ALBERT

Description: Introduction: Cytomegalovirus (CMV) reactivation and disease are important risk factors after allogeneic hematopoietic stem cell transplantation (allo-HSCT), and strongly affect morbidity and mortality after transplant. CMV-specific T cell reconstitution controls CMV reactivation and protects against serious adverse events but a protective level of CMV-specific T cell response or standardized method for its monitoring have not been yet determined. Methods: We designed a prospective, single-center observational study to assess if the kinetic and quality of CMV specific T-cell reconstitution impact the incidence and severity of CMV reactivations. We have enrolled 84 consecutive patients affected by hematological malignancies receiving allo-HSCT followed by Cyclophosphamide and Rapamycin between December 2017 and February 2019. Here we report preliminary data on the first 61 patients. Patients received allo-HSCT from family (siblings=10; HLA haploidentical=24), unrelated HLA-matched (n= 24) donors or cord blood (CB, n=3). The CMV serostatus of host (H) and donor (D) pairs was: H+/D+(n=40, 65%), H+/D-(n=20, 33%) and H-/D+ (n=1, 2%); H-/D-(7% of the overall transplanted population at our center) were excluded. CMV DNAemia was assessed weekly in whole blood (WB). Absolute numbers of polyclonal and CMV-specific T cells were quantified by flow cytometry using Troucount™Tubes (BD) and Dextramer®CMV-Kit (Immudex), respectively, in the graft and fresh WB at days -7, +30, +45, +60, +90, +120, +150, +180 and +360. Dextramer CMV kit includes reagents for the identification of CMV-specific lymphocytes restricted for several HLA class I molecules: A*01:01/*02:01/*03:01/*24:02 and B*07:02/*08:01/*35:01. These alleles allowed the longitudinal evaluation of 54 out of 61 (89%) patients. Results: At a median follow-up of 226 days post-HSCT, 31 (57%) patients experienced a CMV-related clinically relevant event (CRE, median +63 days), including 8 patients (15%) with CMV disease (median +59 days). Univariate analyses showed that the incidence of CMV clinically-relevant reactivation (CRE) was influenced by H/D CMV serostatus (0.90 in H+/D- versus 0.44 in H+/D+pairs, p=0.015) and by previous acute Graft-versus-Host Disease (aGvHD) requiring systemic immunosuppression (0.82 in aGvHD grade II-IV versus 0.52 in aGvHD grade 0-I, p=0.051). The disease status at transplant, the donor type (HLA-matched versus HLA-haploidentical/CB donors), donor's or host's age did not significantly affect the probability to develop CRE. For each time-point, we compared the absolute number of CMV-specific lymphocytes in patients experiencing or not a subsequent CRE. Our data demonstrate that higher levels of CMV-specific CD8+T cells in the donor apheresis and at +45 days after allo-HSCT are associated with reduced risk of subsequent CRE (median CMV-specific CD8+cells/kg in the apheresis=5x103in CRE-positive patients (CRE+) and 5x105in CRE-negative patients (CRE-), p=0.012; median CMV-specific CD8+at +45 days=0.14 cells/μL in CRE+and 1.21 cells/μL in CRE-, p=0.034). Furthermore, patients with any Dextramer positivity at +45 days displayed a lower incidence of CRE compared with subjects who were negative (CRE probability: 0.5 vs 1.0, p=0.003). Conversely, the absolute number of neither polyclonal CD3+CD8+T lymphocytes nor total CD3+T cells correlate with subsequent CRE. Taking advantage of the HLA mismatched-HSCT setting, we then dissected CMV-specific T-cell response according to HLA restriction elements (H/D=shared n=45, D-restricted n=14, H-restricted n=11). In H+/D+pairs, we observed a fast and similar kinetic of reconstitution of CMV-specific lymphocytes restricted by H/D and D HLAs. Conversely, in H+/D-pairs, we detected only CMV-specific CD8+lymphocytes restricted for H/D haplotypes. Host-restricted cells remained undetectable for the first 180 days after HSCT. Conclusion: Early after allo-HSCT and in the donor apheresis, the level of CMV-specific CD8+T cells measured by Dextramer staining differs in patients experiencing or not subsequent CRE. Furthermore, our findings indicate that CMV reactivations can prime H/D-restricted T cells presumably educated in the donor thymus; conversely, D- and H-restricted donor-derived lymphocytes have not yet undergone neither cross-priming nor thymic education respectively. Disclosures Brix: Immudex: Employment. Bonini:Kite/Gilead: Consultancy; Intellia Therapeutics: Consultancy; Intellia Therapeutics: Research Funding; Novartis: Consultancy; GSK: Consultancy; Allogene: Consultancy; Molmed: Consultancy; TxCell: Consultancy; -: Patents & Royalties: Adoptive T cell therapy field.

