Search

Chimeric antigen receptor (CAR) T-cell toxicities resembling hemophagocytic lymphohistiocytosis (HLH) occur in a subset of patients with cytokine release syndrome (CRS). As a variant of conventional CRS, a comprehensive characterization of CAR T-cell–associated HLH (carHLH) and investigations into associated risk factors are lacking. In the context of 59 patients infused with CD22 CAR T cells where a substantial proportion developed carHLH, we comprehensively describe the manifestations and timing of carHLH as a CRS variant and explore factors associated with this clinical profile. Among 52 subjects with CRS, 21 (40.4%) developed carHLH. Clinical features of carHLH included hyperferritinemia, hypertriglyceridemia, hypofibrinogenemia, coagulopathy, hepatic transaminitis, hyperbilirubinemia, severe neutropenia, elevated lactate dehydrogenase, and occasionally hemophagocytosis. Development of carHLH was associated with preinfusion natural killer(NK) cell lymphopenia and higher bone marrow T-cell:NK cell ratio, which was further amplified with CAR T-cell expansion. Following CRS, more robust CAR T-cell and CD8 T-cell expansion in concert with pronounced NK cell lymphopenia amplified preinfusion differences in those with carHLH without evidence for defects in NK cell mediated cytotoxicity. CarHLH was further characterized by persistent elevation of HLH-associated inflammatory cytokines, which contrasted with declining levels in those without carHLH. In the setting of CAR T-cell mediated expansion, clinical manifestations and immunophenotypic profiling in those with carHLH overlap with features of secondary HLH, prompting consideration of an alternative framework for identification and management of this toxicity profile to optimize outcomes following CAR T-cell infusion.

Multiple myeloma (MM) is a malignancy of plasma cells that is nearly always incurable. T cells expressing chimeric antigen receptors (CAR) that target B-cell maturation antigen (BCMA) can recognize and eliminate MM. The murine or other non-human sequences in the single-chain variable fragments (scFv) of many anti-BCMA CARs can elicit recipient immune responses against CAR T cells. We constructed a CAR incorporating an anti-BCMA fully-human heavy-chain variable domain designated FHVH33. FHVH33 lacks the light chain, the artificial linker sequence, and the 2 linker-associated junctions of a scFv, so FHVH33 is smaller than a scFv and is likely to be less immunogenic. The FHVH33-containing CAR utilized in this clinical trial also incorporated a CD8a hinge and transmembrane domain, a 4-1BB domain, and a CD3z domain. The CAR was designated FHVH33-CD8BBZ and was encoded by a gamma-retroviral vector. T cells expressing FHVH33-CD8BBZ were designated FHVH33-T. The FHVH33-T production process was initiated with unsorted peripheral blood mononuclear cells and took 7 days. The treatment protocol was 300 mg/m 2 of cyclophosphamide and 30 mg/m 2 of fludarabine on days -5 to -3 followed by infusion of FHVH33-T on day 0. Twenty-five patients received FHVH33-T infusions. Median age of the treated patients was 62 (range 39-73). Patients received a median of 6 prior lines of therapy (range 3-10). Five dose levels were assessed (Table). Dose level 4, 6x10 6 CAR + T cells/kg was identified as the maximum feasible dose after considering efficacy and manufacturing factors. Twenty-three of 25 patients (92%) obtained objective responses (OR) of partial response (PR) or better. Seventeen patients (68%) attained a best response of stringent complete response (sCR) or very good partial response (VGPR). Thirteen patients have ongoing responses. To date, the median duration of response is 50 weeks for the highest two dose levels. At present, the overall median progression free survival (PFS) is 78 weeks; as responses are ongoing in 13 patients (52%), PFS will likely improve. Nine of 25 patients had extramedullary plasmacytomas at baseline; patients with extramedullary plasmacytomas at baseline were less likely to achieve sCR (P=0.011). All 25 treated patients were evaluable for toxicity. Eighteen patients had grade 1 or 2 cytokine-release syndrome (CRS), and 6 patients had grade 3 CRS. One patient had no CRS. No patients had grade 4 CRS. Five patients received tocilizumab and 4 patients received corticosteroids for CRS. Two of twenty-five patients had grade 3 neurological toxicity possibly attributable to FHVH33-T. No patient had grade 4 neurologic toxicity attributable to CAR T cells. One patient died of influenza pneumonia. We assessed blood CAR+ cells by quantitative PCR. The median peak blood CAR+ cell level was 126.5 cells/µl (range 3-1071 cells/µl), and the median time post-infusion of peak blood CAR + cell levels was 10.5 days (range 7-14). Peak CAR T-cell level was not associated with obtaining a sCR. In contrast, blood CAR+ T cell levels at both 1 and 2 months after infusion were statistically higher for patients obtaining sCR. For the 1-month time-point, blood CAR+ cell levels in cells/mL were 20 for sCR patients and 4 for not sCR patients (P=0.04). Pretreatment serum BCMA was not statistically different when patients obtaining or not obtaining sCR were compared (median serum BCMA in pg/mL: sCR patients 86,243; not sCR patients 261,675, P=0.20). We assessed cell-surface BCMA expression level on MM cells by antibody binding capacity (ABC) flow cytometry. Cell-surface BCMA expression level was not statistically different in sCR versus not sCR patients (median ABC in sites/cell: sCR patients 844; not sCR patients 535, P=0.29). Patients with MM expressing low levels of BCMA obtained durable responses of greater than 2 years duration, which suggests that FHVH33-T can recognize low levels of cell-surface BCMA. Eight patients had extramedullary plasmacytomas at relapse; 4 patients had plasmacytomas biopsied. Two of the biopsied plasmacytomas were BCMA+, and two were BCMA-negative by immunohistochemistry. FHVH33-CD8BBZ CAR T cells caused relatively mild toxicity and a high rate of sCRs in patients with relapsed MM including MM with low cell-surface BCMA expression. Figure 1 Figure 1. Disclosures Brudno: Kyverna Therapeutics: Membership on an entity's Board of Directors or advisory committees. Lam: Kite, a Gilead Company: Patents & Royalties. Kochenderfer: Kite, a Gilead Company: Patents & Royalties: on anti-CD19 CARs, Research Funding; Bristol Myers Squibb: Research Funding.

Abstract Introduction Severe acute respiratory syndrome coronavirus-2 (SARS-CoV2) can induce a strong host immune response. Several groups have investigated the course of antibody responses in patients recovering from SARS-CoV-2 infections but little is known about the recovery of cellular immunity. This study investigated the cellular immune response in people who had recovered from SARS-CoV2 infection. Methods 162 coronavirus disease 2019 (COVID-19) convalescent plasma donors (CCD) and 40 healthy donor (HD) controls were enrolled prospectively in an IRB-approved protocol (Clinical Trials Number: NCT04360278) and provided written informed consent to participate in the study. Using the nCounter platform and host response panel with 785 genes across more than 50 pathways, we compared transcriptomic profiles on RNA samples obtained from the peripheral blood leukocytes of these 162 CCD and 40 HD. Additionally, in 69 of the 162 CCD samples, we evaluated transcriptomic trends at more than one-time point during the convalescent period. Results Age, sex, ethnicity, and body mass index distributions were similar among the CCD and HD. With respect to baseline complete blood counts, hemoglobin, platelets, and absolute basophil and eosinophil counts, all were similar among CCD and HD (Table 1). However, despite sample collections occurring several days after convalescence, mean counts for absolute neutrophil counts, absolute monocyte counts, and absolute lymphocyte counts were significantly higher among CCD compared to HD. 30-90 days after diagnosis, 19 of 773 genes differed (FDR < 0.05) between the average CCD and HD samples. Up-regulated genes included MAFB, CTLA4, PTGS2, and the chemokine signaling genes CXCR4, CXCL5, CXCL2 and CCR4. Down-regulated genes included PTGER2, CASP8, and the interleukins IL36A, IL31, IL20 and IL21 (Figure 1 a,b). Differential gene expression persisted for months. At 90-120 days, 13 genes were differentially regulated, including again MAFB CXCR4, PTGS2, CXCL2 and PTGER2, plus SMAD4. At 120-150 days post-diagnosis, 58 genes were differentially expressed (FDR < 0.05) compared to HD. Pathways with up-regulated genes included Treg differentiation, type III interferon signaling and chemokine signaling. 150-360 days post-diagnosis, 4 genes remained up-regulated on average (FDR < 0.05): PTGS2, PIK3CR, CXCL1 and SMAD4 (Figure 1 c,d). Individual patients varied considerably from the mean trend. Scoring samples by their similarity to the gene expression profile of the mean HD sample, 21 CCD samples from 20 unique patients (12%) were identified as highly perturbed from HD. 84% of these highly perturbed samples were collected > 90 days post-diagnosis. Of these 21 samples, 6 were distinguished by > 2-fold up-regulation of a cluster of interleukin and type-1 interferon genes (Figure 2). Conclusions Overall, our study identified important gene expression trends in CCD compared to HD in the post-acute period. The changes varied with time and among donors. As the expression of T-cell inhibitory molecule CTLA4 fell, the number of differentially expressed increased with the most marked changes occurring 120 to 150 days post-diagnosis in genes in chemokine signaling, type III interferon signaling and Treg pathways. Persistent alterations in inflammatory pathways and T-cell activation/exhaustion markers for months after active infection may help shed light on the pathophysiology of a prolonged post-viral syndrome observed in individuals following recovery from COVID-19 infection. Our data may serve as the basis for risk modification strategies in the period of active infection. Future studies may inform the ability to identify druggable targets involving these pathways to mitigate the long-term effects of COVID-19 infection. Figure 1 Figure 1. Disclosures Danaher: NanoString Technologies: Current Employment, Current holder of individual stocks in a privately-held company.

Introduction Autologous hematopoietic progenitor cell (HPC)-based gene therapy and genome editing are emerging therapeutic strategies for patients with sickle cell disease (SCD). Given the high cell loss incurred during production of genetically modified HPCs, a large number of HPCs (10-15 x10 6 cells/kg) are required. Obtaining sufficient autologous HPCs from patients with SCD is a critical rate limiting step for the success of these novel genetic therapies. Plerixafor, a CXCR4 antagonist, has been shown to be safe and efficient for HPC mobilization in patients with SCD. However, impaired red blood cell (RBC) rheology and chronic inflammation associated with SCD make HPC apheresis collection challenging. For example, HPCs can potentially form aggregates with sickled RBCs traveling deep in the buffy coat upon density-based separation. In addition, elevated levels of activated platelets can increase cell clumping during apheresis. Thus, autologous HPC collection using standard apheresis setting is generally less efficient in patients with SCD and often associated with clumping and unstable cell interfaces during apheresis. It has been reported that standard HPC apheresis collections yield Methods Adult patients with SCD who were enrolled on two clinical trials (NCT03226691 and NCT04443907) at St. Jude Children's Research Hospital and the National Heart, Lung and Blood Institute at the National Institutes of Health were included. Hydroxyurea was discontinued at least 2 weeks and an RBC exchange was performed 3-7 days prior to HPC mobilization. The participants received 240 µg/kg of Plerixafor subcutaneously, and underwent HPC collection 2-4 hours later using the Spectra Optia continuous mononuclear cell collection. Acid-citrate-dextrose-solution A was used for anticoagulation and calcium gluconate was infused during apheresis to prevent citrate-induced hypocalcemia symptoms. Up to 4 total blood volumes were processed for each collection with an adjusted collection preference (CP) of 10-40 and AC/Inlet ratio of 6-12. Collection efficiency (CE) 1 was calculated as CD34+ cell yield / [(pre-CD34 count + post-CD34 count)/2 x whole blood volume processed]. CE 2 was calculated as CD34+ cell yield / (pre-CD34 count x whole blood volume processed). Results We performed 46 autologous HPC collections in 26 participants (age of 20 to 50 years). Mean (±1SD) pre-collection blood indices were: hematocrit (Hct) 27.4 ±3.8%, platelet count 226 ± 111 x10 9/L, and hemoglobin S 25.7 ± 6.0%, WBC count 13.82 ± 6.0 x10 9/L, and CD34+ cell count 62 ± 49 /µL. To address the altered HPC sedimentation and clumping associated with SCD, we targeted a deep buffy coat collection by lowing CP from standard 50 to 10-40, and used more anticoagulant by lowering Inlet/AC ratio from standard 12 to 6-12 respectively. The adjustments increased HPC yield, and 5.60 ± 5.48 x10 6/kg CD34+ cells were collected per procedure. The CE1 was 49.21 ± 28.34 % and CE2 42.74 ± 26.24%, which are comparable to autologous HPC CEs for non-SCD individuals. A deep collection led to highly efficient collections, although this tended to increase RBC contamination of the products (product Hct 6.0 ±2.2% versus < 3% for standard collection). Nevertheless, the participants experienced minimal loss of RBCs with a post-collection Hct of 26.6 ± 2.9%. Platelet loss was also relatively low, as demonstrated by a post-collection platelet count of 152 ± 69 x 10 9/L. No participants required platelet or blood transfusions, had bleeding or severe hypocalcemia-associated symptoms during and after the collections. Conclusion The adjusted apheresis settings with a low buffy coat collection and increased anticoagulant dosage allowed safe and efficient autologous HPC collection in the patients with SCD. Further optimization of apheresis settings is required to increase HPC collection efficiency. Disclosures Sharma: Novartis: Other: Salary support paid to institution; Spotlight Therapeutics: Consultancy; Medexus Inc: Consultancy; CRISPR Therapeutics: Other, Research Funding; Vindico Medical Education: Honoraria; Vertex Pharmaceuticals/CRISPR Therapeutics: Other: Salary support paid to institution. Hankins: Bluebird Bio: Consultancy; UpToDate: Consultancy; Vindico Medical Education: Consultancy; Global Blood Therapeutics: Consultancy. Weiss: Cellarity Inc.: Consultancy; Novartis: Consultancy; Forma Therapeutics: Consultancy; Beam Therapeutics: Current holder of stock options in a privately-held company.

Introduction Natural Killer (NK) cells are an emerging form of cancer immunotherapy currently being tested in clinical trials world-wide. NK cells are innate immune cells that can kill tumor cells via release of cytotoxic granules and via surface expression of the death receptor ligands tumor-related apoptosis-inducing ligand (TRAIL) and Fas ligand. We and others have recently shown that the proteasome inhibitor bortezomib sensitizes tumor cells to NK cell TRAIL-mediated killing by upregulation of death receptor 5. In a recent phase I NK cell dose-escalation study conducted at the NIH (NCT00720785) we have attempted to exploit TRAIL sensitization by administering ex vivo-expanded autologous NK cells to patients with solid tumors or hematologic malignancies that have been pretreated with bortezomib. Ex vivo cultures used to expand clinical grade NK cells for this trial utilize irradiated EBV-LCL feeder cells and IL-2 containing media which upregulates surface expression of TRAIL, substantially augmenting NK cell killing of bortezomib-treated tumors in vitro. Here we characterize the impact of specific expansion conditions used to generate high numbers of NK cells for clinical use on NK cell TRAIL expression. Methods To generate clinical grade ex vivo-expanded NK cells, we first isolated NK cells from patient apheresis products by CD3+ depletion followed by CD56+ selection, and stimulated these enriched NK cells with irradiated EBV-LCL feeder cells at a ratio of 1:10 in X-VIVO 20 supplemented with 10% inactivated human AB serum and recombinant human IL-2 (500 IU/ml). The clinical trial evaluated 8 escalating NK cell dose levels (Figure 1). Cohorts 1-4 received a single infusion of ex vivo-expanded NK cells on day 0 in a dose-escalating fashion (3-6 pts per cohort) and cohorts 5-7 received 1 x 108 NK cells/kg on day 0 and a second escalating dose of NK cells infused on day +5. A "closed bag" Baxter PL732 culture system was used for cohorts 1-7 which was later changed to a GREX500-CS (Wilson Wolf) system in cohorts 7-8. Using flow cytometry, we monitored surface expression of TRAIL on the day NK cells were harvested and infused fresh into patients. We also assessed TRAIL expression on NK cells from a single patient cultured at 6 different cell densities (range: 2.03-16.95 x 106/cm2) using culture conditions mimicking the phase I trial. Results A total of 137 NK cell cultures were harvested and administered fresh to 32 patients. NK cells on the day of harvest expanded a median of 198-fold, 895-fold, and 3637-fold on culture days 14-16, 19-22, and 24-27, respectively. NK cells at harvest contained a median of 99.7% CD3−/CD56+ NK cells, were 68.65% CD16+ and had a median of 88% viability. TRAIL was assessed by mean fluorescence intensity (MFI) with a median surface expression of of 1245 (range 132-4913) at the time of infusion (Figure 1). Expansions for cohort 8 generated 10-14 x109 (1 vessel) and 50-70 x109 NK cells (4-5 vessels) for fresh infusion, enough to support the target dose level of 1x108 (1st harvest) and 5x108 (2nd harvest) NK cells/kg. Remarkably, NK cells grown at higher cell density to reach the target cell numbers for cohort 8 exhibited substantially reduced TRAIL expression (median: 255, range 132-691). Subsequent experiments conducted on NK cells expanded in vitro for 14 days at different cell densities/concentrations showed TRAIL expression (MFI range: 319-1627) inversely correlated with both cell density and concentration (Figure 2). NK cells grown at the highest cell density (16.95 x 106/cm2) and concentration (4.23 x 106/mL) expressed the least amount of TRAIL (MFI 319), in contrast to those cultured at the lowest cell density (2.03 x 106/cm2) and concentration (0.51 x 106/mL), which demonstrated a TRAIL MFI of 1627. Conclusions Although ex vivo cultures using feeder cells make it possible to expand large numbers of NK cells for clinical use in humans, the higher concentrations and density of cells in these cultures reduce NK cell surface expression of TRAIL. In vitro, TRAIL expression appears to inversely correlate with cell density. These data highlight the need to avoid overly concentrating ex vivo expanded NK cells to maximize TRAIL surface expression as a method to potentiate the anticancer effects of adoptively infused NK cells. Disclosures No relevant conflicts of interest to declare.

