Maximum accepted dose for lungs ranges from 6 to 12 Gy. Additionally, in some centers lenses (14%), thyroid gland (7%), larynx (4%), kidneys (4%) and/or salivary glands (2%) are shielded. Introduction: Neurological complications (NC) may appear after allogenic hematopoietic stem cell transplantation (allo-HSCT), but its incidence and impact are not well defi ned. We retrospectively analyzed the occurrence, characteristics and mortality of NC in the setting of allo-SCT. Materials (or patients) and methods: From January 2000 to September 2013, 506 patients underwent an allo-HSCT in our centre. We evaluated 44 patients with signifi cant neurological complications in the setting of allo-HSCT. NC were defi ned as those involving the central nervous system (CNS) (brain, cerebellum, brainstem and spinal cord) and those involving the peripheral nervous system (PNS) (lower motor neuron, motor and sensory roots, dorsal root ganglia, nerve plexus, peripheral nerves, neuromuscular junction and muscles). Median age was 41 years (4-67) and 25 (57%) were male. Underlying diseases were 21 Acute Leukemia, 9 MDS, 5 NHL, 1 HL, 3 Multiple Myeloma and 5 other. Median prior treatment was 2 (0-10) and 22 patients (50%) were in CR, 15 (34%) in PR and 7 (16%) had refractory disease. 14 (32%) had failed a prior HSCT (9 autologous and 5 allogenic S34 41 (63%) of the survivors (63%) had no active GvHD, and only 9 Patients (13%) had an extensive chronic GvHD. Within the overall follow up period, relapse/progression occurred in 35 patients. Thirty of them were treated with donor-lymphocyte infusions (DLI) (median 3 DLIs, range 1-6) and/or a second allogeneic transplantation (n=13). Nineteen of those were at the last follow up alive and 15 hat a complete remission. The estimated OS of all relapsed patients after a median follow up of 46 months (range 4-62 months) beginning from the time of relapse was 55%. Discussion: This update of a prospective trial using reduced intensity conditioning followed by allogeneic stem cell transplantation for myelofi brosis confi rmed a very good long-term OS. Relapse remains the main problem late after transplantation. However, with adoptive immunotherapy using DLI or even second allogeneic transplantation a second remission with long term survival can be induced in about 50% of the relapsed patients. Developing methods for remission monitoring and early prediction and treatment of relapse should be the focus of future studies. Disclosure of Interest: None Declared.

Introduction: In patients with refractory chronic lymphocytic leukemia (CLL), the potential to cure is unique to allogeneic stem cell transplantation (HSCT). Clearance of minimal residual disease by suspected graft-versus-leukemia eff ects has been described elegantly and documented in several independent cohorts of patients. Yet, data from large numbers of patients which support the concept of cure by allogeneic transplantation have not been published. With the advent of new targeted therapies for patients with advanced chemo-refractory CLL this information becomes crucial for clinical decision making and patient counseling. Therefore, we aimed at the description of long-term survival outcomes and the estimation of excess mortality compared to the age-and sex matched general population. Materials (or patients) and methods: Data from patients with CLL who received a fi rst allogeneic HSCT between 2000 and 2010 and were registered with the EBMT database were analyzed. Survival probabilities were calculated by means of Kaplan-Meier estimator in the total population and in patients who passed the 2 and 5-year landmark without previous relapse or progression. Excess mortality of the landmark populations compared to an age-, sexand calendar year-matched general population was estimated with a Cox Regression model. Results: In total 2599 patients were included into the analysis. The median follow-up of patients alive was 4.0 years. The median age at HSCT was 55 years (range, 12 to 74 years). 159 patients (6.1%)

were below the age of 40 years at the time of transplantation. Seventy-four percent of patients were male. The status at HSCT was complete remission in 15%, partial remission in 47%, stable or progressive disease in 32% and missing in 7% of the patients. Fifty percent of the patients had an HLA-matched sibling donor and 75% received reduced-intensity conditioning. For the whole cohort of patients the 5-and 10-year overall survival (OS), progression-free survival (PFS), non-relapse mortality (NRM) were 46%, 35%, 36%, and 35%, 28%, 40%, respectively. The cumulative incidence of relapse was 21% at two years, 28% at fi ve years, and 32% at ten years. 1029 patients and 397 patients were alive without relapse or progression and in follow-up at two and fi ve years after HSCT. Five years after the patients had passed the 2-and 5-year landmark, progression-free survival was 62% (95%CI, 59% to 66%), and 79% (95%CI, 73% to 85%) respectively. Compared to the general population excess mortality of the 5-year landmark population in the subsequent fi ve years was estimated to be 3.1% for male patients at an age of 45 years, 9.0% for male patients at an age of 55 years, and 22% for male patients at an age of 65 years. For female patients in this 5-year landmark population the corresponding excess mortality rates were 4%, 10%, and 26%. (p=0.75) . Signifi cantly, in the PR cohort there was a superior 5-yr OS in the TKI treated patients. However, the eff ect of duration of disease (DOD) had a diff erent infl uence on outcome that was dependent on TKI therapy pre-SCT. Calculating the EBMT score without DOD pre-SCT, and then performing a stratifi ed analysis with DOD <1yr vs >1yr as a prognostic variable, showed that while disease duration <1yr was associated with a signifi cant increase in OS in patients who had not received a TKI pre-SCT (in every risk group), in patients who had received a TKI pre-SCT DOD >1yr was associated with a signifi cant increase in OS only in IR patients (Table) . Discussion: The time to SCT no longer bears the same signifi cance since the introduction of TKI therapy in the prognosis of CML patients undergoing SCT. For non-TKI patients, duration of disease <1yr is favourable for all EBMT groups, whilst for TKI patients, disease duration has no infl uence on good or poor risk score patients, but disease duration >1yr is associated with a signifi cant increase in survival probability in intermediate risk patients. Patients with a poor EBMT score have an improved outcome when treated with a TKI prior to SCT. Disclosure of Interest: None Declared.

Internal Medicine, 11 Hematology, AMC, Amsterdam, Netherlands

Introduction: Allogeneic stem cell transplantation (alloSCT) is the only potentially curative treatment for relapsed CLL patients with fl udarabine refractory disease and/or a deletion 17p del(17p). Retrospective analyses identifi ed bulky disease and lack of response to last treatment as predictors for poor survival.

There is an unmet need for active salvage regimens for these highrisk patients, but so far prospective data on salvage regimens prior to alloSCT are lacking. Here we report our results with R-DHAP remission-induction studied in a prospective international multicenter phase 2 study. The rationale for R-DHAP is formed by in vitro studies showing that platinum induces p53 independent cell death. Materials (or patients) and Methods: Patients up to the age of 70 years with fl udarabine refractory CLL (defi ned, according to the EBMT consensus, as relapse within 1 year after fl udarabine monotherapy or within 2 years after fl udarabine plus a monoclonal antibody) and/or with relapsed CLL with a del(17p) , and with an indication for treatment were eligible for inclusion. Patients received at least 3 cycles of R-DHAP salvage therapy (rituximab on day 1 (375 mg/m 2 1 st cycle, 500 mg/m 2 later cycles), cisplatinum 100mg/m 2 (day 1), cytarabine 2×2000 mg/m 2 (day 2) and dexamethasone 40mg days 1-4) every 4 weeks. Response to R-DHAP was determined according to 2008 IWCLL guidelines. Conditioning for alloSCT was with Flu-TBI (Seattle regimen). Results: Forty patients have been included from February 2009 till April 2012. Two patients were ineligible because of Richter's syndrome at the time of inclusion. Median age was 59 years (range 43-69) and the median number of prior therapies was 2 (range 1-5). Twenty-four patients completed at least 3 cycles of R-DHAP. Due to long donor search. two of these received 4 cycles, two 5 and 1 six cycles. Nine patients received only one or two cycles of R-DHAP because of toxicity, and one patient never received R-DHAP due to poor performance. Overall response rate (ORR) for all 38 eligible patients was 58%: 6 CR (16%), 16 PR (42%). Six patients had SD (16%) and 2 PD (5%). ORR rates in patients with bulky lymphadenopathy (>5 cm, n=17) and in those having del(17p) (n=16) was 59% and 56% respectively. In the fi rst 16 patients four infection related deaths occurred (3 septic shock and 1 encephalitis). After an amendment optimizing infection prophylaxis, no additional severe bacterial or fungal infections or infection related deaths were observed. Twenty-six patients proceeded to alloSCT (68%). With a median follow-up of 16 months (range 6-42 months) 2-years progressionfree survival (PFS) and overall survival (OS) after transplantation is 65% and 72% respectively. Sixteen are free from progression and 3 progressed but are still alive. Seven patients died (3 from acute GVHD, 2 due to encephalopathy, 2 other reasons). Discussion: R-DHAP is an eff ective remission-induction regimen for fl udarabine-refractory CLL patients even in those having bulky lymphadenopathy and/or del(17p), enabling a high percentage of patients to proceed to alloSCT resulting in a high PFS and OS at 2 years. In March 2014 the fi nal analysis of all 50 included patients will be presented. Disclosure of Interest: None Declared. . Myeloablative conditioning (CT) regimen was used in 94% of cases, and the most frequent CT used was busulfan(BU)+cyclophosphamide(CY) (76%), followed by BU+fl udarabine(FLU) (15%), FLU+treosulfan (6%) and FLU+melphalan (3%). TBI was included in 2% of the CT and ATG in 67%. Most pts received cyclosporine (CsA) associated with methotrexate (43%) or CsA alone (20%) for GVHD prophylaxis. The cumulative incidence (CI) of neutrophil engraftment at day+60 was 97% (98% for MSD, 94% for related CB, 100% for MUD and 72% for haplo transplants) with a median time of 19 days. Three percent (n=15) of pts failed to engraft and 2% (n=9) had secondary graft failure. Grade II-IV acute GVHD occurred in 16% of pts and chronic GVHD in12%. Two-year OS was 94% (94% for MSD, 98% for related CB, 91% for MUD and 78% for haplo transplants). Thirty-six pts died, and the reported causes of death were GVHD (28%), infection (28%), sinusoidal obstruction syndrome (6%), hemorrhage (8%) and other causes (30%). Discussion: The fi ndings of this survey provide an overview of HSCT outcomes for SCD and may serve to assist medical professionals to improve criteria and indications of HSCT for SCD patients. This excellent results might be further improved if patients undergo allogeneic HSCT with less comorbities and with contemporary conditioning regimen and suffi cient GVHD-and rejection prophylaxis. Also, it might serve as basis for the expansion of HSCT for SCD worldwide and to provide data for new prospective protocols. Disclosure of Interest: None Declared.

Introduction: This study aims to describe outcomes of allogeneic cord blood transplantation (CBT) in leukodystrophies. Leukodystrophies are rare inborn errors of metabolism in which the development and maintenance of brain myelin is primarily aff ected. They are characterized by rapid neurological deterioration. Hematopoietic stem cell transplantation (HSCT) can halt disease progression for selected leukodystrophies, including adrenoleukodystrophy (ALD), Krabbe disease (globoid cell leukodystrophy) and metachromatic leukodystrophy (MLD). Unrelated cord blood has the advantage of being rapidly available. Materials (or patients) and Methods: All patients undergoing CBT for leukodystrophies in EBMT centers between 1996 and 2012 were included. HSCT data were collected from the EUROCORD database. An additional questionnaire on disease-specifi c clinical outcomes was sent to participating centers. Kaplan-Meier estimates were used to calculate overall survival (OS) and cumulative incidence methods for secondary outcomes, including neutrophil and platelet engraftment, incidence of graft-versus-host disease (GvHD) and disease-specifi c characteristics at most recent follow-up. The two-sided log-rank test was used for univariate comparisons. Results: Seventy patients (31 ALD, 5 Krabbe, 34 MLD) were available for analysis, with a median age at transplant of 6.5 years (range 0-43 years). Median follow-up for survivors was 46 months. OS at 4 years was 60% (±6%). Cumulative incidences of neutrophil and platelet engraftment were 88% (±4%) at day 60, and 73% (±6%) at day 180, respectively. Acute-GvHD (grade II-IV) occurred in 16 patients (22% ±5%) at day 100 and chronic GvHD in 9 patients (13% ±5%). Out of the 24 patients who died, 18 (75%) died of transplant-related causes and 6 (25%) died of disease progression. Higher survival was seen in patients who had no or 1 HLA mismatch (OS 81%), compared to patients who received a CBT with 2 HLA mismatches (OS 47%, p 0.02).

On 37 patients (53%), information on disease-specifi c characteristics was available. OS among these patients was 68% (±8), which was not signifi cantly diff erent from patients without disease-specifi c information (p=0,307) . Nineteen patients (51%) were asymptomatic at transplant, 14 (38%) had mild disease and 4 (11%) were severely aff ected. Overall survival was worse in patients with severe disease at transplant; none of these 4 patients survived, versus 79% of mildly aff ected and 76% of asymptomatic patients. At most recent follow-up, disease status was stable in 15 (57.7%), had improved in 2 (7.7%) and worsened in 9 (34.6%) surviving patients. 

(PGD), combined with human leukocyte antigen (HLA) matching has recently emerged as a therapeutic tool for stem cell transplantation especially in the last decade in couples who already have a child aff ected with a malignant disorder or a genetic disorder. Here we report four pediatric thalassemia major patients who underwent HSCT using grafts from their siblings, who ar selected before implantation to be both unaff ected and HLA-matched donors. Materials (or patients) and Methods: Four pediatric TM patients (two female, two male) between 4-16 years old transplanted at our centre between February 2008 and July 2013 with PGD/HLA typing. They had been treated with regular RBC transfusions and chelation therapy and all of them were assigned to class II according to classifi cation proposed by Pesaro group. Stem cell source in all of the transplantations were bone marrow (BM) and umblical cord blood. Two patients (4 years-old and 11 years old females) received busulfan (Bu) and cyclophosphamide ( mining approach for data analysis. It is commonly applied in fi nancial and technological settings, where data scenarios are complex. We hypothized that given the complexity of allo-HSCT patient data, applying a data mining approach, may yield improved outcome prediction models, as compared to current risk score, which are based on a conventional statistical approach.

Our aim was to develop prediction models for 100 days post allo-HSCT overall mortality (OM) and non-relapse related mortality (NRM 

Introduction: Allogeneic stem cell transplantation (SCT) with both myeloablative (MAC) and reduced intensity conditioning (RIC) is eff ective therapyin AML and MDS. There is paucity of data on the long term outcome (beyond 10 years) following RIC due to the relative recent introduction of this approach. Materials (or patients) and Methods: We have previously reported on the role of dose intensity in a group of 112 patients (pts) with AML (n=85) and MDS (n=17) given SCT with diff erent regimens between 1999 and 2004 (Leukemia 2006 . Overall survival (OS) was similar with MAC and RIC in pts given SCT in remission, but was inferior in pts given RIC in active disease due to high relapse rates. We have now updated SCT outcomes in the same cohort with a median follow up of 10 years (8.5-12.5) in order to verify the original fi ndings with long follow up.