Description: Introduction Allogeneic stem cell transplantation (allo-HCT) survivors are at a defined relevant risk of developing long-term complications: the prevalence of chronic health conditions approaches 75% among HCT survivors. The endocrine system is one of the most frequent targets of complications, providing justification for a long-term and continuous follow-up (LTFU) to assure a timely and appropriate treatment. The aim of our study is to evaluate the incidence of endocrinopathies in survivors in respect of sex, age, donor type, conditioning regimen and GvHD occurrence. Methods A standardized LTFU is applied at our center. We here analyze data consecutively collected in an Institutional database, starting from 2006, including 402 adult patients (pts) who underwent an allo-HCT between 1992 and 2016 at our Institution. A written consent was given by pts allowing the use of medical records for research in accordance with the Declaration of Helsinki. We considered for the analysis pts with an overall survival 〉/=1y. We reviewed pts chapters with a focus on occurrence, management and treatment of diabetes, thyroid disfunctions, dyslipidemia and osteopenia / osteoporosis: diagnosis and follow-up were performed according to guidelines for long-term HCT survivors. Results With a median follow-up of 7y (r 2-25y) 328 pts were evaluable; donor was a match unrelated donor in 107 cases, HLA identical sibling in 88, haploidentical relative in 129 and cord blood in 4. The 5y-incidence of diabetes type 2 was 3%, with a median time after allo-HCT of 1138 days (r 5-4181 days); 13/22 developed diabetes after diagnosis of GvHD (median time 884 days, r 30-3753 days). All pts received indication for diet modification, 11 pts were treated with insulin and 8 pts with metformin. Thyroid disfunction was documented in 38 pts (5y-incidence 8,5%): 2 pts were diagnosed with hyperthyroidism and treated with methimazole or radioiodine treatment. Hypothyroidism was documented in 36 pts (median time after allo-HSCT 799 days, r 65-5021 days). Thirteen pts developed hypothyroidism following the diagnosis of GvHD (median time 1236 days, r 166-3540 days). Only 2 pts did not receive a specific treatment, while all the others received substitutive therapy with levothyroxine. Furthermore 1 pt was diagnosed with a papillary thyroid cancer. The 5y-incidende of dyslipidemia was 30% with a median time after allo-HSCT of 1433 days (r 366-7629), 47 pts developed dyslipidemia after the diagnosis of GVHD (median time 1425 days, r 134-7403 days). Diet-therapy was recommended to all the pts, 29 pts received a statin-based pharmacological treatment, 20 pts a polyenoic-fatty-acids based treatment, while a nutraceutical compound was given in 13 pts. Osteopenia was documented in 120 pts (median time after allo-HSCT 994 days, range 31-6605 days) with a 5y-incidence of 36%. Seventy-nine pts presented osteopenia after diagnosis of GvHD (median time after GvHD diagnosis 707 days, r 13-6379 days). Eight pts did not receive a specific treatment. Two pts received treatment with biphosphonates plus oral vitamin D and calcium supplementation, the 110 remaining pts received oral vitamin D +/- calcium supplementation only. Sixty-four pts developed osteoporosis (median time after allo-HSCT 1000 days, r 60-8836 days), the 5y-incidence was 26%. Forty-four pts developed osteoporosis following the diagnosis of GvHD (median time 724 days, r 28-8280 days). Only 4 pts did not receive any specific therapy; 31 pts received therapy with bisphosphonates, 2 pts denosumab and 27 pts oral vitamin D and calcium supplementation. In univariate analysis no relationship between host sex, age at transplant, TBI exposure, donor or history of GvHD and development of diabetes, thyroid disfunction and dyslipidemia was outlined. Otherwise, osteopenia development was strongly associated with GvHD occurrence and osteoporosis was strongly associated with age, sex and GvHD occurrence (table 1). Conclusions Allo-HCT survivors are at relevant risk of endocrinopathies after transplantation, providing justification for specific monitoring to individualize treatment and follow-up. Of note, classical transplant-related variables are not enough to justify the occurrence of endocrinological disfunction: a further deeper evaluation of a misdiagnosed donor-mediated autoimmune predisposition will be essential. Disclosures Bonini: Intellia Therapeutics: Research Funding.