Introduction Severe Aplastic Anemia (SAA) is a life-threatening bone marrow failure disorder associated with pancytopenia, serious infections, and transfusion dependence. Although long term survival for SAA patients (pts) can be achieved with immunosuppressive therapy (IST), one quarter to one third will fail to respond and about half of responders relapse. Many refractory SAA pts lack an HLA matched donor for salvage allogeneic stem cell transplantation. Although umbilical cord blood transplantation (UCBT) is an alternative approach for SAA pts, the procedure is associated with delayed engraftment and high rejection rates. Further, a substantial portion of UCB units have an insufficient number of TNCs/CD34+ cells for optimal transplant outcomes. To address the need to improve UCBT outcomes for SAA, we instituted a phase II trial exploring the use of nicotinamide (NAM) ex-vivo expanded UCB [omidubicel] to transplant pts with refractory SAA. Methods Eligible pts were 4-55 years and had transfusion dependent SAA. Other criteria included a) failure or intolerance to IST, b) lack of an HLA matched donor, c) having a ≥ 4/8 HLA matched UCB unit with a minimum of 1.8 x 109 and at least 1.8x107/kg TNCs and at least 8 x 106 CD34+ cells, and d) absence of donor specific antibodies to mismatched alleles on the UCB unit. The study was designed to enroll up to 6 pts receiving omidubicel and CD34+ selected haploidentical cells (cohort 1), followed by enrollment of up to 23 pts receiving omidubicel only (cohort 2). The conditioning regimen consists of hATG (40 mg/kg) on D -11 to -8, cyclophosphamide (60 mg/kg) on D -7 and -6, fludarabine (25 mg/m2) on D -5 to -1, and 2 Gy TBI on D -1. Tacrolimus and mycophenolate mofetil were given as GVHD prophylaxis. The primary end point of the study is sustained early engraftment. Secondary endpoints include treatment-related mortality, and standard transplant outcomes. Results A total of 8 pts with SAA refractory to IST and eltrombopag with a median age of 23 years (range 6-45) have been transplanted to date. Pts were heavily transfused (median ferritin 3745 µg/dL), had severe neutropenia with a median pre-transplant ANC of 115 (range 0-680), with 5 (63%) pts having HLA alloantibodies and 1 pt having an invasive fungal infection treated with multiple antifungals and granulocyte transfusions in the peri-transplant period. In cohort 1, the first 3 pts received omidubicel and ~3 x106 CD34+ cells/kg from a haploidentical donor. Since all 3 pts in cohort 1 had sustained engraftment with omidubicel, the study moved to cohort 2, where so far 5 pts have received omidubicel alone. UCB units were matched at a median of 5/8 HLA alleles (range 4-6). Before expansion, units contained a median 1.6 x105 CD34+ cells/kg. At time of transplantation, the expanded units contained a median 75.3 x105 CD34+ cells/kg, representing a median 42-fold expansion. With a median follow-up of 10 months (range 1-35), 7/8 (88%) pts have had early and sustained cord engraftment. One pt in cohort 2 failed to engraft and was salvaged with a haploidentical transplant. Another pt in cohort 2 died on D+ 62 from disseminated adenovirus infection. This pt also developed grade 2 acute GVHD and his IST was stopped early in an effort to boost immunity against adenovirus. The other 7 pts are alive without evidence for acute or chronic GVHD and are transfusion independent. CMV reactivation requiring treatment occurred in 3/6 (50%) of pts at risk. Neutrophil and platelet recovery have been remarkably quick, occurring at a median of 10 days (range 6-14) and 31 days (15-40) respectively. Among the 7 pts with sustained engraftment, six pts had ≥ 95% cord myeloid chimerism by D+ 14 and ≥ 95% T-cell chimerism by D+ 26. Immune recovery has also been remarkably brisk: at D+ 100, the median absolute CD4 count was 186/µL (IQR, 118-340) and mean (±SD) IgG level was 522(±161) mg/dL, respectively (Figure). Conclusions These results provide the first evidence that omidubicel can result in rapid engraftment and sustained hematopoiesis in heavily transfused and allo-immunized SAA pts who are at high risk for graft failure with conventional UCBT. Data showing rapid recovery of IgG levels and rapid reconstitution of CD4, CD8, and NK cell subsets suggests immune recovery may occur quicker with omidubicel compared to conventional UBCT. Taken altogether, our preliminary data suggest omidubicel represents a promising new alternative graft source for SAA pts who lack HLA matched donors. Figure Disclosures No relevant conflicts of interest to declare.

T cells expressing chimeric antigen receptors (CAR) that target B-cell maturation antigen (BCMA) recognize and eliminate multiple myeloma (MM). BCMA is expressed by nearly all cases of MM. BCMA has a restricted expression pattern on normal cells. To reduce the risk of recipient immune responses against CAR T cells, we used a novel, fully-human, heavy-chain-only anti-BCMA binding domain designated FHVH33 instead of a traditional single-chain variable fragment (scFv). The FHVH33 binding domain lacks the light chain, artificial linker sequence, and 2 associated junctions of a scFv. We constructed a CAR designated FHVH33-CD8BBZ. FHVH33-CD8BBZ was encoded by a γ-retroviral vector and incorporated FHVH33, CD8α hinge and transmembrane domains, a 4-1BB costimulatory domain, and a CD3ζ domain. T cells expressing FHVH33-CD8BBZ are designated FHVH-BCMA-T. On this clinical trial, patients received 300 mg/m2 of cyclophosphamide and 30 mg/m2 of fludarabine on days -5 to -3 followed by infusion of FHVH-BCMA-T on day 0. Twenty-one FHVH-BCMA-T infusions have been administered on 5 dose levels (DL), 0.75x106, 1.5x106, 3x106, 6x106 and 12 x106 CAR+ T cells/kg of bodyweight. DL4 (6 x 106 CAR+ T cells/kg) was identified as the maximum feasible dose (MFD) after weighing toxicity, efficacy and manufacturing factors. Patients are now being enrolled on an expansion phase to test the MFD. One patient (Patient 11) received 2 treatments. Four patients have been enrolled who were not ultimately treated. The median age of the patients enrolled is 64 (range 41-72). Patients received a median of 6 prior lines of therapy (range 3-12). Of the 20 FHVH-BCMA-T treatments evaluable for response, 18 (90%) resulted in objective responses (OR). Twelve treatments resulted in VGPR, complete remission (CR) or stringent complete remission (sCR). Ten patients (50%) have ongoing responses that range between 0-80 weeks (6 sCR/CRs, 3 VGPRs, 1 PR). At the highest two DLs (8 patients), 7 patients (88%) have ongoing responses (median duration 20 weeks, range 0+ to 35+ weeks); progressive MM occurred in only 1 patient who had evidence of spinal cord compression on day +5 due to a rapidly expanding plasmacytoma, which required early intervention with high-dose corticosteroid and radiation therapy. Of the 8 patients evaluated for response who had high-risk cytogenetics at baseline, 7 had ORs. Responses are ongoing in 2 patients with TP53 mutations and 1 patient with t(4;14) translocation. Ten treated patients came off study due to progressive MM (9 patients) or death from other causes (1 patient, influenza). Two of 4 patients who had plasmacytomas evaluated for BCMA expression at relapse had evidence of BCMA-negative MM. Four patients had bone marrow aspirates evaluated for BCMA-expression before treatment and at the time of relapse; 3 of these patients had evidence of loss of BCMA expression at relapse. Of 21 FHVH-BCMA-T treatments administered, 20 (95%) were followed by cytokine release syndrome (CRS) with 16 (76%) cases of grade 1 or 2 CRS, 4 cases (19%) of grade 3 CRS, and no cases of grade 4 CRS. Three patients received tocilizumab. The median peak C-reactive protein after all 21 treatments was 196.9 mg/L. Of 21 total treatments, 8 (38%) were followed by neurologic toxicity; there were 5 cases of grade 1-2 neurologic toxicity (headache, dysarthria, confusion, delirium), 2 cases of grade 3 neurologic toxicity (confusion), and 1 patient with grade 4 spinal cord compression due to progressive MM. Two patients received corticosteroids to manage neurologic toxicities. A median of 3.0% (range 0-95%) of bone marrow T cells were CAR+ when assessed by flow cytometry 14 days after FHVH-BCMA-T infusion. We assessed blood CAR+ cells by quantitative PCR. The median peak level of CAR+ cells was 121 cells/µl (range 3-359 cells/µl) and the median day post-infusion of peak blood CAR+ cell levels was 12 (range 7-14). The results from this phase 1 trial demonstrate that FHVH-BCMA-T cells can induce deep and durable responses of relapsed MM with manageable toxicities. Assessment of durability of responses at the maximum feasible dose is a critical future plan. Accrual to the expansion cohort continues. Table Disclosures Manasanch: Novartis: Research Funding; Adaptive Biotechnologies: Honoraria; GSK: Honoraria; JW Pharma: Research Funding; Merck: Research Funding; Quest Diagnostics: Research Funding; Takeda: Honoraria; Sanofi: Honoraria; BMS: Honoraria; Sanofi: Research Funding. Rosenberg:Kite, A Gilead Company: Consultancy, Patents & Royalties, Research Funding. Kochenderfer:Kite, a Gilead company: Patents & Royalties, Research Funding; Celgene: Patents & Royalties, Research Funding; bluebird, bio: Patents & Royalties. OffLabel Disclosure: cyclophosphamide 300 mg/m2 fludarabine 30 mg/m2 Conditioning chemotherapy prior to CAR T-cell infusion

Long-lived, self-renewing, multipotent T memory stem cells (TSCM) can trigger profound and sustained tumor regression but their rareness poses a major hurdle to their clinical application. Presently, clinically compliant procedures to generate relevant numbers of this T-cell population are undefined. Here, we provide a strategy for deriving large numbers of clinical-grade tumor-redirected TSCM starting from naive precursors. CD8(+)CD62L(+)CD45RA(+) naive T cells enriched by streptamer-based serial-positive selection were activated by CD3/CD28 engagement in the presence of interleukin-7 (IL-7), IL-21, and the glycogen synthase-3β inhibitor TWS119, and genetically engineered to express a CD19-specific chimeric antigen receptor (CD19-CAR). These conditions enabled the generation of CD19-CAR-modified CD8(+) TSCM that were phenotypically, functionally, and transcriptomically equivalent to their naturally occurring counterpart. Compared with CD8(+) T cells generated with clinical protocols currently under investigation, CD19-CAR-modified CD8(+) TSCM exhibited enhanced metabolic fitness and mediated robust, long-lasting antitumor responses against systemic acute lymphoblastic leukemia xenografts. This clinical-grade platform provides the basis for a phase 1 trial evaluating the activity of CD19-CAR-modified CD8(+) TSCM in patients with B-cell malignancies refractory to prior allogeneic hematopoietic stem cell transplantation.

Natural killer (NK) cells can enhance engraftment and mediate graft-versus-leukemia following allogeneic hematopoietic stem cell transplantation (HSCT), but the potency of graft-versus-leukemia mediated by naturally reconstituting NK cells following HSCT is limited. Preclinical studies demonstrate that activation of NK cells using interleukin-15 (IL-15) plus 4-1BBL upregulates activating receptor expression and augments killing capacity. In an effort to amplify the beneficial effects of NK cells post-HSCT, we conducted a first-in-human trial of adoptive transfer of donor-derived IL-15/4-1BBL-activated NK cells (aNK-DLI) following HLA-matched, T-cell-depleted (1-2 × 10(4) T cells/kg) nonmyeloablative peripheral blood stem cell transplantation in children and young adults with ultra-high-risk solid tumors. aNK-DLI were CD3(+)-depleted, CD56(+)-selected lymphocytes, cultured for 9 to 11 days with recombinant human IL-15 plus 4-1BBL(+)IL-15Rα(+) artificial antigen-presenting cells. aNK-DLI demonstrated potent killing capacity and displayed high levels of activating receptor expression. Five of 9 transplant recipients experienced acute graft-versus-host disease (GVHD) following aNK-DLI, with grade 4 GVHD observed in 3 subjects. GVHD was more common in matched unrelated donor vs matched sibling donor recipients and was associated with higher donor CD3 chimerism. Given that the T-cell dose was below the threshold required for GVHD in this setting, we conclude that aNK-DLI contributed to the acute GVHD observed, likely by augmenting underlying T-cell alloreactivity. This trial was registered at www.clinicaltrials.gov as #NCT01287104.

Introduction: Approximately 7% of unrelated hematopoietic stem cell (HSC) donors are asked to donate a subsequent time to the same or different recipient. In a recent large CIBMTR study of second time donors, Stroncek et al. incidentally found that second peripheral blood stem cell (PBSC) collections had lower total CD34+ cells, CD34+ cells per liter of whole blood processed, and CD34+ cells per kg donor weight. Based on smaller studies, the time between the two independent PBSC donations (inter-donation interval) as well as donor sex, race and baseline lymphocyte counts appear to influence CD34+ cell yields at subsequent donations. Our objective was to retrospectively evaluate factors contributory to CD34+ cell yields at subsequent PBSC donation amongst NMDP donors. Methods. The study population consisted of filgrastim (G-CSF) mobilized PBSC donors through the NMDP/CIBMTR between 2006 and 2017, with a subsequent donation of the same product. evaluated the impact of inter-donation interval, donor demographics (age, BMI, race, sex, G-CSF dose, year of procedure, need for central line) and changes in complete blood counts (CBC), on the CD34+ cell yields/liter (x106/L) of blood processed at second donation and pre-apheresis (Day 5) peripheral blood CD34+ cell counts/liter (x106/L) at second donation. Linear regression was used to model log cell yields as a function of donor and collection related variables, time between donations, and changes in baseline values from first to second donation. Stepwise model building, along with interactions among significant variables were assessed. The Pearson chi-square test or the Kruskal-Wallis test compared discrete variables or continuous variables, respectively. For multivariate analysis, a significance level of 0.01 was used due to the large number of variables considered. Results: Among 513 PBSC donors who subsequently donated a second PBSC product, clinically relevant decreases in values at the second donation were observed in pre-apheresis CD34+ cells (73.9 vs. 68.6; p=0.03), CD34+cells/L blood processed (32.2 vs. 30.1; p=0.06), and total final CD34+ cell count (x106) (608 vs. 556; p=0.02). Median time interval between first and second PBSC donations was 11.7 months (range: 0.3-128.1). Using the median pre-apheresis peripheral blood CD34+ cell counts from donation 1 as the cut-off for high versus low mobilizers, we found that individuals who were likely to be high or low mobilizers at first donation were also likely to be high or low mobilizers at second donation, respectively (Table 1). This was independent of the inter-donation interval. In multivariate analyses, those with an inter-donation interval of >12 months, demonstrated higher CD34+cells/L blood processed compared to donors donating within a year (mean ratio 1.15, p Pre-apheresis peripheral blood CD34+ cells on Day 5 of second donation were also affected by the inter-donation interval, with higher cell counts associated with a longer time interval (>12 months) between donations (1.23, p6.9) (1.3, p Conclusions: In this large retrospective study of second time unrelated PBSC donors, a longer inter-donation interval was confirmed to be associated with better PBSC mobilization and collection. Given hematopoietic stem cell cycling times of 9-12 months in humans, where possible, repeat donors may be chosen based on these intervals to optimize PBSC yields. Changes in BMI are also to be considered while recruiting repeat donors. Some of these parameters may be improved marginally by increasing G-CSF dose within permissible limits. In most instances, however, sub-optimal mobilizers at first donation appear to donate suboptimal numbers of HSC at their subsequent donation. Disclosures Pulsipher: CSL Behring: Membership on an entity's Board of Directors or advisory committees; Miltenyi: Research Funding; Bellicum: Consultancy; Amgen: Other: Lecture; Jazz: Other: Education for employees; Adaptive: Membership on an entity's Board of Directors or advisory committees, Research Funding; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Medac: Honoraria. Shaw:Therakos: Other: Speaker Engagement.