Results: The median age was50 years (18-70). 58 pts hadactive disease at SCT. The donor was HLA-matched sibling (n=58) or matched-unrelated (n=54). 45 pts meteligibility criteria for MAC and were given intravenous busulfan (ivBu, 12.8 mg/kg) and cyclophosphamide(BuCy). 67 pts were considered non-eligible for standardMAC and were given RIC with fl udarabineand ivBu (6.4 mg/kg, FB2, n=41) or reduced toxicity conditioning (RTC) with fl udarabine and myeloablative dosesof ivBu (12.8 mg/kg, FB4, n=26) . In all, 38 ptsare alive and 74 have died, 48 relapse, 26 non-relapse mortality (NRM). MAC and RIC/RTC had similar outcomes when leukemiawas in remission at SCT; 10-year OS been 47%, 50% and 47% after BuCy, FB4, and FB2, respectively (p=0.97). OS rates of pts with active disease at SCT was 43%, 19% and 0%, respectively (p=0.01) suggesting an advantage for more intense regimens in this setting. Relapse rates were higher after RIC/RTC than MAC throughout the follow-up period. The rate was 30% and 18%, 1 year after SCT (p=0.03), 37% and 20% after 2 years (p=0.08), 49% and 27% after 5 years (p=0.02) and 51% and 29% after 10 years (p=0.02), respectively. NRM rates were higher after MAC than RIC/RTC in the initial 2 years after SCT but approached each other in the late post SCT course. NRM rate was 22% and 9%, 1 year after SCT (p=0.05), 22 and 10% after 2 years (p=0.08), 22% and 15% after 5 years (p=0.27), and 27% and 19% after 10 years (p=0.35), respectively. Thus, OS was similar within the fi rst 2 years after SCT, 56% and 52% after MAC and RIC/RTC, respectively (p=0.86), but there was a trend for better OS after MAC later on, 51% and 36%, 5 years after SCT (p=0.26) and 44% and 31%, 10 years after SCT (p=0.22), respectively. 47 pts were alive 5 years after SCT (42%) and 9 died later on. The expected OS for the next 5 years was 86% and 87% after MAC and RIC/RTC, respectively (p=0.76). Discussion: With long-term follow-up of more than 10 years, RIC/RTC is an acceptable alternative to MAC in ineligible pts. NRM is lower after RIC/RTC in the early post SCT period, but late NRM negates this early advantage. Relapse rates are higher after RIC/RTC throughout the course. Due to these observations it seems an advantage of MAC may become apparent 5-10 years after SCT. Pts who are alive 5 years after SCT can expect similarly good further OS with both approaches. RIC/RTC studies may need to be revisited with long (> 10 years) follow up. Disclosure of Interest: None Declared. Introduction: Rates of obesity have substantially increased in recent years. Pharmacokinetics (PK) of drugs including chemotherapy is diff erent in obese pts. due to alteration in the clearance and volume of distribution. Concerns about toxicity or overdosing in obese pts., based on the use of ABW, are unfounded. Moreover, there is a paucity of information addressing the PK of high dose chemotherapy in obese pts. undergoing HSCT. Materials (or patients) and Methods: For this reason, the ALWP of the EBMT constructed an electronic survey for assessing current practice of dose adjustment of chemotherapy in pts. undergoing HSCT, in transplant centers and for planning retrospective analysis and prospective studies in the future. Results: 56 EBMT centers from 27 countries fi lled the online survey. Among the 56 centers, the percentage of obese pts. was less than 10% in 22 centers (40%), between 10 to 19% in 23 centers (42%) and more than 20% in 10 centers (17%). 45 centers declared they adjust chemotherapy dose for obese pts. (80.5%) and only 11 (19.5%) declared they do not adjust dose. Among centers which adjust dose, most uses BMI as the parameter for defi ning obesity (28 centers, 62%), others use percentage over the actual body weight (ABW) as the basis for defi ning obesity (11 centers, 24 .5%), both BMI and ABW (3 centers, 6.7%) or other parameter (3 centers, 6.7%). Most of the centers that use BMI for adjusting dose defi ne BMI >30 kg/m 2 as the cut-off value (formal defi nition for obesity), only one center uses morbid obesity (BMI > 40 kg/m 2 ), and the remainder uses other cut-off values. Among 11 centers who use ABW, 9 use ABW more than 120% of ideal body weight for adjustment. 84% of the centers use one level of obesity for adjustment while the rest uses 2 levels. The method for determining the weight for chemotherapy calculation was actual body weight (ABW) in 16 centers, ideal body weight (IBW) in 10 centers, IBW + 25% of diff erence between IBW and ABW (IBW+0.25*(ABW-IBW)) in 16 centers and other methods in the rest. Among centers that use dose adjustment, 44% also cap the dose at 2 m 2 for chemotherapy dose based on BSA while 56% do not cap. On the contrary, most of the centers (9/11) that do not adjust dose for weight also do not cap the BSA at 2 m 2 . 79 percents of responding centers use the same approach to dose adjustments for myeloablative, reduced intensity (RIC) or non myeloablative (NMA) conditioning, while 21% reduce the dose less for RIC or NMA conditioning. For Busulfan dose only 7 centers monitor pharmacokinetics (pk). Eleven centers use ideal body weight for calculation, 17 centers use actual weight and 18 centers correct weight according to percentage over actual body weight. Discussion: This EBMT survey reveal large diversity among transplant centers regarding dose adjustment practice for high dose conditioning chemotherapy. Most of the EBMT centers use dose adjustment for obese pts. and about half of them also cap BSA at 2 m 2 , while capping is uncommon in the centers that do not adjust dose. Thus, the range of the fi nal dose is very wide. Even for Busulfan where dose is calculated normally according to ideal body weight, the diversity of dose given for obese pts. is wide. Our next step is to analyze outcomes of transplantation according to dose adjustment practice and subsequently to formulate a methodology for future prospective studies. Disclosure of Interest: None Declared. Introduction: The risk of hemolytic PNH (hPNH) remains one of signifi cant unfavorable outcomes in AA. Here we report analysis of hPNH incidence and risk factors in a large cohort of AA patients treated with immunosuppressive therapy (IST). Materials (or patients) and Methods: We performed retrospective and prospective analysis of clinical and laboratory signs of intravascular hemolysis with serial measurements of PNH clone size by highly-sensitive fl ow cytometry in AA patients initially treated with IST. Cumulative incidence of hPNH was calculated using early death and BMT as the competing risks. Univariate and multivariate Cox regression analyses were used to assess independent predictors of hPNH. Results: After exclusion of 10 cases with initial AA/hPNH, a total number of 238 patients (129 M and 109 F, median age 20, 1-66) with moderate (83), severe (96) and very severe (59) AA after IST with ATG and CsA (199) or CsA alone (39) were included in the study. Using receiver operating characteristic analysis of 482 simultaneous measurements of the PNH clone size and LDH level in 110 PNH+ patients, we identifi ed a 10% threshold of granulocyte clone size as a predictor for LDH level >1.5 upper limit of normal (ULN) (sensitivity 94%, specifi city 91%). Thus, the PNH granulocyte clone size >10% and stable LDH level >1.5 were used as criteria of hPNH in our cohort. After a median follow-up of 55 months (range, 4-202), 22 (9.2%) patients developed hPNH with the cumulative incidence of 10.1% (95% CI, 6.3-16.2) and 18.3% (95% CI, 11.9-28.2) at 5 and 10 years respectively (Fig.1) . According to the PNH status at AA diagnosis, we estimated the cumulative incidence of hPNH in three patient groups: PNH-(87); PNH+ (120); Unknown (31) (Fig.2) . The cumulative incidences of hPNH at 10 years were 0%, 29.2% (95% CI, 18.9-44.9 %) and 20.3% (95% CI, 7.9-51.6%) in the fi rst, second and third groups, respectively (P=0.00014 independently predicted the higher risk of hPNH. Age ≥ 18 was identifi ed as the risk factor in univariate analysis. None of the other potential baseline characteristic (gender, AA severity, baseline blood counts, HLA DRB1*15 positivity, IST protocol) did not infl uence the cumulative incidence of hPNH. Discussion: The risk of developing hPNH after IST is closely related to the baseline PNH clone presence. Progression of subclinical PNH into hemolytic form is most likely in patients with both the initial clone size ≥1% and LDH level ≥0.94 ULN. Thus, these early signs can be considered to assess the risk of hPNH for a more focused monitoring and treatment planning, including allogeneic BMT as a curative approach. Disclosure of Interest: None Declared. Discussion: This extended report from the EBMT ADWP database confi rms that activity in HSCT for severe ADs is sustained despite the introduction of biological therapies. In some diseases such as MS, SSc and IBD where long-term responses are confi rmed, activity has been increasing, whereas in others, particularly IA, activity has fallen due to availability of biological therapies and less sustained responses to HSCT. Activity has been highly country specifi c and the infl uencing factors need further analysis. The activity of allogeneic HSCT remains relatively low and largely restricted to paediatric practice. The impact of the recent 2011 ADWP guidelines on activity will be analysed when data collection for 2013 is complete. Disclosure of Interest: None Declared. S44 eff ector cells, followed by transplantation of hematopoietic progenitor cells (Alexander Blood, 2009) . Recently, gene expression profi ling in peripheral blood revealed an up-regulation of type I interferon (IFN) related genes (as IFN signature) in systemic lupus erythematosus (Bennett J Exp Med 2003) and their fi rst-degree relatives (Niewold Genes Immun, 2007) . Here, we aimed to analyze the IFN signature in SLE patients after receiving HSCT to evaluate whether IFN signature is normalized after "immune reset" or is permanently present, resembling a heritable trait. Materials (or patients) and Methods: Since 1998, ten patients with treatment-refractory SLE received a CD34 + -selected autologous stem cell transplantation after immunoablation with antithymocyte-globulin (ATG-Fresenius, 30mg/kg) and cyclophosphamide (200 mg/kg) as part of a monocentric phase I/II clinical trial. Siglec-1 expression on CD14 + monocytes (as IFN surrogate) was analyzed longitudinally using multicolor fl ow cytometry. In addition, gene expression profi le was analyzed with Microarray (Aff ymetrix) from purifi ed CD14 + monocytes (FACSAria cell sorter) and compared to active SLE and healthy donors. Results: Clinical remission (SLEDAI ≤3) was achieved in all patients despite immunosuppressive drug withdrawal, associated with disappearance of anti-dsDNA antibodies. Two patients died from transplant-related infections; from the remaining eight patients, fi ve are in long-term clinical remission for up to 15 years after HSCT, while three patients suff ered a relapse of SLE. Siglec-1 expression on monocytes, which is increased in active SLE, completely normalized in responding patients after HSCT and was only elevated during viral infections. Based on gene expression profi ling, the number of IFN signature probe-sets with a cut-off of fold change ≥2 or ≤2 was decreased down to 24% (77/320) in patients after HSCT compared to active SLE and completely clustered with healthy controls suggesting a normalization of type I IFN signature. Discussion: The normalization of type I IFN signature observed in responding patients after HSCT suggests that up-regulation of IFN-related genes is not a predisposition but rather the consequence of chronic autoimmunity in SLE. Apparently, immunoablation resets the innate immune system into a self-tolerant state where even up-regulation of IFN during viral infections is not associated with reactivation of SLE. Disclosure of Interest: None Declared. Introduction: Treosulfan is a bi-functional alkylating agent which causes less veno-occlusive disease than busulfan and does not require PK monitoring or anticonvulsant prophylaxis. We previously published results of 70 children who received treosulfan in combination with either cyclophosphamide or fl udarabine prior to Haematopoietic Stem Cell Transplantation (HSCT) for Primary Immunodefi ciency (PID) or Severe Immune Dysregulation (1). We showed that toxicity was lower and T cell chimerism was better in those that received fl udarabine. Materials (or patients) and Methods: The study has now been expanded and focused to include 160 children receiving Treosulfan (42 g/m 2 , 36 g/m 2 or 30 g/m 2 ), with Fludarabine 150 mg/m 2 plus Alemtuzumab (in 124) between 2006 and July 2013 at the Great North Children's Hospital, Newcastle upon Tyne (89) or Great Ormond Street Hospital, London (71) (40 are in the previously published series). Results: Median age at transplant was 12 months (range1.2-219); 49% patients were 12 months or younger. Donors were: matched unrelated (72), 1-2 antigen mismatched unrelated (56), matched sibling (11), other matched family (17), haploidentical (4). Stem cell source was: PBSC (71), BM (48), CB (41). Median follow up was 30 months (range 3-92). Nine deaths occurred before D+100 (94% survival). Overall survival was 84%. Of 25 deaths, 9 patients had HLH, many of whom were not in remission at the time of HSCT. There was no VOD and no toxicity related deaths. 71 (44%) had GVHD, only 12 (7.5%) > grade II. Four patients rejected their graft, 2 were successfully retransplanted, (2 died of infection),a further 2 were successfully retransplanted following poor immune reconstitution; 68 of 102 (67%) more than 1 year post transplantation had 100% donor chimerism in all cell lineages. Increased use of PBSC favoured improved donor myeloid chimerism without increasing the incidence of signifi cant GVHD. High levels of donor chimerism are of particular importance in diseases such as Wiskott Aldrich Syndrome and Chronic Granulomatous Disease. Eight infants with severe combined immunodefi ciency diagnosed at birth were transplanted at ≤3months age and are all alive and well with up to 6 years follow up. Discussion: The combination of Treosulfan, Fludarabine and Alemtuzumab is an ideal choice of conditioning for HSCT in PID, associated with good lymphoid and myeloid engraftment and low regimen related toxicity. Long-term follow up is required to determine the gonadotoxic eff ects of this approach in comparison to Busulfan-containing regimens. 

Introduction: Outcome of HSCT in SCN is not known because it is based on case reports and small series.EBMT data base off ers an unique opportunity to address some unanswered questions in a large cohort of subjects. Materials (or patients) and Methods: All patients registered in the EBMT data base aff ected with SCN were considered eligible to the study. Data regarding HSCT and outcome were derived from the "MED B EBMT form", while additional data regarding history before HSCT, were collected through a specifi c CRF (Med C) sent to all the participating centers. Results: A total of 119 patients from 19 countries entered the study; sixtysix percent of patients originated from Western Europe and 34% from Eastern Europe/Eastern Mediterranean area. Females were 51% of the cohort. Median age at diagnosis of neutropenia was 0.35 years (0-35.4yrs), while median age at fi rst transplant was 4.8 years (range 0.2-43y). Four patients were aff ected with MDS at time of transplant. The cell source was bone marrow (BM) in 56%, peripheral blood (PB) in 26% and cord blood (CB) 18%. Fiftythree percent of patients were engrafted from a matched related and 47% from an matched unrelated donor. Conditioning regimen was myeloablative in 86% and at reduced intensity (RIC) in 14% of the cohort. Engraftment was documented in 91% of subjects: 5% had primary graft failure and 4% lost the engraftment. Overall 22 patients (18.5%) died while 97 were alive (81.5%). Causes of death were: GVDH in 23% of patients, Infections in 23%, organ failure in 18% and combination of previous causes in 27%. relapse/progression of the disease in 9%. Transplant related mortality occurred in 17% of thecohort. Acute GVDH grade 1-2 was documented in 31%, grade 3-4 in 14%.

The 5-year OS and EFS (death,relapse,primary and secondary graft failure being the events) were 77% and 70 % respectively .The 5year OS according to donor type was 84% and 79% (p=0.9) for HLA identical family vs unrelated donor. Also the 5-year OS according to source of cells did not signifi cantly diff erent (CB 95%, PB 79% and BM 62%> p=0.13). The same held true also for 5-year EFS (CB 85% BM 75% and PB 52%> p=0.077).

Five year OS was signifi cantly higher in patients transplanted before age 10 years (83% in age 0-5yrs, 83 % in age 5-10 yrs and 60% in age above 10 yrs-p=0.046), whereas no signifi cant diff erences were seen between HSCT performed before or after year 2000 (64% before year 2000 and 80 % after year 2000 -p=0.159). No diff erence was also seen in OS and EFS according to myeloablative conditioning regimen and RIC. Discussion: This is the largest study ever conducted on a population of patients aff ected with SCN. It shows that the 5-year survival in transplanted SCN patients is close to 80% with no diff erence between matched related and unrelated donor. TRM, close to 17%, is still not negligible. Survival improved after year 2000 probably because of better donor typing techniques. Survival was signifi cantly worse in patients transplanted at age>10 years and a less favourable trend was observed in patients transplanted with peripheral blood vs cord blood and marrow cells.

Overall this study provides the fi rst solid evidence that HSCT in SCN, either from family or unrelated HLA matched donor, is a reasonably good rescue option in patients poorly responding to G-CSF or who transformed into MDS/AL. Disclosure of Interest: None Declared. Introduction: Haploidentical family donors represent a solution to off er to every patient with high-risk hematological disease lacking a HLA-matched donor the potential cure of hematopoietic stem cell transplantation (SCT). Historically extensive application of haploidentical SCT (haplo-SCT) was limited by high rate of late transplant mortality (TRM) and relapse incidence (RI) associated with the delayed immune-reconstitution secondary to the procedures for severe graft-versus-host-disease (GvHD) prevention. Today several strategies of haplo-SCT without T-cell depletion (TCD) are available, but follow-up is still short to evaluate long term outcomes and in particular the impact of chronic GvHD in such a mismatched setting. Here we are reporting the evaluation of a 7 years long term survival in our haplo-SCT TCD experience. Materials (or patients) and Methods: We included for analysis 56 adult patients (pts) who underwent an haplo-SCT for hematologic malignancies, between 2004 and 2012. Data were collected from the local database. All consecutive pts receiving graft after selection of peripheral CD34+ cells-CliniMacs one-step procedure -were selected. No post-transplant immune-suppression was introduced after transplantation as GvHD prophylaxis. In vivo T-cell depletion with ATG (Fresenius) was administered in all pts. In-vivo B-cell depletion with Rituximab 500 mg was administered in selected cases. Pts median age was 52 (range, 17-66) years. The cohort included 45/56 pts with a diagnosis of acute leukemia, 5 Hodgkin lymphoma, 4 non-Hodgkin lymphoma, 2 myelodysplastic syndromes.

Thirty-six/56 pts were in complete remission (CR) at time of transplant. All pts were conditioned with a myeloablative regimen. Donor lymphocytes genetically engineered to express the suicide gene herpes simplex thymidine kinase (TK-DLI) to induce early immune-reconstitution, while selectively controlling GvHD, were infused in 27/56 pts (Ciceri, Bonini et al, Lancet Oncol 2009) . Results: The overall survival (OS) for the entire population at 2-y was 36% and 33% at 5y; the TRM at 100-days was 21%, 41% at 2y (last event recorded at 1y post transplant); the 2y RI was 25% (last event at 19 months). The median follow-up for survivors was 80 months. According to disease risk stratifi cation (DRI) (Armand et al, Blood 2012) the 2y and 5y OS for patient scored as low-intermediate risk was 51% -51%, for patient scored as high and very high 16% and 12% respectively (p 0,0019, HR 3). Day 100 TRM for pts in the intermediate-low DRI category was 16% (29% at 2y) and RI at 2y 19%. All patients long-term disease-free survivors are free of chronic GvHD and immune suppressive treatment (IST) with the exception of two pts who experienced cGvHD after boost of T-repleted donor stem cell for poor graft function. Discussion: These results indicate that TCD haplo-SCT -with exploitation of a strategy aimed to enhance immune reconstitution post transplant -is safe and able to provide long term IST-free survival with low RI and acceptable TRM particularly in pts in CR at transplant. An international randomized Phase III trial based on infusion of TK-DLI, for high-risk leukemia patients undergoing haplo-SCT is currently recruiting in Europe and the United States (TK008 trial, NCT00914628). Introduction: ASCT in MM especially with the use of hematopoietic progenitors (HP) of peripheral blood is a procedure with an acceptable morbidity and low mortality. Both are related with toxicity, bacterial infections and diff erent complications related to neutrophil recovery as the engraftment syndrome (ES). In this sense, the administration of G-CSF is a standard practice to accelerate the leukocyte implant and reduce the risk of infection. However, there are reports that associate G-CSF and complications such as ES. For this reason, the clinical utility of the G-CSF in patients with MM undergo ASCT could be controversial. We compare the clinical outcomes of two consecutive cohorts of patients with MM in who the ASCT was performed with G-CSF (group A) or without G-CSF (group B) during the aplasia of the procedure. Materials (or patients) and methods: 52 consecutive patients (31 males, median age 59 years) with MM treated with ASCT were included between April 2010 and March 2013 on a single institution. The conditioning regimen was melphalan 200 mg/m 2 and the source of HP was peripheral blood in all cases. In 25 patients (48%) ASCT was performed in a home care program. The group A (n=27) received G-CSF 5 mcg/kg/day from day +7 until neutrophil recovery (absolute neutrophil count >1.0×10 9 /l on two consecutive days). In the group B (n=25) the procedure was performed without G-CSF. Results: Baseline characteristics of the patients (age, sex, disease status at transplant, previous treatment and cellularity CD34+ infused) were similar in both groups. In overall series, median days of hospitalization were 20 days (range, [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] and median days with neutrophil count <0.5 x10 9 /l was 8 days (range, 5-16) . Twenty-nine patients (56%) had neutropenic fever and 20 of them (39%) had a microbiologic isolation. The incidence of mucositis and intestinal toxicity grades II-IV was 21% and 25%, respectively and 20 patients (39%) had ES. None of the patients died during the ASCT. The analysis of the groups (G-CSF vs. no-G-CSF) is summarized in table 1. There was a signifi cant increase in the incidence of neutropenic fever and ES in the group of patients receiving G-CSF (p=0.0001 and p=0.001, respectively). Discussion: The non-use of G-CSF during the aplasia period in patients with MM undergoing ASCT was safe and it has been associated with a signifi cant reduction of neutropenic fever and ES without signifi cant impact on the duration of hospitalization or immediate mortality associated with the procedure. Disclosure of Interest: None Declared. Introduction: Application of treosulfan (TREO) to a myeloablative conditioning prior to allogeneic HSCT is currently evaluated in randomized phase III clinical trials with reference to busulfan. TREO is a pro-drug that at pH above 5 undergoes a non-enzymatic sequential activation to intermediate monoepoxide (S,S-EBDM), and then diepoxide (S,S-DEB). The aim of the study was to investigate penetration of TREO and its biologically active epoxides into brain and cerebrospinal fl uid. Materials (or patients) and Methods: The studies were carried out in 48 juvenile (10 days old, JR) and 48 young adult (34-35 days old, YAR) male and female CD® rats applying a serial sampling design upon approval of the ethical committee. The animals received an intravenous bolus of TREO at the dose of 500 mg/kg b.w. Blood was withdrawn via a heart puncture in the JR and from retrobulbar venous plexus in the YAR under isofl urane anesthesia. Following the blood sampling, the animals were sacrifi ced under ether anesthesia for withdrawal of cerebrospinal fl uid from Cisterna magna (in the YAR only) and collection of the brain (in all the rats). The samples were immediately acidifi ed by citric acid in order to avoid the artifi cial conversion of TREO. Concentrations of TREO and S,S-EBDM in the rat plasma, cerebrospinal fl uid and brain were determined by the validated HPLC-MS/MS method, whereas S,S-DEB was quantifi ed by the validated HPLC-UV method. Pharmacokinetic parameters, such as maximal concentration (C max ), area under the curve (AUC) and biological half-life (t 0.5 ) were calculated in WinNonlin 6.2 (Pharsight) using a non-compartmental analysis. Results: Both TREO and S,S-EBDM were found in the rat blood, brain and cerebrospinal fl uid, whereas concentrations of S,S-DEB in all the analyzed samples were below the limit of quantifi cation. Mean ratios brain/plasma C max and AUC of TREO in the JR (0.04 and 0.13) were higher than in the YAR (0.02 and 0.08). The JR demonstrated also a higher brain exposure to S,S-EBDM as the mean ratios brain/plasma C max and AUC of the compound in this group were 0.40 and 0.50, respectively, in comparison to 0.22 and 0.22 observed in the older animals group. Elimination of the pro-drug and its active monoepoxide from plasma as well as brain proceeded slower in the JR, compared to the YAR, as proved by the higher values of their t 0.5 . In the YAR cerebrospinal fl uid both TREO and S,S-EBDM reached higher levels than in the brain: mean ratios cerebrospinal fl uid/plasma C max and AUC were 0.04 and 0.12 for TREO, and 0.52 and 0.65 for S,S-EBDM, respectively. Applying the descriptive statistics no sex-related diff erences in the pharmacokinetic parameters of TREO and S,S-EBDM were found. Discussion: In the rat model TREO and S,S-EBDM showed rather limited sex-independent penetration into brain tissue and cerebrospinal fl uid. Somewhat better capability of S,S-EBDM to cross a blood-brain barrier, when compared to TREO, may result from its higher lipophilicity. Our results confi rm that penetration of TREO into brain is signifi cantly lower in comparison to busulfan for which brain/plasma C max ratio 0.74 was noted. This fact may account for the low neurotoxicity of TREO reported in the clinical trials. However, our studies also revealed that in very young children an exposure of the central nervous system to TREO and S,S-EBDM might be higher than in adults. Disclosure of Interest: F. Główka Confl ict with: The studies were sponsored by medac GmbH (Wedel, Germany), M. Romański Confl ict with: The studies were sponsored by medac GmbH (Wedel, Germany), J. Baumgart Confl ict with: Dr. Baumgart is employed in medac GmbH (Wedel, Germany), S. Böhm Confl ict with: Dr. Böhm is employed in medac GmbH (Wedel, Germany). Introduction: Two apheresis devices (Spectra Optia and COBE Spectra) have been recently compared on mononuclear cell (MNC) collection for extracorporeal photochemotherapy (ECP). 1 However, the signifi cantly lower total blood volume processed with the Spectra Optia device in that study may determine the observed diff erences of the signifi cantly lower platelet, white blood cell (WBC), granulocyte, and MNC absolute counts, as well as red blood cell (RBC) contamination in the Spectra Optia collections. Materials (or patients) and Methods: To overcome the bias of different processing volumes potentially aff ecting results, we compared the effi ciency and collection parameters of the two MNC systems using a cross-over design. We prospectively evaluated No G-CSF (n=25) p value Days with neutrophil <0,5x10 9 /l, median (range) 7 (5-11) 10 (5-16) 0.02 Fever, n (%) 23 (85%) 6 (24%) 0.0001 Microbiologic isolation, n (%) 11 

200 consecutive ECP apheresis procedures that had been performed by invariably processing two times the patient´s blood volume. Between March 2012 and September 2013, 14 patients underwent ECP due to diff erent diseases. Each patient was consecutively assigned to the semiautomated COBE Spectra (Terumo BCT, Lakewood, CO, USA; version 7.0) and the automated Spectra Optia v9 (Terumo BCT). Results: According to the equivalent groups design, we did not detect signifi cant diff erences in the variables analyzed before and after the apheresis procedure in both groups (WBC, MNC, RBC, and platelet counts), thus assuming any potential diff erence in product data to be due to the apheresis system being used. Collection effi ciencies (CE) with Spectra Optia (n=100) S48 as well as platelet loss, collection effi cacy (CE1=[CD34+ cells collected]/[mean of CD34+ cells processed]), throughput (CR/time), and collection rate (CR, CD34+/kgBW in product/ CD34+ pre-apheresis) did not diff er between both protocols. There was a slight tendency towards higher CE1, CR and throughput on day 4 collections. In 3 out of 22 patients in the day 5 group and 3 out of 13 in the day 4 group a second apheresis procedure was required (p=0.6). Discussion: PBPC apheresis on day 4 seems as feasible and eff ective as collection on day 5. Collection on day 4 produced lower white blood cell content in the product and allows a reduction in G-CSF exposure to healthy donors. Disclosure of Interest: None Declared. We divided patients according to baseline cytokines level below and above median into 'low' and 'high' expressors. Only VEGF baseline level infl uenced mobilization effi cacy. The group of VEGF 'low'expressors had longer median time of G-CSF treatment before fi rst apheresis than 'high' expressors (12 vs 10 days, p=0.013) . Additionally we observed that maximal number of CD34+ in peripheral blood occurred later in 'low' expressors (Me=13 days) than in 'high' expressors (Me=10.5 days, p=0, 01) . Baseline VEGF level correlated adversely with duration of G-CSF treatment before fi rst apheresis (R=-0.5, P=0.002) and with the day when maximal number of CD34+cells were present in the peripheral blood (R=-0.53, p=0.001).