Description: Abstract 203 INTRODUCTION: Occurrence of a robust Graft versus Leukemia (GvL) effect is at the basis of the curative efficacy of Hematopoietic Stem Cell Transplantation (HSCT). Although leukemia-specific antigens have been characterized, most of the antileukemic potential of transplantation resides in alloreactivity towards patient-specific antigens, such as minor and major histocompatibility antigens. This is particularly relevant in the context of transplantation from related haploidentical and matched unrelated donors (MUD), in which mismatched HLA molecules are potent targets for alloreactive donor T cells. Still, upon in vivo selective pressure by donor T cells, acute myeloid leukemia can undergo genomic rearrangements which result in loss of the patient-specific HLA haplotype, a mechanism which our group recently demonstrated to be frequently responsible for leukemia relapse after haploidentical HSCT (Vago et al. NEJM, 2009). METHODS: 103 patients who underwent a partially HLA-matched transplantation for Acute Myeloid Leukemia (AML), MyeloDysplastic Syndrome (MDS), or Chronic Myelomonocytic Leukemia (CMML) at the San Raffaele Hospital from 2002 to present, were included in our analysis. For 67 patients, 26 of whom transplanted in complete remission, the stem cell donor was related haploidentical. For the remaining 35 patients, 23 of whom transplanted in complete remission, the donor was unrelated and mismatched for an average of 2/12 HLA alleles. All patients received donor T cells as part of the transplantation protocol. Post-transplantation follow-up comprised monthly bone marrow examination, with Short Tandem Repeat (STR) chimerism analysis and HLA typing performed in parallel on the marrow aspirate samples. The same analyses were performed also for an additional patient, referred to our attention for relapse of AML after two MUD transplantations performed in another center, both from HLA-C and DPB1-mismatched donors, and several donor lymphocyte add-backs. In cases of relapse with suspected loss of the mismatched HLA alleles, STR chimerism and HLA typing were performed also on purified leukemic blasts. RESULTS: Disease relapse occurred in 28/67 and 8/35 patients after haploidentical or MUD transplantation performed at our center, respectively. After haploidentical transplantation, 10/28 relapses (35.7%) were due to mutated leukemic blasts which had lost the patient-specific HLA haplotype. Median time to relapse in these patients was 281 days after HSCT (range 67–390), and all patients had received a high dose of donor T cells (median 289×106 CD3+ cells/kg, range 90–583). Upon detection of the mutated leukemic blasts, five of these ten patients were enrolled to receive a subsequent transplantation from a different stem cell donor, mismatched for the remaining HLA haplotype. None of the 8 relapses after MUD transplantations performed at our center displayed loss of the mismatched HLA. However, in the additional patient referred to our attention after two transplantations from partially HLA-mismatched unrelated donors, we documented loss of the HLA haplotype carrying mismatched alleles in the leukemic blasts responsible for leukemia relapse. CONCLUSIONS: With respect to our original series of five relapses with leukemic loss of the mismatched HLA haplotype, we describe here five additional cases based on the same mechanism, further consolidating the utmost clinical relevance of this escape mechanism from the GvL effect mediated by haploidentical donor T cells. Prompt detection of the mutant leukemic blasts allowed us to avoid infusion into the patients of donor T cell add-backs, predictably inefficacious in controlling these relapses, and to quickly enroll these patients into a salvage transplant from a different donor, mismatched for the remaining HLA haplotype. Moreover, we demonstrate that even in the case of fewer HLA mismatches, such as in MUD transplants, the genomic rearrangement we described can lead to clinical relapse. Given the constant increase in the number of transplants performed from partially HLA-mismatched donors worldwide, we recommend that post-transplantation HLA typing of bone marrow samples should be routinely performed at disease relapse to detect mutant leukemic blasts, and to guide therapeutic strategies targeted against them. Disclosures: Bonini: MolMed S.p.A.: Consultancy.