Chimeric antigen receptor (CAR) T cells expressing B-cell maturation antigen (BCMA) can target and kill multiple myeloma (MM). BCMA was chosen as a target for MM because it is expressed by almost all cases of MM but has a restricted expression pattern on normal cells. CAR antigen-recognition domains made up of monoclonal antibody-derived, single-chain-variable fragments (scFv) are potentially immunogenic. To reduce the risk of recipient immune responses against CAR T cells, we used the sequence of a novel anti-BCMA, fully-human, heavy-chain-only binding domain designated FHVH33. The FHVH33 binding domain sequence was from TeneoBio, Inc. FHVH33 is smaller than a scFv. FHVH33 lacks the light chain, artificial linker sequence, and 2 associated junctions of a scFv, so it is predicted to be less immunogenic than a scFv, especially murine-derived scFvs. We constructed a CAR incorporating FHVH33, CD8α hinge and transmembrane domains, a 4-1BB costimulatory domain, and a CD3ζ T-cell activation domain. The CAR, FHVH33-CD8BBZ, is encoded by a γ-retroviral vector. FHVH33-CD8BBZ-expressing T cells (FHVH-BCMA-T) exhibited a full range of T-cell functions in vitro and eliminated tumors and disseminated malignancy in mice (Lam et al, Blood (ASH abstract) 2017 vol 130: 504). We are conducting the first clinical trial of FHVH-BCMA-T. Patients receive conditioning chemotherapy on days -5 to -3 with 300 mg/m2 of cyclophosphamide and 30 mg/m2 of fludarabine followed by infusion of FHVH-BCMA-T on day 0. This dose-escalation trial has 5 planned dose levels (DL). Twelve patients have received FHVH-BCMA-T on 3 DLs, 0.75x106, 1.5x106 and 3x106 CAR+ T cells/kg of bodyweight. Three patients were enrolled on the trial but not treated. The median age of patients enrolled was 63 (range 52-70); patients received a median of 6 lines of anti-myeloma therapy (range 3-10) prior to treatment with FHVH-BCMA-T. Ten patients out of 12 patients have achieved objective responses (OR). Five patients have obtained CRs or VGPRs to date. One patient achieved a partial remission (PR) 26 weeks after FHVH-BCMA-T infusion through a continued decrease in a measurable plasmacytoma. Five out of 7 patients who had myeloma with high-risk cytogenetics had an OR (Table). ORs occurred in patients with large soft-tissue plasmacytomas. Loss of BCMA expression on myeloma cells after treatment was documented in 2 patients. Two patients who developed progressive MM after CAR T-cell infusion had evidence of minimal residual disease in bone marrow 1-2 months post infusion of CAR T cells (patients 7,8). Eleven out of 12 patients had cytokine release syndrome (CRS); CRS grades ranged from 1-3 (Lee et al. Biol Blood Marrow Transplant 25 (2019) 625-638). The median peak C reactive protein (CRP) of the patients with CRS was 156.3 mg/L. Of 12 patients, 1 received the interleukin-6-receptor antagonist tocilizumab on day +6 to treat grade 3 CRS with hypotension requiring low-dose pressor therapy, grade 2 ejection fraction (EF) decrease and elevation of creatinine kinase (CK). All parameters returned to baseline by day +10. Patient 12 had a grade 3 decrease in EF which resolved by day +29. Two patients had grade 2 neurotoxicity that resolved without intervention: patient 3 had headaches, dysarthria and word-finding difficulties that resolved after 6 days while patient 6 had headaches on day +4. Patient 12 had grade 3 neurotoxicity with confusion on day +2; she was given dexamethasone with improvement in mental status the same day. After attaining a response, patient 6 died from influenza complications 6 weeks after FHVH-BCMA-T infusion. A median of 10.6% (range 1.1-46) of bone marrow T cells were CAR+ when assessed 14 days after FHVH-BCMA-T infusion. We assessed blood CAR+ cells by quantitative PCR. The median peak level of CAR+ cells was 76.5 cells/µl (range 3-347 cells/µl) and the median day post-infusion of peak blood CAR+ cell levels was 13 (range 9-14). The results from this phase 1 trial demonstrate that FHVH-BCMA-T cells can induce responses at low dose levels. Patients who had no CRS or low-grade CRS achieved objective responses. Toxicity was limited and reversible. Accrual to this trial continues. A maximum tolerated dose has not been determined yet. These results encourage further development of FHVH CAR-T. Table Disclosures Manasanch: Janssen: Honoraria; Sanofi: Honoraria; Takeda: Honoraria; Merck: Research Funding; Skyline Diagnostics: Research Funding; Sanofi: Research Funding; Quest Diagnostics: Research Funding; Celgene: Honoraria. Trinklein:Teneobio, Inc.: Employment, Equity Ownership. Buelow:Teneobio, Inc.: Employment, Equity Ownership. Kochenderfer:Kite and Celgene: Research Funding; Bluebird and CRISPR Therapeutics: Other: received royalties on licensing of his inventions. OffLabel Disclosure: Cyclophosphamide and fludarabine are used in combination for conditioning chemotherapy prior to CAR T-cell infusion

In experimental models, ex vivo induced T-cell rapamycin resistance occurred independent of T helper 1 (Th1)/T helper 2 (Th2) differentiation and yielded allogeneic CD4+ T cells of increased in vivo efficacy that facilitated engraftment and permitted graft-versus-tumor effects while minimizing graft-versus-host disease (GVHD). To translate these findings, we performed a phase 2 multicenter clinical trial of rapamycin-resistant donor CD4+ Th2/Th1 (T-Rapa) cells after allogeneic-matched sibling donor hematopoietic cell transplantation (HCT) for therapy of refractory hematologic malignancy. T-Rapa cell products, which expressed a balanced Th2/Th1 phenotype, were administered as a preemptive donor lymphocyte infusion at day 14 post-HCT. After T-Rapa cell infusion, mixed donor/host chimerism rapidly converted, and there was preferential immune reconstitution with donor CD4+ Th2 and Th1 cells relative to regulatory T cells and CD8+ T cells. The cumulative incidence probability of acute GVHD was 20% and 40% at days 100 and 180 post-HCT, respectively. There was no transplant-related mortality. Eighteen of 40 patients (45%) remain in sustained complete remission (range of follow-up: 42-84 months). These results demonstrate the safety of this low-intensity transplant approach and the feasibility of subsequent randomized studies to compare T-Rapa cell-based therapy with standard transplantation regimens. This trial was registered at www.cancer.gov/clinicaltrials as #NCT 00077480.

Although peripheral blood stem cells (PBSCs) have replaced bone marrow (BM) as the most common unrelated donor progenitor cell product collected, a direct comparison of concurrent PBSC versus BM donation experiences has not been performed. We report a prospective study of 2726 BM and 6768 PBSC donors who underwent collection from 2004 to 2009. Pain and toxicities were assessed at baseline, during G-CSF administration, on the day of collection, within 48 hours of donation, and weekly until full recovery. Peak levels of pain and toxicities did not differ between the 2 donation processes for most donors. Among obese donors, PBSC donors were at increased risk of grade 2 to 4 pain as well as grade 2 to 4 toxicities during the pericollection period. In contrast, BM donors were more likely to experience grade 2 to 4 toxicities at 1 week and pain at 1 week and 1 month after the procedure. BM donors experienced slower recovery, with 3% still not fully recovered at 24 weeks, whereas 100% of PBSC donors had recovered. Other factors associated with toxicity included obesity, increasing age, and female sex. In summary, this study provides extensive detail regarding individualized risk patterns of PBSC versus BM donation toxicity, suggesting donor profiles that can be targeted with interventions to minimize toxicity.

Therapies with novel mechanisms of action are needed for multiple myeloma (MM). B-cell maturation antigen (BCMA) is expressed in most cases of MM. We conducted the first-in-humans clinical trial of chimeric antigen receptor (CAR) T cells targeting BCMA. T cells expressing the CAR used in this work (CAR-BCMA) specifically recognized BCMA-expressing cells. Twelve patients received CAR-BCMA T cells in this dose-escalation trial. Among the 6 patients treated on the lowest 2 dose levels, limited antimyeloma activity and mild toxicity occurred. On the third dose level, 1 patient obtained a very good partial remission. Two patients were treated on the fourth dose level of 9 × 10(6) CAR(+) T cells/kg body weight. Before treatment, the first patient on the fourth dose level had chemotherapy-resistant MM, making up 90% of bone marrow cells. After treatment, bone marrow plasma cells became undetectable by flow cytometry, and the patient's MM entered a stringent complete remission that lasted for 17 weeks before relapse. The second patient on the fourth dose level had chemotherapy-resistant MM making up 80% of bone marrow cells before treatment. Twenty-eight weeks after this patient received CAR-BCMA T cells, bone marrow plasma cells were undetectable by flow cytometry, and the serum monoclonal protein had decreased by >95%. This patient is in an ongoing very good partial remission. Both patients treated on the fourth dose level had toxicity consistent with cytokine-release syndrome including fever, hypotension, and dyspnea. Both patients had prolonged cytopenias. Our findings demonstrate antimyeloma activity of CAR-BCMA T cells. This trial was registered to www.clinicaltrials.gov as #NCT02215967.

Introduction: As clinical applications for Chimeric Antigen Receptor (CAR) T-cell therapy expand, cell manufacturing incorporating closed-system, automated instruments are supplanting traditional open-system, labor-intensive culture methods. At our institution and others, the CliniMACS Prodigy (Miltenyi Biotec), a closed-system automated device, has demonstrated success in the production of CAR T-cells from T-cell enrichment, activation, viral transduction, and expansion to downstream harvest for cryopreservation/fresh infusion. However, the duration of T-cell activation/viral transduction, and total T-cell culture duration are variable across centers (2-5 days and 7-13 days) and merit evaluation prior to routine use. Methods: Following the opening of our clinical protocol (Clinicaltrials.gov NCT03448393), CD19/CD22 Bispecific CAR T-cell products were manufactured on the Prodigy for 4 patients (Original Method, OM) but CAR T cell manufacturing was felt to be suboptimal. Consequently, we investigated a modified processing method (Modified Method, MM), for the manufacture of clinical grade products using the Prodigy. Specifically, TransAct CD3/CD28 reagent mediated T-cell activation/stimulation and lentiviral transduction (MSCV-CAR1922-WPRE; Lentigen Inc.) was terminated with a wash step at Day 3 (instead of the wash step at Day 5, as in the OM). Overstimulation of the relatively more sensitive patient cells was proposed as a likely cause of suboptimal cell viability and expansion in the OM runs. Final cell harvest was planned between culture days 7-12. A total of 4 apheresis products were evaluated using this MM and compared with the 4 prior runs using the OM. All products were obtained from live or deceased patients with disease profiles similar to patients on the clinical trial. Other process parameters (enrichment for CD4/CD8 subsets, in-process media changes with GMP-TexMACS Medium supplemented with human IL-2 (200IU/mL) and 3% human AB serum) were kept unchanged across the 2 methods. Transduction by Protein L expression, viability and cell phenotype (CD3, CD4/CD8) were measured by flow cytometry. Results: From ~0.1x109 CD4/8 enriched T-cells placed into the Prodigy culture chamber on day 0, the mean viable Total Nucleated Cells (TNC) obtained in the final product was 1.93x109 ± 0.27x109 in the 4 MM runs. This cell dose was accomplished by culture Day 7. In contrast, in the OM runs, the mean viable TNC obtained in the final products between Days 9 and 12 was 0.8x109 ± 0.7x109 (Figure 1a). Viable CAR transduced CD3+ fold increase was calculated for days 0-7 of the MM cultures and 0-9 and 0-12 of the OM cultures depending on day of harvest and for the 4 MM products the average fold increase was 15.3 ± 4.2 by Day 7 (Figure 1b) which was ~3 fold greater than OM products harvested on day 9 or 12. Viability of transduced cells was >80% throughout MM culture. In contrast, viability was about 31% during manufacturing of one of the OM products (Figure 1c). On the day of harvest, >99% of the cells were CD3+ T-cells for all 4 MM products (Figure 1d) with no remaining CD19+CD22+ cells. The CD4/CD8 ratio was as expected and favored CD4 T-cells over CD8 T-cells (Figure 1e). Transduction efficiency based on Protein L binding was >70% for the MM products and passed clinical release criteria (Figure 1f). A head-to-head comparison of the 2 methods from the same starting fraction in one patient product (Figure 1, 3A, 3B) also confirmed all findings above. Conclusions: Our data demonstrate that the modified CD19/CD22 Bispecific CAR T-cell manufacturing method (MM) which terminated T-cell activation/transduction by culture Day 3, resulted in reproducible and robust CAR T cell production, even in the relatively more sensitive patient cells. Viability, Viable TNC recovery, CD3% and Protein L expression were consistently higher with the MM compared to the OM. All final products in the MM met product release criteria. In addition, final product dose requirements were consistently met by culture Day 7 when using the MM, augmenting process efficiency. Consequently, we have adopted the MM for the manufacture of clinical CD19/CD22 Bispecific CAR T cells. However, to determine if this change effects CAR T cell potency, studies have been initiated to compare differences in T-cell subsets, activation/exhaustion/senescence and differentiation markers, and the metabolic activity of cells manufactured by the 2 methods. Disclosures No relevant conflicts of interest to declare.

Anti-CD19 chimeric antigen receptor (CAR) T cells have powerful activity against B-cell lymphoma, but improvement is clearly needed. Toxicity, including cytokine-release syndrome (CRS) and neurologic toxicity, occurs after anti-CD19 CAR T cell infusions. Most CAR T-cell toxicity is caused, either directly or indirectly, by cytokines or other proteins that are secreted from CAR T cells. The structure of a CAR is an extracellular antigen-recognition domain connected by hinge and transmembrane (TM) domains to intracellular T-cell signaling moieties. In vitro, T cells expressing CARs with hinge and TM domains from the CD8-alpha molecule released significantly lower levels of cytokines compared with T cells expressing CARs with hinge and TM domains from CD28; however, T cells expressing CARs with hinge and TM domains from CD8-alpha retained sufficient functional capability to eradicate tumors from mice (Alabanza et al. Molecular Therapy. 2017. 25(11) 2452). To reduce cytokine production with a goal of reducing clinical toxicity, we incorporated CD8-alpha hinge and TM domains into an anti-CD19 CAR. The CAR also had a human antigen-recognition domain, a CD28 costimulatory domain, and a CD3-zeta domain. This CAR was designated Hu19-CD828Z and was encoded by a lentiviral vector. Hu19-CD828Z was different from the FMC63-28Z CAR that we used in prior studies. FMC63-28Z had hinge and TM domains from CD28 along with a CD28 costimulatory domain, a CD3-zeta domain, and murine-derived antigen-recognition domains. Twenty patients with B-cell lymphoma were treated on a phase I dose-escalation clinical trial of Hu19-CD828Z T cells (Table). Patients received low-dose cyclophosphamide and fludarabine daily for 3 days on days -5 to -3. Two days later, on day 0, CAR T cells were infused. The overall response rate (ORR) after 1st treatments with Hu19-CD828Z T cells was 70%, and the complete response (CR) rate 55%; the 6-month event-free survival was 55%. The anti-lymphoma activity of Hu19-CD828Z T cells in the current trial was comparable to the anti-lymphoma activity of FMC63-28Z T cells in a similar prior trial that also enrolled patients with advanced B-cell lymphoma. In the prior trial, we observed a 73% ORR, a 55% CR rate, and a 6-month event-free survival of 64% in 22 patients treated with FMC63-28Z T cells (Kochenderfer et al. Journ. Clin. Oncology. 2017 35(16) 1803). In our previous clinical trial of FMC63-28Z T cells, the rate of Grade 3 or 4 neurologic toxicity among 22 patients treated was 55%. Strikingly, in our trial of Hu19-CD828Z T cells, the rate of Grade 3 or 4 neurologic toxicity was only 5% (1/20 patients). In addition, the rate of Grade 2 or greater neurologic toxicity with FMC63-28Z T cells was 77.3% while the rate of Grade 2 or greater neurologic toxicity with Hu19-CD828Z T cells was 15%. To explore the mechanism for the difference in neurologic toxicity in patients receiving FMC63-28Z T cells versus Hu19-CD828Z T cells, we assessed serum levels of 41 proteins in patients treated with these CAR T-cells. This comparison is valid because the same Luminex methodology was used for the serum protein analysis for both trials, and controls of known amounts of each protein were assayed to ensure that protein levels were comparable on the different trials. Lower levels of several serum proteins that might be important in CAR toxicity were found in patients treated with Hu19-CD828Z T cells versus patients treated with FMC63-28Z T cells: Granzyme A (P Disclosures Rossi: KITE: Employment. Shen:Kite, a Gilead Company: Employment. Xue:Kite, a Gilead Company: Employment. Bot:KITE: Employment. Rosenberg:Kite, a Gilead Company: Research Funding. Kochenderfer:Kite a Gilead Company: Patents & Royalties: CAR technology, Research Funding; Celgene: Research Funding.

