ALBERT

Description: Background: In allogeneic BMT patients, the presence of allo-reactive donor CD4+ T cells in the graft were reported to be the primary cause of GvHD. Moreover, donor T-cells are required to promote the stem cell engraftment and to decrease the disease relapse. A number of studies also reported that a subset of CD4+CD25+ T cells usually generated de novo from the thymus that expressed FoxP3 regulate the T cells allo-reactivity in vivo. Thus, to establish a therapeutically useful adoptive T-cells immunotherapy, we depleted the CD4+ T cells from the graft and transplanted along with T cell depleted (TCD) BM cells in clinically relevant parent to F1 experimental allogeneic BMT model. Our hypothesis is that CD4-depleted graft will not cause GvHD, preserve the thymic function, homeostatically produce donor BM-derived CD4+ T cells along with FoxP3+CD4+CD25+ regulatory T cells with beneficial anti-opportunistic infection and anti-tumor effects. Methods: We used a parent (C57BL/6) to (C57BL/6 X BALB/c)CB6F1 allogeneic BMT model with a combination of TCD BM and splenocytes as the hematopoietic graft. CD4+ or CD8+ cells were selectively depleted from the splenocytes of C57BL/6 donor mice using MACS column. 1×106 CD4-depleted splenocytes or a mixture of 2×106 CD8-depleted and 1×106 CD4-depleted splenocytes and/or grafts containing 10×106 unfractionated splenocytes along with 5×106 TCD BM cells harvested from the congeneic C57BL/6 donor mice, were adoptively transferred to lethally irradiated (11Gy) CB6F1 mice. GvHD was monitored twice weekly by weight loss and other clinical signs. After 50 days post transplant recipients mice were bled or sacrificed and lymphocytes isolated from blood and different organs were analyzed by multicolor FACS. Results: Within 50 days of transplant the recipients of CD4-depleted splenocytes had 100% survival without GvHD whereas recipients of mixture of CD4- and CD8-depleted splenocytes or unfractionated splenocytes suffered from severe GvHD (%weight loss below 20%) with 50% survival. Surprisingly, very significantly expansion of total CD4+ T cells (37% ± 7% of lymphocytes, CD4:CD8 ratio 6:1) occurred in the blood of recipients of CD4-depleted splenocytes. In contrast the recipients of mixture of CD4- and CD8-depleted splenocytes DLI or whole splenocytes had only few CD4+ T cells (~2% ± 2% of lymphocytes, CD4:CD8 ratio 1:2). Over 90% of the CD4+ T cells in the blood of recipients of CD4-depleted splenocytes were from the donor BM and included significantly higher number of CD25+CD4+ T cells compared with the recipients of mixture of CD4- and CD8-depleted splenocytes or unfractionated splenocytes. Similarly, significantly increased numbers of FoxP3+CD25+CD4+ regularity T cells were also found in the spleen and thymus of recipients of CD4-depleted splenocytes compared with the recipients of mixture of CD4- and CD8-depleted splenocytes or unfractionated splenocytes (p

Description: Host antigen-presenting cells (APC) persist after high-dose chemotherapy and hematopoietic stem cell transplantation (HSCT) and initiate graft-versus-host disease (GvHD) in mouse models of HSCT. The role for donor APC on transplant outcomes is less clear. In clinical allogeneic HSCT from HLA-matched siblings, larger numbers of donor plasmacytoid dendritic cell (DC) precursors were associated with more relapse, and worse survival. Depletion of CD11b+ cells from bone marrow (containing CD11b+ DC) modestly augmented graft-versus-leukemia (GvL) activity in murine allogeneic HSCT. In this study, using allogeneic MHC mis-matched HSCT (C57BL/6→B10.BR) of mice bearing a lymphoblastic leukemia (LBRM), recipients of FACS–purified CD11b− donor DC plus FACS–purified HSC and T-cells had dramatically improved long-term survival (45% alive at 〉100 days) compared to d 5% survival among recipients of HSC and T-cells, or HSC, T-cells and CD11b+ DC (p