Description: Abstract 4596 Background: consolidation of complete remission (CR) with high-dose chemotherapy and autologous (auto) stem cell transplantation (SCT) can be an alternative option for AML and MDS patients (pts) when allogeneic SCT is not feasible. Elderly pts. are not usually offered a SCT for the high risk of letal toxicities; moreover, their long term survival is less then 10–15% also when a CR after induction chemotherapy has been obtained. New treatment strategies are warranted to improve the outcome of these poor prognosis pts. Treosulfan can induce a myelotoxicity comparable to busulfan, without CNS and mucosal toxicities, known to occur with busulfan. The fludarabine-cytarabine (FLA) association is effective as induction treatment of acute myeloid leukaemia (AML) and myelodisplastic syndrome (MDS). We have combined high-dose treosulfan with FLA in a new conditioning regimen prior to autoSCT for AML and MDS elderly pts. Aim: evaluation of preliminary data on toxicity and efficacy of a treosulfan-based conditioning regimen prior to autoSCT in MDS and AML pts. Methods: period 7/2006 to 2/2010, 10 pts, median age 68 (60-76). Diagnosis and risk factors: 5 AML with MDS-related changes (1 complex and 4 normal karytotype), 1 AML with minimal differentiation, 1 AMML with hyperleucocytosis (〉100.000/microl), 1 AMML with normal karyotype and 1 MDS (RAEB2) with del5q. Chemo cycles before autoSCT: median 2 (range 2–3), at least 1 containing HDara-C (8 gr/sqm total dose). All pts were in CR at autoSCT, 57 days (median, range 44–132) after first CR documentation. Conditioning regimen (FLAT): treosulfan 10 gr/sqm for 3 days, fludarabine 30 mg/sqm for 5 days, cytarabine 2 gr/sqm for 5 days, PEG filgrastim 1 s.c. vial after autoSCT (8 pts). Transplant: PBSC, median 6.2×10e6 CD34+/kg BW (range 3.8–9). Results: the overall median toxicity grade (CTC) was 1 (0-3), with hepatic (3 cases) and cardiac (2 cases of reversible atrial fibrillation) involvement. No deaths due to treatment occurred. Grade 4 neutropenia and thrombocytopenia median duration was 12 (range 9–37) and 27 (range 11–71) days, respectively. Grade 4 neutropenia in pts who received PEG filgrastim lasted 10.5 days (range 9–25). Four out of 10 pts did not experience either fever or sepsis. After autoSCT, 5 pts relapsed (4 haematologic, 1 only cytogenetic) and 4 of them died. Six pts are currently alive, 5 in first CR and 1 in second CR (after allogeneic SCT for cytogenetic relapse) 1025 days (median, range 60–1448) after autoSCT. Overall, EFS from autoSCT was 368 days (median, range 60–1448) with a 45% probability of survival at 3 years. Conclusions: the FLAT regimen is myeloablative, as demonstrated by the prolonged cytopenia in all pts. It proved to be feasibile anyway, remarkably in these elderly pts. Furthermore, prolonged survival free from disease was achieved. This new regimen is promising for AML/MDS pts as it could be an effective option to intensify the consolidation of their initial CR, with limited extra-hematologic toxicites. A phase II study in pts older than 65 is ongoing at our Center to confirm these data. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 4438 BACKGROUND: Graft Failure (GF) occurs in 5–27% of patients (pts) after allogeneic hamatopoietic stem cell transplant (HSCT) and is associated with high morbidity and mortality related to infections and hemorrhagic events. Graft function may be poor as result of graft rejection, primary disease relapse or Poor Graft Function (PGF). The incomplete recovery of blood counts is defined primary PGF and the decreasing blood counts after successful engraftment secondary PGF. Several factors may determine GF: disease risk and status, conditioning regimen, HSC source, HLA compatibility, T cell content, immunosuppression, GvHD, viral infections, drugs. GCSF and Rhu-EPO are readily available and effective in PGF but with no effects on platelets. Second transplantation from the same donor, with or without conditioning therapy, can boost the haematopoietic recovery in pts with GF. Unfortunately, both a second peripheral CD34+ mononuclear cells (MNC) mobilization and a marrow harvest in the operating room may be contraindicated early after the first donation as not safe for donors. Intrabone SCT can overcome the risk of graft failure even with a low number of CD34+ MNC, as it has been demonstrated in cord blood transplant. Here we investigate in three adult pts with GF a bone-to-bone boost (BBB) with a small marrow harvest from respective donors, unfit for a second conventional donation. AIM: to evaluate the feasibility of the BBB technique in 3 pts with graft failure. METHODS: pts were 2 males (57, 53 y) with PGF with a diagnosis of AML and CMML, respectively, and a female (44 y) with graft rejection and AML relapse. In the first two patients prolonged pancytopenia and hypoplastic marrow were documented, with diagnosis of primary PGF and secondary PGF, respectively, donor chimerism ranging from 80–100% (STR and HLA), without evidence of leukemia. In the third patient, after prolonged pancytopenia an AML relapse was documented with 89% blasts on bone marrow aspirate. In PGF patients no conditioning regimen was administered before the boost at day 30 and 72 after SCT, respectively. In the patient with AML relapse Melphalan 200 mg/mq was given 48 h before the infusion, at day 35 after SCT. The 3 donors were related, haploidentical. For the BBB procedure small quantities of bone marrow (〈 200ml) were collected from the posterior iliac crest bilaterally of the donors, at the bedside, during deep sedation and analgesya. Shortly after the unmanipulated marrow harvested was infused in superior-posterior iliac crest mono- or bilaterally, depending on the volume, during deep sedation and analgesya. In pt 1 Mononucleated cell (MNC) dose was 0.9 × 10^8/Kg for a volume of 166 ml. In pt 2 MNC dose was 0.4 × 10^8/Kg for a volume of 88 ml. In pt 3 MNC dose was 0.3 × 10^8/Kg for a volume of 140 ml. RESULTS AND CONCLUSION: In this cases the BBB technique proved feasible and safe for both the donor and the patient. Patient 1 received a second PBSC boost, without conditioning, 3 months after the BBB, and he's now alive, in CR, 13 months after the first transplant. Patient 2 died 3 months after the first transplant for pneumonia and sepsis. Patient 3 is alive, in CR, 4 months after the first HSCT. This practice can give the chance of HSC boost to patients with GF without the need of a GCSF mobilization for donors, with a minimal invasive operation. This can give the option to overcome and resolve infectious and hemorrhagic complications, bridging patients to further therapies and procedures. The intra-bone SCT may be a facilitating tool for microenvironment reconstitution, seeding and subsequent differentiation and may as well have a tolerogenic effect, through the mesenchymal stromal cells infused with the harvest. Further studies are necessary to assess the efficacy of this procedure. Disclosures: No relevant conflicts of interest to declare.