Patients were also divided according to median cytokines levels at apheresis into 'low' and 'high' expressors. Introduction: Suicide gene therapy applied to allogeneic hematopoietic stem cell transplantation (allo-HSCT) is one of the widest clinical applications of gene therapy. By the infusion of donor lymphocytes transduced to express the Herpes Simplex Virus Thymidine Kinase (TK) suicide gene, patients achieve a rapid immune reconstitution and substantial protection against tumor recurrence. TK-cells are promptly eliminated in case of graft versus host disease (GvHD), with complete resolution of the adverse reaction. In the present work we characterize the immunological profi le of a cohort of long-term survivors after suicide gene therapy and we studied the long-term fate of TK-cells to shed light on memory T cell dynamics after transplantation. Materials (or patients) and methods: We studied 14 adult patients who underwent allo-HSCT (haploidentical HSCT: n=11; HLAidentical HSCT n=3) and infusion of purifi ed suicide-gene modifi ed donor T cells (median dose: 1.9×10 7 cells/kg) for high-risk hematologic malignancies between 1995 and 2010. Results: At a median follow-up of 9 years (range 3-17), all patients are in complete remission. Five out of 14 patients experienced GvHD in the early phase post immune reconstitution; in all cases, ganciclovir administration proved eff ective in abrogating the adverse reaction. No symptoms or complications related to GvHD were observed during the long-term follow up, and none of the patient is receiving immunosuppressive drugs. We observed a complete recovery of the T-cell compartment, comprising of physiological levels of circulating naïve and memory CD4 and CD8 cells. TK-cells were detected in the majority of analyzed patients (90%), at low levels (median=0.43%±6.9%). Ex vivo selection of pure TK-cells after polyclonal stimulation and NGFR-purifi cation confi rmed the presence of functional transduced cells, thus directly demonstrating the ability of memory T cells to persist for years. The proportion of TK-cells detectable at the longest follow-up did not correlate with the number of infused cells, nor patients or donors' age, but instead with the peak of TK-cells observed within the fi rst months after infusion, suggesting that antigen recognition is dominant in driving in vivo expansion and persistence of memory T cells. Of notice TK-cells could be retrieved also in patients successfully treated with ganciclovir for GvHD, thus confi rming the selective action of ganciclovir only on proliferating TK-cells. Accordingly, ganciclovir sensitivity was preserved in long-term persisting TK-cells, independently from their diff erentiation phenotype. While infused TK-cells displayed a predominant eff ector memory phenotype, gene modifi ed T cells persisting long-term were enriched for central memory (CD45RA-CD62L+) and stem memory (CD45RA+CD62L+CD95+) phenotypes, suggesting the higher ability of these T cell subsets to persist and shape the immunological profi le long-term in treated patients. Discussion: These data show that a complete donor-derived immune system is restored in adult surviving long-term after suicide gene therapy. After infusion, gene modifi ed cells persist for up to 14 years in treated patients. Further studies on TK-cell TCR repertoire and vector integrations are currently being performed to elucidate the in vivo dynamics of infused memory T cells. Disclosure of Interest: None Declared. Introduction: Allogeneic stem cell transplantation (SCT) can be a curative treatment for hematological malignancies. The therapeutic eff ectiveness is attributed to the graft-versus-tumor (GVT) eff ect, mediated by alloreactive T cells and natural killer (NK) cells. Although T cells can induce a potent GVT eff ect, they can also induce graft-versus-host disease (GVHD), causing high morbidity and mortality. Interestingly, after non-myeloablative allogeneic SCT, early NK cell repopulation has been associated with decreased relapse rates, without increasing GVHD incidence, illustrating a possible role for donor NK cell adoptive transfer after allogeneic SCT. However, isolation of suffi cient numbers of activated NK cells from donor origin is challenging. Therefore, we established a cytokine-based culture system with the capability of generating GMP-compliant NK cell products from hematopoietic progenitor cells (HPC) with high cell numbers, purity and functionality. Here, we investigated whether the aryl hydrocarbon receptor (AHR) antagonist StemReginin1 (SR1) can boost the ex vivo generation of NK cells from bone marrow (BM) or mobilized peripheral blood (PB) derived stem cells. Materials (or patients) and Methods: CD34+ cells were isolated using immunomagnetic beads from BM samples or G-CSF mobilized apheresis material. These CD34+ cells were expanded during 14 days in medium containing SCF, Flt3L, TPO, IL-7 and IL-15 and subsequently diff erentiated into NK cells using IL-15 and IL-2. HPC-NK cell expansions were performed with or without SR1, and HPC-NK cells were assessed for their cytolytic functions against AML and MM cell lines. In addition, expression levels of typical NK-activating receptors and diff erentiation markers were analyzed by fl ow cytometry. Results: Interestingly, initial cultures showed that addition of SR1 during the whole culture process improved the expansion and diff erentiation and functionality of the NK cells generated.

In the presence of SR1 we were able to expand BM-and PBderived CD34+ cells up to 800 fold in 6 weeks. The SR1-generated HPC-NK cell products contained 80 -90% NK cells which express high levels of activating NKG2D and natural cytotoxicity receptors. Furthermore functional analysis showed effi cient lysis of AML and multiple myeloma (MM) cell lines at very low NK-target ratios. Currently we are investigating the in vivo potency of these SR1-generated HPC-NK cells in our intra-femoral tumor models in NOD/SCID/IL2Rg null mice. Discussion: Addition of the AHR antagonist SR1 in our culture system facilitates the generation of high numbers of functional NK cells from BM or G-CSF mobilized PB CD34+ cells, which hold great promise for future donor NK cell-mediated therapy after allogeneic SCT. Disclosure of Interest: None Declared. Introduction: Advanced Ewing sarcomas (ES) are associated with poor prognosis. Despite multimodal and highly toxic therapeutic approaches including surgery, irradiation, (high-dose-) chemotherapy (with or without allogeneic stem cell transplantation; allo-SCT), overall survival is still unsatisfactory. Materials (or patients) and Methods: We generated clonal ES specifi c allo-restricted T cells for adoptive transfer to enhance effi cacy and to decrease toxicity of donor lymphocyte infusions (DLI) after allo-SCT. However, inclusion of these cells into current therapy protocols is limited due to high production complexity and low cell numbers. In order to overcome these obstacles and to facilitate good manufacturing practice (GMP) accredited off -the-shelf products in the future, we generated HLA-A*0201-restricted T cell receptor (TCR) transgenic T cells directed against ES associated antigens EZH2 666 and CHM1 319 by retroviral transduction.

Results: IL21±IL7±IL15±IL2 based expansion protocols revealed suffi cient TCR transgenic T cell numbers expressing a CD62L + / CD45RO + central memory phenotype up to 80%. These cells maintain specifi c recognition and killing of HLA-A*0201 + ES cell lines expressing the antigen. In addition, adoptively transferred TCR transgenic T cells caused inhibition of local tumor growth and metastasis in Rag2 −/− γ C −/− mice. Discussion: The infusion of TCR transgenic donor lymphocytes may serve to render DLI after allo-SCT more effi cacious and less toxic. Disclosure of Interest: None Declared. Introduction: Patients with primary induction failure and chemoresistant relapse are poor candidates for hematopoetic stem cell transplantation (HSCT). Additional attempts to induce remission with various combinations of chemotherapy alone will unlikely improve the outcome and will contribute to excess toxicity. Immunotherapy with natural killer cells (NK) and related innate lymphocytes before HSCT has the potential to induce hematologic response and improve outcome of HSCT. Use of lymphodepleting chemotherapy and agents able to sensitize leukemia cells to immune attack might further improve the effi cacy of NK therapy. Materials (or patients) and Methods: In 2013 Three patients with acute myeloid leukemia (NRAS, FLT3, t (10;11)) with primary induction failure (n=1), refractory relapse (n=1) and refractory relapse after HSСT (n=1) (all male, median age 10, 9 years (1,8-13,8) ) have been included in the study. Preconditioning consisted of fl udarabine 30 mg/m 2 day 1-4, cyclophosphamide 30 mg/m 2 day 1-4, decitabine 20 mg/m 2 day 1-5, bortezomib 1.3 mg/m 2 day 2,5, valproic acid 25mg/kg day 1-6, infusion of CD56+ cells from haploidentical donor (Clini-MACS single step CD56 selection) on day 6 and interleukin-2 (IL-2) on days 6 to 16. The median dose of infused CD56+ cells was 23,6×10 6 /kg (21.6-42.5). All patients received HSCT from the same haploidentical donor after conditioning with ATG, Fludarabin, Treosulfan, Melphalan (n=2) or ATG, Clofarabine, Melphalan, Thiotepa, Vp16 (n=1). TCR alpha/beta and CD19+ depletion was used for graft processing. The median dose of CD34+ cell was 8.1×10 6 /kg, TCR alphaeta -22×10 3 /kg. Tacrolimus + short MTX were used as GVHD prophylaxis in 2 patients, tacro monotherapy in 1 pt. Results: All patients developed bone marrow aplasia. In two patients blasts were cleared from the bone marrow by the day of conditioning start. None of the 3 patients treated developed signifi cant acute organ toxicity beyond IL-2-associated fever. Primary engraftment occurred in 3/3 pts, the median time to reach neutrophil and platelet recovery was 20 days and 16 days, respectively. All patients remained in full donor chimerism on days +30, 60, 90, 120, 180, 210. All patients received additional post-transplant NK-DLI. None of the patients developed acute GvHD after preconditioning and transplantation. Two patients developed moderate GVHD after DLI. All patients are alive, in complete remission and off immunosuppression with a median follow up of 9 (8-10) months. Discussion: Adoptive transfer of allogeneic haploidentical innate lymphocytes before HSCT was safe and reduced tumor burden in refractory leukemia. It could be safely combined with intensive conditioning and HSCT. We plan to evaluate this strategy in a pilot phase II trial. . While the complete editing (CE) procedure currently requires multiple manipulation steps, 'single TCR editing' (SE), based on the disruption of a single endogenous TCR chain, followed by gene transfer of the tumor specifi c TCR, enables the generation of redirected T cells devoid of their natural TCR repertoire during a single round of T cell activation, improving the feasibility of its clinical translation. This might be particularly useful to reduce the risk of Graft versus Host Disease (GvHD) after allogeneic hematopoietic stem cell transplantation, since the expression of the endogenous TCR repertoire is completely abrogated. Materials (or patients) and methods: We exploited a HLA-A2 restricted TCR specifi c for NY-ESO-1, a cancer testis antigen expressed by solid tumors and hematological malignancies, to directly compare the safety and effi cacy profi le of unedited TCR transferred T cells (TR), SE and CE T cells. Genetically modifi ed T cells were tested in vitro for their surface phenotype and killing ability as well as in vivo in NSG mice engrafted with the myeloma cell line U266 (HLA-A2+ and NY-ESO-1+). Results: Gene editing does not detectably aff ect the phenotype, function and proliferative potential of engineered lymphocytes. Our protocols ensured the maintenance of an early diff erentiated memory phenotype, with enrichment in central memory and CD45RA+/CD62L+/CD95+ memory stem T cells. Upon lentiviral transfer of the NY-ESO-1-specifi c TCR, we observed signifi cantly higher levels of the tumor-specifi c TCR expression, evaluated as NY-ESO-1 specifi c dextramer binding, in edited versus transferred T cells (relative fl uorescence intensity to untransduced cells: CE: 37; SE: 31; TR: 19; p<0.01). Edited T cells were more effi cient than unedited-TCR transferred T cells in killing NY-ESO-1-pulsed cell lines (half maximal eff ective peptide concentration in a 51 Cr release assay: 310, 210, 186 nM for TR, SE and CE T cells respectively) and NY-ESO-1 + myeloma cell lines naturally processing the antigen. Importantly our SE and CE T cells displayed no activity against NY-ESO-1targets. In NSG mice, NY-ESO-1 redirected SE and CE T cells completely eliminated an NY-ESO1 + HLA-A2 + , WT1myeloma cell line, that, on the contrary, expanded in bone marrow in the presence of WT1-redirected CE T cells. Most importantly, mice infused with CE or SE T cells did not experience xenogeneic GvHD. Discussion: Our results demonstrate that the TCR single editing approach is eff ective in redirecting T cell specifi city as evidenced by the potent anti-tumor eff ect both in vitro and in vivo while potentially eliminating the risk of GvHD associated with the infusion of donor-derived lymphocytes. Moreover, the relative speed and simplicity of the TCR single editing protocol should facilitate its clinical application to patients with hematological malignancies. Introduction: CD33/34 lineage specifi c chimerism analysis (CD33/34 LCA) can improve the predictive value of chimerism monitoring after allogeneic SCT in AML. Recently, real time PCR (qPCR) has been proposed for highly sensitive and accurate quantitation of chimerism. However, data to compare clinical impact of qPCR versus gold standard conventional STR-PCR was not available. In 2011, we reported on the impact of STR chimerism monitoring in a pediatric AML multicenter study (Rettinger, Willasch et al., Blood 2011; 118(20) :5681-88). Materials (or patients) and Methods: We present a prospective analysis on CD33/34 LCA in bone marrow (BM) and mononuclear cell (MNC) chimerism in peripheral blood (PB) by qPCR in the pediatric AML cohort mentioned above. Pre-existing STR-PCR data were used for comparison. This analysis for the fi rst time refers to the clinical impact of a combination of CD33/34 LCA and qPCR technology.

Results: Monitoring of 75 transplantations, performed in 72 patients at 11 German pediatric transplant centers between 05/2005 and 04/2009, included 26 (0-84) PB and 5 (0-21) BM samples per patient and covered 1.5 (0.1-4.3) years (median). PB was analyzed weekly and BM at days 30, 60, 100 and 6, 9, 12, 15 and 18 months post-transplant. Lowering the quantifi able limit by qPCR (here ≤10E-3 (0.1%)) can result in detection of recipient derived (autologous) signals (AS) in virtually every sample. The challenge was, to defi ne a clinically relevant cut off level for diagnosis of complete (CC) or mixed chimera (MC). For qPCR, cut off levels of 0.1 and 0.5% for AS were evaluated and chimerism was defi ned according to tab. 1. Based on this defi nition, risk groups and profi les were specifi ed (tab. 2). Allocation into risk groups 2 to 4 was done at day of fi rst MC in either compartment. Risk profi le I (RP I) defi ned each sample with MC in either compartment as HR, whereas risk profi le II (RP II) defi ned each sample with MC in MNC as HR. 56/75 transplant follow-ups allowed for analysis of predictive value of chimerism (tab. 3). Sensitivity of RP I ranged from 100% (qPCR cut off 0.1) to 80% (STR-PCR); notably specifi city was low and false positive rate high, if this highly sensitive approach was applied. In RP II sensitivity was signifi cantly lower compared to RP I, but specifi city was higher and false positive rate lower. Relapse was predicted earlier by qPCR with a median of 33 days compared to STR-PCR in 43% if 0.1% cut off for defi nition of MC was applied. If 0.5% cut off was used, this rate dropped to 22%. Discussion: QPCR as well as STR-PCR allowed for monitoring of chimerism. Each method revealed its specifi c pros and cons according to its high (qPCR, ≤10E-3) or moderate (STR-PCR, 10E-2) quantifi able limit. For the fi rst time we describe clinically applicable cut off levels for qPCR to defi ne complete and mixed chimeras. LCA in CD33/34 subpopulations increased the sensitivity of chimerism monitoring. Early and sensitive detection of impending relapse was possible by qPCR. However, the substantial proportion of false positives had to be kept in mind. Disclosure of Interest: None Declared.

Introduction: Double UCBT(dUCBT) has extended the use of UCB for heavier patients (pts) who require a higher cell dose. In a recent randomized study in children, outcomes after dUCBT were similar compared to single UCBT. For most children a single unit can be used if TNC>3×10 7 /Kg, but for heavier pts dUCBT is an option. Outcomes and risk factors analysis after dUCBT in children have not been described. Materials (or patients) and Methods: We analyzed 211 children who underwent a dUCBT between 2002 and 2012 in 63 EBMT centers. Analysis was performed separately for pts with malignant (MA, n=173) and non malignant (NM, n=38) diseases. Results: Among pts transplanted for malignancies, 96 had ALL, 51 AML, 9 MDS, 2 CML and 15 chronic lymphoid diseases; median age at UCBT was 14y and median weight was 49 kg; disease status (DS) was 1 st CR (43%), ≥2 nd CR (38%) or advanced (19%). In this group, 148 pts received a MAC and 25 a RIC regimen. Cyclopho sphamide(Cy)+fl udarabine(Flu)+TBI was administrated in 42% of the cases; 64% received ATG before day 0. Median number of collected TNC was 6,4×10 7 /kg and 59% of the grafts had ≥2 HLA mismatches. GvHD prophylaxis was CSA-based in 89% of the pts (52% CSA+MMF). Median follow-up was 24 months. CI of PMN and PLT engraftment was 87% and 62%, with a median time of 23 and 45 days, respectively. Better engraftment was associated with pts's negative CMV serology (94 vs 78%, p=0.01) and diagnosis of AML vs ALL vs MDS (96 vs 83 vs 67%, p=0.003). CI of aGvHD grade II-IV was 48%; 30/133 pts at risk developed cGvHD (17 extensive; CI at 2y: 17%). CI of 2y-relapse was 26%. In 147 pts transplanted for acute leukemia (AL), CI of relapse was 28% for recipients in 1 st CR, 18% in ≥2 nd CR and 42% for advanced diseases (p=0.3). 2y-TRM was 29%; 82 pts died of relapse (n=40), infections (n=14), GvHD (n=9) and other causes (n=19). 2y-EFS was 44% (45% for ALL, 40% for AML, 78% for MDS and 38% for others). According to DS (1 st CR, ≥2 nd CR and advanced), EFS was 49%, 47% and 22% in ALL and 41%, 38% and 38% in AML. 2y-OS was 49%. In multivariate analysis, higher EFS and OS were associated with the use of CyFluTBI conditioning (p=0.009 and p=0.016, respectively). Among pts with NM diseases, 24 had bone marrow failure (BMF: 10 FA, 13 SAA and 1 other inherited), 2 hemoglobinopathies, 7 immune defi ciencies and 5 metabolic disorders; median age at dUCBT was 11y and median weight was 40 kg. In this group, 27 pts received RIC and 11 MAC, 30% were TBI-based and 76% fl ubased regimens; 91% had ATG in the conditioning. GvHD prophylaxis was CSA-based in 76% of the pts (29% CSA+MMF, 21% CSA+prednisone); 60% of the grafts had ≥2 HLA mismatches. Median follow-up was 34 months. Of these 38 pts, 28 engrafted with a median time of 23 days. Among pts who did not engraft, 3 had an autologous reconstitution (1 alive) and 3 had a subsequent alloHSCT (2 alive). 14 pts developed aGvHD grade II-IV; 12/31 pts at risk had cGvHD (4 extensive). 2y-TRM was 49% and 20 pts died of disease progression (n=1), infections (n=5), GvHD (n=4), rejection (n=4) and other causes (n=6). 2y-OS was 44% (30% for BMF, 50% for immune defi ciencies and 80% for metabolic disorders). Discussion: This survey suggests that dUCBT is feasible in children for whom a single unit cell dose is not suffi cient. Despite higher incidence of aGVHD, EFS and OS seem to be comparable with other alternative stem cell sources in children lacking a matched donor. Disclosure of Interest: None Declared. Results: The total group of patients showed a pEFS at 3 years of 55.7%; cumulative incidence (CI) of relapse (CIR) and CI of treatment related mortality (CI TRM) were 36.68% and 11.56%. The following pre transplant factors did not infl uence outcome: age, gender, relapse site, cytogenetics, donor type, and stem cell source. EFS probability for patients who received their transplant in CR2 was 39%, for patients in >CR2 44% and for patients who were not in remission 0% (p=0.01). Also immune phenotype of the leukemia infl uenced survival: pEFS was 63% in pre-B ALL patients, 19% in T-ALL patients, and 0% in patients where no information was available. pEFS was 22% for patients with very early relapse, 53% for patients with early relapse, and 83% for patients with late and 41% in patients in CR3 or non remission (p<0.001). MRD values post transplant were analyzed diff erent time points. Patients were grouped according to their MRD level: Patients were defi ned as MRD high: >10E-3; MRD low: positive, <10E-3, and MRD negative. pEFS at 3 years was for MRD high, MRD low and MRD negative patients at day +30: 0.0%, 41.7%, and 59.23%, respectively (p=0.009). At day +60 for MRD low and MRD negative patients: 47.0%, and 65.0%, respectively (p<0.001). At day +90 for MRD high, MRD low and MRD negative patients: 0.0%, 50.0%, and 64.0%, respectively (p<0.001). At day +180 for MRD high, MRD low and MRD negative patients: 0.0%, 60.0%, and 79.0%, respectively. And at day +365 for MRD low and MRD negative patients: 0.0%, and 95.0% respectively. Discussion: When analyzed either for the diff erent time points or for the highest MRD value post transplant, MRD results had always signifi cant infl uence on survival. All patients who developed a MRD load of >10E-3 at any time after transplant relapsed and died. However, 78% of patients who became MRD positive <10E-3 did not develop frank relapse. Hence, dynamic evolution needs to be considered as 8/15 patients with an MRD load of <10E-3 at day 180 survived. However, if MRD was not cleared by day 365 all patients fi nally relapsed. As all patients who developed MRD >10E-3 at any time post transplant fi nally relapsed, MRD may serve as an endpoint for further pre-emptive therapies. Disclosure of Interest: None Declared.