Description: Background: Patulin (4-Hydroxy-4H-furo 3,2-C-pyran-2(6H)-one) is a mycotoxin produced by various species of fungi. Previous literature reports have suggested that Patulin is toxic to the immune system and is believed to kill T cells by covalently bonding to intermediates in the p38 map kinase pathway. We hypothesized that Patulin may have selective cytotoxicity against T-cell and B-cell malignancies. Method: The cytotoxicity of Patulin was tested against normal peripheral blood mononuclear cells, malignant hematopoietic cell lines and primary leukemia cells. Malignant and normal mononuclear cells were incubated with serial dilutions of Patulin for 24 hours. Growth inhibition was assessed using an MTT assay and cytotoxicity and apoptosis was measured using flow cytometry following staining with 7AAD and annexin V. The concentration producing half maximal cytotoxicity (IC50) was determined for each target cell population. To further examine on the mechanisms responsible for the cytotoxic effect of Patulin, caspase pathway activation was examined by Western blot analysis. Results: Twelve leukemia/lymphoma cell lines, 8 samples of blood containing primary leukemia and lymphoma, and mononuclear cells from 9 normal donors were tested. Patulin induced cell death of normal and malignant hematopoietic cells in a dose-dependent manner at micromolar concerntrations. The T-cell H9 cell line is 〉 100 times more sensitive to Patulin than normal T cells (IC50, 0.5μM vs. 〉50μM). The IC50 of other cell lines can be found in the Table. Annexin-V positivity, indicating apoptosis, was seen in 15%, 70% and 90% of H9 cells exposed to Patulin 0.3, 1.6 and 32μM 24hours, respectively. The fractions of viable primary tumor cell were also markedly reduced after treatment, with IC50 between 0.3 to 2.3 for acute lymphoblastic/myeloid leukemia. As shown in the Fig, western blotting of cell extracts from patulin-treated H9 cells revealed that treatment with Patulin for 24 hours results in activation of caspase-8, -9 and -3 into cleaved forms. PARP, the substrate of caspase-3, was significantly degradated after Patuin treatment. Conclusion: Our findings suggest that Patulin can kill primary leukemia and leukemia cell lines by triggering apoptosis and by activating both intrinsic and extrinsic caspase pathways. The selective cytotoxicity of Patulin on T/B-cell malignant cell lines and primary leukemia cells suggests potential therapeutic activity in acute lymphoblastic leukemia/lymphoma. Figure Figure Patulin IC50 of leukemia/lymphoma cell lines and primary tumor cells IC50 (μM) for mallignant cell lines T-cell Jurkat MJ G11 H9 HH HUT78 Average 2.2 7.6 0.5 0.9 1.2 2.5 B-cell Daudi Ramos JVM2 GRANTA519 1.5 1.8 1.5 7.0 3.0 Myeloid HL60 KG1 K562 1.8 2.4 13 5.7 IC50 (μM) for primary leukemia/lymphoma cells T-ALL B-ALL Mantle cell lymphoma AML Pt 001 Pt006 Pt 005 Pt014 Pt002 Pt004 Pt007 Pt009 2.3 0.5 0.3 0.4 6.0 1.0 0.4 0.3 IC50 (μM) for T-cells from normal donors CD3+ (n=9) 〉50

Description: Background: The use of UCB transplants is limited by the difficulty of finding units with sufficient cells to provide reliable engraftment, and the slower kinetics of hematopoietic engraftment compared to other hematopoietic progenitor cell grafts. Grafts consisting of two unrelated UCB result in accelerated engraftment compared to single unit UCB transplants, but the mechanism for enhanced engraftment is unknown. Methods: We extracted data from published studies on myeloid engraftment following transplantation of single (9 studies, 1542 engrafted patients) or double (4 studies, 94 engrafted patients) UCB units (Table). A weighted average of median days (and estimated SD) to engraft for single and double UCB transplants was calculated. The cumulative incidence of myeloid engraftment for single or double UCB transplants was modeled based the assumption that the time to engraft was described by a normal distribution. For double and triple UCB transplants, the probability of myeloid engraftment at any time post-transplant was calculated as 1 minus the probability of non-engraftment for a single UCB unit squared or cubed, respectively. Results: The weighted average percentage for myeloid engraftment following single UCB transplants was 89%, with a median day of engraftment of 26.4 days. The weighted average for the rate of myeloid engraftment following double UCB transplants was 96%, with a median day of engraftment of 22 days. The calculated median days to engraft from single or double UCB transplants, using the assumption of normal distributions, were 22 and 26 days, respectively. Applying the