Description: Evolution of the chronic lymphocytic leukemia (CLL) clone upon treatment with the BTK-inhibitor ibrutinib has never been systematically assessed in a prospective way, though changes in the tumor genetic composition have been anecdotally reported by retrospective backtracking of mutations acquired in ibrutinib-refractory patients. The IOSI-EMA001 observational study (NCT02827617) enrolls patients treated with ibrutinib in the clinical practice, and aims at prospectively assessing the dynamics of the clonal architecture of high risk CLL upon therapy with ibrutinib by taking advantage of leukemia samples homogeneously collected at pre-specified timepoints during treatment course. Ibrutinib induces a transient lymphocytosis by displacing CLL cells formerly residing in the tissue compartments into the peripheral blood (PB). By analyzing samples collected after two weeks of ibrutinib treatment, here we tracked early genetic changes of the CLL clone occurring at the time of redistribution lymphocytosis. This analysis includes the first 10 patients enrolled in the study (median age=76 years; Binet A=10%, B=60%, C=30%; IGHV unmutated=60%, mutated=20%, NA=20%; 17p deletion=40%). PB samples were collected before ibrutinib treatment start (day -28 to 0) and after two weeks of treatment (week 2, +/-3 days). Tumor genomic DNA was isolated from FACS sorted CLL cells (CD19+/CD5+ elements; 〉99% purity). T cells were also sorted as source of germline material to confirm the somatic origin of mutations and filter out sequencing noise. The HaloPlex High Sensitivity library preparation protocol and ultra-deep next generation sequencing (coverage ~40.000x) on MiSeq (Illumina) were used to track TP53, BTK and PLCG2 mutations. By uniquely molecular barcoding each DNA library fragment, and by targeting both DNA strands, this approach allowed to reach a sensitivity of 10E-4. The CAPP-seq library preparation protocol and deep next generation sequencing (sensitivity ~5x10E-3) were used to track mutations of a panel of genes known to be recurrently affected in CLL (ASXL1, ATM, BIRC3, BRAF, EGR2, FBXW7, IKZF3, IRF4, KRAS, MAP2K1, MGA, MYD88, NFKBIE, NFKB2, NOTCH1 including 3'UTR, PAX5 enhancer, POT1, RPS15, SAMHD1, SF3B1, TP53, XPO1, ZMYM3). At baseline before treatment start, 33 non-synonymous somatic mutations (allele frequency: 0.65-99%) were identified in 9 of the 10 cases (one patient was devoid of mutations of the above-mentioned genes). The mutational profile was consistent with that expected in a high risk population meeting the current indication for ibrutinib treatment in the clinical practice(TP53=60%; NOTCH1=40%; MGA=30%; ZYMYM3=20%; ATM=10%; BIRC3=10%; EGR2=10%; IRF4=10%; IKZF3=10%; MYD88=10% NFKBIE=10%; SF3B1=10%; XPO1=10%) (Fig. 1A). Before treatment start, high sensitive ultra deep next generation sequencing did not disclose any non-synonymous somatic mutation of BTK and PLCG2 in CLL cells circulating in the blood. The genetic composition of the circulating leukemia clone did not significantly change after two weeks of treatment despite redistribution of CLL cells in the blood compartment (median lymphocyte count at baseline 50.9x10E9/L vs 77.5x10E9/L at week 2) (Fig 1B). Minor genetic changes between baseline and week 2 included the appearance of a small clone mutated in SAMHD1 and the disappearance of a small clone mutated in BIRC3 in 2 different patients. At week 2, no mutations of BTK and PLCG2 became evident in the circulating compartment. With the limitation of the current sample size, these data suggest that redistribution lymphocytosis does not mobilize CLL high risk clones harboring mutations of clinical relevance into the blood compartment. Figure 1 Figure 1. Disclosures Rossi: Gilead: Honoraria, Research Funding; Janssen: Honoraria; AbbVie: Honoraria.