Introduction: Evaluation of EBMT registry data of Ewing Tumors to explore trends in indication and outcomes over 30 years. Materials (or patients) and Methods: Since 1980, 3695 patients (pts) with Ewing tumors (2186 males, 1499 females) have been registered in the EBMT data base from 142 European centers in 24 countries with only 70 allogeneic (allo) SCTs of the total. MAT indications were primary metastatic disease or high-risk local disease (tumor size and/or poor response) in 2411pts and relapse in 719pts. The median age is 15 years (yrs) (range, 1 to 65) with 2568pts <18yrs. Peripheral blood stem cells were used in 3143 pts. Busulfan based combinations were used in 1392 pts, Melphalan based regimens in 421pts, Treosulfan containing regimens in 182 pts, Total body irradiation (TBI) based regimens in 131pts whilst 258 pts had other various combinations. The median survival time is 0.5 yrs after allogeneic and 2.8yrs after autologous SCTs. Results: The 5-year OS (5-yr OS) was 44% in 3521pts with autologous (auto)SCT (49% during primary tx. and 31% after relapse) and 12% for alloSCT (p<0.001). In the whole cohort a steady improvement in 5-yr OS is observed: ≤1990 23%, 1990 to ≤2000 39% and > 2000 48% (p<0.001). 5-yr OS according to age show 48% for ≤10yrs. 43% for >10 to ≤18yrs and >18yrs 38% (p<0.001). In the auto SCT group localized disease 5-yr OS is 53% and 41% for stage 4pts. The remission status prior to SCT has a major impact in autoSCT: 5-yr OS in CR1 (1343pts) is 58%, for PR (836pts) is 40% but only 20% in 146 SD or primary refractory 146pts (p<0.001). In relapse pts a second CR prior to MGT still results in a 46% 5-yr OS whilst pts with residual disease do signifi cantly worse with< 20% (p<.000). A single SCT procedure (1556pts) results in 51% 5-yr OS whereas repetitive regimes, likely to refl ect less favorable pts (355pts), in 41% (p<0.003). In primary treatments TBI regimens are worse with 5-yr OS of 38% whilst nonTBI regimens result in 50% (p<0.026). The 5-yr OS for the MAT regimens as used after year 2000 are as follows in primary treatment: Busulphan based (684 pts) 60%, Melphalan based (148 pts) 37%, Treosulfan based (95pts) 19% and others (133pts): 55%. A Cox proportional hazards regression model identifi es age, response status, stem cell source and MGT regimens as independent risk factors. Discussion: This EBMT Ewing data set holds important information for decision making processes and suggests exploring in more depth the results and roles of Busulphan versus Treosulfan in front line trials. Disclosure of Interest: None Declared. Introduction: Hematopoietic stem cell transplant-associated thrombotic microangiopathy (HSCT-TMA) is a challenging posttransplant complication without good therapeutic options. In the most severe form of HSCT-TMA with systemic vascular injury, mortality rates approach 90%. Targeted therapy is urgently needed for these high risk patients. We recently observed that dysregulation of the complement system may be involved in the pathogenesis of HSCT-TMA. These fi ndings suggest that the complement inhibitor eculizumab could be a therapeutic option for this severe HSCT complication; however, the effi cacy of eculizumab in children with HSCT-TMA and its dosing requirements still need to be determined. Materials (or patients) and Methods: We treated eleven children with severe HSCT-TMA using eculizumab. There are no eculizumab dosing recommendations for HSCT-TMA, thus we measured eculizumab serum levels and adjusted the dose to achieve a therapeutic level of >99 μg/ml based on recommendations for children with aHUS. Total hemolytic complement activity (CH50) was measured in serum during eculizumab therapy as a pharmacodynamic marker of eculizumab induced complement blockage at the same time points as eculizumab drug levels. Pharmacokinetic and pharmacodynamics analyses were performed to correlate eculizumab drug levels with clinical response and the degree of complement blockade as measured by CH50. Results: All 11 patients had severe multivisceral TMA unresponsive to calcineurin therapy withdrawal and/or therapeutic plasma exchange. Terminal complement activity as measured by sC5b-9 level was elevated in all patients at TMA diagnosis. Seven of eleven children (64%) achieved resolution of HSCT-TMA and recovery of organ function after achieving therapeutic eculizumab levels and complete complement blockade but required higher doses and/or more frequent eculizumab infusions to achieve therapeutic drug levels than currently recommended for children with aHUS. Six responding patients were able to discontinue therapy after receiving a median of 9 doses of eculizumab (range 4-17) and remain clinically well with median follow up time of 1 year after treatment. One patient remains on eculizumab maintenance therapy. Four critically ill patients failed to reach therapeutic drug levels even after dose escalation and they subsequently died after receiving a median of 3 doses of eculizumab (range 2-7). There were no side eff ects attributable to eculizumab. CH50 activity below 5% directly correlated with therapeutic eculizumab level and clinical response. Discussion: Our observation indicates that eculizumab is a promising therapeutic option for pediatric patients with severe HSCT-TMA, but HSCT patients appear to require higher medication dosing than recommended for other conditions and respond better when therapy is initiated early in the disease course. We propose to guide eculizumab dosing in HSCT-TMA based on pharmacodynamic monitoring and suggest that measuring CH50 may aid in timely dose adjustments, especially in highly catabolic patients who will likely require more intense drug dosing. Controlled studies of this approach in both children and adults after HSCT are warranted. Disclosure of Interest: None Declared.

Introduction: A meta-analysis of sibling unstimulated marrow vs G-PB allograft trials demonstrated faster hematologic recovery and increased cGvHD with G-PB but no survival benefi t. We completed a randomized trial comparing G-PB with G-BM in sibling allografts for hematologic malignancies (Couban et al, ASH, 2013) and report the association of allograft cell counts with hematologic recovery and cGvHD and the pattern of organ-specifi c cGvHD. Materials (or patients) and Methods: A phase III randomized trial of matched sibling G-PB vs G-BM allografts in adults with hematologic malignancies was conducted. The primary endpoint was time to treatment failure defi ned as the fi rst occurrence of extensive cGvHD, disease progression/relapse or death. Adaptive stratifi cation (minimization) balanced treatment arms by centre, disease, early vs late disease and conditioning. Donors received G-CSF 5 micrograms/kg/d sc for 4 or 5 days with apheresis on Day 5 and, if necessary, Day 6 or BM harvest on Day 5. Recipients between 16 and 65 years received myeloablative conditioning and cyclosporine/methotrexate as GvHD prophylaxis. Results: 230 sibling donor-recipients were randomized; 7 were unevaluable. Indication for allograft was AML (109; 49%), ALL (57; 25%); MDS/MPD (31; 14%), CML (13; 6%), lymphoproliferative disorder(12; 5%) and biphenotypic leukemia (1) . Median total nucleated cells (TNC)/kg were 7.59×10 8 (range: 2.73-18.64) and 5.92×10 8 (range: 0.94-13.23) in the G-PB and G-BM arms (P<0.0001). Median CD34 cells/kg were 5.50×10 6 (2.03-23.94) and 3.22×10 6 (0.29-8.71) (P<0.0001). At a median follow-up of 36 months (range 9.6-48), using Cox proportional hazards multivariable modelling adjusted for the 4 minimization factors, time to treatment failure did not diff er signifi cantly between arms (HR 0.91; 95% CI 0.68-1.22; P= 0.52). There was also no OS diff erence .8% vs 60.4%; G-PB vs G-BM; p=0.90). Neutrophil and platelet recoveries were related to CD34 cells/kg in G-PB (neutrophils: regression slope (RS) -0.39; platelets: RS -0.21) and G-BM (neutrophils: RS -0.93; platelets RS -0.49). There was no correlation of CD34 cells/kg or TNC/kg with overall or extensive cGvHD. More hepatic cGvHD was reported with G-PB. Discussion: In this large, randomized trial, there were no diff erences in time to treatment failure, overall survival or cumulative incidence (CI) of aGvHD, overall or extensive cGvHD despite substantive differences in the allografts. Correlation of CD34 cell dose with hematologic recovery was expected. The fi nding of no diff erence in the CI of overall and extensive cGvHD between the two arms, despite an anticipated 10-fold diff erence in T-cell dose and the lack of correlation between CD34 cell dose and cGvHD suggest that factors other than T-cell and CD34 dose must aff ect the likelihood Introduction: The relative importance of the resolution level of HLA typing has not been fully defi ned for related donor transplantation. Materials (or patients) and Methods: To address this question, we retrospectively evaluated patients who underwent a fi rst related HSCT from 2000 to 2010 from an HLA high-resolution matched (MRD, n=2244), high-resolution 1 locus-mismatched (HR-MMRD, n=116), or low-resolution 1 locus-mismatched related donor (LR-MMRD, n=396) in the graft-versus-host direction with the available data of high-resolution HLA typing at 3 loci including HLA A, B and DRB1 using the database of the Japan Society for Hematopoietic Cell Transplantation. Results: The median age was 40 years (0-74). The median followup duration of surviving patients was 950 days. The cumulative incidences of grade II-IV and III-IV acute GVHD in the HR-MMRD and LR-MMRD groups were signifi cantly higher than those in the MRD group (HR-MMRD 39%, LR-MMRD 48%, and MRD 30% for grade II-IV acute GVHD, and HR-MMRD 20%, LR-MMRD 20%, and MRD 10% for grade III-IV acute GVHD, Figure A) . In multivariate analyses, both LR-MMRD and HR-MMRD were associated with an increased risk of grade II-IV and grade III-IV acute GVHD. There was no statistically signifi cant diff erence in the incidence of acute GVHD between the HR-MMRD and LR-MMRD groups in multivariate analyses. The cumulative incidences of nonrelapse mortality (NRM) at 2 years in the HR-MMRD and LR-MMRD groups were signifi cantly higher than that in the MRD group (HR-MMRD 20%, LR-MMRD 26%, and MRD 14%). Finally, the presences of LR-MM and HR-MM were associated with a signifi cantly inferior overall survival (OS) at 2 years (HR-MMRD 54%, LR-MMRD 49%, and MRD 62%, Figure B ). In multivariate analyses, both LR-MMRD and HR-MMRD were associated with an increased risk of NRM and an inferior OS. Again, there was no statistically signifi cant diff erence in the incidence of NRM and the probability of OS between HR-MMRD and LR-MMRD groups in multivariate analyses. Discussion: In conclusion, both LR-MM and HR-MM have similar adverse impact on the outcome in related HSCT as in unrelated HSCT. Disclosure of Interest: None Declared. 1 Eurocord APHP, Paris, France, 3 Ospedale San Raff aele, Milano, 4 Universitò Roma Tor Vergata, Roma, 5 Ospedale San Martino, Genova, Italy, 6 IPC, Marseille, France, 7 Ospedale Careggi, Firenze, Italy, 8 BMT, Pekin, China, 9 BMT, Athens, Greece, 10 BMT, Antalya, Turkey, 11 BMT unit, Lyon, France, 12 BMT unit La Fe, Valencia, Spain, 13 BMT Unit, Pescara, Italy, 14 Rigshospitalet, Copenhagen, Denmark, 15 Hospital Saint Pau, Barcelona, Spain, 16 conditioning regimen and GVHD prophylaxis), and prospective trials are needed to consolidate these fi ndings. Disclosure of Interest: None Declared. Irradiation. Anti-thymocyte globulin (12 mg/kg over 3 days, from -5 to -3) and rituximab (200 mg/m 2 on day -1) were used as prophylaxis of GVHDand EBV-related lymphoproliferative disorders, respectively. Results: Sustained primary engraftment occurred in 49/50 pts, the remaining child, who had experienced secondary graft failure, being successfully re-transplanted from the other parent.

The median time to reach neutrophil and platelet engraftment was 13 days (9-18) and 11 days (8-20), respectively. None of these pts developed gut or liver acute GVHD. Thirteen pts experienced skin-only grade I-II GVHD, this leading to a cumulative incidence (CI) of 26%. Only 2 out of the 48 pts at risk developed skin limited chronic GVHD, the CI being 4%. Two pts died for causes other than disease relapse (1 idiopathic pneumonia and 1 multi-organ failure), the CI of transplantation-related mortality (TRM) being 4%. Nine pts relapsed, the CI of disease recurrence being 19%. With a median follow-up of 18 months (range 6-36), the 2-year Kaplan Meier estimate of leukemia-free survival (LFS) was 77%; this value was 79% for pts with ALL. LFS of pts who did or did not experience skin-only acute GVHD was 85% and 74% (p=NS). Patients given TBI had a higher LFS (87%) compared with pts who received only chemotherapy (47%, p<0.01). The CI of viral infections (mainly CMV reactivations) was 53%. In the fi rst 4 weeks after HSCT, circulating T cells mainly consisted of the gd+ subset, while ab+ T lymphocytes predominated thereafter. Intriguingly, CMV reactivation promoted the expansion of Vd1+ T cells. Mature NK cells expressing, including also the alloreactive cells, were detectable at 1 month in peripheral blood after HSCT. S59 are alive. Pts receiving fl u-cy-TBI (n=17) had a signifi cantly higher CI of NE (82% vs 40%, p=0.03); a trend towards better NE was noted in pts given a RIC conditioning (68% vs 41%, p=0.1). Ten pts developed acute GvHD grades II-IV, and 7 of 26 at risk had chronic GVHD (6 limited and 1 extensive). Fifteen pts relapsed (3 AML) in a median time of 7 months (range 1-31); CI of relapse at 2 years was 36±8%. The CI of TRM at 1 year was 36±2%. Overall 25 patients died, of which 8 before day 100. Causes of death were relapse (n=10), infection (n=6), EBV-related lymphoproliferative disease (n=2), GvHD (n=3), others (n=4). OS at 2 years was 44±8%. It was not diff erent for pts with JAK2 mutation at diagnosis (OS: 41% vs 42%). OS was 63% vs 29% for pts receiving fl u-cy-TBI or other conditioning (p=0.1), 63% vs 24% for pts who did not receive or received ATG (p=0.06). Among pts who underwent RIC, 11 were TBI based without ATG: 5 of these are alive. EFS at 2 years (for relapse and graft failure) was 29±8%, with signifi cant improvement in pts receiving fl u-cy-TBI (52% vs 10%, p<0.01) and lacking ATG before day 0 (46% vs 10%, p=0.02). Discussion: Graft failure was a major concern in these pts. Introduction: HSTL is a rare and extremely aggressive peripheral T cell lymphoma subtype characterized by primary extranodal disease with involvement of spleen and liver. The response of HSTL to standard chemotherapy regimens is poor, and most patients die within the fi rst two years after diagnosis. Although anecdotal cases of successful allogeneic stem cell transplantation (alloSCT) for HSTL have been described, a structured analysis of the impact of SCT on the course of HSTL is missing. The objective of this study was to analyze the outcome of patients who underwent allogeneic or autologous SCT (autoSCT) for HSTL. Materials (or patients) and Methods: This is a registry-based retrospective multicentre study including patients 18 years or above with histologically verifi ed HSTL who underwent alloSCT or autoSCT between January 2003 and May 2013 and were reported to the EBMT. Baseline patient, disease, and transplant data were collected from MED-A forms. Centers with potentially eligible patients were contacted to provide additional treatment and follow-up information including a written histopathology report.

Histology reports were reviewed by an expert hematopathologist (H.S.). Statistical analysis was descriptive and employed log rank comparisons for univariate assessment of the impact of baseline characteristics on survival endpoints. Results: 74 patients were identifi ed in the database who fulfi lled the inclusion criteria. Additional information upon center request was provided for 36 them. Twelve of these had to be excluded after histopathology review, leaving 24 patients in the fi nal study sample. There was a predominance of male patients (58%), and the median age was 38 (19-67) years. 79% of the patients were diagnosed with stage IV disease, and systemic symptoms at diagnosis were present in 75%. Splenomegaly was reported in 92%, hepatomegaly in 71%, and bone marrow involvement in 58%. A history of Crohn's disease was reported for 9% of the patients, and one patient had received a previous solid organ transplant. Eighteen (75%) received an alloSCT (56% identical sibling, 6% mismatched relative, 39% unrelated donor), and 6 patients an autoSCT. The median time from diagnosis to SCT was 5 (2-52) months. After a median of 2 (1-4) pretreatment lines, half of the patients were in CR, 38% in PR, and 12% refractory at the time of SCT. Conditioning was TBI-based in 50% of the alloSCT group. (and considered as myeloablative in 78%). With a median follow-up of 36 months, 10 patients (56%) are alive after alloSCT and 1 patient (17%) after autoSCT, with non-relapse mortality (6 patients) and relapse (4 patients) being the leading cause of death after alloSCT and autoSCT, respectively. In contrast, only 2 patients (12%) experienced relapse after alloSCT. 3-year disease-free survival was 55% after alloSCT. Discussion: Although still preliminary, this study provides the fi rst evidence that alloSCT can result in long-term survival in a signifi cant proportion of patients suff ering from this otherwise inevitably fatal disease. The lack of effi cacy of autoSCT suggests that the benefi t of alloSCT is due to graft-versus-lymphoma activity rather than to the cytotoxic eff ects conferred with high-dose therapy. Disclosure of Interest: None Declared.

Introduction: Allogeneic stem cell transplantation (allo-SCT) is a well-established consolidation treatment for patients with multiply relapsed HL. Although reduced intensity conditioning regimens (RIC) were initially associated to a non-signifi cant improvement in progression-free survival (PFS) through a decrease in non-relapse mortality (NRM), they were also associated to a higher relapse rate (RR) after allo-SCT. The objective of this analysis has been to compare the results of RIC-allo vs myeloablative conditioning protocols (MAC) At relapse, patients were treated with either non immunologicbased strategy (chemotherapy, radiotherapy) or immunomodulation (DLI and/or discontinuation of immunosuppressive therapy), or both. We compared their outcomes according to the treatment they received at relapse. Results: Relapse occurred at a median time of 2.7 months after transplant (from 0.5 to 18.6 months) and 14 patients had a localized cutaneous relapse (7 MF, 3 anaplasic, 2 T-nos, 1 AITCL and 1 NK/T). Among the 13 patients who received DLI (DLI alone: 5, DLI + radio/chemotherapy: 8), 9 obtained a response (7CR, 2 PR) and 6 are still alive, with a median follow up of 3 years. For the 38 who did not receive DLI, immunosuppressive therapy was tapered or stopped for 23 and led to a prolonged remission, for 2 of them, along with extensive chronic GVHD. In the non-DLI group, 22/38 received a radio/chemotherapy treatment with a response rate of 50% and a 1-year survival rate of 25%. Among the 8/38 patients who experienced a response in this group, fi ve developed a chronic GVHD. Median overall survival was 144 days in the non-DLI-group and 557 days in the DLI-group (p=0.01). In univariate analysis, time from transplant to relapse (p=0.015), chronic GVHD after relapse (p=0.036) and receiving DLI (p=0.016) were associated to a better overall survival (OS). In multivariate analysis, DLI was the only remaining factor associated with OS (p=0.005). Disease status at transplant, regiment's intensity, patient's age and relapse localization (cutaneous or disseminated, p=0.08) were not found to be associated with OS. Discussion: The benefi t of DLI in case of relapsing T-cell lymphomas after transplant is suggested in this study and represent an argument in favour of the existence of GVL eff ect in these diseases. These fi ndings could lead to the set up of an active immunomodulation strategy after transplant for patients with high-risk disease, even if the short delay from transplant to relapse, currently observed, can limit this therapeutic option. Disclosure of Interest: None Declared.