mathematical model, the predicted cumulative engraftment for a double UCB graft is 99%, with a the curve for cumulative engraftment kinetics quite similar to the observed engraftment in Barker 2005 and the calculated curve for double UCB transplants based upon analysis of extracted data (Figure). Conclusion: UCB units do not appear to facilitate engraftment in grafts containing multiple units. A stochastic statistical model, in which different UCB grafts engraft independently, appears to account for the observed 4-day decrease in the median time to achieve myeloid engraftment following a double versus single UCB transplant. The use of a triple UCB graft is predicted to shorten the median time to achieve neutrophil engraftment to a median of 20 days. Published experiences describing myeloid engraftment following single and double UCB transplants Single UCB Transplant N, % engrafted Median day to engraft Double UCB Transplants N, % engrafted Median day to engraft Laughlin NEJM 2001 N=61, 90% 27 Barker Blood 2005 N=50, 100% 23 Hamza BJH 2004 N=28, 86% 29 Yoo ASH 2005 N=12, 100% 23 Barker Blood 2001 N=31, 88% 26 Kai ASH 2004 N=11, 82% 21 Thomson Blood 2000 N=30, 81% 25 Ballen ASH 2005 N=21, 90% 20 Sanz Blood 2001 N=22, 100% 22 Terakura BBMT 2007 N=148, 81% 22 Wagner Blood 2002 N=102, 86% 22 Migliaccio Blood 2000 N=200, 91% 28 Stevens Blood 2002 N=1111, 90% 27 Figure Figure

Description: It has been previously reported that increases in relapse and graft rejection can be associated with a low busulfan area under the curve (AUC) and hepatic toxicities are associated with a high AUC. This has lead to strategies to adjust busulfan dosing to achieve a target AUC. A retrospective analysis was performed in patients with severe obesity, defined as a Body Mass Index (BMI) greater than 40, to evaluate the reliability of IV busulfan dosed using the package insert dosing strategy to reach a targeted AUC. The study included 11 females and 8 males who received IV busulfan as part of their hematopoietic progenitor cell transplant preparative regimen. There were 11 autologous and 8 allogeneic transplants for lymphoma (10) and leukemia (9). The mean and median ages were 38 and 41 years (range 19–51). The mean and median BMIs for this population were 47.8 and 46.2 (range 40.4–62.7). The mean and median weights were 139.5 kg and 142.5 kg (range 107–192.3). TDM was performed with the first dose of busulfan along with subsequent dosage adjustments to achieve a target AUC of 1150–1350 micromol X min/L. Busulfex package insert dosing recommendations are for 16 doses of 0.8 mg/kg (12.8 mg/kg) using an adjusted ideal body weight (AIBW) for obese patients. AIBW = 0.25 × (actual weight − ideal body weight) + ideal body weight. Using the AIBW dosing, the mean starting dose to actual body weight ratio was 0.48 mg/kg (range 0.40–0.55 mg/kg). Based on the AUC analysis, the predicted mean AUC using the AIBW dosing was 972 (range 702–1356). Six AUCs were below 900, 10 AUCs were above 900 but below 1150, 2 AUCs were within the target range and 1 AUC was greater than 1350. Thirteen of the patients had repeat AUC analyses following a subsequent dose to confirm that the dose adjustments achieved the targeted AUCs. Seven of these patients had further dose adjustments following the repeat AUC analysis. Five patients with a busulfan mean half-life of 235 minutes (range 218–247 minutes) had higher AUCs than the rest of the group. The mean and median AUCs for these 5 patients were 1177 and 1223 (range 997–1357). The remaining 14 patients with a mean busulfan half-life of 177 minutes (range 151–201) had mean and median AUCs of 899 and 903 (range 702–1105). The total therapy delivered ranged from 12.7 to 21.4 mg/kg based on the AIBW. When calculated on actual weight, the range was 6.9 to 11.24 mg/kg. Conclusion: The AIBW dosing strategy of IV busulfan is inadequate to achieve the targeted AUC of 1150–1350 in severely obese patients. Severely obese patients require dose increases in the majority of cases, while those with delayed clearance could be overdosed if higher doses are administered without TDM. Severely obese patients require TDM and the majority will require dosage adjustments to administer IV busulfan in the targeted therapeutic range.