Description: Abstract 46 The broader application of haploidentical stem cell transplantation (haplo-HCT), is limited by the delayed immune reconstitution (IR) secondary to the procedures for GvHD prophylaxis. This ultimately results in a high-rate of infectious complications and non-relapse mortality. We dynamically analyzed immunoreconstitution (IR) in patients undergoing haplo-HCT for acute leukemias enrolled in two different phase I-II clinical trials aimed at improving IR. In the first trial (TK007), 28 patients (out of 50 enrolled) received suicide-gene transduced donor T cells at day +42 after a T-cell depleted graft, in the absence of post-transplant immunosuppression. In the second trial (TrRaMM), 40 patients received an unmanipulated graft and a rapamycin-based GvHD prophylaxis. T-cell immune reconstitution was more rapid in TrRaMM than in TK007 patients, with a threshold of CD3 cells〉100/μl reached at days +30 and +90, respectively. In both trials IR was mainly composed of Th1/Tc1 lymphocytes with an inverted CD4/CD8 ratio. While in TrRaMM patients we observed an early expansion of naïve and central memory T cells, producing high amounts of IL-2, in TK patients IR was mainly composed of activated effectors. Furthermore, in TrRaMM patients we detected high levels of CD4+CD25+CD127- T regulatory cells (up to 15% of circulating T lymphocytes) that persisted after rapamycin withdrawal, and was significantly superior to that observed in TK patients and in healthy controls. Interestingly, in contrast to the different kinetics of T-cell reconstitution, no differences were observed in time required to gain protective levels of CMV-specific T cells, as shown by ψIFN ELISPOT analysis. Protective frequencies of CMV-specific lymphocytes were observed 3 months after HCT in both groups, a time-point that in TrRaMM patients corresponds to the average time of rapamycin withdrawal. In both trials the number of circulating CMV-specific T cells was inversely correlated to the number and severity of subsequent CMV reactivations and days of antiviral therapy. GvHD was diagnosed in 16 TrRaMM patients (40%) and in 10 TK patients (35% of patients who received TK cells). Severity of GvHD was different in the two cohort of patients with 5 TrRaMM patients (12,5%) and only 2 TK patients (7%) with grade III-IV GvHD. Of interest, in the TrRaMM group CMV-specific immunity was significantly hampered by the immunosuppressive treatment required to treat GvHD. On the contrary, in the TK group, the administration of ganciclovir was able to activate the suicide machinery and control GvHD without impairing viral-specific T-cell immunocompetence. These results matched with the kinetics of CMV reactivations. We observed that while in TrRaMM patients 80% of viral reactivations occurred after the immunosuppressive therapy, in TK patients no significant differences could be assessed before and after therapy. IFN-ψ ELISPOT might thus be a relevant and predictive test to guide patient-specific clinical monitoring and antiviral treatment. Overall, these results show that early immune reconstitution can be promoted in haplo-HCT by different strategies associated with a wide range of alloreactive potential. The risks and benefits associated with alloreactivity should guide the therapeutic choice tuned on patient disease status and co-morbidities. Disclosures: Bordignon: Molmed Spa: Employment.