Introduction: Mantle cell lymphoma (MCL) is considered incurable by current standards due to biology of disease characterized by relapsing course and eventual resistance despite initial responsiveness to treatment. The common fi rst line treatment for young and fi t patients in advanced disease is immunochemotherapy for remission induction followed by the high-dose chemotherapy with or without total body irradiation (TBI) and autologous hematopoietic stem cell transplantation (HCT) in the fi rst complete (CR) or partial remission (PR). To date, the best conditioning regimen before HCT remains unknown and the role of TBI is one of the main issues. Introduction: Blood and marrow transplantation is a low volume, high risk specialty. Insitutional practices vary considerably. Consensus practice is important for physician and family confi dence, so as to optimise patient outcomes during stem cell transplantation with best practice and as a platform for research. The 12 paediatric / adolescent transplant centres of the UK and Ireland are increasingly collaborative over the last years as they look to support each other's practice, achieve such consensus and continue and establish research studies. This report summarises the eff orts of that groups and commends them to other national paediatric BMT groups. Materials (or patients) and Methods: The Group's activities over the last 7 years are reviewed in the following areas: a) National Audit Meeting in which all transplants performed from the previous year are in a single national meeting attended by physicians, nurses and data managers from each centre. There have been 9 national vidoe MDT, held monthly. 47 patients have been discussed -75% pre-transplant patients where indication, conditioning or donor were the focus of discussion. The remainder were transplanted patients with complications. All centres have attended. There have been 2 such conferences with metabolic disease focus -one discussing patient specifi c X-ALD questions and one discussing similar questions in MLD. Parallel to these increasing face-to-face contacts there is increasing email contact between the group about diffi cult patients between meetings. Discussion: We have found these national, increasingly frequent joint meetings to be extremely useful and we commend them to other national groups. We also believe that the opportunity to discuss individual cases with such a national group is not only helpful to patient management but is easily seen -for families and professionals -to be important to patient management and generates confi dence in the complicated transplant process. Disclosure of Interest: None Declared. Introduction: In vivo serotherapy with T-cell depleting antibodies is common prior to (mis)matched unrelated allogeneic stem cell transplantation. Both anti-thymocyte globulin (ATG) and alemtuzumab are used, but an optimal dose has not been defi ned. A frequently applied total dose for alemtuzumab in children is 1.0 mg/kg. Recovery of T-cells is signifi cantly slower after alemtuzumab in comparison with ATG and measuring the alemtuzumab concentration in sera might be useful for optimising individual dosing. Materials (or patients) and methods: The concentration of alemtuzumab was measured in 650 sera of 62 children, transplanted between 2003 and 2012. Thirty-seven patients received a median total dose of 1 mg/kg (range 0.8-1.5 mg/kg), whereas 25 patients received a median total dose of 0.6 mg/kg (range 0.4-0.6 mg/kg). Samples were tested by quantitative fl ow cytometry as well as by ELISA with an anti-idiotypic antibody. Lymphocyte subsets were analysed frequently post transplantation by fl ow cytometry. Results: The ELISA showed a good correlation with the more timeconsuming fl ow cytometry, but ELISA was more sensitive. The serum concentration of alemtuzumab showed a large individual variation at day 1 post HSCT (median 3.4 μg/mL; range 0.1-10.9 μg/mL, Figure 1 ) and was signifi cantly higher in patients receiving a higher dose. The median half-life was 11 days. Four patients developed aGvHD and all had a serum concentration from day +10 far below the P10 of the total group ( Figure 1 ). In three of these patients anti-drug antibodies were demonstrated. The time to reach <0.1 μg/mL, considered a threshold below which immune recovery can occur, varied from day -2 to day +96 post HSCT. To get more insight in the infl uence of alemtuzumab on immune recovery, we compared patients at three weeks post HSCT (the median time to engraftment). A high alemtuzumab concentration correlated with delayed engraftment (day +28 vs day +19, p=0.001), delayed T-cell recovery (day +100 vs day +55; p=0.03) and delayed NK-cell recovery (day +35 vs day +23; p=0.03).

There was no signifi cant diff erence in overall survival or eventfree survival, but in patients transplanted for a malignancy a high alemtuzumab dose correlated with a lower relapse rate (50% vs 25%, p=0.02). A signifi cantly higher alemtuzumab concentration was measured in older children (BW 55 kg vs 21 kg; p=0.002).

The median time for T cells to exceed 100 cells/μL was 87 days. Most patients showed T-cell recovery when the alemtuzumab concentration was below 0.1 μg/mL. However, NK-and B-cell recovery was frequently seen at concentrations up to 2.2 μg/mL (Fig 2) .

Discussion: Children with higher bodyweight had relatively higher levels of alemtuzumab and delayed immune recovery. Dosing in mg/kg might result in overdosing in children with a high bodyweight. Pharmacokinetic and pharmacodynamic modelling may help to establish an optimal individual dosing. A study in adults resulted in an optimal total dose of 10 mg, with a very fast recovery of NK cells (Gartner 2013). CD52 is expressed on T-, B-and NKcells but the recovery of NK-and B-cells (even in the presence of comparatively high drug levels) was more rapid than recovery of T-cells. CD52 negative T-cells have been seen in the fi rst months post HSCT (Garland, 2006) , but their function is unknown and will be studied further. Materials (or patients) and Methods: We, therefore, established a donor-patient-specifi c NOD/SCID/IL2Rγ null (NSG) xenotransplantation model to study donor-specifi c NK cell alloreactivity towards pediatric BCP-ALL. To this aim, we performed adoptive transfer studies with donor-derived mature NK cells in BCP-ALL-bearing NSG mice exemplifying the situation of supportive cell transfer as an adjunct to haplo-identical transplantation. In addition, we humanized NSG mice (huNSG) with donor-derived SCs and studied the potential alloreactivity of immature NK cells emerging in the early post-transplantation period. Results: Adoptive transfer experiments with mature NK cells demonstrated a distinct GvL reactivity when defi ned KIR-KIRL mismatch constellations between donor and recipient were chosen. NK cell alloreactivity was hereby associated with increased CD107a but not IFN-g synthesis ("split response"). Experiments performed in huNSG mice revealed that patient-specifi c BCP-ALL was rejected only in those huNSG mice that had been transplanted with a donor that exhibited a KIR-KIRL mismatch towards the injected patient-specifi c specimen. As recently published in vitro data suggested that the DNA-demethylating drug 5-Aza-cytidine -usually applied as a direct cytotoxic agent in MDS or AML patients -might modulate NK cell function, we next tested the eff ect of a longterm, low dose 5-Aza-cytidine treatment in our xenotransplantation model (week 6 onwards post HSCT). We found that the extent of BCP-ALL burden was dramatically reduced in the 5-Aza-cytidine treated huNSG but not in control mice lacking donor-specifi c surrogate hematopoesis. However, NK cell receptor analysis revealed that in these sets of huNSG mice, the expression of inhibitory KIRs was negligible and 5-Aza-cytidine therapy had not resulted in changes of either inhibitory (KIRs, NKG2A) or activating (NKp44, NKG2D) NK cell receptor expression indicating that the immature, KIR -NK cell compartment had conferred alloreactivity. In line with this, sorted KIR -NK cells of KIR-KIRL matched donors displayed a high level of alloreactivity towards pediatric BCP-ALL in vitro that was again accompanied by increased CD107a but not IFN-g synthesis.

[PH-O116]

Discussion: We here provide evidence that both mature KIR-KIRL mismatched NK cells but also "unlicensed" immature, KIR -NK cells may convey a signifi cant anti-tumor reactivity towards pediatric BCP-ALL in vivo. Pre-transplantation Fresenius ® ATG was administered to all recipients (10 mg/kg/day from day -6 to -3). Due to graft rejection requiring a second HSCT, two patients received a second conditioning regimen based on the same drug combination, without TBI in case it had been administered for the fi rst HSCT. G-CSF-mobilized donor peripheral blood stem cells were T-cell depleted by means of CD34 + cell positive selection as sole GVHD prophylaxis and infused 72 hours after completion of the preparative regimen. Results: The T-depleted grafts contained a median of 19.1×10 6 (range 5.6-35) CD34 + cells, 1.4×10 4 (range 0.35-50) CD3 + lymphocytes and 4.7×10 4 (range 1.8-21) CD19 + lymphocytes per kg recipient body weight. Engraftment occurred in 9 out of 12 patients (75%), the cumulative incidence of graft rejection being 17% (95% CI, 5-59). The two patients who rejected their fi rst allograft achieved a complete engraftment after a second HSCT using the other haploidentical parent.

Cumulative incidences of grade II-IV acute and chronic GVHD were 17% (95% CI, 5%>59%) and 35% (95% CI, 14-89), respectively. The conditioning regimen was well tolerated, with no fatal toxicity and 3 cases of grade III toxicity. The cumulative incidence of transplantation-related mortality was 17% (95% CI, 5-59). With a median follow-up of 2.3 years (range 0.6-11.5), the Kaplan-Meier estimates of OS and DFS were both 83% (95% CI, 62-100), while the EFS probability was 67% (95% CI, 40-93). Discussion: These data demonstrate that a reduced intensity, fl udarabine-based conditioning regimen, followed by infusion of high doses of T-depleted HSC, is able to guarantee engraftment with good OS and DFS, confi rming the feasibility of haploidentical HSCT in FA. Given our promising results, we suggest that haploidentical HSCT may currently be regarded as a feasible and eff ective option for FA patients with an indication for allogeneic HSCT lacking an HLA-matched donor identifi ed within a reasonable time frame. Introduction: High dose chemotherapy (HDCT) with tandem or triple carboplatin and etoposide course is currently the fi rst curative choice for GCT patients (pts) who relapse to front-line CT or who are cisplatin refractory. Very limited information is available on the indication to proceed to HDCT courses for pts having a progressive disease post-induction CT/CD34+ cell collection and immediately prior to HDCT. Materials (or patients) and Methods: We reviewed our two Institution series of pts who underwent a HDCT program whereby induction CT followed by CD34+ cell mobilization and one or multiple HDCT courses where administered, and in various settings. All pts received induction CT for CD34+ mobilization and debulking purposes. Induction/mobilizing regimens were HD-CTX (n=32), cisplatin, etoposide, ifosfamide (PEI, n=37), paclitaxel-ifo (n=13), and mixed regimens (n=12 p=0.527] ). In the pure second-line setting, 2/5 progressing pts (40%) maintained a remission at +32 and +14 months after HDCT vs 16/34 responding pts (47%). Discussion: In the present series we were still unable to identify an independent and signifi cant association of disease chemosensitivity to induction CT with PFS and OS in pts receiving HDCT for advanced GCT. No conclusion should be stated thus far, and unresponsive patients should not be excluded by HDCT. Results are limited by small numbers.

In the absence of a roadmap coming from the available and present data, an ongoing retrospective study of the EBMT-STWP is aimed to assess the prognostic role of this variable in a large patient series. Disclosure of Interest: None Declared. End point of the study: To assess the outcome of this program in patients with acute myeloid leukemia (AML) and refractory anemia with excess blasts (RAEB). Materials (or patients) and Methods: Patients. We studied 62 patients with AML (n=46) or RAEB (n=16), who received a fi rst allo graft in our Unit from family donors, mismatched for one full HLA haplotype. All patients received CyA from day 0, MMF from day +1 and PT-CY 50 mg/kg day +3 and day +5. The graft source was unmanipulated marrow in all cases. The median age was 51 (18-74); 27 patients were in CR1 (43%), 13 in CR2 (21%) and 22 had active leukemia at transplant (36%). The conditioning regimen consisted of total body irradiation (10-12 Gy) and fl udarabine 30 mg/m 2 x4 (n= 24 patients, median age 39, range or thiotepa (10 mg/kg), busulfan 9.6 mg/kg and fl udarabine (150 mg/kg) (TBF) (n= 38 patients, median age 58 , range 17-74). For 13 patients over the age of 65 the total dose of busulfan was cut at 6.4 mg/kg (2 days). Results: Engraftment was achieved in all patients. Acute graft versus host disease (GvHD) grade II-IV developed in 11% , chronic GvHD (moderate+ severe) in 10%). Transplant related mortality (TRM) was 14%, (10% for CR1+CR2 and 20% for active disease).

The 3 year cumulative incidence of relapse was 13% for CR1+CR2 patients and 43% for patients with active leukemia. The 3 year actuarial survival is 64% for CR1+CR2, and 28% for patients with active disease (p=0.002). There was no diff erence in survival when comparing the TBI regimen (48%) vs the TBF regimen (52%), despite a large diff erence in patients age (39 vs 58 years, p=0.0001) and more patients with advanced disease in the TBF (20% vs 50%, p=0.02). Donor lymphocyte infusions (DLI). Patients with molecular relapse, identifi ed as increased WT1 expression above 100 copies/10 4 abl) (n=10), received DLI in doses ranging from 1×10 5 to 1×10 7 : the complete response rate (reduction of WT1 levels <100 copies) was 45% and 2 years survival 38%. Patients with hematologic relapse (n=10) received chemotherapy followed by DLI (doses ranged from 1×10 5 to 1×10 7 /Kg: complete remission was achieved in 3 patients, and the actuarial 2 years survival was 33%. Discussion: Conclusions Myeloablative HLA-haploidentical marrow transplantation with PT-CY is associated with a low rate of acute and chronic GVHD, a very low rate of NRM and encouraging survival in a setting of adults with high risk AML/RAEB. The combination of thiotepa busulfan fl udarabine (TBF) is not inferior to full dose TBI and can be performed also in older adults up to over 70 years of age. Disclosure of Interest: None Declared.

planted in CR1 (30%), 55 pts in CR2 or higher CR (27%) and 86 pts in more advanced disease (43%). Median age at time of haplo-SCT was 42 years (range, 18-66). 109 pts out of 202 received a reducedintensity conditioning regimen (RIC, 54%). 129 pts received antithymocyte globulins (64%) as part of the conditioning regimen and 5 pts received Alemtuzumab (5%). Stem cell source was bone marrow for 83 pts (BM, 41%) and peripheral blood for 119 pts (PB, 59%). Median dose of infused mononuclear cells was 4.4×10 e8 /kg (range, 1-62.87) for BM. Median dose of infused CD34+ cells was 7.25×10 e6 /kg (range, 1.06-15.26) for PB. Donor median age was 39 years (range, . 53 male pts received a graft from a female donor (26%). CMV donor/host serostatus was as follow: neg/neg for 28 pts (14%), pos/neg for 16 pts (8%), neg/pos for 36 pts (18%) and pos/pos for 118 pts (60%), missing for 4 pts. The 60-days cumulative incidence of engraftment was 90+/-2%, with a median time for reaching ANC >0.5×10 e9 /L of 17 days (range, 6-63). The cumulative incidence of acute GvHD was 32±3% for grade II or higher and 15±3% for grade III or higher. The 3-years incidence of chronic GvHD was 39±3% (n=76: 51 limited, 24 extensive). At 3 years, the estimates of leukemia-free survival (LFS) for pts transplanted in CR1, CR2 or in advanced disease were 46±7%, 36±6%, 12±4%, respectively. The estimates of overall survival (OS) were 58±7%, 47±7% and 13±4%, respectively. The cumulative incidences of relapse (RI) were 30±6%, 31±6% and 59±5%, respectively. Non-relapse mortality (NRM) incidences were 24±6%, 33±6% and 30±5%. Discussion: These data suggest that unmanipulated graft haplo-SCT is a valid treatment option in acute leukemia. The current LFS and NRM rates support the use of such transplant approach as part of the treatment algorithms for adult acute leukemia patients with an indication to allogeneic SCT but lacking an HLA-identical sibling or matched-unrelated donor. Disclosure of Interest: None Declared. Introduction: The poor outcome of AML elderly patients is not only due to the low complete remission (CR) rate and high Treatment Related Mortality (TRM), but also to an unacceptably high Relapse Incidence (RI). In this setting the adverse biological features of the disease play the main role, but also the lack of an eff ective post consolidation therapy could negatively aff ect the RR. Materials (or patients) and Methods: We prospectively evaluated two post-consolidation strategies in 72 AML elderly patients, achieving CR after the 1st induction chemotherapy, including high dose aracytin and antracyclines and fi t for autologous stem cell transplantation (ASCT). The main objective was to assess the feasibility and effi cacy in terms of OS, DFS and RI of two postremission strategies, administered on the basis of the mobilization outcome: ASCT and low-dose gemtuzumab-ozogamicin (GO). According to the study design, those patients collecting ≥3×10 6 CD34+/kg, underwent ASCT, while those collecting <3×10 6 /kg, received GO, 3 mg/m 2 monthly, for at least 3 months. Results: Among the 72 patients in CR1, 55 (76,3%) were able to undergo PBSC mobilization attempt, after 1rst consolidation and 24/55 (44%) collected >3×10 6 CD34+ cells/kg; among the 55 patients eligible for PBSC mobilization, 7 did not receive the planned treatment, 23 underwent ASCT and 25 received GO. With a median follow-up of 70 months (range: 24-124), 20/55 patients are alive, 18 of them in CCR. The 8 years OS and DFS are respectively 35.9% (24-49.8) and 31.2% (21-43.8) Median OS and DFS are 22 and 16 months respectively. A landmark analysis was performed computing survival from CR after fi rst consolidation in the 48 patients who received the assigned treatment mobilization-driven (ASCT or GO). At multivariate analysis post-consolidation treatment and hyperleukocytosis signifi cantly predicted OS and DFS while diagnosis of secondary AML was signifi cantly associated with a higher relapse rate with an 83.4% vs 54% of de novo AML (p=0.01). Patients with WBC>50,000/ml had a signifi cantly lower OS with a 0% 3 years OS vs the 46% observed at 8 years in patients with WBC count< 50,000/ ml (p=0.01). After 8 years, 57% of patients in the GO arm are alive, compared to 25.4% of patients in the ASCT arm, 25% in supermobilizers (collecting >7.1×10 6 /kg CD34+ cells) and 27.3% in normal mobilizers (collecting <7.1×10 6 /kg CD34+ cells) (Figure 1 ). Both normal and super mobilizer patients, receiving ASCT, had respectively 2.36 (95% CI: 1-5.9) (p=0.07) and 3 (95% CI: 1.2-7.8) (p=0.02) RR to die compared to patients receiving GO (p=0.04). Finally, DFS at 8 years was 45.3% in patients receiving GO, compared to 26% observed in patients receiving ASCT with a RR of 2.1 (p= 0.05). Discussion: Our study outlines low feasibility and effi cacy of ASCT in elderly AML patients, while post-consolidation with GO appears safe and eff ective. Larger multicentre studies are warranted in order to confi rm GO effi cacy in this setting. Disclosure of Interest: None Declared.

Introduction: Polyneuropathy, organomegaly, endocrinopathy, dermopathy associated with a paraproteinaemia (POEMS syndrome) is a rare paraneoplastic syndrome secondary to a plasma cell dyscrasia. Eff ective treatment of the underlying plasma cell dyscrasia, including ASCT, can control the disease and often dramatically control symptoms though limited data is available for ASCT in POEMS. The aim of this study was to describe the clinical outcome of ASCT for patients with POEMS syndrome. Materials (or patients) and methods: Patient-, disease-, and transplant-related variables were collected according to the data entries in the EBMT database, including tracking incomplete data entries from participating centers. Systemic involvement and organ-specifi c response to ASCT was detailed utilizing an organ involvement tool pre-and post-ASCT. Results: 127 patients underwent an ASCT between 1997-2010 and satisfi ed the entry criteria. The median age was 50 years (range 26-69) with 51.2% ≤50 years of age. The extent of systematic disease involvement was: peripheral neuropathy in 97.6%, volume overload in 61.7%, organomegaly in 79%, papilledema in 36%, dermopathy in 73% and 60% had sclerotic bone lesions at presentation. The median time from diagnosis to ASCT was 7.5 months (range 1-346) with 31.5% of patients receiving an ASCT >12 months from diagnosis. The graft source was PBSC in 100% of patients. Disease status at ASCT was: 48% CR/PR, 34% SD/MR/ untreated and 18% in PD (missing information in 19% of patients).

The conditioning regimen was Melphalan ≥200mg/m 2 in 63%, Melphalan <200mg/m 2 in 23% with a TBI-containing regimen being utilized in only in only 1 patient. The total rate of engraftment at 3 months was 96.8%, with 2 patients (1.6%) dying before engraftment. Engraftment syndrome was documented in 22.8% of ASCT recipients including peri-engraftment fever in 23% & pulmonary infi ltrates in 7.5%. Haematological complete response (CHR) was characterized in 25% with 17% having progressive disease or died without attaining CHR. CHR was demonstrated in13.3% at 12 month, 20% at 36 months and 23% at 60 months (median time to CHR of 8.6 months) post-ASCT. In a sub-group (n=25), an organ symptom tool to linearize disease impact, demonstrated a signifi cant global improvement in symptom scores (2-way ANOVA p<0.0001) with expansion of the dataset to the full cohort underway. With a median follow-up of 47.7 months (95%CI 38.3, 58.6), the non-relapse mortality at 1, 2 & 5 years was 3.3% (95%CI 0.1-6.4%), 4.4% (95% CI 0.6-8.2%) & 7.7% (1.9-13.6%), respectively. The median progression-free survival (PFS) was 106 months (95% CI 87.8, NR) with a 5-year PFS of 73.5% (95%CI 63.2-83.7%). The 5-year overall survival (OS) was 88.6% (95% CI 81.5-95.8%). The extent of organ involvement did not predict for either PFS or OS. Discussion: The data analyzed in this study, demonstrates that ASCT can be an eff ective & safe therapeutic modality for patients with POEMS syndrome. Due to the systemic infl ammatory nature of the condition, due care should be taken to supporting these patients through this procedure, especially in relation to engraftment syndrome. Disclosure of Interest: None Declared.