Background: Reactivation of latent viruses post-allogeneic stem cell transplant (SCT) negatively affects outcomes and increases non-relapse mortality. We have reported over 80% early CMV reactivation rate and significant additional costs in recipients of T cell depleted (TCD) SCT. Ex vivo generated multi-virus specific T cells (MVSTs) are effective as a therapy of active infection, but have not been evaluated as prophylaxis early after transplant. In a Phase I study (NIH 14-H-0182) we transferred MVSTs targeting immunodominant viral proteins of CMV, Epstein-Barr virus (EBV), BK and adenovirus (Ad) immediately post SCT as a novel approach to prevent viral reactivation post-SCT. Methods: All subjects were enrolled in HLA-matched T cell depleted (TCD) transplant protocol (NIH 13-H-0144). MVSTs cells were manufactured from SCT sibling donors by stimulation of elutriated lymphocytes for 14 days with seven overlapping peptide libraries (pepmixes) pulsed onto autologous dendritic cells (DCs) in presence of IL-7, IL-15 and IL-2. MVSTs were infused as early as possible (day 0 to +60) post SCT. A Phase I 3+3 dose escalation design was used as follows: Cohort 1 - 1x10e5 total nucleated cells (TNC)/kg, Cohort 2 - 5x10e5 TNC/kg, Cohort 3 - 1x10e6 TNC/kg. Three additional subjects received MVST cells manufactured using pepmix-pulsed mononuclear cells as stimulators (Cohort 3A; 1x10e6TNC/kg) under an amended protocol. The primary safety endpoint at day 42 post infusion was the occurrence of dose limiting toxicity (DLT) defined as Grade IV GVHD or any other severe adverse even (SAE) deemed to be at least "probably" or "definitely" related to the MVST infusion. Patients were followed to day +100 post SCT for secondary outcomes of efficacy, immune reconstitution and GVHD biomarkers (ST2, REG3). CDR3 sequencing (ImmunoSEQ) was performed on selected MVST products and peripheral blood samples post MVST and compared to control SCT recipients. GVHD biomarkers were analyzed pre- and post-treatment. Results: Twelve subjects were treated: nine received MVSTs generated using DCs and three subjects using mononuclear cells (cohort 3A). Median time from SCT to MVST administration was 13 days (range D +2 to +52 post-SCT). Median time to MVST for subjects in Cohort 3A was 3 days (range 2-7). There were no immediate infusion-related adverse events or DLT at the highest dose level. De novo grade II-III aGVHD post-MVST infusion was seen in three subjects (one in Cohort 1 and two in Cohort 3A), but GVHD biomarker elevation predated MVST infusion. CMV reactivation post-MVST occurred in 6 out of 12 subjects (50%) vs. 45 out of 52 patients (50% vs 87%) in a historical group of recipients of T cell depleted SCT. In all cases CMV reactivation occurred in the context of high dose steroids and in two subjects MVSTs were derived from CMV seronegative donors with minimal anti-CMV activity. One subject experienced rapidly rising AdV viremia (asymptomatic) and received an additional infusion of MVSTs. We saw self-limiting low level EBV replication in 8 cases and one BK viremia, but no disease. CDR3 sequencing of MVST products and serial peripheral blood from subjects revealed a robust contribution of ex vivo expanded cells to the overall repertoire, in contrast to untreated controls where the repertoire of (sham) MVST cell products generated from the transplant donors did not significantly overlap with the immune repertoire in peripheral blood of TCD-SCT recipients in the early post SCT period. Only at day +180 some convergence of repertoires became visible indicating spontaneous immune reconstitution (Figure). Detailed CDR3 analysis of cytokine-captured CMV pp65 and IE-1 specific CD4+ and CD8+ T cells was performed in a representative subject clearly demonstrating the detrimental effect of high-dose steroids on frequency of anti-viral T cells, precipitating CMV reactivation. Conclusions: This is the first report demonstrating safety and feasibility of using MVST immediately post-SCT to rapidly reconstitute anti-viral immunity and ameliorate the detrimental impact of the early viral reactivation in SCT recipient. No DLTs were seen and a minimal risk of aGVHD was observed, as there was no correlation with GVHD biomarkers. As revealed by serial CDR3 sequencing, MVSTs robustly contributed to the T cell repertoire. Our results suggest efficacy of this strategy in reducing viral reactivation. A Phase II study is warranted. Disclosures No relevant conflicts of interest to declare.

Introduction: Cell enrichment and/or depletion by selection is critical for graft engineering in cellular and gene therapies. The semi-automated CliniMACS Plus Instrument (Miltenyi Biotec) is used in clinical cell processing to enrich or deplete a variety of hematopoietic cells including T, B, monocytes, NK cells, and/or CD34+ hematopoietic stem/progenitor cells (HSPCs). Antibody-conjugated paramagnetic beads attach to the desired cells which are separated from the rest of the apheresis or bone marrow product in a closed-system magnetic separation column. Either the selected cells retained in the separation column (positive selection) or those in the "flow-through" fraction (negative selection) may be used for downstream processing. There is sparse data comparing cell selection efficiency and purity across the various selection methods on the CliniMACS Plus Instrument. Further, the effect of donor demographics and other upstream processing events on final viable cell recovery (VCR)/ cell purity (CP) are unknown. This study systematically compared these parameters across various selection methods. Methods: From 2008 to 2018, data were obtained from 45 clinical processing protocols involving 991 consecutive adult and pediatric donor/patient apheresis products or bone marrow harvests, which underwent cell selections on the Miltenyi CliniMACS Plus device. Cell selections included 1.CD34+ (n=669), 2.CD4+ (n=179), 3.CD3-/56+ (n=79), 4.CD3+/19- (n=6), 5.CD4+/8+ (n=21), and 6.CD25- (n=37) cells. Based on protocol design or logistical needs, apheresis or bone marrow collections were processed/selected on the same day or held overnight for next day processing. Products were also either washed to remove platelets (PW) or loaded onto the selection device directly. Data on donor age, BMI, donor type (autologous vs. allogeneic), sex, race, pre-apheresis peripheral blood cell counts and concentrations, as well as apheresis bag blood counts were evaluated. Outcome variables included post-selection VCR and CP. Correlations and multivariate regression analyses were performed for the largest selection type, i.e CD34+ cell selections. Results: Summary data (means ± 1SD) on method-specific enrichments and depletions are shown in the Table. Median post-selection VCR and CP varied by selection method. VCR was the highest for the CD34+ cell selections (Fig. a). CP was the highest and most consistent for CD4+ selections (Fig. b). Platelet and red cell depletions also varied by selection method and averaged 3.9 ± 1.63 and 2.06 ± 0.74 logs, respectively. Selection procedures used for cell depletion resulted in near complete removal of the undesired cell fractions (Table). In multivariate regression analyses, products processed on the same day (compared to those held overnight) (Fig. c), those that underwent a PW (Fig. d), and had higher pre-apheresis peripheral blood CD34+ absolute cell counts and concentrations (Fig. e) were associated significantly with higher CD34+ CP (adjusted r2=0.2; p Conclusions: This is the largest study, to our knowledge, to systematically summarize and analyze Miltenyi CliniMACS Plus selections for clinical graft engineering. Cell selections on the Miltenyi CliniMACS Plus resulted in robust enrichment with near complete depletion of the undesired cell fractions. CD34+ CP was higher in patients with higher CD34+ cells counts following mobilization. Same day processing, as well as PW further improved CD34+ CP. PW also improved CD34+ VCR. This was likely due to a reduction in non-specific platelet binding to the paramagnetic beads. Higher pre-selection CD34+ cell concentrations in the product marginally worsened VCR. This was possibly due to bead saturation in each selection method. Improved CD34+ VCR over time was likely due to increased platelet washes in the latter years. Further studies are ongoing to assess impact of cell selection variations on downstream manufacturing steps (cell transduction, expansion) and clinical (cell engraftment, GVHD incidence) outcomes. Disclosures Larochelle: Stem Cell Technologies: Patents & Royalties: StemDiff Hematopoietic Differentiation Kit.

T cells are potentially curative for patients with metastatic cancer, but many patients with cancer have T cells that are 'terminally differentiated', a condition associated with treatment failure. We have observed that less differentiated T cells have a greater capacity to proliferate, persist and destroy large cancer deposits. Advances in regenerative medicine might allow the generation of rejuvenated T cells from induced pluripotent stem cells (iPSC). We have previously reported that T cells can be generated from iPSC in vitro by co-culturing them OP9 stromal cells expressing Notch-1 ligand, Delta-like-1 (OP9/DLL1). These cells have limited tumor-specificity but also exhibit unconventional and NK cell-like properties demonstrating lineage diversion into alternative lymphoid development pathways, with unknown consequences for their safety and efficacy. To generate iPSC-derived T cells with more naturalistic tumor-specific T cell programs, we sought to restore physiologic signals for selection, maturation and survival. We employed a novel 3D thymic culture system using fetal thymic tissue and generated a novel type of T cell, 'iPSC-derived thymic emigrants' (iTE). Antigen-specific CD8αβ+ iTE exhibited functional properties in vitro that were almost indistinguishable from natural naïve CD8αβ+ T cells, including vigorous expansion and robust anti-tumor activity. iPSC-derived immature T cells generated using OP9/DLL1 and 'educated' in fetal thymic organoids in a 3D culture system resembled naturally-occurring 'young' T cells, as analyzed using whole genome RNA-seq techniques. iTE recapitulated many of the transcriptional programs of naïve T cells in vivo and revealed a striking capacity for engraftment, memory formation and efficient tumor destruction. Although many milestones remain, our data show that 'Next-Gen' autologous tumor-specific T cells can realistically be generated from iPSC using 3D thymic organ tissue. Our next goal is now to employ these cells to treat patients with metastatic cancer because iPSC-derived T cells have a potentially unlimited capacity for proliferation, engraftment and anti-tumor activity. Disclosures Vizcardo: NCI: Patents & Royalties: International Patent Application PCT/US2017/65986. Klemen:NIH/NCI: Patents & Royalties: International Patent Application PCT/US2017/65986. Restifo:NIH/NCI: Patents & Royalties: International Patent Application PCT/US2017/65986.

Background: Progressive Multifocal Encephalopathy (PML) is a subacute infection of the central nervous system (CNS) mediated by John Cunningham (JC) polyomavirus (PyV). The disease is frequently fatal, unless adaptive immunity to JC virus is restored. The disease typically occurs in immunosuppressed patients (e.g. HIV/AIDS). In the modern era, patients with hematological malignancies treated with rituximab and/or fludarabine or following immunosuppressive therapy for transplant or autoimmunity (e.g. multiple sclerosis (MS) or Crohn's disease) are also at risk. JC-PML can also occur in patients with genetic defects of immunity. Specific treatments do not exist and immunocompromised patients with history of cancer or genetic defects of immunity have no realistic chance for rapid recovery of their ability to fight infections. In subjects with Rituximab-related PML the case-fatality ratio is >90%. Survivors can be left with severe neurological disabilities. Methods: Based on our previously established procedure for generation of multi-virus virus-specific T cells using overlapping peptide libraries (pepmixes) as immunogens, we have developed dedicated polyomavirus-specific T cells (PyVST) targeting polyomavirus BK large T (LT) and Viral Protein 1 (VP1) antigens and highly cross-reactive with the structurally-homologous JC LT and VP1 proteins (see Figure). We hypothesized that adoptive transfer of donor-derived PyVSTs could be safely used for therapy of patients with refractory JC-PML. We have developed a pilot study to test the feasibility and safety of adoptive immunotherapy strategy using PyVSTs from the partially matched healthy 1st degree relative donors (sibling, parent or offspring; NIH study 16-N-0072). Adults diagnosed with PML who have no other treatment options were eligible. Patients with MS or HIV were excluded. PyVSTs were generated from donor blood leukocytes cultured for 14 days in G-rex flasks upon stimulation with BK LT and VP1 pepmixes. Subjects underwent baseline physical examination, MRI and lumbar puncture. Upon enrollment subjects received an intravenous infusion of a fixed dose of 1x10e6 (+/-10%) PyVST cells/kg. Safety monitoring period was 28 days after each infusion. Subjects were eligible for up to two additional infusions (dose 2x10e6 PyVSTs/kg) a minimum of 28 days apart if no toxicities were observed. Serial MRI and lumbar punctures were performed to monitor response. Subjects were followed for 12 months after the last infusion. Results: Nine subjects have been enrolled and treated with at least one infusion of PyVSTs under this protocol and the trial is still accruing subjects. No immediate infusion reactions or adverse events have been observed within the safety monitoring period. No CNS immune system reconstitution syndrome (IRIS) has occurred. One subject received a single infusion and withdrew from the study following Day 14 visit due to unwillingness to travel. One subject received two doses of PyVSTs with stabilization of the disease, not requiring the 3rd infusion. This subject was withdrawn from the study just prior to the 1 year follow up visit, as he was diagnosed with the stage IV lung cancer and entered hospice. Seven subjects received 3 infusions of PyVSTs. 3 of these 7 patients died of refractory PML, all >30 days after the 3rd dose of T cells. At the time of this report one subject completed the study at one year with stable PML. The remaining subjects are in the follow-up phase of the protocol. Conclusions: We report that partially matched T cells targeting PyVs generated from healthy related donors can be safely used for adoptive immunotherapy of severely immunocompromised patients with PML. Multiple infusions of T cells are very well tolerated at the doses of up to 2x10e6 PyVST cells/kg, displaying excellent safety profile without adverse events. Furthermore, our data suggest possible efficacy of this strategy as a life-saving therapy for patients who otherwise face a dismal prognosis. Disclosures No relevant conflicts of interest to declare.

Introduction. Allogeneic hematopoietic stem cell transplantation using UCB is an alternative approach for pts (pts) with hematological malignancies lacking an HLA matched donor. However, for pts with severe aplastic anemia (SAA), UCB transplants are associated with delayed engraftment, high graft failure rates and poor survival. Ex-vivo expanded UCB using nicotinamide (NAM) can engraft in NOD/SCID mice, and in pilot studies in pts with hematological malignancies, results in rapid engraftment and durable hematopoiesis. Here we investigated a novel transplantation approach using NAM-expanded UCB in refractory SAA pts who lacked an HLA matched donor, hypothesizing this regimen would accelerate engraftment and immune reconstitution compared to conventional UCB transplants. PTS AND Methods. Eligible pts had SAA with severe neutropenia (ANC Results. From 2017 to 2018, two SAA pts (22 years male and 45 years female, pre-transplant ANC ≤300/uL, and had failed ATG/CSA/Eltrombopag) were successfully transplanted with a single ≥ 6/8 HLA-matched NAM-expanded UCB unit combined with haplo CD34+ cells from a relative. The UCB units before expansion contained a median total nucleated cell (TNC) dose of 2.8 x 107/kg and 1.7 x 105 CD34+ cells/kg. At transplant, the cultured NAM-expanded units contained a median 6.0 x 107 TNCs/kg and 96.4 x105 CD34+ cells/kg, representing a median post TNC and CD34+ cell expansion of 2-fold and 52-fold, respectively. At 12 months and 5 months post-transplant, both pts survive with stable engraftment, transfusion independence, and without acute or chronic GVHD. The median time to neutrophil recovery (ANC > 500/ μL) was only 6.5 days (range 6-7), and platelet recovery was 35.5 days (31-40); chimerism studies showed that both pts achieved >95% cord donor myeloid chimerism and T-cell chimerism at a median 6.5 (6-7) and 23.5 days (21-26) respectively . Immune recovery in both pts receiving NAM-expanded UCB was brisk (Figure 1): absolute CD4+ count > 200 cells/μL occurred at 17 and 60 days; at day 100, median CD4+ numbers was 382/μL and median IGA was 92 mg/dL. In comparison to 16 SAA pts transplanted sequentially at our institute from 2013-2016 using a single unexpanded CBU combined with haplo CD34+ cells, median cord graft doses were 3.6 x 107 TNCs/kg and 1.2 x105 CD34+ cells/kg; b) median time to ANC and platelet recovery was 10 and 51 days; c) median time to >95% cord donor myeloid chimerism was 63 days; d) at day 100, only 3/16 (19%) unexpanded UCB recipients had CD4+ count > 200 cells/μL, and the median CD4+ number was only 74 cells/μL and the median IGA was only 31 mg/dL. This first in human transplant trial suggests neutrophil engraftment, platelet recovery, and post-transplant immune recovery are superior inSAA pts transplanted with NAM expanded UCB compared to conventional nonexpanded UCB (all P Conclusion. These encouraging preliminary results show for the first time that NAM-expanded UCB results in rapid cord engraftment, sustained hematopoiesis and accelerated immune recovery in treatment refractory, neutropenic SAA pts. The high numbers of transplanted CD34+ progenitor cells in NAM-expanded grafts could potentially overcome graft failure associated with conventional UCB transplantation for SAA, obviating the need for co-transplanting haplo CD34+ cells as a stem cell back-up. Disclosures No relevant conflicts of interest to declare.