Description: The optimal therapy for patients with chemotherapy sensitive relapsed or refractory lymphoma is high dose therapy followed by autologous hematopoietic stem cell rescue. Rituximab (R) has been added to salvage regimens to increase response rate, thereby making more patients eligible for high dose therapy. However, when R is used prior to the salvage regimen, it has been associated with a delay in platelet engraftment (Hoerr et al, J Clin Oncol. 2004 Nov 15;22:4561–6). We have previously noted in a retrospective review of 117 patients with lymphoma treated with high dose therapy and autologous HSC transplant that concurrent treatment with R did not impact stem cell collection or post transplant engraftment (Kaufman et al, BBMT, February 2005, Sup 1, [11.2] 6). AMD3100 (plerixafor) is a CXCR4 inhibitor that, when used with G-CSF, more effectively mobilizes stem cells than G-CSF alone. In order to test the hypothesis that R does not negatively impact stem cell collection or post-transplant engraftment when AMD3100 is used with G-CSF, we performed a prospective trial of the use of AMD3100, G-CSF and R for patients with CD20 (+) relapsed chemosensitive lymphoma versus the use of AMD3100 and G-CSF for patients with CD20 (−) relapsed chemosensitive lymphoma. Patients were treated with 2 cycles of ICE ± R depending on CD20 status of the malignant cell. Patients who had a response proceeded to mobilization with AMD3100 and G-CSF for the CD20 (−) group (Arm A) or AMD3100, G-CSF, and four weekly doses of R at 375 mg/m2 (two doses prior to G-CSF and AMD3100, and two doses after) for the CD20 (+) group (Arm B). After collection, patients were treated with high dose therapy with targeted intravenous busulfan, etoposide and cyclophosphamide followed by autologous HSC transplantation. Patient demographics, mobilization characteristics, graft yield, engraftment data, and toxicity were assessed. 21 patients have been accrued. 11 in Arm A (10 Hodgkin Lymphoma {HL} and 1 with Peripheral T Cell Lymphoma) and 10 in Arm B (2 HL and 7 NHL, and 1 with a composite HL/NHL). The median number of days of collection was 2 for each arm. The median CD34 (+) collected was 4.64 * 106 CD34+ cells/kg in Arm A compared to 5.25 * 106 CD34+ cells/kg in Arm B (p=0.6) The median number of CD34(+)/CD38(−) was similar for both arms. As expected from in vivo B-cell depletion, the percentage of CD19 (+) cells in the product was decreased in Arm B compared to Arm A (2.24% vs. 0.09%, p

Description: Background: Successful reconstitution of cellular immunity following allogeneic HPCT reduces the risk of relapse and confers protection against opportunistic infections. We performed an IRB-approved retrospective analysis of patients who underwent allogeneic HPCT and in whom the content of immune cells were measured in the graft and in post-transplant blood samples. Methods: The study population consisted of 122 patients with hematologic disorders (71 acute leukemia; 14 chronic leukemia; 18 lymphoma, 12 MDS; 3 aplastic anemia; and 4 other) who underwent HPCT with a non-T cell depleted graft from an HLA matched related (73) or unrelated (49) donor. 47 patients had low risk disease (AA, ALL CR1, AML CR1, CML CP1), while 75 had high risk disease (all others). The conditioning regimen was non-myeloablative in 38 (31%), included ATG in 18 (15%), and included TBI in 54 (44%). Peripheral blood was drawn at a median of 101 days post-HPCT and analyzed for T-cell subsets, B-cells, NK cells, and dendritic cells. Subjects were divided into three strata based upon the maximal value for the content of each cell subset in the blood. Univariate and multi-variable stepwise logistic regression analyses were performed to test the association of pre-transplant clinical factors, the cells in the graft, and the numbers of immune cells in the blood post-transplant with overall survival. Results: The estimated three-year survival for all subjects was 53%, with death in 49/122 patients (40%) due to progressive disease (37%), infection (29%), GVHD (20%), and other causes (14%). Univariate factors associated