In multivariate analysis, the following factors were associated with increased incidence of grade II-IV GVHD: use of MAC (HR: 1.4, p=0.013) , absence of ATG (HR: 2.39, p<0.001), ≥ 4/6 HLA mismatches (HR: 1.35, p=0.048) ; whereas MAC (HR: 1.6, p=0.03) and absence of ATG (HR: 3.48, p<0.001) were associated with increased incidence of aGVHD grade III-IV. Presence of aGVHD grade III-IV was associated with higher early NRM (HR: 1.81, p=0.004). Discussion: Incidence of acute and chronic GVHD remains lower in dUCBT setting when compared to published data after HLA matched adult donor. In patients developing aGVHD, skin is the most common organ aff ected, followed by GIT and liver. MAC regimen, ≥ 4/6 HLA mismatches and absence of ATG as part of conditioning regimen were associated with higher incidence of aGVHD (II-IV). Despite the lower incidence, aGVHD (III-IV) is associated with higher early NRM, therefore prospective trials evaluating the role of conditioning regimen and GVHD prophylaxis in dUCBT should be addressed. Disclosure of Interest: None Declared. Introduction: After T cell depleted allogeneic stem cell transplantation (alloSCT), donor lymphocyte infusion (DLI) can elicit an alloreactive T cell response mediating a curative graft versus leukemia reaction (GVL) in the absence or presence of graft versus host disease (GVHD). Materials (or patients) and Methods: To compare response magnitude, strength and specifi city, we selected patients who responded without or with GVHD after receiving DLI for recurrent disease or incomplete donor chimerism after T cell depleted HLAmatched alloSCT. Results: To analyze immune reconstitution, we fi rst measured absolute numbers of activated circulating CD8 T cells pre and post DLI. A signifi cant increase of HLA-DR pos CD8 T cells was observed from 27×10 6 /L (median, range 6-143) to 92×10 6 /L (median, range 19-1548, p=0.03), and subsequent cloning of HLA-DR pos CD8 T cells showed high frequencies of virus specifi c T cell clones targeting EBV epitopes. Anti-viral reactivity was found irrespective of the development of GVHD, indicating a similar capacity for protective immunity after DLI in all patients. Next, to establish whether GVL-inducing alloreactive CD8 T cells also resided in the activated T cell compartment, we fi rst cloned and analyzed both HLA-DR neg and HLA-DR pos CD8 T cells from selected samples. Alloreactive T cell clones, specifi cally recognizing patient and not donor cells, were only found in HLA-DR pos fractions. We next investigated whether the magnitude of alloreactivity diff ered between patients without or with GVHD, by comparing the frequencies of alloreactive T cell clones in multiple samples from all 13 patients. Signifi cantly lower percentages of alloreactive T cell clones were found in patients without GVHD as compared to patients with GVHD (median 3% range 1-18 versus median 31% range 10-68, p=0.02), showing that occurrence of GVHD is associated with a higher response magnitude. To determine whether higher magnitudes of alloreactivity in patients with GVHD were associated with stronger responsiveness of the T cells, we quantifi ed the IFN-γ production per T cell after stimulation with EBV-LCL, representing the intended hematopoietic targets of DLI. T cell clones from patients with GVHD demonstrated stronger responsiveness as compared to clones from patients without GVHD, which may have aff ected in vivo expansion resulting in diff erent response magnitudes. To determine whether development of GVHD was correlated with tissue specifi city of the alloreactive response, we tested all T cell clones for recognition of fi broblasts (FB) representing the non-intended targets of DLI. T cells recognizing FB were exclusively found in patients with GVHD. To mimic pro-infl ammatory conditions, FB were pretreated for 4 days with 100 IU/ml IFN-γ resulting in limited recognition by some T cell clones isolated from patients without GVHD. Discussion: Our data show lower magnitude and strength, and restricted tissue recognition of CD8 T cell alloreactivity in patients responding to DLI without GVHD, which may have prevented high tissue levels of pro-infl ammatory cytokines resulting in selective recognition of hematopoietic cells. In patients with GVHD however, the higher magnitude and strength of the hematopoiesis directed response may have initiated a cascade of events creating an infl ammatory environment leading to destruction of nonhematopoietic tissue resulting in GVHD. Disclosure of Interest: None Declared. Introduction: Chronic graft-versus-host disease (cGVHD) is the major late complication of allogeneic stem cell transplantation (Allo-HSCT) and has a signifi cant impact on quality of life. In severe cases of cGVHD, especially involving in pulmonary system, it becomes life-threatening and can signifi cantly reduce quality of life and increase mortality. Antithymocyte globulin (ATG) has been used to prevent acute GVHD eff ectively, but the protection from cGVHD remains controversial. In our study, we want to investigate the associations between pulmonary cGVHD and usage of ATG, and clarify whether ATG administered before Allo-HSCT will improve survival and reduce the incidence rates of pulmonary cGVHD. Materials (or patients) and Methods: 355 adult patients who diagnosed with acute leukemia and survived over 100 days after Allo-HSCT between 2000 and 2012 were analyzed, with 217 without ATG before Allo-HSCT, and 138 having ATG. Risk factors associated with pulmonary cGVHD, such as age, donor-recipient sex pair, donor type, donor-recipient blood type pair, disease status before Allo-HSCT, prior acute GVHD, overall survival (OS) rates, and cumulative incidence rates of pulmonary cGVHD for the 2 groups were compared. Results: Table 1 shows the patient demographics, indicating a strong similarity between the 2 groups in terms of recipient age, donor age, gender, disease and conditioning regimen, with the exception of a higher proportion of peripheral blood stem cell, ABO mismatch and unrelated donor in the ATG group. 52% of patients receiving ATG developed grade 0 to 1 acute GVHD, and in 62% of patients not receiving ATG (p=0.0317, odds ratio=0.61, 95% CI, 0.39-0.96). Pulmonary cGVHD was signifi cantly higher in patients not using ATG before Allo-HSCT (p=0.0068, odds ratio=0.34, 95% CI, 0.15-0.77). Moreover, using ATG before Allo-HSCT is associated with lower incidence of pulmonary cGVHD (adjusted odds ratio=0.26, 95% CI, 0.12-0.60), after adjustment for the eff ects of graft source. The 5-year OS rate for the group taking ATG was 51.3%; for the other without ATG, the rate was 49.7%. Comparison of cumulative incidence rates of pulmonary cGVHD between the 2 groups indicated that the number for patients without ATG before Allo-HSCT was 34.5%, higher than ATG group (9.3%), although there was no statistical signifi cant diff erence (P=0.077). Trend for ATG dose-response was not found in this study. Discussion: Among patients underwent Allo-HSCT, those who received conditioning regimen with ATG had lower risks and cumulative incidence rates of developing pulmonary cGVHD. To identify risk factors for cGVHD and trend for dose-response of ATG, further study is suggested. Disclosure of Interest: None Declared. Rotterdam, Netherlands, 7 CHU, Poitiers, 8 CHU, Bordeaux, France, 9 UMCU, Utrecht, Netherlands, 10 Leiter der Klinischen Abteilung für Hämatologie, Graz, Austria, 11 Rigshospitalet, Copenhagen, Denmark, 12 CHU de Lille, Lille, France Introduction: To assess the impact of antithymocyte globulins (ATG), on patients' outcome after unrelated cord blood transplantation (UCBT) following a reduced-intensity regimen (RIC), we conducted a retrospective based registry analysis on 661 adults with hematological malignancies. Materials (or patients) and Methods: 661 adults patients underwent unrelated single(s) or double(d) UCBT following RIC [TBI/cyclophosphamide/fl udarabine)] between 2004 and 2011.

Participating EBMT centers were asked to provide information on type, timing and total dose of ATG used. Diagnosis was AML/ALL in 51%, MDS/CML in 19% and lymphoid malignancies in 30%; 28% of patients were transplanted in early disease status, 28% in intermediate and 44% in advanced disease. 30% of patients had a previous autologous transplantation. Single UCBT was used in 226 (34%) patients, while 435 (56%) were transplanted with dUCBT. HLA matching was defi ned as low resolution for HLA-A and HLA-B, and high resolution for HLA-DRB1, and for dUCBT, the highest degree of HLA incompatibility was considered. Therefore, most of the HLA incompatibilities were 4/6 (n=435, 72%). Median number of total nucleated and CD34+ cells collected were 4.4×10 7 /kg and 1.6×10 5 /kg, respectively. All patients received TCF, with TBI 2Gy (86%), TBI 4Gy (12%) and TBI 6Gy (2%). Rabitt ATG (rATG) was used as part of RIC in 82 patients (12.4%) with a median total dose of rATG (Fresenius®) of 20 mg/kg (5-60) and rATG (Genzyme®) of 8 mg/kg (5-15). GVHD prophylaxis consisted of cyclosporine A (CsA)+ mycophenolate mofetil (MMF) in 91%, Csa alone±other in 9%. The median follow-up was 36.3 months. Results: Patients receiving rATG-TCF had more MDS/CML (30% vs 20%, p<0.01), were transplanted earlier (p=0.02) and there was a trend of being transplanted with more advanced disease (53% vs 43%, p=0.06). Type and dose of rATG were not associated with any outcomes. Multivariate (MV) models for outcomes were built adjusting for the diff erences between 2 groups of rATG-TCF (yes and no) and other risk factors that impact outcomes (patients's age>51 years, positive CMV serology, MDS/CML, advanced disease status, year of transplant, HLA 4/6 and ABO incompatibility).

In the fi nal MV models, use of rATG was associated with decreased 

incidence of aGVHD (HR=0.31, 95%, CI=0.17-0.55, p<0.0001), higher incidence of NRM (HR=1.68, 95% CI=1. 16-2.43, p=0.0009) and decreased OS (HR=1.69, 95% CI=1.19-2.415, p=0.003), however it was not associated with engraftment, chronic GVHD and relapse. In rATG-RIC-group, the main cause of death was transplantation related in 51% of cases. Death related to infections was 72% in the rATG-RIC group compared to 38% in the non rATG group. Discussion: In conclusion, in this retrospective and multicentre analysis, use of rATG as part of TCF regimen, was associated with decreased incidence of acute GVHD with increased NRM and decreased OS, probably related to the higher incidence of infections. We suggest that in UCBT for adults with hematological malignancies given a TCF regimen, use of rATG should be avoided, however timing and dose of rATG are still open questions. Disclosure of Interest: None Declared.

Introduction: The Fludarabine-Busulfan-ATG conditioning regimen is widely used for allogeneic stem cell transplantation (SCT) of AML patients. This conditioning regimen, fi rst described by S. Slavin (Slavin et al., Blood 1998) in the context of HLA identical SCT contained two days of oral Busulfan (Flu-Bu2) and cyclosporine A (CsA) alone. The conditioning has since evolved with the use of the intravenous (iv) form of busulfan and variable GVHD prophylaxis: CsA alone versus CsA and mycofenolate mofetyl (MMF) or methotrexate (MTX), with or without ATG. Materials (or patients) and methods: To assess the impact of the intensity of GVHD prophylaxis on the outcome of allogeneic SCT conditioned with the Flu-Bu2 regimen, we conducted a registry-based retrospective study on a population of 144 patients older than 50 years who underwent a fi rst HLA identical SCT with

Introduction: In patients with AML in CR1 with >intermediate risk disease allogeneic HSCT is treatment of choice. Conditioning inten-sity is varied, reduced intensity (RIC) conditioning is given to older patients, whereas young patients traditionally receive myeloablative regimens (MAC). In patients between the ages of 40-60 both types are used with little comparative data. Previous studies had shown that RIC regimens were associated with higher relapse risks but lower risks of transplant related mortality (TRM). We hypothesized that in low-intermediate risk disease based on cytogenetic classifi cation RIC is superior to MAC whereas in high risk leukemia MAC is superior to RIC given higher antileukemic activity. Materials (or patients) and Methods: This study included 2974 of 5388 eligible patients with AML transplanted in CR1 in 2000-2011 classifi ed by cytogenetic risk status at diagnosis. Regimens were classifi ed as MAC (n=1638) or RIC (n=1336). Median followup of surviving patients was 46 and 41 months. Groups diff ered by many variables. MAC recipients were younger (37.6 vs 53.8 years), had a shorter interval from diagnosis to transplantation (143 vs 165 days), were more frequently male (53% vs 48%), had less frequently poor risk cytogenetics 19% vs 22%, received less frequently stem cells from an unrelated donor (20% vs 33%), and had more frequently marrow as a stem cell source (36% vs 7%). Results: Results: Table 1 shows similar overall (OS) and leukemia free survival (LFS) in both groups but a lower relapse incidence (RI) and a higher transplant related mortality incidence (TRM) in the MAC group. Acute grade II-IV GvHD was higher with MAC, incidence of chronic GvHD did not diff er signifi cantly. In univariate analysis overall survival was higher with RIC in cytogenetic good risk AML (55+5% vs 77+7% MAC vs RIC) but not in intermediate risk (61+1% vs 62+2%) or poor risk AML (42+3% vs 40+3%). Relapse incidence was lower with MAC in poor risk AML (36+3% vs 51+3%) and intermediate risk AML (21+1% vs 30+1%) but not in good risk AML (19+4% vs 13+5%). TRM was higher in MAC vs RIC in all three cytogenetic risk groups. Multivariate analysis confi rmed a signifi cant LFS and OS advantage of RIC in good risk but not in intermediate and poor risk leukemia. Table 1 Discussion: Conclusions: In patients aged 40-60 MAC conditioning has no advantage over RIC in spite of RIC transplant recipients being generally in a poorer risk category. We confi rm lower relapse rates but higher TRM risks with MAC compared to RIC. We fail to show superiority of MAC in patients with high risk cytogenetics but there appears to be an advantage for RIC over MAC in the small cohort of patients with good risk leukemia. Disclosure of Interest: None Declared.

While protocols exist for infant ALL with clear criteria for HSCT in CR1 pts, there is no consensus for HSCT in infants with CR1 AML. In this population of pts with aggressive disease, unrelated cord blood is attractive due to its prompt donor availability and high stem cell content. Materials (or patients) and Methods: We retrospectively analyzed 254 pts with ALL (n=159) or AML (n=95) diagnosed within 1 year (y) of age. Pts received single UCBT after myeloablative conditioning (MAC) between 1995 and 2012 in EBMT centers

) with ASCT, followed by maintenance with lenalidomide plus prednisone (RP) or lenalidomide alone (R) in NDMM patients. Materials (or patients) and Methods: Patients with NDMM ≤65 years, eligible for ASCT were enrolled. Patients received 4 28-day cycles of lenalidomide-dexamethasone (Rd) induction (lenalidomide 25mg day 1-21 and low-dose dexamethasone 40mg day 1,8,15,22) followed by stem cell mobilization. Patients were randomized to receive consolidation with 6 28-day cycles of cyclophosphamide-lenalidomide-dexamethasone (CRD, cyclophosphamide: 300 mg/m 2 day 1,8,15; dexamethasone: 40mg days 1,8,15,22; lenalidomide: 25mg days 1-21) or 2 courses of melphalan 200 mg/m2 followed by stem-cell support (MEL200-ASCT). Patients were randomly assigned to receive maintenance with lenalidomide alone (R: 28-day cycles of lenalidomide 25mg day 1-21) or lenalidomide plus prednisone (RP: 28-day cycles of lenalidomide 25 mg day 1-21, prednisone 50mg every other day). Primary study endpoint was progression-free survival (PFS); secondary endpoints were safety and overall survival (OS). Data cut off was October

Median time from enrolment to maintenance was 14 months. In the intention to treat population of patients eligible for maintenance, 2-year PFS from the start of maintenance was 69% for RP and 58% for R patients

However, scant data delineate the success of mobilisation of PBSC (PBSC2) in fi rst relapse after prior ASCT. The multi-centre phase III BSBMT/UKMF Myeloma X Trial demonstrated the clinical utility of a second/salvage ASCT (ASCT2). However, unless remobilisation is viable, such therapy is limited to those patients with stored PBSC from fi rst line. Hence, a secondary objective of the trial was to determine the feasibility, safety and effi cacy of remobilisation at fi rst relapse, and the use of PBSC2 in support of ASCT2. Materials (or patients) and Methods: Eligible patients with MM relapsing after prior ASCT were enrolled. All were re-induced with PAD (bortezomib, doxorubicin & dexamethasone) delivered in 2-4 21-day cycles, then randomized to either ASCT2 or weekly cyclophosphamide (C-weekly). Patients with inadequate stored PBSC1 were scheduled to undergo remobilization; those with adequate PBSC1 were off ered the opportunity to re-mobilize. Results: 297 patients were enrolled

9%, motor neuropathy 0.9%, sensory neuropathy 6.5%, other (all grades) 32.7%; many likely relating to PAD or other prior therapies. 174 patients were randomized (ASCT2 n=89, C-weekly n=85). 42 (47.2%) received PBSC1, 29 PBSC2 and 11 (12.4%) a mixture of the two (PBSCMix)

Discussion: Stem cell harvesting after bortezomib-based reinduction for myeloma relapsing after prior ASCT is both feasible and safe, permitting a second/salvage ASCT for patients without stored PBSC. Although follow-up is currently short, outcomes with either stem cell source, or a mixture, are equivalent

INTERIM ANALYSIS OF A PHASE II CLINICAL TRIAL

High-dose melphalan alone or in combination with TBI is most frequently used for conditioning. In our center we developed an original protocol of tandem autoHSCT, in which fi rst conditioning is based on total marrow irradiation (TMI) alone and second comprises of high-dose melphalan. The intention was to increase the effi cacy by using both radio-and chemotherapy in myeloablative doses and to limit toxicity of irradiation by its targeted administration. The aim of this planned interim analysis of the phase II clinical trial (ClinicalTrials. gov NCT01665014) was to evaluate safety of TMI

who achieved at least partial response after one or more lines of induction chemotherapy and who mobilized at least 5x10 6 CD34+ cells/kg were eligible for the trial. The fi rst conditioning included TMI (total dose 12 Gy, 4 Gy fractions on days -3, -2, -1) applied with the use of helical tomotherapy. Patients with active sites of disease, as revealed by FDG-PET, received additional 'boosts' (total 24 Gy, 4 Gy fractions on days -9 to -4)

TMI +/-additional 'boosts' resulted in absolute neutropenia in all patients. Median time to neutrophil >0.5 G/L and platelet >50

-16) d., respectively. Non-hematologic grade 2-3 toxicity was as follows: infections (gr

3, 10%), nausea (gr. 2, 22%, gr. 3, 2%), vomiting (gr. 2, 6%), diarrhea (gr. 2, 4%

SD (3%). The probabilities of OS and PFS at 16 months were 100% and 91% (+/-5%), respectively. No cases of non-relapse mortality were registered

personalized treatment approach in MM and is characterized by very good toxicity profi le with preserved myeloablative potential. Tandem autoHSCT using TMI in sequence with highdose melphalan appears safe (no cases of NRM) with encouraging rates of early response. Final evaluation of its effi cacy requires longer follow-up

SNPS IN THE PROMOTER REGION OF THE IL17A GENE ALLOW ANTICIPATION OF COMPLICATIONS AFTER SIBLING HLA-IDENTICAL ALLOGENEIC STEM CELL TRANSPLANTATION

Introduction: GvHD is the main cause of morbimortality after allo-SCT. Several SNPs in cytokine genes have shown to be associated with with SCT outcome. IL17 has been implicated in the pathogenesis of various autoimmune diseases but its importance in SCT is not well-known. The objective of this study was to analyse the infl uence of IL17A SNP genotypes on the risk and severity compli

Results: The association between IL17A genotypes and complications after allo-SCT are shown in Table 2. Patients transplanted from donors harboring genotype CC for the SNP rs8193036 show increased risk of grade III-IV acute GvHD (7/26 vs 47/397, p=0.035) and of grade II-IV acute GvHD (13/26 vs 133/409, p=0.048). Patients transplanted from donors harboring allele A in the SNP rs4711998 show increased risk of extensive chronic GvHD (53/161 vs 43/177, p=0.045). Relapse rate was not related with IL17A SNP genotypes. Finally a higher risk of toxicity-related mortality (TRM) was observed in patients transplanted from donors harboring allele A for SNP rs2275913 (78/293 vs 46/227, p=0.048), donors harboring allele G for SNP rs3819024 (78/279 vs 46/242, p=0.011) and donors harboring allele A for SNP rs4711998 (68/250 vs 55/229, p=0.044). Discussion: IL17A SNP genotyping might be useful to anticipate complications after sibling HLA-identical allo-SCT and, therefore, to improve the clinical management of transplanted patients

Interleukin-17A (IL-17A) and IL17F are two highly homologous pro-infl ammatory cytokines that locally direct the infl ux of neutrophils to sites of infection and infl ammation. The role of IL-17 in acute GVHD is still controversial. To date, all studies on Th17 in experimental GVHD were performed using donor T cells from 17a-defi cient animals (IL17a-/-), which are capable of IL-17F production. Since IL-17A-defi ciency might be substituted by the presence of IL-17F, these studies did not fully address the function of IL-17 in GVHD