Conflict of interest may arise when 1 physician serves 2 persons whose medical care is interdependent. In hematopoietic cell transplantation (HCT) from unrelated donors and in the setting of solid organ transplantation from living donors, the standard of care is for donors and recipients to be managed by separate physicians to provide unbiased care. However, the practice patterns of evaluation and care of related donors and recipients are not well described. A survey of HCT centers in the United States was conducted by the Donor Health and Safety Working Committee of the Center for International Blood and Marrow Transplant Research to determine the type of provider involved in medical clearance, informed consent, and medical management of hematopoietic cell collection and the relationship of that provider to the HC transplant recipient. The response rate was 40%. In greater than 70% of centers, transplantation physicians were involved or potentially involved in overlapping care of the HC transplant donor and the recipient. These patterns were similar between transplantation teams caring for adult or pediatric donors and recipients. Among responding centers, medical management of recipients and their related donors by the same provider is common, a practice that has the potential for conflict of interest.

Plerixafor (AMD3100) and granulocyte colony-stimulating factor (G-CSF) mobilize peripheral blood stem cells by different mechanisms. A rhesus macaque model was used to compare plerixafor and G-CSF–mobilized CD34+ cells. Three peripheral blood stem cell concentrates were collected from 3 macaques treated with G-CSF, plerixafor, or plerixafor plus G-CSF. CD34+ cells were isolated by immunoselection and were analyzed by global gene and microRNA (miR) expression microarrays. Unsupervised hierarchical clustering of the gene expression data separated the CD34+ cells into 3 groups based on mobilization regimen. Plerixafor-mobilized cells were enriched for B cells, T cells, and mast cell genes, and G-CSF–mobilized cells were enriched for neutrophils and mononuclear phagocyte genes. Genes up-regulated in plerixafor plus G-CSF–mobilized CD34+ cells included many that were not up-regulated by either agent alone. Two hematopoietic progenitor cell miR, miR-10 and miR-126, and a dendritic cell miR, miR-155, were up-regulated in G-CSF–mobilized CD34+ cells. A pre-B-cell acute lymphocytic leukemia miR, miR-143-3p, and a T-cell miR, miR-143-5p, were up-regulated in plerixafor plus G-CSF–mobilized cells. The composition of CD34+ cells is dependent on the mobilization protocol. Plerixafor-mobilized CD34+ cells include more B-, T-, and mast cell precursors, whereas G-CSF–mobilized cells have more neutrophil and mononuclear phagocyte precursors.

This study describes a novel application of HLAMatchmaker to determine platelet compatibility in 16 alloimmunized patients with aplastic anemia refractory to random donor platelet transfusions. HLAMatchmaker is a software algorithm that predicts HLA compatibility by identifying immunogenic epitopes represented by amino acid triplets in antibody-accessible regions of human leukocyte antigen (HLA) molecules and determines the number of triplet mismatches (TMMs) and highly immunogenic triplet mismatches (HIMMs). Corrected count increments (CCIs) and molecular HLA typing were available for 523 transfusions. Conventional compatibility assessment based on cross-reactive group (CREG) determination was not predictive of transfusion outcome. Low HIMMs and TMMs numbers were associated with a higher likelihood of satisfactory (CCIs ≥ 8) compared with unsatisfactory (CCIs < 8) outcomes (median HIMMs = 4 vs 6, p2 value < .001; median TMMs = 11 vs 13, p2 value < .001). Although receiver operator characteristic curves revealed that HIMMs or TMMs number are not powerful predictors of individual transfusion outcome, a threshold of at least 3 HIMMs or at least 9 TMMs appeared to be associated with successful transfusions. Triplet-matched transfusions were successful, regardless of CREG matching. Our data validate HLAMatchmaker for platelet transfusions and demonstrate its potential to refine and expand donor selection for HLA-alloimmunized patients.

Background: We previously demonstrated that non-T cells in apheresis products collected from patients with cancer, particularly in those with circulating leukemic cells or low CD3 counts, can substantially affect T cell expansion and transduction and, potentially, in vivo efficacy. Flask adhesion and elutriation methods to remove monocytes and granulocytes, and separation using CD3/CD28 activation beads are inefficient at removing tumor and other inhibitory cells. In the context of an ongoing CD22 CAR clinical trial, we introduced magnetic CD4/CD8 selection prior to activation and lentiviral transduction which resulted in enhanced in vivo CAR T-cell potency compared to products generated prior to the introduction of CD4/CD8 T cell selection. Design: Children and young adults with relapsed/refractory CD22+ hematologic malignancies eligible for our phase I dose escalation anti-CD22 CAR protocol were enrolled on study (NCT02315612). All had bone marrow evaluations at baseline, prior to lympho-depleting chemotherapy (Fludarabine 25 mg/m2 x 3 days and Cyclophosphamide 900 mg/m2 x 1 day) and again at day 28 (+/- 4 days) post-CAR infusion. Three dose levels were explored during dose-escalation (3 x 10e5; 1 x 10e6 and 3 x 10e6 transduced CAR T-cells/kg) prior to treating in the expansion phase at the 1 x 10e6 dose level, which is the focus of this report. As a result of a manufacturing failure in an enrolled patient, as of January 2017, the CAR-T manufacturing process was modified to include CD4/CD8 bead selection (TCS) for all subsequent patients. CAR T cell activation, transduction and expansion processes were otherwise unchanged. Results: From December 2014 to July 2017, 30 patients with ALL were treated, of which 22 (73%) were treated at 1 x 10e6 transduced CAR T cells/kg. All patients had active bone marrow involvement at baseline and 18 (82%) had an M2 marrow (>5% blasts) or higher disease burden. Eighteen had a prior transplant and 14 were previously treated with CAR. The first 15 patients were treated using an unselected apheresis product and 7 patients were treated using CD4/CD8 TCS. The first patient treating using TCS initially had a product failure when CAR-T cells were manufactured without TCS, due in part to a low CD3 at the time of collection resulting in substantial numbers of non-T cells in the apheresis product. Following CD4/CD8 TCS, the CD3 T-cell content in the starting product increased from 47% to 90%, T cell proliferation increased from 1.97-fold to 24.2-fold, transduction increased from 2.57% to 33.4% and an MRD negative complete remission was achieved. Following TCS, median transduction efficiencies were significantly higher (33.4% (pre-TCS) vs 40.7%, p=0.02). Cytokine release syndrome (CRS) occurred in 19 of 22 patients and was grade 1 CRS in 12 (4 in the TCS arm); grade 2 in 6 (2 in the TCS arm) and grade 4 in one patient. Amongst those developing CRS, tocilizumab was utilized in 2 of 13 (15%) patients pre-TCS and in 3 of 6 (50%) patients in the TCS arm to prevent higher grade CRS. Steroids were administered to 1 of 13 patients pre-TCS versus 3 of 6 (50%) patients in the TCS arm. Clinical parameters suggestive of a more robust inflammatory response in TCS patients included statistically significantly higher peak IL-6, IL-8, IL-10 and IL-15 levels as well as higher ferritin (Figure), and higher peak CAR-T% (75% vs 82%, p=0.03) with a trend towards higher absolute CAR values. Furthermore, 3 of the 5 responders in the TCS arm had evidence for a secondary expansion with a late rise in ferritin, WBC and CAR-T% with 2 patients having confirmed hemophagocytosis on the bone marrow at day 28, which was not seen in patients enrolled pre-TCS. For the entire cohort, 16 of 21 (76%) patients evaluable for response attained a complete remission (11 of 15 enrolled pre-TCS and 5 of 7 post-TCS). Amongst the TCS group, 5 of 5 who were CD22 CAR naïve attained MRD negative remission; two patients pre-treated with CD22 CAR T at an outside institution experienced poor in vivo CAR expansion possibly due to immune rejection. Conclusion: Ongoing experience on our expansion cohort confirms activity of CD22 CAR T cells resulting in high remission induction rates in relapsed refractory ALL, including responses in patients previously treated with CD19 CAR T cells. T cell selection using CD4/CD8 positive selection led to enhanced in vivo CAR-T expansion resulting in a 100% MRD negative complete remission rate (5 of 5 patients) in CD22 CAR naïve patients. Figure Figure. Disclosures No relevant conflicts of interest to declare.

Delayed donor red cell engraftment and pure red cell aplasia (PRCA) are well-recognized complications of major ABO-incompatible hematopoietic stem cell transplantation (SCT) performed by means of myeloablative conditioning. To evaluate these events following reduced-intensity nonmyeloablative SCT (NST), consecutive series of patients with major ABO incompatibility undergoing either NST (fludarabine/cyclophosphamide conditioning) or myeloablative SCT (cyclophosphamide/high-dose total body irradiation) were compared. Donor red blood cell (RBC) chimerism (initial detection of donor RBCs in peripheral blood) was markedly delayed following NST versus myeloablative SCT (median, 114 versus 40 days;P

We have recently shown that more than 90% of long-term culture initiating cells (LTC-IC) mobilized in the peripheral blood (PB) of normal individuals express HLA-DR and CD38 antigens and can sustain hematopoiesis for only 5 weeks. However, 10% of LTC-IC in mobilized PB are CD34+ HLA-DR- and CD34+ CD38- and can sustain hematopoiesis for at least 8 weeks. We now examine the ex vivo expansion potential of CD34+ HLA-DR+ cells (rich in mature LTC-IC) and CD34+ HLA-DR- cells (rich in primitive LTC-IC) in granulocyte colony-stimulating factor (G-CSF) mobilized PB progenitor cells (PBPC). Cells were cultured in contact with M2-10B4 cells (contact) or in transwells above M2-10B4 (noncontact) without and with interleukin-3 (IL-3) and macrophage inflammatory protein (MIP-1alpha) for 2 and 5 weeks. Progeny were evaluated for the presence of colony-forming cells (CFC) and LTC-IC. When CD34+ HLA-DR+ PB cells were cultured in contact cultures without cytokines, a threefold expansion of CFC was seen at 2 weeks, but an 80% decrease in CFC was seen at week 5. Further, the recovery of LTC-IC at week 2 was only 17% and 1% at week 5. This confirms our previous observation that although CD34+ HLA-DR+ mobilized PB cells can initiate long-term cultures, they are relatively mature and cannot sustain long-term hematopoiesis. In contrast, when CD34+ HLA-DR- mobilized PB cells were cultured in contact cultures without cytokines, CFC expansion persisted until week 5 and 49% and 11% of LTC-IC were recovered at week 2 and 5, respectively. As we have shown for steady state bone marrow (BM) progenitors, recovery of LTC-IC was threefold higher when CD34+ HLA-DR- PBPC were cultured in noncontact rather than contact cultures, and improved further when IL-3 and MIP-1alpha were added to noncontact cultures (96 +/- 2% maintained at week 5). We conclude that although G-CSF mobilizes a large population of "mature" CD34+ HLA-DR+ LTC-IC with a limited proliferative capacity, primitive CD34+ HLA-DR- LTC-IC present in mobilized PB have similar characteristics as LTC-IC from steady state BM: (1) they can be maintained in noncontact cultures containing IL-3 and MIP-1alpha for at least 5 weeks; (2) they are subject to the same proliferation inhibitory influences of contact with stroma. Since the absolute number of primitive LTC-IC (week 8 LTC-IC) per mL of G-CSF mobilized PB is similar to that per mL of steady state BM, these studies further confirm that G-CSF mobilized PBPC may have similar long-term repopulating abilities as steady state BM. ispartof: Blood vol:89 issue:11 pages:3991-7 ispartof: location:United States status: published

Polymorphic structures of the neutrophil Fcγreceptor IIIb (FcγRIIIb) result in alloantibody formation that causes alloimmune neonatal neutropenia and transfusion reactions. Alloantigens located on FcγRIIIb include the antigens NA1 and NA2. In four cases of alloimmune neonatal neutropenia, granulocyte-specific alloantibodies directed against a thus far unknown antigen were detected by granulocyte agglutination and immunofluorescence tests in the maternal sera. By the use of the monoclonal antibody–specific immobilization of granulocyte antigens (MAIGA) assay, the new antigen, termed SH, was located on the FcγRIIIb. Nucleotide sequence analysis of the FcγRIIIb coding region from a SH(+) individual showed a single-base C→A mutation at position 266, which results in an Ala78Asp amino acid substitution. A family study confirmed that this nucleotide difference is inherited, and corresponds to the SH phenotype. Serologic typing of 309 randomly selected individuals showed an antigen frequency of 5% in the white population. The same frequency was found by genotyping, for which a technique based on polymerase chain reaction (PCR) using sequence-specific primers (PCR-SSP) was developed. Typing of all SH(+) individuals for NA1 and NA2, and PCR-restriction fragment length polymorphism analysis of the NA-specific PCR products from five SH(+) individuals using the SH-specific endonuclease SfaN I showed that SH antigen is very probably the result of an additional mutational event in the NA2 form of the FcγRIIIB gene. Immunochemical studies also demonstrated that the SH determinants reside on the 65- to 80-kD NA2 isoform of the FcγRIIIb. Our findings show the existence of an additional polymorphism of the FcγRIIIb, which can result in alloantibody formation causing alloimmune neonatal neutropenia.