Description: Abstract 2385 Background. Allogeneic haemopoietic stem cell transplantation (HSCT) is the best therapeutic option for high-risk hematological disease (HRHD), particularly acute myeloid and lymphoblastic leukemia (AML/ALL). A widely application of HSCT is limited by lack of availability of a suitable donor for every candidate patients (pts). In order to offer a donor to every candidate pts, several centers had developed in the last decade alternative strategy of HSCT, such as umbilical cord blood (UCB) or family haploidentical HSCT (haplo-HSCT). With the aim to treat high-risk leukemia in the ideal appropriate time by allogeneic HSCT, we adopted a policy relying upon HLA typing at entry of every pts diagnosed with HRHD, followed, in absence of a MRD, by the prompt activation of the MUD search. Patients lacking a MRD or a MUD at the appropriate timing are proposed for haplo-HSCT. Methods. Here we report an intention-to-treat analysis of alternative donor HSCT at our Institution for pts diagnosed with high risk AML or ALL. Data were obtained from local institutional database. Results. Between Jan-2004 and July-2010 241 pts (median age 48y, r 15–72) with diagnosis of ALL (60 pts; median age 33y, r 15–64; male 39) or AML (181 pts, median age 51y, r 19–72; male 83) were defined as “high risk status” and received an indication to HSCT according to EBMT (European Group for Blood and Marrow Transplantation) and Northern Italy Leukemia Group (NILG, www.nilg.it) recommendations. In the ALL group 3 pts died before proceed to HSCT, 5 pts are waiting for the identification of a donor. Overall, 52/60 (86%) pts underwent HSCT, the status of disease at transplant was of complete remission (CR) for 27 pts, persistence of disease (PD) for 25. In the AML group 6 pts died before proceed to HSCT, 15 pts are waiting for the identification of a donor. Overall, 160/181 (88%) pts underwent HSCT, the status of disease at transplant was of CR for 97 pts, PD 63. Overall 92 pts activated the research of a MUD, 42 proceed to transplant, 7 received a UCB, 26 received a haplo-HSCT due to absence of a suitable donor in the appropriate time frame or failure to met the criteria to engage a MUD donor. The median time from diagnosis to registry activation was 69 days (r 5–876), the median time from activation to transplant 84 days (r 28–348). In the group of pts in CR at transplant, 37 underwent HSCT from a MRD and 36 from a MUD, 4 pts received a UCB and 47 a haplo-HSCT. Seventy-two pts are alive and in CR at last follow-up, 3/72 after a second transplant from a different donor (haplo-HSCT) and 3/72 after chemotherapy and adoptive immunotherapy to treat hematological relapse. Fifty-two pts died and the causes of death were: relapse (27), infection-related (19), graft-versus-host-disease (GvHD; 5), acute myocardial infarction (1). In the group of pts in PD at transplant, 16 underwent HSCT from a MRD and 6 from a MUD, 3 pts received UCB and 63 a haplo-HSCT. Fifteen pts are alive, 14 in CR, 1 pts under adoptive immunotherapy; 1/14 pts received a second transplant from a different donor (haplo-HSCT) to treat hematological relapse. Seventy-three pts died and the causes of death were: relapse (43), infection (23), GvHD (7). Overall, the median survival is 382 days and the median follow-up for pts alive is 548 days. The 1-year overall survival (OS)/transplant related mortality (TRM)/relapse incidence (RI) are 50,76% 26,96% and 40% respectively. For pts transplanted in CR the 1y OS/TRM/RI are 77,2%, 20,01% and 25% respectively. The outcome analysis per donor source (MRD vs MUD vs haplo-HSCT) is comparable (p=ns). For pts transplanted in PD the 1y OS/TRM/RI are 19,7%, 41% and 67% respectively. The outcome analysis per donor source (MRD vs MUD vs haplo-HSCT) shows a trend of lower RI and TRM in the haplo-SCT vs MRD. Conclusion. The policy adopted provided an allogeneic HSCT in 〉 80% of candidate high-risk acute leukemia patients. No significant differences were registered in outcome for patients transplanted from matched-related, unrelated or family haploidentical donors. Further evaluation and a long-term follow-up will add important evaluation in term of long-term disease control and long-term toxicities. Disclosures: Bonini: MolMed: Consultancy. Bordignon: Molmed: Employment.

Description: Background Antithymocyte globulins Fresenius (ATG-F) are purified, concentrated preparations of polyclonal immunoglobulin G from hyperimmune serum of rabbits immunized with human thymus activated lymphocytes. These immunoglobulins induce immunosuppression through T-cell depletion and immune modulation. The polyclonal nature of ATG-F is responsible for its effects on the immune system: T-cell depletion in blood and peripheral lymphoid tissues through complement-dependent lysis and T-cell activation and apoptosis; modulation of molecules involved in leukocyte-endothelium interactions; induction of apoptosis in B-cell lineages; interference with dendritic cells. ATG-F administered before allogeneic hematopoietic stem cell transplantation (allo-HSCT) reduces the risks of graft rejection and graft-versus-host disease (GvHD), but the slow clearance of the xenoserum might delay immune reconstitution, increase the risk of disease relapse and impair the activity of a donor lymphocyte infusion (DLI) performed early after allo-HSCT. Methods We studied 24 patients with hematologic malignancies, who underwent allo-HSCT from