Materials (or patients) and Methods: Experiments were performed as BMT recipients and wildtype

Experiments with IL17af-/-animals as BMT-recipients ruled out a contribution of host derived protective IL-17A and F production after GVHD-induction. As compared to controls, recipients of IL-17af-/-donor T cells showed an onset of severe diarrhea early after BMT. Also serum analysis at day 14 after BMT showed a higher level of infl ammatory cytokines in these animals. These results, the known fact that IL17-receptor is expressed on gut epithelial cells, and that destruction of the epithelial barrier aggravates the cytokine storm after BMT, led us to the hypothesis that donor derived IL-17A and F essentially contribute to recipient's gut-epithelial integrity after BMT. Mechanistically, FITC-dextran assays demonstrated a loss of the intestinal epithelial integrity in recipients of IL17af-/-donor T cells. Furthermore, we observed signifi cantly decreased epithelial cell proliferation in the intestine of these animals, as confi rmed by BrdU incorporation staining. Interestingly, when we compared animals after BMT with either IL17a-/-or IL-17af-/-donor CD4 + T cells, GVHD-lethality was significantly higher in recipients of double-defi cient T cells. Therefore, according to our hypothesis, it seems that IL17A-defi ciency can be substituted by IL-17F. Discussion: Here we demonstrate that defi ciency of donor derived IL17A and IL17F increased GVHD severity by alteration of recipient's intestinal barrier. Furthermore we showed that IL17A-deficiency can be substituted by IL-17F in our experimental model. In summary, the pro-infl ammatory TH17-cytokines IL-17A and IL-17F Interest: None Declared. PH-O140 DOUBLE CORD BLOOD TRANSPLANTATION: INCIDENCE, ORGAN INVOLVEMENT AND RISK FACTORS OF ACUTE GRAFT VERSUS HOST DISEASE

Hôpital Saint

Daniel den Hoed Cancer Centre

Saint-Louis Hospital

Tel-Hashomer, Israel, 15 Bone marrow transplantation unit

Double umbilical cord blood transplantation (dUCBT) has been used as a strategy to circumvent the problem of cell dose in adult patients. It is related with higher incidence of acute GVHD (aGVHD) when compared with single UCBT. The association of aGVHD incidence with the number of cells infused, HLA disparities or GVHD prophylaxis is yet to be determined in a larger series of patients

One-hundred day cumulative incidence (CI) of aGVHD grade II-IV was 36%±2% and III-IV 15%±2%, with a median onset of 28 days. Of those pts with grade II-IV aGVHD, 82% had skin involvement, 66% gastro-intestinal tract (GIT) and 25% liver. Ninety-eight percent of pts with aGVHD grade III-IV received steroid based treatment. At day 60, CI of engraftment was 84%. Of pts who did not engraft, 21% developed aGVHD (mostly grades I or II). Two-hundred and eighteen out of 712 pts at risk developed chronic GVHD (cGVHD) with a CI at 2 yrs of 25%. One-hundred and two pts (47%) had previous aGVHD grade II-IV and 116 (53%) had de novo cGVHD

time from diagnosis to transplantation (median 176 vs 173 days), gender of donor, WBC at diagnosis (median 6.3 vs 7.3x10 9 /L), cytogenetic risk (intermediate in 58 vs 61% and poor in 28% vs 22%, with < 5% missing data) and FLT3 mutation status (33% vs 34%, with 50% of missing data). Source of SCT was peripheral blood in 93% of cases in both groups. The transplantation was performed more recently in the group with the triple immunosuppression

With a median follow up of 15.2 months, in univariate analysis, as compared to the patients who received relatively reduced immunosuppressive GVHD prophylaxis, those having received the triple drug association had a signifi cantly higher risk of relapse

73), a trend towards reduced leukemia free survival (LFS) (38±9% vs 54 ± 6%, p=0.08) and overall survival (OS)

Discussion: These data suggest that the use of a triple GVHD prophylaxis associating ATG, CsA and MMF or MTX for AML patients allografted in CR1 with a sibling donor and conditioned with Flu-ivBu2 might unfavorably impact the outcome of allogeneic SCT by increasing the risk of relapse

Giebel 20 , M. Mohty 21 and Acute Leukemia Working Party of the European Blood and Marrow Transplant Group

Erasmus MC-Daniel den Hoed Cancer Centre

Cumulative incidence of Gd 2-4 acute GvHD by D+100 was 17.8% (5/28), cumulative incidence of Gd 3-4 acute GvHD by D+100 was 7.1% (2/28). 64% of patients (18/28) are surviving, at median follow up of 651 days (range 7-3841). 36% (10/28) have died: relapse in 4/10; CMV pneumonitis in 2/10; multiorgan failure in 2/10. The mortality rate at D+28 was 3.8 % (1/28) and at D+100 was 19.2% (5/26).Only one patient succumbed to IFD progression, and one death was related to uncontrolled sepsis. In general GT were well tolerated, with no adverse reactions in 68% (19/28), fevers in 18% (5/28) and mild respiratory symptoms in 11% (3/28) patients. One patient had transfusion-associated CMV infection (no CMV negative donors available) but has survived. Discussion: In our study, prophylactic single donor GT led to a successful outcome of HSCT in a very high risk group of children Introduction: Allogeneic hematopoietic cell transplantation (HCT) is one of the most curative therapeutic modalities for hematological malignancies. But relapse is the leading cause of treatment failure and some long-term survivors after HCT suff er late relapse of the underlying disease. Many studies have shown that graftversus-host disease (GVHD) can be used as surrogate markers for graft-versus-tumor (GVT) eff ects. In this study, we analyze the infl uence of GVHD on late relapse after allogeneic HCT. Materials (or patients) and Methods: From the database of JSHCT, we extracted the data of patients with hematological malignancies who received fi rst allogeneic HCT between 1974 and 2011. A total of 29,431 recipients were included in this study. The infl uence of GVHD on risk of late relapse was assessed by including acute and chronic GVHD as time-dependent covariates in Cox models adjusted for disease risk, donor-recipient sex disparity, HLA disparity, conditioning intensity, age at transplant, year of transplant and type of transplant

DF5", n=7,224) was 3.0%. In multivariate analysis, standard risk at transplant, disease (MDS or ML), myeloablative conditioning and age at transplant =<15 y/o were signifi cant factors for reduced risk of relapse in DF2, and disease (ALL, AML or MDS) and donor-recipient sex (female-to-male) were in DF5. Cumulative incidence of acute GVHD (grade II-IV) at day 100 was 33.5%. Cumulative incidence of chronic GVHD (limited and extensive) at 1 year was 28.6%. Grade II-IV acute GVHD was associated with reduced risk of relapse for patients with AML

R. Ladenstein 1, 2, * , , U. Poetschger 2 , E. Glokova 4 , H. Jürgens 5 , U. Dirksen 5 , J. Michon 6 , O. Oberlin 7 , R. Luksch 8 , I. Yaniv 9 , H. van der Berg 10 , J. Wheelan 11 , I. Lewis 12 , M. Bregni 13 , C. Peters 14 , F. Lanza 15 and for the EBMT Pediatric and Solid Tumor Working Groups 1 Paediatric Haematology/Oncology, St. Anna Kinderspital, 2 Studies and Statistics on Integrated Research and Projects, St. Anna Kinderkrebsforschung e.V., Vienna, Austria, 3 Paediatric Haematology/Oncology, Institut Gustave Roussy, Villejuif, France, 4 Children's Cancer Resaerch Institute, Vienna, Austria, 5 11 , J. Sierra 12 , S. Nguyen 13 , N. Maillard 14 , , W. Linkesch 16 3 Hopital Saint Antoine, Paris, 4 Centre Hospitalier Lyon Sud, lyon, France, 5 

Introduction: In CN-AML, risk groups based on molecular markers are well established after conventional therapy. We conducted a registry-based analysis to defi ne the prognostic role of molecular subgroups in conjunction with classical risk factors after alloSCT. Materials (or patients) and methods: 1006 pts. fulfi lled the inclusion criteria (fi rst alloSCT for CN-AML, matched sibling or unrelated donor, known mutational status of FLT3-ITD and NPM1). Median age was 52y, stage at SCT was CR1 (n=702), CR2/3 (n=161), primary induction failure (PIF, n=51) and relapse (n=92). Results: With a median follow up of 25mo, 2y OS/LFS were 70±2/64±2% after SCT in CR1, 61±5%/46±5% in CR2/3, 42±7%/37±7% in PIF, and 29±5%/24±5% in relapse. In 702 pts. transplanted in CR1, OS, LFS, cumulative incidence of relapse (CRI) and non-relapse mortality (NRM) were calculated according to molecular subgroups. Whereas NPM1 mut had no infl uence on outcome, FLT3-ITD was associated with increased CRI (p=0.0001), and decreased LFS (p=0.0008) and OS (p=0.002) . See tab 1 for outcome of subgroups based on NPM1 and FLT3-ITD mutations.The mutational status of CEBPA was known in 151 CR1 pts. with NPM1 mut /FLT3-ITDneg (double negative) genotype. 2y OS/LFS in 13 pts. with CEPBA mutation was 100%/92±3%, 138 pts. without mutations in either NPM1, FLT3 or CEBPA achieved a 2y OS/LFS of 77±3%/72±3% (p=0.05 for OS/0.09 for LFS). A multivariate Cox model of predefi ned risk factors for alloSCT in CR1 (age, WBC at dx, time to CR1, donor, conditioning, year of SCT, and molecular subgroups) was performed. Age >median was the only risk factor for NRM (HR=3.06, p<0.001), the presence of FLT3-ITD was the only risk factor for CIR (HR=2.56, p<0.001 for FLT3-ITD/NPM1 wt , HR=2,78, p=0.001 for FLT3-ITD/NPM1 mut ). Both age >median and FLT3-ITD were associated with inferior LFS (HR=1.80, p<0.001 for age; HR=0.49, p=0.001 for FLT3-ITD/NPM1 wt , HR=0.41, p=0.003 for FLT3-ITD/NPM1 mut ) and, most importantly, inferior OS (HR=2.37, p<0,001 for age; HR=0.48, p=0.001 for FLT3-ITD/NPM1 wt , HR=0.36, p=0.002 for FLT3-ITD/NPM1 mut ). The latter data allowed to identify 3 prognostic groups with 2y OS of 85±3% (younger age AND FLT3-ITDneg), 71±2% (older age OR FLT3-ITD) and 49±5% (older age AND FLT3-ITD), p<0.0001). In contrast, donor type (sibling vs. MUD) or intensity of the conditioning (MAC vs. RIC) did not play a signifi cant role in any molecular subgroup. Discussion: We identifi ed age and FLT3-ITD as major risk factors for OS after alloSCT for CN-AML in CR1, independent from other factors as donor type or conditioning. In pts. with FLT3-ITD, the protective role for NPM1mut seems less prominent after alloSCT. An excellent outcome after alloSCT in CR1 was observed among patients with double negative genotype, the majority also being negative for CEBPA mutations. Encouraging results were found after SCT in PIF, relapse or advanced CR. Disclosure of Interest: None Declared. 1 1 Hematology/Hematopoietic Stem Cell Transplantation, 2 Radiation Oncology, 3 Biostatistics, City of Hope, Duarte, 4 HCT, Los Angeles, 5 Cytogenetics, 6 HLA, 7 Radiation Physics, City of Hope, Duarte, United States Introduction: The overall survival (OS) for relapsed (RL) acute leukemia patients or those considered induction failure (IF) treated with HCT is 16-19% (Duval et al., JCO 2010) . While randomized studies have shown a dose response relationship, with higher doses of radiation resulting in decreased relapse, this benefi t is off set by increased treatment related mortality. Materials (or patients) and Methods: To explore the safety and tolerability of targeted radiation treatment in the context of HCT, a phase I trial is being conducted in which escalated doses of targeted whole body radiation is delivered to marrow bearing and lymphoid areas, while sparing non hematopoietic (vital) organs. The transplant preparative regimen is as follows: TMLI on days -10 to -6; etoposide 60mg/kg [adj bw] on day -5 with cyclophosphamide 100mg/kg [ideal bw] on day -3. The radiation dose was started at 1200cGy delivered in 150 cGy fractions twice a day. Initially the dose of radiation was escalated in increments of 150 cGy to 1500 cGy using standard 3x3 design and then 100 cGy incrementally to a maximum of 2000 cGy using rolling 6 design. Dose limiting toxicity is defi ned according to the Bearman and CTCAE 3.0 (for hematologic toxicity) scales. Liver and brain dose was kept at 1200 cGy. Median normal organs received 16-60% of the marrow dose (lung 44%, esophagus 33% and oral cavity 28%). All patients received peripheral blood stem cells on day 0. GVHD prophylaxis consisted of tacrolimus and sirolimus. 

0.0003 0.002 0.0009 0.75 S70 WBC at HCT median 1.5 (0.1-14.9 ) % blasts (blood) at HCT median 2% (0-85%) % blasts (marrow) at HCT median 46% (10-90%) Two patients presented with extramedullary disease at time of HCT. With a median follow-up for alive patients of 9.9 months (1.0-53.4), the OS and cumulative incidence of relapse/progression at 1 year are 52% (95%CI: 32.9-68.1) and 45% (95%CI: 30.9-65.5) respectively. Two patients are currently being treated at 1900 cGy dose, MTD has not been declared. All patients treated at ≥ 1700 cGy achieved CR at the day 30 post transplant evaluation. Twenty (49%) of patients developed acute GvHD; 6(30%) of these developed grades 3-4. The day 30 and day 100 NRM was 0% and 5.3% respectively. The most common toxicity across the dose levels tested is grade 1 GI and grade 2 stomatitis (Bearman Scale). Causes of death were disease progression/persistent disease n=14, GvHD n=1 and infection n=1. Discussion: These results are encouraging and suggest that 1) doses of TMLI can be safely escalated to 1800 cGy in combination with etoposide and cyclophosphamide (MTD not reached, dose escalation will continue to 2000 cGy); 2) all patients treated with ≥1700 cGy achieved a CR at the day 30 evaluation and that 3) a reduction in relapse/progression compared to published reports can be achieved without increasing NRM using targeted whole body radiation in this high risk population. Disclosure of Interest: None Declared. (Vago et al, N Engl J Med, 2009 ). Based on this discovery, here we investigate the biologic bases of other cases of relapse after allo-HSCT. Materials (or patients) and methods: Serial AML samples at time of diagnosis, relapse after chemotherapy and relapse after allo-HSCT were collected from 9 patients. AML blasts were FACS-purifi ed and gene expression profi le was analyzed using Illumina Human HT12 arrays. Deregulated genes were identifi ed by pairwise LIMMA analysis. Gene Ontology and Gene Set Enrichment Analysis curated databases were interrogated to identify deregulated processes. Validations were performed by molecular HLA typing, locus-and allele-specifi c qPCR, immunophenotypic analysis and functional ex vivo assays. Results: Comparative gene expression profi ling of leukemic blasts harvested at diff erent time-points during the disease history of the 9 patients evidenced the selective deregulation of immunerelated processes in AML at post-transplantation relapse as compared to diagnosis and relapse after sole chemotherapy. Amongst the downregulated transcripts, we found all the genes related to the HLA Class II presentation pathway, comprising both subunits of HLA-DR, -DQ, -DP and the accessory molecules HLA-DM, -DO and CD74. Class II downregulation was confi rmed by qPCR and multiparametric fl ow-cytometry, and occurred in 6 out of the 9 patients analyzed, which interestingly were those transplanted from partially HLA-mismatched donors. The decrease in HLA Class II surface expression resulted in impaired recognition by allogeneic CD4 T cells of leukemic cells harvested at relapse, suggesting that this phenomenon might represent a novel mechanism of immune evasion. We investigated the molecular bases of Class II downregulation, fi nding no genomic alterations in the HLA or accessory genes, and the signifi cant transcriptional downregulation of the HLA Class II master regulator CIITA. CIITA expression is in turn tightly regulated through methylation of its promoter: experiments are currently ongoing to verify the methylation status of CIITA promoter at post-transplantation relapse and to investigate whether demethylating agents could revert CIITA silencing and recover HLA Class II expression in leukemia at relapse. Discussion: Our data provide further evidence on how tightly relapse and immune evasion are linked, by demonstrating that beside the previously described genomic HLA loss, also transcriptional HLA Class II downregulation can be at the basis of AML relapses after partially mismatched HSCT. Moreover, this CIITAdependent mechanism might have important clinical implications, by providing a novel biological rationale for the documented effi cacy of demethylating agents in association to donor lymphocyte infusions as salvage therapy for relapses after allo-HSCT. Disclosure of Interest: None Declared. Introduction: We recently demonstrated that ruxolitinib (INCB018424), the fi rst approved JAK1/JAK2 inhibitor for treatment of myelofi brosis (MF), exerts potent anti-infl ammatory activity. This may at least in part explain higher infection rates observed in ruxolitinib-treated patients. NK cells are critical for cancer-immune surveillance and cytokine-mediated signals are central for proper NK cell activation. We here aimed to characterize in detail the eff ects of JAK1/2 inhibiton on human NK cells. Materials (or patients) and Methods: Highly purifi ed CD56 + NK cells were isolated from human peripheral buff y coats by magnetic bead isolation and subsequently exposed to increasing concentrations of ruxolitinib (0.1-10 μM). Cytokine (1000U/ml IL-2, 25ng/ml IL-15)-induced NK cell proliferation was analyzed by CFSE dilution. Phenotypic and functional NK cell activation markers (NKp46, NKG2D, Granzyme B, CD16, and CD69) were analyzed by fl ow cytometry (including CD107 expression for degranulation). NK cell function was tested by fl ow-cytometry-based killing assays and quantifi cation of IFN-γ production upon stimulation with either MHC class I-defi cient K562 target cells or cytokines (IL-12, IL-18). In addition, phenotypic and functional analyses were also tested during NK receptor activation via plate-bound activating NKp46 antibodies. Signaling events were analyzed by Western Blot analysis to detect phosphorylation of JAK1 and JAK2 as well as by applying phospho-fl ow technology to evaluate ruxolitinib-mediated changes of cytokine-dependent signalling cascades (pS6, pSTAT1, pSTAT3, pSTAT5, pERK, pAKT, pP38, and pZAP70). Results: Our results demonstrate provide fi rst evidence that ruxolitinib profoundly aff ects cytokine-induced NK cell activation. This includes a signifi cant and dose-dependent reduction of NK cell proliferation, reduced induction of activation-associated surface markers (including NKp46, NKG2D, Granzyme B, CD16, CD69) as well as impaired killing activity against the classical NK target cell line K562. In addition, all main functional activities of NK cells are down-regulated as shown by reduced cytotoxic capacity, impaired degranulation and IFN-γ production. After wash-out, the inhibitory eff ects of ruxolitinib on NK cells are fully reversible, as shown by proper re-activation by cytokines. In contrast to cytokine-mediated NK cell activation, stimulation via the NK-specifi c receptor NKp46 are not aff ected by ruxolitinib. Of note, ruxolitinib does not aff ect NK cell viability. On a molecular level, phosphofl ow analyses revealed that cytokine associated signaling events, such as phosphorylation of STAT5 and S6 were dose-dependently reduced by ruxolitinib in primary human NK cells. Discussion: Ruxolitinib strongly inhibits NK cell activation leading to impaired proliferation and functional activity. Experiments verifying these eff ects in patients are currently ongoing and will be presented at the meeting. Our fi ndings may have important clinical implications, when considering the application of ruxolitinib as GvHD therapy, because NK cells are critically involved in the GvL eff ect after allogeneic stem cell transplantation.