Relapsed pre-B acute lymphoblastic leukemia (ALL) portends a poor prognosis even with hematopoietic stem cell transplantation (HSCT). CD19 chimeric antigen receptor (CAR) T cells have shown promise in early studies although morbidity related tohigh gradecytokine release syndrome (CRS) and/or neurotoxicity could limit its wide applicability in patients with high disease burden. The lympho depleting chemotherapy regimen may affect both toxicity and response and has not been well studied. Relapse rates among complete responders to CD19 CAR therapy occur in nearly half of patients in the first year. We report outcomes from our completed clinical trial of 53 children and young adults with relapsed/refractory ALL (n=51) or lymphoma (n=2) with a median follow up (mF/U) of 18.7 months. The first 21 patients received a low dose fludarabine (25 mg/m2/day Days -4 to -2) and cyclophosphamide (900 mg/m2 Day -2) preparative regimen (LDflu/cy) and results are reported in Lancet 385:517-28. The regimen for the subsequent 32 patients, who all received 1x106 CAR+ T cells/kg, was stratified based on disease burden. Subjects with low burden ALL (lowALL; 25% marrow blasts or lymphomatous disease) received an alternative regimen [FLAG (n=6), ifosfamide/etoposide per AALL0031 (n=2) or fludarabine (30mg/m2/day Days -6 to -3) and cyclophosphamide (1200 mg/m2/day Days -4 and -3) (HDflu/cy; n=8)] in an attempt to mitigate severe CRS risk and improve response. Four highALL subjects received LDflu/cy due to comorbidities including Trisomy 21. CRS was graded and anti-cytokine therapy was instituted as per Blood 124:188-95. Date for data cutoff was July 31, 2016. Of the 53 subjects 11 had primary refractory ALL, 5Ph+, 3 with Trisomy 21, 4 with CNS2 and 2 with CNS3 ALL including one with extensive leptomeningeal and parenchymal involvement. Cells were manufactured in 7-11 days and none underwent a test expansion. One patient was not infused due to rapidly progressive fungal pneumonia but was accounted for in all analyses. Of 51 ALL patients, 31 (60.8%) achieved a complete response (CR) with 28/31 (90%) of responders negative for minimal residual disease (MRD-). All 6 subjects with CNS ALL were rendered into CNS1 status with resolution of leptomeningeal enhancement, where appropriate, and CAR cells in CSF. The median leukemia free survival (mLFS) of MRD- CR responders is 18 months with a 49.5% probability of LFS beginning at 18 months (mF/U 22.6 months). Grade 3 (n=5) and 4 (n=2) CRS combined for a severe CRS incidence of 13.5%. Three grade 3 neurotoxicities(1 each: dysphasia, delirium, headache) and 2 seizures (one grade 1, one grade 2) occurred. There were no grade 4 neurotoxicities, even in the subject with extensive CNS disease. Subjects with low ALL had a significantly higher CR rate (18/21; 85.7%) than those with high ALL (13/32; 40.6%) (p=0.0011) and use of a flu/cy regimen correlated with higher response (29/44; 65.9% vs 2/8; 25%; p=0.0301). Overall survival in all subjects receiving a flu/cy regimen was 13.3 months with a 34.7% probability of survival beginning at 38 months (mF/U 18.7 months), which is significantly longer than those who did not receive a flu/cy regimen (5.5 months, no survivors beyond 11 months). The hazard ratio (HR) of not receiving a flu/cy regimen was 6.35 (1.906-21.14; p=0.0026). mLFS of subjects with MRD- CR who received a flu/cy regimen was not reached with a 53.3% probability of LFS beginning at 18 months (mF/U 22.6 months). Of the 28 subjects achieving MRD- CR, 21 had a subsequent HSCT with a median time to HSCT of 54 days from CAR infusion. 8/28 (28.6%) relapsed with CD19+ (n=2), CD19-/dim (n=5), CD19 unknown (n=1) blasts. Relapse was significantly more common in subjects who did not have a HSCT after CAR therapy (6/7; 85.7%) compared to those who did (2/21; 9.5%) (p=0.0001). Even accounting for transplant related mortality, them LFS in the HSCT group was not reached with a 62% probability of LFS beginning at 18 months. This is significantly longer than them LFS of 4.9 months in MRD- CR subjects who did not proceed to HSCT (p=0.0006) with a HR of 16.9 (3.37-85.1) of not having a subsequent HSCT. In all, CD19 CAR T cell therapy was effective and safe with a low incidence of severe CRS and neurotoxicity. In this nonrandomized series, the rate of durable remission was higher when a flu/cy preparative regimen was used and consolidation HSCT was employed. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures Lee: Juno: Honoraria. Kochenderfer:bluebird bio: Patents & Royalties, Research Funding; Kite Pharma: Patents & Royalties, Research Funding. Rosenberg:Kite pharma: Research Funding. Mackall:NCI: Patents & Royalties: B7H3 CAR.

Background: Pain, symptoms and adverse events (AEs) after first donation of bone marrow (BM) or mobilized peripheral blood stem cells (PB) from unrelated donor subjects have been well documented. However, little is known about the safety of second donation experiences. Methods: Second donations of NMDP unrelated donors performed between 2004 and 2013 were evaluated (n=720). Two groups were studied; those where 1) the second donation was PB (PB-PB, n=362 or BM-PB, n=240), compared with the first donation experience of those making only one PB donation (PB, n=16,095), or 2) the second donation was BM (BM-BM, n=42, or PB-BM, n=76), compared with the first donation experiences of those making only one BM donation (BM, n=5,829). Recovery, modified toxicity criteria (MTC), pain, collection yield and serious adverse events (SAEs) were compared. Results: Most second donations were for the same recipient (n=515, 72%); BM-PB donors (85%) vs PB-PB donors (68%, p For second PB donations there was no difference in the use of a central venous collection catheter, need for a 2-day collection, G-CSF dose or procedure duration among the 3 donor groups. The total CD34+ cell yield for PB donors who only donated once (median = 657x106) was significantly higher (p The major finding was that for donors who gave second donations of either PB or BM there were no significant differences in skeletal pain, common toxicities and recovery compared to donors who donated once (Figure 1a and 1b). In the second PB donation group multivariate analysis found that maximum skeletal pain (grade 2-4), maximum MTC (grade 2-4) and time of recovery were dependent on donor sex and age but not donor group. Maximum skeletal pain was also dependent on BMI and maximum MTC was also dependent on BMI, race and collection year. In the second BM donation group maximum skeletal pain was dependent on race, sex and CMV status and time of recovery was dependent on sex and age at donation. Neither was dependent on donor group. Prognostic factor analysis of PB-PB and BM-PB donations revealed several important points concerning second PB donations (Table 1). a) There is an increased risk of MTC or skeletal pain if the first donation was BM and second was PB. b) High maximum MTC with first donation predicts high maximum MTC or pain with the second donation. c) High pain with first donation predicts similar pain with second donation. d) High BMI predicts high MTC with the second donation. e) >12 months between donations, high maximum MTC with the first donation and slower recovery after the first PB donation predicts slower recovery after the second donation. Conclusions: Second PB and BM donation experiences were similar to first donation experiences for pain, MTC, recovery and SAEs, but yields of second grafts were overall lower for both PB and BM collections. Low first donation collection yield was associated with the need for a second donation. This data provides reassurance that second donations of either BM or PB do not increase risk to unrelated donors. Knowledge of the donor's first experience can help predict what to expect for their second donation and allows for appropriate counseling. Disclosures Pulsipher: Novartis: Consultancy, Other: Study Steering Committee; Jazz Pharmaceutical: Consultancy; Chimerix: Consultancy; Medac: Other: Housing support for conference.

Background: Allogeneic stem cell transplant (SCT) recipients suffer from a defective T cell mediated immunity causing potentially fatal reactivation of latent viruses. After T cell depleted SCT we have observed over 80% CMV reactivation and significant additional costs ($58-74k/patient). Despite aggressive monitoring and pre-emptive therapy, reactivation and/or positive CMV serology brings a significantly higher risk for non-relapse mortality (NRM). Adoptive transfer of ex vivo generated virus specific donor T cells is effective as a treatment of infection post-SCT but it has not been tested as a prophylaxis of early reactivation. Here in a Phase I study we transferred multi-virus specific T cells (MVSTs) immediately post SCT, targeting CMV, Ebstein-Barr virus (EBV), BK and adenovirus (Ad) as a novel strategy to prevent viral reactivation in the recipients of T cell depleted sibling HLA-matched SCT. Methods: Subjects were eligible if enrolled in HLA-matched T cell depleted transplant protocol (13-H-0144) and deemed at risk for CMV reactivation. MVST cells were manufactured from SCT sibling donors. Elutriated lymphocytes were stimulated with autologous dendritic cells (DCs) pulsed with seven overlapping peptide libraries (pepmixes) spanning the length of immunodominant proteins from CMV (pp65 and IE1), EBV (BZLF1 and EBNA1), BK (LT and VP1) and Ad5. Cultures were maintained in G-Rex flasks for 14 days in presence of IL-7, IL-15 and IL-2 (after 72hrs), tested for sterility, phenotype, potency and cryopreserved. MVST cells were thawed and administered intravenously as early as possible (day 0 to +60) post SCT. A Phase I 3+3 dose escalation design was used at the following dose levels: Cohort 1 - 1x10e5 total nucleated cells (TNC)/kg, Cohort 2 - 5x10e5 TNC/kg, Cohort 3 - 1x10e6 TNC/kg. The primary safety endpoint at day 42 post infusion was the occurrence of dose limiting toxicity (DLT), (Grade IV GVHD or any other severe adverse even (SAE) deemed to be at least "probably" or "definitely" related to the MVST infusion. Patients were followed to day +100 post SCT for secondary outcomes, including efficacy (Figure) and immune reactivity (for donor/recipient pairs). Results: MVST cells recognized the majority of pepmixes, were polyfunctional and robustly proliferated in response the cognate antigens, but minimally against allogeneic targets- suggesting a limited ability to induce GVHD. CDR3 sequencing of T cell repertoire showed a significant reduction in diversity and a striking dominance of a limited number of clonotypes in the final MVSTs. Nine subjects were enrolled and treated with MVST cells. MVSTs were successfully generated for all subjects, meeting the release criteria. Median time from SCT to MVST administration was 16 days (range D +6 to +52 post-SCT). Two subjects received MVST after day +30 due to cardiac instability and scheduling. There were no immediate infusion-related adverse events or DLT by day 42. One subject in cohort II developed a self-limiting grade I cytokine release syndrome in the setting of low-level EBV reactivation. One patient (cohort 1) developed de novo grade III aGVHD post-MVST infusion. CMV reactivation post-MVST occurred in 4 out of 8 evaluable subjects (50%) who completed D+100 post-SCT vs. 45 out of 52 patients (50% vs 87%; p value=0.031) in a historical cohort of recipients of T cell depleted SCT. In all cases CMV reactivation occurred during treatment with high dose steroids. In two cases MVST were generated from CMV seronegative donors and showed minimal activity against pp65 and IE1. In eight evaluable subjects who reached D+100 post-SCT there was no EBV-related disease, but we saw self-limiting low level EBV replication in 6 out of 8 cases. There were no cases of BK or Ad-related disease or viremia. ELISPOT analysis at D+100 revealed robust reconstitution of anti-viral immunity in analyzed recipients (vs. donors, Figure, B) Conclusions: This is the first report demonstrating that it is safe and feasible to use adoptively transferred allo-MVST immediately post-SCT to rapidly reconstitute anti-viral immunity and ameliorate the detrimental impact of the early viral reactivation in SCT recipient. No DLTs were seen. MVSTs had a markedly reduced allo-reactivity and carried a minimal risk of GVHD. Our results also suggest efficacy of this strategy in reducing viral reactivation. A Phase II portion of this study is currently enrolling patients. Figure. Figure. Disclosures Sabatino: Kite: Employment, Equity Ownership.

INTRODUCTION: Ex-vivo T cell depletion strategies have been widely used to reduce the incidence of graft versus host disease (GVHD) in allogeneic stem cell transplantation (allo-SCT). Although several options of ex-vivo graft manipulation strategy are available, direct comparison between strategies along with relevant biomarkers has been lacking. Here we evaluated cellular and plasma biomarkers in two separate graft manipulation strategies, CD3-CD19 depletion versus CD34+ selection using the Miltenyi CliniMACS and their association with clinical outcomes. METHODS: Forty two subjects with hematological malignancies underwent HLA matched sibling allo-SCT at a single center between 2012 and 2015 and received either an ex-vivo CD3-CD19 depleted, CD34+ negatively selected graft (CD3/19D, n=20) or an ex-vivo CD34+ cell positively selected graft (CD34S, n=22). Both cohorts were treated with the same conditioning regimen of cyclophosphamide, fludarabine, and total body irradiation (600-1200 cGy) and GVHD prophylaxis of low dose cyclosporine. Peripheral blood mononuclear cells and plasma samples were collected at days 14 or 30, 60, 100 post-transplant. Post-transplant cellular immune reconstitution was evaluated by multi-color flow cytometry immunophenotyping, characterizing the subsets of memory T cells, regulatory T cells (Tregs), natural killer (NK) cells, and B cells with various functional markers. The plasma levels of ST2, Reg3α, and sTNFR1 were measured using enzyme-linked immunosorbent assay (ELISA). RESULTS: The median age at transplant was 48 years (range 17-70) in CD3/19D and 45 years (11-73) in CD34S. At a median follow up of 37 months in CD3/19D and 22 months in CD34S, the major clinical outcomes were similar between two groups; the overall survival (70% and 86%), non-relapse mortality (5% and 4.5%), and cumulative incidence of relapse (35% and 39%) at 2 years, respectively. Two subjects in CD3/19D developed late engraftment failure before day 100 but all other subjects achieved primary neutrophil and platelet recovery. Unexpectedly, the cumulative incidence of grade II-IV acute GVHD was higher in CD3/19D (61%) in comparison to the incidence in CD34S (32%, P=0.07, Figure). The cumulative incidence of extensive chronic GVHD was 33% in CD3/19S and 24% in CD34S. The fraction of Helios negative Tregs post-transplant was significantly lower in CD3/19D (median [interquartile range]: 10.4% [7.1-16.4] at day 30; 4.9% [3.0-8.3] at day 60) compared to CD34S (23.8% [10.7-35.8], P=0.03 at day 30; 8.8% [6.8-18.4], P=0.01 at day 60, Figure). Plasma ST2 levels were significantly higher in CD3/19D (45ng/mL [27-67] at day 14; 33ng/mL [27-62] at day 28) in comparison to CD34S (29ng/mL [19-40], P=0.03 at day 14; 25ng/mL [14-33], P=0.03 at day 28, Figure). In addition, significantly higher CD4 naive T cells, lower effector memory and PD-1 bright CD4 T cells were observed in CD3/19D in comparison to CD34S. NK and B cell profiles were not significantly different between the two groups. CONCLUSION: Both methods of ex vivo TCD were associated with extremely low NRM rates (~5%).We observed a higher cumulative incidence of acute GVHD in the recipients of CD3/19 depleted grafts, accompanied with the distinct biomarker profiles of poor Treg reconstitution and high level of ST2. CD3/19 depletion may have disproportionately depleted Tregs in the graft, leading to uncontrolled tissue damage and GVHD evidenced by higher ST2 levels. Further validation is required to confirm the utility of monitoring Treg reconstitution and ST2 level as biomarkers to predict the outcomes of T cell depleted allo-SCT. Figure 1. Figure 1. Disclosures Battiwalla: NIH/NHLBI: Employment.

Background: Chimeric antigen receptors (CARs) are fusion proteins that combine antigen-recognition domains and T-cell signaling domains. T cells genetically modified to express CARs directed against the B-cell antigen CD19 can cause remissions of B-cell malignancies. Most CARs in clinical use contain components derived from murine antibodies. Immune responses have been reported to eliminate CAR T cells in clinical trials, especially after second infusions of CAR T cells (C. Turtle et al., Journal of Clinical Investigation, 2016). These immune responses could be directed at the murine components of CARs. Such immune responses might limit the persistence of the CAR T cells, and anti-CAR immune responses might be an especially important problem if multiple infusions of CAR T cells are administered. Development of fully-human CARs could reduce recipient immune responses against CAR T cells. Methods: We designed the first fully-human anti-CD19 CAR (HuCAR-19). The CAR is encoded by a lentiviral vector. This CAR has a fully-human single-chain variable fragment, hinge and transmembrane regions from CD8-alpha, a CD28 costimulatory domain, and a CD3-zeta T-cell activation domain. We conducted a phase I dose-escalation trial with a primary objective of investigating the safety of HuCAR-19 T cells and a secondary objective of assessing anti-lymphoma efficacy. Low-dose chemotherapy was administered before HuCAR-19 T-cell infusions to enhance CAR T-cell activity. The low-dose chemotherapy consisted of cyclophosphamide 300 mg/m2 daily for 3 days and fludarabine 30 mg/m2 daily for 3 days on the same days as cyclophosphamide. HuCAR-19 T cells were infused 2 days after the end of the chemotherapy regimen. Patients with residual lymphoma after a first treatment were potentially eligible for repeat treatments if dose-limiting toxicities did not occur with the first treatment. Repeat infusions were given at the same dose level as the first infusion or 1 dose level higher than the first infusion. Findings: A total of 11 HuCAR-19 T-cell infusions have been administered to 9 patients; 2 patients received 2 infusions each. So far, there is an 86% overall response rate (Table). Grade 3 adverse events (AEs) included expected cytokine-release syndrome toxicities such as fever, tachycardia, and hypotension. Corticosteroids were used to treat toxicity in Patient 3. The interleukin-(IL)-6 receptor antagonist tocilizumab was used to treat toxicity in Patient 4, and both tocilizumab and corticosteroids were used to treat toxicity in Patient 8. Only 1 of 8 evaluable patients, Patient 3, has experienced significant neurological toxicity to date. This patient experienced encephalopathy that was associated with a cerebrospinal fluid (CSF) white blood cell count of 165/mm3. Almost all of the CSF white cells were CAR T cells, and the CSF IL-6 level was elevated. All toxicities have resolved fully in all patients. In Patient 1, tumor biopsies revealed a complete loss of CD19 expression by lymphoma cells after 2 HuCAR-19 T-cell infusions, which to our knowledge is the first documented complete loss of CD19 expression by lymphoma after anti-CD19 CAR T-cell therapy. This loss of CD19 expression was associated with lymphoma progression. After first CAR-19 T-cell infusions, HuCAR-19 cells were detectable in the blood of every patient. The median peak number of blood CAR+ cells was 26/microliter (range 3 to 1005 cells/microliter). Blood HuCAR-19 cells were detected after second infusions in the blood of both patients who received second infusions. Patient 1 obtained a partial response after a second infusion after only obtaining stable disease after a first infusion. We detected elevations of inflammatory cytokines including IL-6, interferon gamma, and IL-8 in the serum of patients experiencing clinical toxicities consistent with cytokine-release syndrome. Interpretation: T cells expressing HuCAR-19 have substantial activity against advanced lymphoma, and infusions of HuCAR-19 T cells caused reversible toxicities attributable to cytokine-release syndrome. Disclosures Kochenderfer: Kite Pharma: Patents & Royalties, Research Funding; bluebird bio: Patents & Royalties, Research Funding.