with death included high risk, age, the use of reduced intensity conditioning regimen, the use of TBI, the use of ATG during conditioning and the measurement of lower numbers of total T-cells, CD4+ T-cells, CD8+ T-cells, γδ T-cells, DC1 and DC2 in the peripheral blood during the first 200 days post-transplant. A multi-variable Cox model identified non-myeloablative conditioning (HR 2.2, 95% CI 1.2–3.9), TBI (HR 1.9, 95% CI 1.1–3.3), transplant risk strata (HR 1.9, 95% CI 1.0–3.6), and a blood CD3+ T-cell count of less than 600 cells/mcL (HR 1.8, 95% CI 1.2–2.5) as independent risk factors for post-transplant death. The presence of acute GVHD (all grades) or graft constituents was not significantly associated with survival. Limiting the study population to those subjects who survived at least 100 days showed that blood CD3+ T-cells, non-myeloablative conditioning, the use of TBI remained significantly associated with survival. Conclusions: Higher CD3+ counts in the early post-transplant period predict better survival. Patients who fail to achieve a blood CD3+ T-cell count of 〉600/mcL in the first 200 days post-transplant may be appropriate subjects for adoptive cellular immunotherapy. Low Risk Patients Low Risk Patients High Risk Patients High Risk Patients

Description: We are investigating methods to reduce the graft-versus-host disease (GVHD) potential of donor T-cells while retaining graft-versus-leukemia (GVL) activity in allogeneic HSCT. Previous investigations by our group and others in have shown that naive CD4 T-cells induce severe acute GVHD, while memory CD4 T-cells do not induce GVHD but retain GVL activity in murine transplant models. These findings have led to studies for the development of methods to increase the number of memory T-cells available for transplant. The calcium ionophore, ionomycin, is a T-cell activating agent and mitogen. By increasing intracellular Ca2+ levels, ionomycin is induces T-cell activation through signaling mechanisms including phospholipase C activation, hydrolysis of phosphoinositides, and activation of protein kinase C. Differences in memory and naive T-cell responses to ionomycin have been attributed to resistance of memory T-cells to increases in Ca2+. Memory T-cells lack intracellular Ca2+ stores, and are also resistant to influx of Ca2+. Brief low dose ionomycin exposure (20min, 2μM) of T-cells, leading to increased density of naive T-cells, has previously been exploited as a method for separating memory and naive T-cells by Percoll gradient separation. Since ionomycin exposure induces T-cell activation through native Ca2+ dependent signaling mechanisms, we hypothesized that ionomycin-treated T-cells would shift to an activated/memory T-cell phenotype. Murine splenic T-cells were treated with 1.3μM ionomycin for 4hr. Memory and naive T-cell subsets and activation markers were analyzed by flow cytometry. 75% and 85% of untreated CD4 and CD8 T-cells, respectively, had the CD62L+ naive phenotype. These numbers were dramatically reduced to 7% and 17% after ionomycin exposure, representing a shift to the memory T-cell phenotype. Viability of T-cells was not significantly affected. The majority of remaining CD62L+ naive T-cells expressed activation markers CD25 and CD69. The fraction of CD4+CD25+Foxp3+ regulatory T-cells was also determined by intracellular staining of the transcription factor and co-expression of surface markers. CD4+CD25+Foxp3+ regulatory T-cells represented 4% of untreated CD4 T-cells and 3% of ionomycin-treated CD4 T-cells. While ionomycin has been used for many years in studies of T-cell activation, to our knowledge this is the first demonstration of a rapidly-induced shift of naive T-cells to a memory phenotype. A pilot experiment was conducted testing the GVHD activity of ionomycin-treated splenocytes (SP) in B6→ (B6 × Balb/C)CB6F1 recipients. 