Introduction: Despite the advances in the treatment of multiple myeloma using new targeted therapies and autologous HSCT the disease remains largely incurable. Several prospective studies comparing planned tandem autologous-reduced intensity allogeneic HSCT (auto-allo) to single/tandem autologous HSCT have shown no overall survival advantage despite improvements in PFS and lower relapse rates with reduced intensity allograft, mainly due to increased NRM rates. Only two prospective EBMT and Italian studies reported PFS and overall survival benefi ts in favor of the auto-allograft. Currently allogeneic HSCT is recommended within the context of clinical trials and only in high risk multiple myeloma patients who continue to have a very poor outcome with autologous HSCT. While such clinical trials are ongoing there remains a need to address the role of autologous HSCT prior to reduced intensity allogeneic HSCT. The objective of this retrospective study is to evaluate the role of upfront cytoreductive autologous HSCT prior to allograft in the outcomes of patients who have undergone reduced intensity conditioning (RIC) allograft following induction therapy. Materials (or patients) and Methods: We performed a retrospective analysis of the EBMT database on ITT basis comparing the outcomes of patients who were planned to receive auto-allograft to those who underwent early RIC without a prior autologous HSCT. From 1996 to 2013 a total of 690 patients were registered as reduced intensity allograft. Of 517 patients registered as planned auto-allograft; 472 patients received their planned allograft. A total of 173 patients received early RIC allograft without prior auto-HSCT. Median age at fi rst transplant was 53 yrs (20-72) in the auto-allo and 51 yrs (31-77) in the early RIC group. Median time from diagnosis was 6.6 mos (2-156 mos) in the auto-allo and 7.7 mos (2.7-11.9) in the early RIC group. 9 Eurocord International Registry, 10 Hematology, Hopital Saint Antoine, Paris, FranceIntroduction: Currently, an increasing number of patients with AML in fi rst complete remission (CR1) receive their allograft after reduced intensity conditioning (RIC) because of age or pre-existing comorbidities, which may increase the risk for non-relapse mortality (NRM). NRM can be predicted by the hematopoietic cell transplantation comorbidity index (HCT-CI) and the European Group for Blood and Marrow Transplantation (EBMT) score, that both take different parameters into account. We set out to compare these scores and to integrate the parameters of both scores into a novel score in patients with AML in CR1 receiving a RIC allograft. Materials (or patients) and Methods: Both the HCT-CI and the EBMT-score were assessed prior to transplantation in 812 patients with AML in CR1, who were allografted after RIC between 2000 and 2011. Both scores were evaluated individually and their parameters were subsequently included in a combined analysis with the addition of patient and donor CMV serology. Multivariable analysis was performed with endpoint NRM at 2 years. Parameters with an hazard ratio of >1.2 were incorporated into a novel score, that was further internally validated by bootstrapping. C statistics were used to assess the discriminative ability of the models. Results: The median age at transplantation was 58 (range 20-76) years. Patients received an unrelated graft in 47% of the transplants. The majority of underlying AMLs exhibited an intermediate-risk karyotype. The median follow-up time was 34 (range 3-138) months. Active infection, cardiac disease an moderate pulmonary disease were the most frequently observed comorbidities with a prevalence of more than 15%. Both the HCT-CI and the EBMT-score showed relatively weak predictive value with c statistics of 0.54 and 0.55, respectively. Reassessment of the parameters of the HCT-CI and EBMT-score allowed to identify 16 predominant parameters with weights based on their attributed hazard ratios, that were subsequently applied in an integrated score. The predominant parameters included 11 comorbidities, but also age, donor type, time from diagnosis to transplant and CMV serology of the patient and donor. The integrated score allowed to identify 3 groups with 2 year NRM estimates of 8±2%, 17±2%, and 38±4% for low-, intermediate-, and high-risk patients, respectively. The c statistic for the integrated score predicting 2 year NRM was 0.67, indicating strong predictive power. Of note, cumulative incidences of NRM continued to increase at 5 years to 12±2%, 21±2% and 44±5% in the low-, intermediate-, and high-risk subgroups, respectively. The integrated score did not predict for relapse, but proved to be associated with overall survival. Discussion: In conclusion, multivariable analysis of all parameters from both the HCT-CI and the EBMT-score allowed to develop a simple, integrated score with improved prediction of NRM in a specifi c group of patients with AML in CR1. These results suggest that prediction of NRM in subgroups of patients may be improved by a specifi c, dedicated score instead of an universal "one size fi ts all" score, that embraces diff erent diagnoses and conditioning regimens. S82 GM-CSF and G-CSF for neutropenic patients undergoing allogeneic stem cell transplantation to fi nd the antifungal eff ect of GM-CSF and G-CSF. The clinical trial was registered at www. clinicaltrial.org as NCT01232504. Materials (or patients) and Methods: Patients aged 14 to 60 years old without evidence of fungal diseases were randomized to receive once daily subcutaneously granulocyte macrophage colony stimulating factor (GM-CSF) (5-7ug/kg/d), combination of rhGM-CSF and rhG-CSF (2-3ug/kg/d each), or granulocyte colony stimulating factor (rhG-CSF) (5-7ug/kg/d) (starting on day 5 post transplant), treatment was continued until recovery from neutropenia. All patients received antimicrobial and antifungal prophylaxis with levofl oxacin 500mg per os daily and fl uconazole 200mg per os daily. Results: A total of 187 patients from 5 centers of China were evaluated for effi cacy. Baseline characteristics among the three groups were well balanced, except the GM-CSF group had more male subjects than the G+GM group (77.97% vs. 53.73%, p=0.011).The cumulative incidences of proven and probable IFD were 8.71%, 9.09% and 27.12% from 4 th week to 100 days respectively for GM-CSF, GM-CSF+G-CSF and G-CSF. GM-CSF and GM-CSF+G-CSF group had signifi cantly lower incidence of IFD (p=0.01 and 0.008 respectively) than G-CSF group. Moreover, the response rates of antifungal treatment in GM-CSF (60%) and G-CSF+GM-CSF (50%) groups were higher than that in G-CSF group (43.75%) p<0.05. The median time for neutrophil recovery to 0.5×10 9 /L was slightly prolonged on the GM-CSF arm (12.57 days) comparing with G-CSF (11.25 days) and G-CSF+GM-CSF group (11.37 days) (p = 0.042). GM-CSF and GM-CSF+G-CSF groups had shorter time of platelet recovery to 50×10 9 /L than G-CSF group (18.00 days, 18.37 days and 23.67 days respectively, p=0.027).The overall survival (OS) and disease free survival (DFS) at 100 days post transplant of GM-CSF group and GM-CSF+G-CSF group were both signifi cantly higher than those in G-CSF arm (p<0.05)(fi gure 1 and 2). After a median follow up of 601 (10~1096) days, 140 (74.47%) patients survived. The OS and DFS at 180 days, 360 days were not signifi cantly diff erent among the three groups. Infection related mortality was lower in GM-CSF arm than that in G-CSF arm (p=0.018). There were no signifi cant diff erences in relapse, GVHD, hemorrhage related mortality among the three groups. Discussion: Comparing with G-CSF, GM-CSF for neutropenic recipients of allogeneic stem cell transplantation could decrease incidence of IFD and infection related death at day 100 post transplant; Furthermore, GM-CSF could increase the OS and DFS at day 100, but did not increase acute GVHD. Our trial supports the use of GM-CSF in recipients of allogeneic stem cell transplantation during neutropenic period. Disclosure of Interest: None Declared.

Introduction: Infection and reactivation of human cytomegalovirus (CMV), Epstein-Barr virus (EBV) and adenovirus (ADV) are frequent and severe complications of hematopoietic stem cell transplantation (HSCT) and solid organ transplantation (SOT). In recent years increasing numbers of viral pathogens expanding from herpes-simplex virus (HSV), polyoma virus BK, herpes virus (HHV)-6, respiratory syncytial virus (RSV), parainfl uenza and infl uenza virus have been shown to cause substantial morbidity and mortality in the immunocompromised host. Adoptive immunotherapy with virus-specifi c cytotoxic T cells (CTLs) can eff ectively reconstitute antiviral immunity with limited acute toxicity or risk of GvHD. In patients receiving an allogeneic cord blood (CB) transplant or a transplant from a virus-seronegative donor and since donor blood is generally not available for solid organ recipients, allogeneic related or unrelated third party T-cell donors would off er an alternative option. Recent studies have shown an impaired functionality of antiviral memory T cells during and after mobilization with granulocyte G-CSF. Materials (or patients) and Methods: Frequency assessments of virus-specifi c T cells in more than 400 HLA-typed healthy donors as well as in HSCT/SOT donors using high throughput T-cell assay were performed over a period of 4 years at Hannover Medical School. To identify the most effi cient antigens for immunotherapy, we assessed the frequencies of CMV, EBV, ADV, HHV6 and BK-specifi c memory T cells in >400 HLA-typed donors using HLArestricted peptides and 17 diff erent overlapping peptide pools (2 for CMV, 2 for ADV, 9 for EBV, 2 for BK, 2 for HHV6) by short in vitro stimulation in IFN-g EliSpot. Results: For each virus, we identifi ed at least >60% potential CTL donors with highly signifi cant diff erences in frequencies of T cells against viral antigens of diff erent species as well as against diff erent viral epitopes from one species. Among the antigens tested frequencies for CMV pp65, EBV BZLF1, and HHV6 U90 peptide pools were highest. Overall frequencies of antiviral T cells detected by Tcell receptor staining were lower than those of the corresponding peptide pools. Especially in the case of ADV a donor response to a certain peptide may not be determined without prior short-term in vitro peptide stimulation. In addition, confi rmatory testing for CMV serology using western blot technique revealed approximately 15% false-positive results, possibly infl uencing future analysis and selection of potential stem cell and T-cell donors. Discussion: Third-party partially HLA-matched virus-specifi c T cells from seropositive individuals may be an option for patients receiving an allogeneic CB transplant, a transplant from a virusseronegative donor or a transplant from a cadaveric donor. The results of the study were used to establish a registry of potential T-cell donors ("T cells of interest" registry, TOI registry). In the TOI registry the donors' HLA type (class I and II high resolution), virus serology (ADV, CMV and EBV), virus-specifi c T-cell frequencies, best T-cell detection method, and results of functional and alloreactivity assays are documented. This registry of HLA-typed allogeneic T-cell donors profi led for virus-specifi c T cells will facilitate rapid availability of T cells for adoptive immunotherapy of virus-associated diseases in transplant recipients without an adequate T-cell donor. Disclosure of Interest: None Declared. Introduction: Allogeneic haematopoietic stem-cell transplantation (HSCT) is an eff ective cure for many paediatric malignant and providing protection from advanced bacterial and fungal disease. GT were well tolerated and side eff ects were minimal. The fi nding of a low incidence of Gd 3-4 acute GvHD is interesting, with GT possibly reducing translocation of lipopolysaccharide across the intestinal mucosa and therefore limiting the fi rst stage of GvHD pathophysiology. Disclosure of Interest: None Declared. Introduction: Aim of this study was to evaluate the incidence of infections and the related mortality in a group of patients who underwent an unmanipulated G-CSF-primed bone marrow transplantation from an haploidentical family donor (Haplo) compared to patients undergoing transplantation from HLA-identical sibling (ID-SIB). Materials (or patients) and Methods: We analyzed 2 groups of patients (58 Haplo and 58 ID-SIB) matched for demographic characteristics, underlying disease, conditioning regimen (myeloablative conditioning-MAC or reduced intensity conditioning-RIC) and disease status (early or advanced) at transplant. All patients were conditioned with an identical regimen consisting of the Thiotepa, Fludarabine and i.v. Busulphan combination and an identical antimicrobial prevention and monitoring. The data regarding infectious episodes and related mortality during the fi rst year after transplantation were retrospectively collected from the patients' medical records. Results: Over 1 year of observation, the number of febrile episodes and documented infections were, respectively, 297 and 270 in the Haplo group and 209 and 178 in ID-SIB group (Table) .

At 1 year, the cumulative incidence of infection related mortality (IRM) was, respectively, 17% in ID-SIB and 31% in Haplo (p=ns), while the probability of overall survival (OS) was 63% and 50% (p=ns) in ID-SIB and Haplo, respectively. By analyzing the subgroups of our patient series, 1-year IRM was higher in Haplo MAC respect to Haplo RIC (33% vs 15%; p=0.07) and in Haplo either transplanted in early or advanced phase (early: 20% vs 6%, p=ns; advanced: 48% vs 35%, p=ns). Discussion: Haplo had higher incidence of febrile and infectious episodes and experienced a signifi cant higher incidence of haemorrhagic cystitis. Haplo had a higher IRM, which could justify the lower 1-year OS in Haplo with respect to ID-SIB. Disclosure of Interest: None Declared. Introduction: Loss of immune competency due to immunosuppression of patients after transplantation might result in the reactivation of BK and JC polyoma viruses. In the absence of eff ective immunity, reactivation of BK and JC polyoma virus may lead to severe complications related to urogenital system and the central nervous system (CNS) named hemorrhagic cystitis (HC) and lethal progressive multifocal leukoencephalopathy (PML). There is an urgent need for curative treatments like adoptive transfer of virus-specifi c T cells used to fi ght cytomegalovirus (CMV) infections. Thus, identifi cation of novel, preferentially immunodominant T cell epitopes are of crucial importance. Materials (or patients) and Methods: In a mixed lymphocyte peptide culture (MLPC), we evaluated T cell response in healthy donors towards HLA-A2 restricted immuno-dominant peptides JCVp36 and p100. By ELISPOT and by tetramer staining we detected specifi c T cell responses.Results: Titration studies demonstrated that p100-specifi c T cells were of high avidity. Peptide-specifi c CD8 + T cells were successfully expanded and maintained for up to 5 weeks. P100-specifi c CD8 + T cells expressed the T cell activation marker CD137 on the cell surface.In an approach to identify novel T cell epitopes derived from BK and JC polyomavirus in an HLA independent manner, we used a set of overlapping peptides spanning whole VP1 protein. Using matrix pools and/or individual peptide stimulation we detected a number of novel T cell specifi cities. Natural processing was verifi ed by stimulation with VP1-derived virus like particles. Restriction analysis was performed using K562 cell lines as target cells transfected with various MHC class I molecules.Several novel BKV and JCV protein derived CD8+ T cell epitopes could be detected which will hitherto broaden our therapeutical armamentarium. Discussion: Our long term aims includes enrichment and isolation of immuno-dominant T cell specifi cities by MHC-Streptamer R and usage in an adoptive T cell transfer in the clinical setting. Disclosure of Interest: None Declared.[ Results: The mortality rate was signifi cantly higher in unrelated HSCT and haploidentical HSCT than HLA-matched related HSCT (P=0.000); but there was no signifi cant diff erence in mortality rate between unrelated HSCT and haploidentical HSCT (P=0.577). Besides, the mortality rate was signifi cantly higher in patients receiving HSCT before 2010 than after 2010 (P=0.010) ( Figure 1 ). During the follow-up period, 663 patients died. Out of them, the most common causes of mortality were infection (44.6%), relapse (29.0%), GVHD (5.4%), bronchiolitis obliterans (BO) (2.7%), graft failure (2.6%), diff use alveolar hemorrhage (DAH) (2.4%), transplantation-associated thrombotic microangiopathy (TMA) (2.3%) ( Figure 1 ). There was no statistical diff erence in causes of mortality among three groups, except for liver failure (P=0.023). Besides, there was also no statistical diff erence in causes of mortality between receiving HSCT before 2010 and after 2010, except for TMA (P=0.007). Furthermore, multivariate analysis showed that a high risk of disease status pre-HSCT (P=0.000), unrelated (P=0.007) or haploidentical HSCT (P=0.003), lack of chronic GVHD (P=0.000), grade 2 -4 acute GVHD (P=0.000) and receiving HSCT before 2010 (P=0.000) were all signifi cantly correlated with higher mortality rate post-HSCT. Discussion: This study for the fi rst time compared mortality rate and causes of mortality after HLA-matched related, unrelated and haploidentical HSCT, and showed that compared with HLA-matched related HSCT, unrelated and haploidentical HSCT were associated with higher mortality rate (P=0.000). Besides, our study suggested that infection, relapse and GVHD were the three most common causes of mortality after allo-HSCT. Furthermore, this study confi rmed that a high risk of disease status pre-HSCT, unrelated or haploidentical HSCT, lack of chronic GVHD, grade 2-4 acute GVHD and receiving HSCT before 2010 were all signifi cantly correlated with higher rate of mortality after allo-HSCT. Introduction: Bone loss is a common complication for survivors of allogeneic haematopoietic stem cell transplant (HSCT); osteoporosis is reported in up to 50% of transplant recipients. Multiple risk factors contribute to this bone loss including chemotherapy, radiotherapy, GVHD, gonadal failure and glucocorticoid use. Older age, poor nutrition, weight loss and physical inactivity also increases the risk for declines in bone mineral density (BMD) in this population. Previous studies have demonstrated that bone loss occurs within the fi rst 6-12 months after HSCT; the appropriate time to initiate BMD testing is not known. Delay in BMD testing may delay needed preventative strategies in patients who may already have or be at risk for osteoporosis. Few studies have addressed early changes in BMD in the 100 days post HSCT. BMD testing before and early after HSCT could identify both patients with low pre-existing bone density and those with rapid loss post transplant and allow for early interventions to prevent or reverse bone loss. Materials (or patients) and Methods: We reviewed charts of patients who underwent an allogeneic HSCT between 2011 and 2013. Included were all patients available for BMD testing between 2011 and 2013. BMD was measured at the lumbar spine, total hip, and femoral neck using dual energy x-ray absorptiometry (DXA) before HSCT and at day 100 post HSCT. Risk factors for osteoporosis were evaluated, including personal history of fracture, family history of hip fracture, alcohol, smoking, rheumatoid arthritis, weight changes, transplant conditioning therapy, and history of steroid use. Results: A total of 91 patients, 56 male and 35 female were reviewed. Mean age was 48. The hematologic diagnoses were AML: 35; ALL; 14; CML: 7; CLL: 10; NHL: 9; RAEB: 5; other: 11. At day 100 post HSCT patients experienced a mean decline of 3.27%± 4.19% in lumbar spine BMD, 4.43% ± 4.83% in total hip BMD and 4.53% ± 5.25% in femoral neck BMD (P=<0.0001 for lumbar spine, total hip, and femoral neck). The 42 patients who received glucocorticiods for GVHD had a signifi cantly greater decline in BMD at the total hip (5.79% vs. 3.27% p=0.012) and femoral neck (5.96% vs. 3.27% p=0.015) sites compared to the patients not receiving glucocorticoids. Weight changes over the fi rst 100 days post transplant were not signifi cantly associated with changes in BMD.Since least signifi cant change in BMD at a good DXA facility may be around 3%, we queried how many patients had a signifi cant decline in BMD over the 100 days post transplant. At spine, 46 patients (50.4%); at total hip 56 patients (61.5%); and at femoral neck 55 patients (60.4%) had a greater than 3% decline. Discussion: We demonstrate that HSCT recipients have significant declines in both hip and spine BMD in the 100 days postallogeneic HSCT. The declines in femoral sites imply rapid cortical bone loss as well as the expected trabecular bone loss. Activation of RANKL by local cytokines produced by the transplant marrow are likely responsible for these rapid changes. Glucocorticoid and weight loss eff ects are less prominent. Early detection of changes in BMD may help to target osteoporosis therapies to patients at the greatest risk of bone loss and eventual fracture. Disclosure of Interest: None Declared.

Introduction: Levels of cytokines, chemokines and soluble molecules fl uctuate after allogeneic hematopoietic stem cell transplantation (alloHSCT). These biomarkers are possible to be a diagnostic and prognostic value for transplantation-associated coagulopathy (TAC 

Health, Labour and Welfare in 2008. Membranous TM has many aspects to inhibit infl ammation via activated Protein C (APC), lectin-like domain, and activated thrombin-activatable fi brinolysis inhibitor as well as to inhibit coagulation by binding factor IXa and Xa through APC. We established SIGHT (SOS, sinusoidal obstruction syndrome; infection; GVHD, graft versus host disease; hemophagocytic syndrome; TMA, thrombotic microangiopathy) study group to explore the association between TAC and other complications following alloHSCT. In the present study, we investigated the eff ects of rTM on levels of the biomarkers and whether we could prevent TAC using rTM among registered patients in the SIGHT study group. Materials (or patients) and Methods: SIGHT study group covered nationwide institutes more than 70 in Japan. Patients whoever were eligible for alloHSCT to treat hematopoietic disorders by physicians decision were permitted and written informed consent was requested to register in the study. Blood samples were collected from the patients before and after transplantation through day 28. Levels of cytokines (IL-6, TNF-α, HMGB1), chemokines (MCP-1, RANTES), and soluble molecules (VCAM-1, E-selectin, PAI-1, PDMP) were measured by ELISA. rTM was administered as a prophylactic therapy for TAC in day 4-14 after alloHSCT. Control group was received heparin or no anti-coagulation therapy as prophylaxis during same schedule as rTM group. Patients were not randomized to these three arms but allowed to select one by their physicians decision. The primary endpoint was a comparison of fl uctuation of these biomarkers in three arms. Secondary endopoint was a comparison of the rate of developing TAC. Results: The subjects were 228 patients who underwent alloHSCT in SIGHT study group. 183 patients out of them were analyzed to measure level of these biomarkers. There was no significant diff erence in sex, age, donor source, and the proportion of total body irradiation received as conditioning regimen between two groups characteristic. MCP-1 and IL-6 exhibited more signifi cant elevations on day 7 after alloHSCT, while HMGB1 did on the day of alloHSCT. In contrast, the levels of TNF-α, sE-selectin, sVCAM-1, PAI-1 and PDMP exhibited increment since around day 7 after alloHSCT. Signifi cant improvements in TNF-α, sE-selectin, sVCAM-1, HMGB1, PAI-1 and PDMP was shown after rTM-treatment, but not shown in control group. Regarding complications following alloHSCT, the rate of developing acute GVHD was signifi cantly lower in rTM group than without it (12.7% vs 30.6%, p=0.0368), while SOS or TMA did not diff er (9.9% vs 11.1%, p=0.8260 ). Discussion: The present fi ndings suggest that rTM has the possibility to play a therapeutic role for TAC following alloHSCT contributing to ameliorate acute GVHD. Endothelial and pro-coagulant biomarkers were released delayed compared with infl ammatory biomarkers. Further studies are needed whether rTM should administer later or longer to improve SOS/TMA than the present duration after alloHSCT. Disclosure of Interest: None Declared.