INTRODUCTION: Haploidentical allogeneic stem cell transplantation (haplo-SCT) incurs a risk of bidirectional immune reactions with either severe acute graft versus host disease (aGVHD) or graft rejection. Induction of immune tolerance with post-graft cyclophosphamide dramatically improves the outcomes of haplo-SCT. However the optimal duration and the combination of systemic immunosuppressive agents in haplo-SCT remain controversial. Ultra-low dose interleukin 2 (ULD IL-2) preferentially expands regulatory T cells (Tregs) and natural killer (NK) cells, promoting both GVHD prevention and graft versus leukemia (GVL) effects. These properties suggest that ULD IL-2 could play a useful role in haplo-SCT. Here we report the outcomes of a pilot study to determine the safety and feasibility of ULD IL-2 as GVHD prophylaxis in haploidentical allo-SCT (14-H-0180, Clinical Trials.gov ID: NCT02226861) METHODS: Ten subjects with high risk hematological malignancies received a myeloablative conditioning regimens of fludarabine 120mg/m2 (day -10 to day -8) and total body irradiation (TBI; 600-1200 cGy, day -10 to day -6). Thereafter the subjects received donor lymphocyte infusion (DLI) products (2x108 CD3+/kg) on day -6, followed by post-DLI cyclophosphamide 120mg/kg on days -3 and -2. CD 34+ selected, peripheral blood stem cell product was infused on day 0. Sirolimus was initiated on day -1 with goal trough level of 5-12ng/mL until day+60. ULD-IL2 (aldesleukin, 100,000 IU/m2) was given subcutaneously daily for 12 weeks starting day +1. Peripheral blood mononuclear cells (PBMC) and plasma samples were collected at days 14, 28, 60, 84, 100 post-transplant. Multi-color flow cytometry immunophenotyping assay were performed to characterize the subsets of memory T cells, Tregs, NK cells, and monocytes with various functional markers. Plasma levels of biomarkers (ST2, Reg3α, sTNFR1, ANG1, IL-6) were measured using a multiplex microfluidic channel assay. RESULTS: The median age at transplant was 35 years (range 20-66). Most subjects had a high risk EBMT score (median 4, range 2-7) and HCT co-morbidity index (median 4, range 2-7). All subjects achieved successful engraftment (neutrophil >500/uL; median 13 days, platelet >20k/uL; median 15 days) and rapid full donor myeloid and lymphoid chimerism by day 21. At median follow up of 6 month, the overall survival was 71%. One subject died of hepatic veno-occulusive disease (VOD) on day 32 and one subject died of relapse on day 178. All evaluable subjects tolerated ULD-IL2 without significant toxicities. Four subjects experienced either de-novo or rapid exacerbation of acute GVHD after discontinuation of ULD IL-2, resulting in the cumulative incidence of grade II-IV aGVHD of 61% and grade III-IV aGVHD of 36% (Figure A). ULD IL-2 expanded and maintained Helios+FoxP3+Tregs population (pre-transplant, 4.7%±3.1%; day 30, 36.2%±23.1%; day 84, 17.4%±10.6%) as well as CD56brightNK cells population (pre-transplant, 10.7%±13.7%; day 30, 49.7%±10.8%; day 84, 26.1%±6.8%). However on discontinuation of ULD IL-2 both populations decreased to low levels within one week. The timing of aGVHD correlated with a fall of %Tregs in PBMC and a sharp increase of ST2 level in plasma (Figure B). CONCLUSION: ULD IL-2 can be safely administered as GVHD prophylaxis after haplo-SCT. Rebound GVHD after discontinuation of ULD IL-2 implies that donor-derived Tregs acquired dependency to exogenous ULD IL-2. Our study is proof of principle that ULD IL-2 induces immune tolerance through Tregs expansion in haplo-SCT, inspiring further clinical and basic researches in human IL-2 biology. Figure. Figure. Disclosures No relevant conflicts of interest to declare.

Introduction: Allogeneic hematopoietic stem cell transplantation (HSCT) is curative for SAA unresponsive to immunosuppressive therapy (IST). For SAA patients (pts) with a HLA-identical sibling donor, the use of peripheral-blood (PB) grafts have been reported to have increased risk of graft-versus-host disease (GVHD) compared with using the bone marrow (BM) as the stem cell source. For pts lacking an HLA-identical donor, transplantation using unrelated cord blood (UCB) or haplo-identical donors has historically been associated with high graft failure rates and poor survival. Here we report and compare the clinical outcomes for two novel transplantation strategies under investigation at our center aimed at improving engraftment and survival and reducing GVHD in high-risk SAA pts refractory to IST: combined haploidentical and cord-blood (haplo-cord) transplants vs. HLA-matched sibling (matched-sib) donor PB HSCT with partial T-cell depletion. Methods: Forty-seven pts with SAA or SAA evolved to MDS that was unresponsive to IST underwent HSCT at a single center between 2008 and 2016, and received either a haplo-cord transplant (n= 28; pt eligibility included lack a HLA-matched donor but availability of a haploidentical relative, a ≥4/6 HLA matched UCB unit and severe neutropenia with an ANC Figure Figure. Disclosures Young: GSK/Novartis: Research Funding.

Introduction: Little is known about the lethality of acute GVHD (aGVHD) in T -cell depleted (TCD) allogeneic hematopoietic stem cell transplantation (HSCT). We examined the incidence of aGVHD and its relative contribution to non relapse mortality (NRM) in a cohort of consecutive TCD HSCT at a single institute. Methods: We report 132 consecutive subjects who had undergone TCD HLA-identical sibling HSCT between 2006 and 2016 for hematologic malignancies. All subjects received conditioning with Fludarabine 125mg/m2, Cytoxan 120 mg/kg and 1200 cGy total body irradiation (TBI) ( Thirty-six subjects were over the age of 55 years, 66 were females and 66 were males. Transplant indications were acute leukemia (92), myelodysplastic / myeloproliferative syndromes (26), lymphoid malignancies (7) and chronic myelogenous or myelomonocytic leukemia (7). Fifty-three percent of subjects were at a high risk of relapse. The median follow-up post HSCT was 4.3 years. Results: The incidence of Grade 1-2 and 3-4 aGVHD were 51.6% and 19%, respectively. Of these, 15% of aGVHD was steroid refractory and was treated with infliximab/basiliximab or with mesenchymal stromal cells. Extensive or limited chronic GvHD was observed in 30% and 24% of subjects. 74% of those at risk developed cytomegalovirus (CMV) reactivation. The Kaplan-Meier (KM) estimate for overall survival (OS), NRM and cumulative incidence of relapse (CIR) for the entire cohort was 71%, 14.6% and 24% respectively at one year, and 52%, 32.5% and 39% at 5 years. We considered the impact of CMV reactivation, slow donor lymphoid chimerism and steroid refractory aGVHD on NRM. However there was no significant impact from CMV reactivation or slow (> 31 days) achievement of complete donor lymphoid chimerism. Significantly improved outcomes were noted for those transplanted beyond 2012: OS, NRM and CIR being 82%,6.2% and 20% at one year, and 68%, 6.2% and 41% at 3 years. Cox proportionate hazard modeling identified steroid refractory aGVHD (HR 4.0, P=0.007) and transplant prior to 2012 (HR 6.7, P=0.001) as significant factors impacting NRM. Conclusions: T cell addback after ex vivo TCD HSCT was associated with a significant burden of aGVHD. Steroid refractory aGVHD impacted NRM, but slow lymphoid engraftment, disease risk, CMV reactivation and age did not. Significant improvements in NRM in the current era suggest greatly improved salvage of steroid refractory aGVHD. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.

We have previously described a 24-year-old woman with quinine-dependent antibodies that reacted with neutrophils, red blood cells (RBCs), platelets, and T lymphocytes. The drug-dependent neutrophil antibody was found to react with 85- and 60-Kd neutrophil membrane molecules. In these studies, we further characterized these molecules and found that both were glycosyl-phosphatidylinositol (GPI)-linked and contained sialic acid residues and N-linked carbohydrate side chains, but neither contained O-linked carbohydrates. The protein backbone of the 60-Kd molecule was 45 Kd, and the 85 Kd glycoprotein (GP) was made up of 33- and 31-Kd proteins. While some GPI-anchored neutrophil GPs are released by stimulated neutrophils, neither the 85- nor the 60-Kd GP was released by neutrophil stimulated with C5a, f-met-leu-phe (FMLP), or phorbol myristate acetate (PMA). Neutrophil-specific antigen NB1 is located on a 58- to 64-Kd GP. To determine if the quinine-dependent antibody and anti-NB1 recognize the same GP, immunoprecipitation studies were performed with the quinine-dependent antibody using neutrophils with varying NB1 phenotypes. The 60-Kd GP was detected on NB1-positive neutrophils from 11 of 12 donors tested, but not on NB1- negative neutrophils from two donors tested. After solubilized 125I- labeled neutrophils were absorbed with anti-NB1, the quinine-dependent antibody immunoprecipitated the 85-Kd GP, but not the 60-Kd GP. These results indicate that anti-NB1 and the quinine-dependent antibody identified the same GP. The 85-Kd GP was detected on neutrophils from all 14 donors tested. The electrophoretic mobility of the 85-Kd GP was similar to the electrophoretic mobility of the major 125I-labeled neutrophil protein.

In this study, we investigated the psychosocial effects of unrelated marrow donation. Survey questionnaires were administered pre-donation, shortly post-donation, and 1 year post-donation to all donors through the National Marrow Donor Program over a 3-year period. Univariate, bivariate, and multivariate analyses were then performed. Donors were generally quite positive about the donation 1 year post-donation: 87% felt it was “very worthwhile” and 91% would be willing to donate again in the future. Marrow donors were more likely than kidney donors to feel better about themselves as a result of the donation (P < .001). Donors with longer collection times, in general, had less positive psychosocial outcomes from the donation. Donors who experienced lower back pain or difficulty walking as a result of the donation were more likely to experience the donation as more stressful and painful than expected, but no more likely to experience it as less positive emotionally than donors who did not experience these side effects.

A 23-year-old woman experienced six distinct episodes of severe combined neutropenia and thrombocytopenia. At least one of the episodes was accompanied by hemodialysis-requiring acute renal failure and fragmentation hemolysis (hemolytic uremic syndrome [HUS]). In retrospect, all episodes were probably associated with the ingestion of quinine. Quinine-dependent antibodies to platelets, neutrophils, T lymphocytes, and red blood cells (RBCs) were detected in the patient's serum. By a monoclonal antibody antigen capture assay, the patient's serum contained IgG antibodies, which in the presence, but not absence, of quinine reacted with platelet glycoprotein (GP) complexes Ib/IX and IIb/IIIa, but not Ia/IIa. By immunoprecipitation assay, the serum, after addition of quinine, reacted strongly with an 85-Kd neutrophil membrane protein and weakly with 130- and 60-Kd moieties. Serum adsorbed with RBCs in the presence of quinine continued to react with platelets and neutrophils, and serum that was absorbed with platelets continued to react with neutrophils and RBCs, indicating that the antigenic targets were different on platelets, neutrophils, and RBCs. Since platelets and endothelial cells share some antigens, we tested patient serum for antibodies to human umbilical vein endothelial cells (HUVEC); no quinine-dependent antibodies to HUVEC were detected. However, her quinine-dependent antibodies not only bound to platelets and neutrophils, but also activated neutrophils. Thus, the patient's serum with quinine aggregated neutrophils, but neither agent alone caused activation. Moreover, the patient's serum with quinine (but not without) augmented the adherence of neutrophils to HUVEC. Treatment of the patient's serum with staphylococcal protein A removed the quinine neutrophil aggregation cofactor, suggesting it was due to IgG. In both neutrophil aggregation and adherence assays, decomplementation of the patient's serum markedly blunted its effect. Furthermore, the patient's serum failed to aggregate formalin-inactivated neutrophils, suggesting neutrophil activation, probably by activated complement, was necessary for aggregation and adhesivity to endothelium. We conclude that our patient's neutropenia, thrombocytopenia, lymphopenia, and anemia were due to quinine-dependent antibodies, and that these antibodies recognized epitopes that were different in the three target cell populations. We further suggest that HUS was likely secondary to the activation and adhesion of neutrophils to endothelium.

Antibodies to the neutrophil-specific antigens NA1 and NA2 are associated with alloimmune neonatal neutropenia (ANN), autoimmune neutropenia of childhood, and acute pulmonary transfusion reactions. These antigens have been found to be located on the neutrophil Fc-gamma receptor III (FcRIII). The mother of a child with ANN was found to lack both NA antigens and to produce an antibody that reacted with all normal neutrophils tested. We used maternal antibody and a CD16 monoclonal antibody (MoAb) that has specificity for FcRIII to immunoblot and immunoprecipitate neutrophil membranes of various NA phenotypes. Both antibodies immunoblotted an approximately 40- to 70-Kd glycoprotein (GP) on NA1, NA2-positive membrane, an approximately 40- to 55-Kd GP on NA1-homozygous membranes, and an approximately 55- to 70- Kd GP on NA2-homozygous membranes. Both antibodies also immunoprecipitated a 50- to 80-Kd GP from NA1, NA2-positive cells, a 50- to 60-Kd GP from NA1-homozygous cells, and a 55- to 80-Kd GP from NA2- homozygous cells. To further examine the specificity of the maternal antibody, sequential immunoprecipitation studies were performed using maternal antisera and a CD16 MoAb. After extracts of 125I surface- labeled neutrophils were precleared with maternal serum, CD16 MoAbs no longer immunoprecipitated any GP. Neither the CD16 MoAb nor a rabbit polyclonal antibody specific for FcRIII detected any GP in maternal neutrophil membranes by immunoblotting. Neutrophil FcRIII is a glycosyl- phosphatidylinositol anchored membrane GP as is decay accelerating factor and both are absent from neutrophils of patients with paroxysmal nocturnal hemoglobinuria (PNH). Maternal neutrophil membranes were probed with antibody specific for DAF and an 80-Kd GP was detected. This woman also has had no clinical evidence of PNH. These studies provide further evidence that the NA1 and NA2 antigens are on FcRIII and identify a healthy person whose neutrophils lack not only the neutrophil specific antigens NA1 and NA2 but multiple other epitopes of FcRIII and, therefore, likely lack FcRIII entirely.

Neutrophil-specific alloantibodies and the antigens they recognize are important in clinical medicine, but little is known about the structure of these antigens. Alloimmunization to the antigen NB1 is a clinically important cause of neonatal neutropenia and febrile transfusion reactions. To study the immunochemistry of the NB1 antigen, we prepared neutrophil plasma membranes and granules by nitrogen cavitation and differential centrifugation and then analyzed them by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting with alloantibodies to several neutrophil-specific antigens. Two different antisera to the neutrophil-specific antigen NB1 identified an approximately 55-Kd protein by immunoblotting on neutrophil membranes from four NB1-positive donors but not on neutrophil membranes from five NB1-negative donors. Four anti-NB1 antisera immunoprecipitated a 58- to 64-Kd protein from extracts of NB1- positive neutrophils surface-labeled with 125I using lactoperoxidase, but not from similarly treated NB1-negative neutrophils. Normal human serum did not immunoprecipitate or immunoblot any proteins from these same neutrophil preparations. The NB1 antigen was detected by immunoblotting in secondary granules but was not found in primary granules. The electrophoretic mobility of the antigen was decreased slightly by reduction, suggesting that intrachain disulfide bonds were present. After reduction, the antigen could no longer be recognized by anti-NB1 antisera, but treatment of the antigen with periodate had no effect on the ability of anti-NB1 antisera to recognize the antigen, suggesting that it is not a carbohydrate. The data suggest that the neutrophil-specific antigen NB1 is present on a 58- to 64-Kd surface glycoprotein that is also present in secondary granules, and that the NB1 epitope is not a carbohydrate but probably resides in the tertiary structure of the protein backbone.