5 × 106 T-cell depleted bone marrow cells (TCD-BM) were transplanted along with 10 × 106 treated or untreated SP. Mice that received untreated SP all died from acute GvHD by 34 days after transplant, while all recipients of ionomycin-treated SP survived until the experiement was terminated at day 49 (average weight loss was 25%, data not shown). Continuing experients will refine the dose to further reduce GVHD symptoms and also test GVL activity of the treated cells. Treatment of donor T-cells with ionomycin may represent a clinically applicaple method to engineer donor lymphocyte infusions that are safer for HSCT patients. Figure 1. Survival of CB6F1 recipients after transplant with 5 million B6 TCD-BM and 10 million B6 splenocytes that were either untreated or stimulated ex-vivo with a combination of PMA, ionomycin and brefeldin-A for 4 hours. 5 recipient animals per group. The experiment was terminated at day 49. Figure 1. Survival of CB6F1 recipients after transplant with 5 million B6 TCD-BM and 10 million B6 splenocytes that were either untreated or stimulated ex-vivo with a combination of PMA, ionomycin and brefeldin-A for 4 hours. 5 recipient animals per group. The experiment was terminated at day 49.

Description: Background: High dose therapy and autologous transplant (HDT) benefits many patients with myeloma, but the addition of other chemotherapeutics or TBI to high dose melphalan (HDM) does not improve outcomes. Bortezomib (B) is a proteasome inhibitor that synergizes with chemotherapy due to its effects on DNA repair enzymes. Recent data have shown that B up-regulates the anti-apoptotic protein MCL-1, which would suggest that the sequence of administration may be critical to the combination of B and HDM. We hypothesize that B followed by M will be inferior to M followed by B. To test this hypothesis, we designed a randomized phase I trial combining escalating doses of B and Melphalan 200 mg/m2 (Mel200) in order to determine the toxicity, optimal dose and optimal sequence of administration. Methods: Patients were randomized to receive either B 24 hours before Mel 200 (ARM A) or B 24 hours after Mel 200 (ARM B). Doses of B escalated from 1.0mg/m2 up to 1.6mg/m2 as defined using the Escalation with Overdose Control (EWOC) method, a Bayesian phase I design. Standard transplant criteria were used for enrolling myeloma transplant patients with the additional criterion of measurable disease at the time of transplant (〉5% plasma cells by biopsy or M-protein 〉1.0). Enrolled patients underwent BM aspirate on day -4 (before B) and day 0 (before PBSC infusion). Bone marrows were tested for annexin V staining on the plasma cell population, and myeloma cells were sorted for subsequent protein analysis. Routine demographics, toxicity, and engraftment data were also collected. Results: Twenty one patients have been enrolled to date, with 15 evaluable for response assessment. B doses range from 1.0–1.6mg/m2 with the current dose at 1.6mg/m2. Age range was 48–74 years. One patient had resistant disease at the time of transplant. Median time to WBC and Plt engraftment were 13 days and 15 days, and not different based upon B dose, sequence of administration, or historical patients treated with M alone. 10 patients have been randomized to the B.→ M arm, and 11 to the M.→B arm. 11 of 15 (73%) evaluable patients achieved a VGPR or better at 6 months post transplant, with an overall response rate of 14/15 achieving PR or better (93%). There were no significant differences in myeloma cell annexin V staining on day -4 between the 2 groups, however on the day 0 bone marrow, there was an increase in the percent of Annexin V (+) MM cells for the group randomized to the M.→B arm compared to the B.→M arm (45.3% vs 30%, respectively). To date there is no difference in bone marrow IL6 or VEGF levels between the 2 randomized groups though only 10 patients have been enrolled at the highest dose level. Conclusion: The combination of B and MEL 200 is safe with a toxicity profile and engraftment kinetics similar to that seen in a historical cohort receiving MEL 200 alone. Toxicity has been no different from our previous experience with MEL 200 alone. Preliminary lab data suggests that the M.→B arm may be superior to the B.→M arm. Data on DNA repair and additional cytokines will be presented.