familiar or unrelated donors, after a conditioning regimen based on myeloablative treosulfan and fludarabine. As graft rejection and GvHD prophylaxis, 17 patients received ATG Fresenius at the dose of 10 mg/kg over 16 hours at day -4,-3 and -2 before HSCT; 5 patients received in vivo T-cell depletion at the dose of 20 mg/kg at day -4,-3 and -2 before HSCT; 2 more patients received ATG Fresenius at the dose of 10 mg/kg much earlier before allo-HSCT (day -14, -13 and -12) since the treatment protocol included a donor DLI at day +3 after transplantation. We collected serum samples at different timepoints, from the first ATG dose to at least 3 weeks after HSCT. We used a flow cytometry-based assay to detect the concentration of free T-cell specific rabbit IgG (SRIgG) which corresponds to the serum biological activity against human T-lymphocytes, as opposed to the levels of unspecific rabbit IgG (RIgG) by ELISA, which lack anti-T-cell function. Results In our cohorts of patients we observed a concentration peak at 64 hours after the first ATG administration, corresponding to the end of the last immunoglobulin dose administration. Interestingly, patients who received the 20 mg/kg dose of serum reached four times higher levels of SRIgG, suggesting a non-linear correlation between the administered dose and the measured plasmatic peak concentrations. Moreover, the terminal elimination half life of SRIgG is significantly shorter than the one of RIgG: 14 vs 67 days (data not shown), indicating that, for the dose of 10 mg/kg, 10 days after HSCT, SRIgG titre has already reached sub-therapeutic levels. Conclusions Previous results indicated that in vivo specific activity of rabbit ATG Fresenius disappears from circulation around day +7 after their last administration (at the dose of 10 mg/kg, day -4, -3, 2). Based on our current data, we can suggest that residual specific anti-lymphocyte activity is dependent upon i. timing and ii. dosage of ATG administration, although this correlation might not be linear. According to these results, current policies of a fixed schedule for DLI infusion should be revisited and possibly adapted to each patient, based on specific conditioning protocols. Disclosures: Seitz: Fresenius Biotech GmbH: Employment. Martinius:Fresenius Biotech GmbH: Employment.

Description: Introduction: Post transplant cyclophosphamide (PT-Cy) has recently emerged as a very promising pharmacological strategy to overcome human leukocyte antigen (HLA) barriers in the setting of haploidentical hematopoietic cell transplant (HCT). We recently reported a promising preliminary experience on the use of PT-Cy and sirolimus as graft-versus-host disease (GVHD) prophylaxis in matched allo-HSCT (Greco R et al, Blood 2016). Herein we describe long-term outcomes of matched allogeneic HSCT, using treosulfan-based conditioning, and GVHD prophylaxis with PT-Cy and sirolimus. Methods: In our center, we collected 104 adult patients (pts) receiving matched HSCT for high-risk hematological malignancies, mainly acute myeloid leukemia (n=43). Donor was matched related (MRD) for 45 pts, 10/10 matched unrelated (MUD) for 39 pts and 9/10 MUD for 20 pts. Median age was 48 years (range 19-78). At HSCT, 51% of patients were not in complete remission (CR), 39% were in CR1 and 11% in subsequent CR. Graft source was mainly PBSCs (95%). All pts received a conditioning regimen based on treosulfan and fludarabine; 89% received an intensified conditioning with the addition of melphalan. All pts received PT-Cy (50 mg/kg/day) on days 3 and 4. Sirolimus was given from day 5, and withdrawn 3 months after HSCT. Mycophenolate mofetil (MMF) was added from day 5 to day 30, only in MUD. All patients were treated according to current institutional programs upon written informed consent for transplant procedures. Results: Median follow up was over 16 months (range 3-51). Median CD34+ and CD3+ cell doses were 5.6x10^6/Kg (range, 1.5-10.9) and 2.0x10^8/Kg (range, 0.2-8.0), respectively. All the recipients of allo-HSCT experienced a sustained donor cell engraftment. The cumulative incidence of grades II-IV and III-IV acute GVHD at 100 days was 21% and 9%, respectively. The cumulative incidence of chronic GVHD was 25% at 2 years; we observed severe chronic GVHD only in 4% of pts. The cumulative incidences of relapse and non-relapse mortality (NRM) were 33% and 8% at 2 years, respectively. Two-year overall survival (OS) was 67% and progression free survival (PFS) 59%. The composite end point of GVHD-free/relapse-free survival (GRFS) was 52% at 2 years, in which events include grade 3-4 acute GVHD, systemic therapy-requiring chronic GVHD, relapse, or death. There was a longer OS, 2-year OS was 77% (p = 0.05), and a trend towards higher PFS and GRFS, 63% (p=0.07) and 58% (p=0.06) respectively, for pts with CR status at HSCT. We did not found a significant effect of HLA-matching (9/10 versus 10/10) or donor type (related versus unrelated) on major transplant outcomes (NRM, PFS, GRFS, relapse, acute and chronic GvHD). Conclusion: These outcomes confirmed that matched allogeneic HSCT using treosulfan-based chemotherapy, PT-Cy and sirolimus, is associated with low NRM and acceptable severe GVHD, providing relevant long-term survival in high-risk diseases. A randomized trial comparing this strategy with other kind of in-vivo T-cell depletion (i.e. ATG) is warranted. Disclosures Vago: Moderna TX: Research Funding; GENDX: Research Funding.