Introduction: Allogeneic hematopoietic stem cell transplantation (alloHSCT) can provide a curative therapy for hematological malignancies but may result in complications such as relapse, infection, and acute and chronic graft versus host disease (GVHD). Two divergent approaches to GVHD prophylaxis (post-HSCT depletion of donor lymphocytes vs. suppression of immune activation) in a reduced-intensity, matched-unrelated donor setting were compared in a randomized, open label, phase 2 prospective trial (NCT00520130), powered to assess the incidence of severe cGVHD using NIH criteria. Methods: Hematological malignancy patients received disease-specific induction chemotherapy DA-EPOCH-FR or FLAG (Salit et al, JCO 2012; 30:830) for disease control and host lymphodepletion to CD4+ cell target Results: 81 pts (NHL=25, HL=8, CLL=18, AML/MDS=10, CML=3, CTCL/PTCL=5, ALL=4, MM=2, other=6), median age 50 yrs (range, 21-71) were included in the study (AC=42, TMS=39). The two arms were similar in age, gender, disease, relapse risk (Kahl), HCT-comorbidity index, and donor HLA match (8/8 or 7/8). Median time to neutrophil engraftment was 9 vs. 11 days in AC vs. TMS, respectively (p=0.017). There were no differences in platelet recovery (p=0.96). One case of graft failure occurred in a myeloma patient on the AC arm. D100 mortality probabilities were 12% (95% CI, 5-25) and 10% (95% CI, 4-24) in AC and TMS, respectively (p=0.20). Median survival in AC was 18.8 mo and 41.7 mo in TMS, with a median follow-up of 53 mo in AC and 50.6 mo in TMS. 3yr OS was comparable: AC 42% (95% CI, 28-57) vs. TMS 58% (95% CI, 42-73) (p=0.20). The 3yr malignancy progression rate was higher in the AC arm (AC 51% (95% CI, 34-65) vs. TMS 21% (95% CI, 10-35), p= 0.0062). 3yr relapse related mortality rates were 29% (95% CI, 16-44) vs. 14% (95% CI, 5-29) (p=0.067) and non-relapse mortality 29% (95% CI, 16-43) vs. 28% (95% CI, 14-43) (p=0.75) in AC vs. TMS, respectively. The most common grade ≥3 adverse events (CTCAE 4.03) within 100 d post-transplant were infections (22%) with more viral infections in the AC arm (p=0.0007). Reactivation of CMV occurred earlier in the AC arm, incidence 58% (95% CI, 42-71) vs. 24% (95% CI, 12-38) by D100 (p=0.035). Rates of aGVHD were similar; Gr II-IV at 6 mo in AC 38% (95% CI, 23-53) vs. TMS 41% (95% CI, 26-57) (p=0.59); Gr III-IV at 6 mo AC 21% (95% CI, 11-35) vs. TMS 13% (95% CI, 5-26) (p=0.61). In contrast, significantly lower rates of any grade cGVHD occurred in the AC arm compared to TMS at 36 mo (27% (95% CI, 14-41) vs. 59% (95% CI, 40-74)) (p=0.0076). The incidence of severe cGVHD was strikingly different: AC 5% (95% CI, 1-15) vs. TMS 31% (95% CI, 16-47) (p=0.0007). In the Cox model, the only prognostic factor for severe or any cGVHD was the TMS treatment arm, HR 6.8 (95% CI, 1.5-30.3, p=0.012) and HR 2.3 (95% CI, 1.1-4.8, p=0.026), respectively. Lymphocyte recovery (ALC 500/µL) was markedly delayed in AC, median 76 vs. 16 d (p Conclusions: This prospective, randomized trial demonstrates that the use of AC when compared with TMS led to a significant reduction in incidence of severe and overall cGVHD. These two GVHD prophylaxis regimens had similar incidences of aGVHD but very different effects on post-alloHSCT immune reconstitution, infection and relapse. Future strategies for cGVHD prevention will need to further address these issues. Disclosures Off Label Use: There is currently no FDA approved product for GVHD prevention or therapy. The GVHD prophylaxis regimens described in this study (Alemtuzumab-Cyclosporine and Tacrolimus-Methotrexate-Sirolimus) are used off-label as GVHD prophylaxis regimens in reduced-intensity allogeneic HSCT.

Chronic Graft Versus Host Disease (CGVHD) remains the main source of non-relapse mortality and morbidity among recipients of allogeneic hematopoietic stem cell transplantation (allo-HSCT). Although our lab and others have identified infiltrates of Th1/Tc1 and Th17 effectors in skin and oral mucosa, CGVHD targets multiple organs and no common factor or pathway has been demonstrated to reflect the broad range of CGVHD inflammatory and fibrotic manifestations. To identify the systemic cytokine pathways supporting the development and persistence of CGVHD, we chose to profile gene expression in circulating monocytes; monocytes up-regulate distinct patterns of gene expression in response to different cytokines, acting as in situ reporters. The NIH Natural History Study of CGVHD ([NCT00092235][1]) has primarily enrolled moderate to severely affected patients. Microarray analysis was performed on sorted monocytes from 10 normal controls (ND) and 26 patients selected from this cohort based on severe cutaneous involvement. Two interrelated pathways, each containing multiple genes, were consistently up-regulated across a cutaneous CGHVD spectrum ranging from lichenoid infiltrates to extensive sclerosis: (1) Interferon (IFN)-inducible genes including those involved in signaling, lymphocyte homeostasis and trafficking, apoptosis and antigen uptake and presentation ( STAT1, CXCL10, TNFSF13B, TNFSF10, TAP1 ), and (2) innate immune receptors for pathogens and cellular damage that can trigger IFN production and inflammasome assembly ( TLR2, TLR4, TLR7, AIM2, DDX58, CLEC4E ). Using multiplex RNA gene expression assays (Nanostring) to verify these pathways, we found significant up-regulation of IFN-inducible and damage-response genes in 69 CGVHD patients with a broad range of organ involvement, as compared with 14 allo-HSCT patients never developing CGVHD, or with 19 normal controls (Figure 1A, B). These pathways were further substantiated in plasma ELISA assays showing elevated levels of IFN-induced chemokines (CXCL9, CXCL10) in both lichenoid and severe sclerotic patients. Immunohistochemistry substantiated expression of Type I IFN-induced factors (MxA) in inflammatory infiltrates in CGVHD-targeted organs: lichenoid and sclerotic skin, oral mucosa and salivary gland. Consistent with induction of Type I IFN by activation of TLR and RIG-I receptors, levels of expression of DDX58 and TLR7 correlated with up-regulation of Type I IFN inducible genes ( OAS1, IFIT1, XAF1 ). Finally, multiplex RNA assessments on monocytes collected from 18 patients over serial time courses following NCI allo-HSCT protocols ([NCT00520130][2] and [NCT00074490][3]) substantiated a pattern of parallel up-regulation of multiple IFN-inducible and damage responsive genes at CGVHD onset, and of decline upon therapy and resolution (Figure 1C). A key point is that comparable up-regulation of these pathways was found in patients with extensive lichenoid or sclerotic CGVHD, both in the established CGVHD patients in the initial microarray and in the serial time courses of CGVHD development. These results support a model that IFN and inflammasome activation induced by the innate immune systemOs response to damage initiates an inflammatory process in CGVHD; IFN then can induce damage receptors, chemokines, cytokines and enhanced antigen presentation that sustain CGVHD. These interlocking analyses of gene expression patterns, plasma analytes and tissue are the first to support a unifying hypothesis of induction of IFN by innate response to cellular damage as a mechanism for initiation and persistence of CGVHD. ![Figure 1.][4] Figure 1. Disclosures No relevant conflicts of interest to declare. [1]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT00092235&atom=%2Fbloodjournal%2F126%2F23%2F922.atom [2]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT00520130&atom=%2Fbloodjournal%2F126%2F23%2F922.atom [3]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT00074490&atom=%2Fbloodjournal%2F126%2F23%2F922.atom [4]: pending:yes

Background: Adoptive cellular therapy with genetically modified T-cells using viral-based vectors to express chimeric antigen receptors targeting the CD19 molecule have demonstrated dramatic clinical responses in patients ALL. However, not all patients respond and CD19-negative escape has been observed. Thus, additional targets are needed. CD22 is a B-lineage-restricted, transmembrane phosphoglycoprotein of the Ig superfamily that is widely expressed on B-precursor ALL. Therefore, CD22 represents a promising target. This phase I dose-escalation study represents the first in human testing of anti-CD22 CAR adoptive cell therapy. The primary objectives were to determine the feasibility of producing anti-CD22 CAR cells meeting established release criteria and to assess the safety of administering escalating doses of anti-CD22-CAR engineered T cells in children and young adults with relapsed or refractory CD22-expressing B cell malignancies following a cyclophosphamide/fludarabine preparative regimen. Secondary objectives include determination of anti-leukemia effects, measurement of persistence of adoptively transferred anti-CD22 CAR T cells, and evaluation of cytokine profiles. Design: Children and young adults with relapsed/refractory CD22-expressing hematologic malignancies were eligible. Study endpoints included toxicity, feasibility, and antigen-specific immune and clinical responses. All enrolled subjects underwent autologous leukopheresis for peripheral blood mononuclear cells. Cells were then CD3+ enriched and cultured in the presence of anti-CD3/-CD28 beads followed by lentiviral vector supernatant containing the anti-CD22 (M971BBz) CAR, with culture duration of 7-10 days. On Day-4 (cell infusion=Day 0), subjects began induction chemotherapy with fludarabine 25 mg/m2 on Days-4, -3 and -2 and cyclophosphamide 900 mg/m2 on day-2. The first dose level started at 3 x 105 transduced T-cells/recipient weight (kg). Results: 6 subjects, aged 7-22 years, with ALL have been treated to date. All enrolled subjects had previously undergone at least one prior allogeneic hematopoietic stem cell transplant and all had received treatment with CD19 directed CAR-T cell therapy. Five subjects had a CD19 negative antigen escape, and one subject was a non-responder to prior CD19 CAR therapy. All subjects had demonstration of CD22 expression on > 99% of their ALL, although the antigen binding capacity had variability from < 900 to > 13,000 sites/cells. All subjects underwent successful culture, expansion and infusion of anti-CD22 CAR T-cells at the first dose level. The second subject enrolled met criteria for dose-limiting toxicity by virtue of grade 3 diarrhea which led to dose expansion at the first dose-level to treat a total of 6 subjects. Two subjects had grade 1 cytokine release syndrome (CRS), one subject had grade 2 CRS and in two subjects CRS was not seen. Evidence for CAR-T cell expansion was seen in peripheral blood, bone marrow and cerebrospinal fluid (Table). Clinical responses were evaluated at day 28 (+/- 4 days) post-infusion and included two subjects who had disease progression, two with disease stabilization and one subject who attained a minimal residual disease (MRD) negative complete remission. Flow cytometric CAR persistence was detected out to 47 days post-infusion in the responding patient with remission maintained for 3 months post-infusion. One patient is actively undergoing treatment and is too early to evaluate. Conclusions: This first-in-human anti-CD22 CAR T-cell therapy is safe, feasible and clinically active in patients who have undergone previous CAR therapy. Understanding mechanisms which may determine clinical efficacy are being explored. Accrual to the next dose level at 1 x 106 transduced T cells/kg is planned. Table 1.#Age/SexPrior HCTPrior anti-CD19 CARCD19 neg relapse?CD22 site densityPre-HCT disease burden (% leukemia in aspirate)Maximum CD22 CAR expansion (flow)CRSBest ResponsePBMarrowCSF122/MYYY208495-100%00n/aNonePD220/FY (2)YY134525%52.3%19.5%0%1MRD neg CR322/MYYY846>90%73%36%32%1SD422/MYYN258995%6%1%0%2SD57/FYYY283932%0%1.3%0%NonePD617/FYYY21851%n/an/an/an/an/aSD: stable disease; PD: progressive disease Disclosures Mackall: Juno: Patents & Royalties: CD22-CAR.

B-cell maturation antigen (BCMA) is a protein expressed by normal and malignant plasma cells. We are conducting a phase I clinical trial of an anti-BCMA chimeric antigen receptor (CAR-BCMA) that incorporates an anti-BCMA single-chain variable fragment, a CD28 domain, and a CD3-zeta T-cell activation domain (Carpenter et al. Clinical Cancer Research 2013). Autologous T cells are genetically modified to express the CAR with a gamma-retroviral vector. Patients receive a single infusion of CAR-BCMA T cells. Before the CAR T-cell infusions, patients receive a chemotherapy regimen of 300 mg/m2 of cyclophosphamide and 30 mg/m2 of fludarabine with each chemotherapy agent given daily for 3 days. The purpose of the chemotherapy is to enhance activity of the CAR T cells by depleting endogenous leukocytes. Twelve patients have been enrolled, and 11 patients have been treated on one of 4 dose levels, 0.3x106, 1x106, 3x106, and 9x106CAR+ T cells/kg of bodyweight. Patients had advanced multiple myeloma (MM) with a median of 7 prior lines of therapy. Of the 6 patients treated on the lowest 2 dose levels, one patient had a transient partial remission (PR) of 2 weeks duration; the other 5 patients had responses of stable disease (SD). On the 3rddose level, 2 patients obtained responses of stable disease, and one patient obtained a response of very good PR (VGPR) with complete elimination of MM bone disease on positron emission tomography (PET) scan, normalization of serum free light chains, and clearance of bone marrow plasma cells. Toxicity among patients on the first 3 dose levels was mild and included cytopenias attributable to chemotherapy, fever in 3 patients, and signs of cytokine release syndrome including tachycardia and hypotension in Patient 8 who had a VGPR. Two patients have been treated on the highest dose level of 9x106CAR+ T cells/kg. The first patient on this dose level, Patient 10, had MM making up 90% of total bone marrow cells before treatment. Starting 4 hours after infusion of CAR T cells, Patient 10 exhibited signs of cytokine release syndrome including fever, tachycardia, dyspnea, acute kidney injury, coagulopathy, hypotension requiring vasopressor support, and muscle damage manifesting as an elevated creatine kinase level and weakness. His neutrophil count was less than 500/µL before the CAR-BCMA T-cell infusion and remained below 500/µL for 40 days after the CAR T-cell infusion before recovering. He also experienced prolonged thrombocytopenia. Patient 10’s myeloma was rapidly eliminated after CAR-BCMA T-cell infusion. By immunohistochemistry staining for CD138, bone marrow plasma cells decreased from 90% before treatment to 0% one month after the CAR T-cell infusion. The serum M-protein decreased from 1.6 g/dL before treatment to undetectable 2 months after treatment. The serum and urine immunofixation electrophoresis tests were negative 2 months after the CAR T-cell infusion. Patient 10’s current myeloma response is stringent complete remission. The second patient treated on the 9x106CAR+ T cells/kg dose level, Patient 11, had IgG lambda MM with 80% bone marrow plasma cells before treatment. Patient 11 experienced signs of cytokine release syndrome with toxicities including fever, tachycardia, hypotension, delirium, hypoxia, and coagulopathy. Patient 11’s M-protein decreased from 3.6 g/dL before treatment to 0.8 g/dL 4 weeks after treatment. His serum lambda free light chain decreased from 95.9 mg/dL before treatment to 0.15 mg/dL 4 weeks after treatment. Four weeks after CAR T-cell infusion, bone marrow plasma cells were undetectable. T cells containing the CAR-BCMA gene were detected in the blood of all 10 patients evaluated with peak levels of 0.04 to 18.2% of blood mononuclear cells. Patient 10 had the highest peak absolute number of blood CAR T cells with 51 CAR+ T cells/µL. Blood levels of IL-6 and other inflammatory cytokines were highest in patients with clinical signs of cytokine release syndrome, and the 3 patients with the highest serum IL-6 levels also had the most impressive anti-myeloma responses. Before treatment, the mean serum BCMA level of treated patients was 243 ng/mL. In responding patients, serum BCMA levels decreased after treatment. Toxicities in patients receiving CAR-BCMA T cells were similar to toxicities in leukemia patients treated with anti-CD19 CAR T cells. Our findings demonstrate strong anti-myeloma activity in the first clinical trial of a CAR targeting BCMA. Disclosures: Wang: Celgene: Research Funding. Kochenderfer:bluebird bio Inc.: Research Funding. Off Label Use: Use of cyclophosphamide and fludarabine as a conditioning regimen for adoptively-transferred T cells will be part of the presentation.