Description: Advanced myelodysplastic syndrome and secondary or relapsed AML are difficult to treat, with CR rates following standard induction chemotherapy below those for de novo AML. We are examining a novel induction approach using the anti-CD33-calicheamicin conjugate gemtuzumab ozogamicin (GO) in combination with conventional chemotherapy. Methods and Patient Population: We present data from an ongoing phase I/II study, demonstrating the activity and tolerability of the combination in patients with MDS and AML. 32 pts (19 pts from phase I, 13 from phase II) were assigned to receive cytarabine 2 g/m2 over 4 hours once daily, and Topotecan 1.5 mg/m2 by continuous infusion daily on days 1–5. GO was then given on day 6 at a dose of 3 mg/m2 (7 pts), 4.5 mg/m2 (16 pts), or 6 mg/m2 (7 pts); 2 additional pts did not receive the GO because of intercurrent events. The MTD of GO was established at 4.5 mg/m2, which was used in phase II. The study population consisted of 13 pts with advanced MDS (IPSS 2–3), 13 pts with secondary AML, and 6 pts with relapsed AML. Median age was 64 (34–77). Results: ATGO was generally well tolerated. Bacterial infections, including 9 bacteremic episodes, were the most common non-hematologic complications. Two cases of reversible but unexplained grade 3 delirium were seen at 6 mg/m2 GO, meeting study criteria for dose-limiting toxicity. Only one case (3%) of clinical venoocclusive disease (VOD/SOS) occurred, in the single subject who received a second (half) dose of GO on day 16. Asymptomatic grade 3 hepatic transaminase elevations occurred in 8 pts (25%). There were 4 deaths (12.5%) prior to day 40: 2 infectious, 1 CNS bleed, and 1 sudden death. Based on intent to treat, 17 pts (53%) achieved complete remission, with or without full platelet recovery (15 CR, 2 CRp) after one cycle of induction. Following ATGO induction, 12 of the 17 responders went on to receive post-remission therapy consisting of allogeneic hematopoietic cell transplantation (HCT) for 8 pts (7 using reduced intensity conditioning), or standard consolidation alone (3 pts). Post-remission therapies, including HCT, were not associated with VOD/SOS or peculiar toxicities. As expected, achievement of CR after ATGO was associated with improved overall survival (p=0.043). Estimated 2 year OS for all pts is 34% (49% for responders vs. 14% for non responders). There was a trend toward improved survival in younger pts. Cytogenetic class, type of disease (MDS v. AML), and administration of prior chemotherapy were not significant predictors of either CR or OS. 15 pts (47%) are alive, with a median follow-up of 242 days (37–1182); 7 are currently in CR, including 5 of the 8 pts who received allogeneic HCT. Conclusions: Our data show that ATGO is well tolerated and without associated VOD after a single delivered dose of GO. The pathophysiology of the reversible delirium in the highest GO dose cohort remains unclear, but no neurologic effects were observed at the lower dose levels. The CR rate is promising, as is the ability to deliver post-remission therapy including HCT, in this challenging patient population.