Search

Background Carfilzomib (K), a second generation proteasome inhibitor that is approved for patients (pts) with relapsed refractory multiple myeloma (RRMM) is associated with increased cardiovascular (CV) adverse effects (AE), in particular hypertension (HTN), dyspnea and cardiac failure (CCF), based on the ENDEAVOR (Dimopoulos, MA. et al. Lancet 2016) and ASPIRE studies (Stewart, K.et al. NEJM 2015). The clinical characteristic and underlying mechanism of K induced CVAE have been poorly elucidated, and the limited published data on the utility of cardiac enzymes as biomarkers have not been revealing. We conducted a retrospective single centre review of K-treated pts who have undergone systemic serial cardiovascular and cardiac enzyme assessment, to profile the nature of CVAE and patterns in cardiac enzymes that might have predictive utility in K-induced CVAE. Method: Between January 2016 to June 2018, all pts who were treated with K (D1,2,8,9,15,16 in a 28 day cycle)-based regimen for RRMM at St.Vincent's Hospital Melbourne underwent systematic cardiovascular assessment including documentation of baseline cardiac risk factors (RF) and transthoracic echo (TTE). Serial troponin, creatinine kinase (CK) and B-type Natriuretic Peptide (BNP) were performed at baseline (C1D1 or 2), mid cycle (D8 or 9) and end of treatment (D16) of every cycle. Repeat TTE was done at physicians' discretion. Systematic documentations of CVAEs (graded according to CTCAEv5) were retrieved from medical records. The incidence and severity of CVAE was collated and correlated with serial cardiac enzyme levels, the association of which was tested using Student's t-test. Results 76 pts (67% male; median age 66(46-86) years) with RRMM who received a median of 6 cycles (1-26) of K-based treatment were included in this analysis. At baseline, 16 pts (21%) had at least 1 CVRF including HTN, history of ischaemic heart disease, hyperlipidaemia, diabetes or smoking. The incidence of HTN prior to treatment was 39%. On treatment, the incidence of HTN was 85% (28% grade≥3). 17% of pts required the addition of antihypertensive medications and 11% had K dose reduction due to HTN. K-related dyspnea occurred in 53% of pts (28% grade≥3), 52% of who required K dose reduction/interruption. CCF occurred in 14.5% (grade≥3, 13%) that all necessitated K interruption/dose reduction. 69 pts had baseline TTE that showed a median RVSP (right ventricular systolic pressure) of 32mmHg (range 20-48). On treatment, 38 pts had repeat TTE, mainly driven by dyspnea with a median RVSP of 41(20-93mmHg). Of the 18 pts with grade≥3 dyspnea who had repeat TTE, 83% had RSVP ≥39mmHg and 33% had RVSP≥50mmHg, compared to 70% with RSVP ≥39mmHg and 10% RVSP≥50mmHg in the group with grade 1-2 dyspnea. Troponin_I and CK level did not change significantly throughout K-treatment. However BNP level rose and fell within each cycle, typically with peaks at mid cycle (D9 and D16) and troughs at the beginning of a new cycle (D1 or 2). 87.1% of pts had at least one elevated BNP (>100ng/L) with a median maximal level of 300ng/L (range 114-2320ng/L). There was a significant difference in the incidence of grade ≥3 CVAEs (mainly HTN) in pts whose mid cycle BNPs (D9 or 16) were elevated in ≥50% of measurements during the first 4 cycles of K compared to the rest (65.5% vs. 25.0%, p=0.0084). In pts with dyspnea, persistent rise in mid cycle BNPs that does not normalise prior to next treatment cycle trended towards an increased incidence of raised RVSP of ≥39mmHg (21.7% vs. 7.9%, p=0.0543). Conclusion. The incidence of CVAEs in pts treated with K is higher in the real-world setting compared to that reported in the Endeavor and ASPIRE studies. Dyspnea is frequent, the degree of which correlates with the degree of raised RVSP, thus indicating pulmonary hypertension as a contributor to dyspnea in the setting of K treatment. BNP level typically peaks during mid cycle and troughs at the beginning of next cycle, indicating a temporal relationship to K infusions. In pts with dyspnea, the association of persistent rise in mid-cycle BNP (rather than troponin or CK) with raised RVSP and HTN perhaps indicate transient endothelial dysfunction as the mechanism for K-induced CVAE rather than myocyte injury. Disclosures Quach: Amgen: Consultancy, Research Funding; Sanofi Genzyme: Research Funding; Janssen Cilag: Consultancy; Celgene: Consultancy, Research Funding. Tam:Roche: Honoraria; AbbVie: Honoraria, Research Funding; Pharmacyclics: Honoraria, Travel funding; Beigene: Honoraria, Other: Travel funding; Pharmacyclics: Honoraria; Gilead: Honoraria; Janssen: Honoraria, Research Funding; AbbVie: Honoraria, Research Funding; Beigene: Honoraria, Other: Travel funding; Gilead: Honoraria; Roche: Honoraria.

T cell malignancies represent a class of devastating hematologic cancers with high rates of relapse and mortality in both children and adults for which there are currently no effective or targeted therapies. Despite intensive multi-agent chemotherapy regimens, fewer than 50% of adults and 75% of children with T-ALL survive beyond five years. For those who relapse after initial therapy, salvage chemotherapy regimens induce remissions in 20-40% of cases. Allogeneic stem cell transplant, with its associated risks and toxicities, is the only curative therapy. T cells engineered to express a chimeric antigen receptor (CAR) are a promising cancer immunotherapy. Such targeted therapies have shown great potential for inducing both remissions and even long-term relapse-free survival in patients with B cell leukemia and lymphoma7-9. Thus, a targeted therapy against T cell malignancies represents a significant unmet medical need. However, several challenges have limited the clinical development of CAR-T cells against T cell malignancies. First, the shared expression of target antigens between T effector cells and T cell malignancies results in fratricide, or self-killing, of CAR-T cells. Second, harvesting adequate numbers of autologous T cells, without contamination by malignant cells is, at best, technically challenging and prohibitively expensive. Third, the use of genetically modified CAR-T cells from allogeneic donors may result in life-threatening graft-vs.-host disease (GvHD) when infused into immune-compromised HLA-matched or mismatched recipients. We hypothesized that deletion of CD7 and the T cell receptor alpha chain (TRAC) using CRISPR/Cas9 in CAR-T targeting CD7 (UCART7) would result in the efficient targeting and killing of malignant T cells without significant effector T cell fratricide or induction of GvHD. To generate the CD7 CAR, the anti-CD7 single chain variable fragment (scFv) was created using commercial gene synthesis and cloned into the backbone of a 3rd generation CAR with CD28 and 4-1BB internal signaling domains. The construct was modified to express CD34 via a P2A peptide to enable detection of CAR following viral transduction. Human primary T cells were activated using anti-CD3/CD28 beads for 48 hours prior to bead removal and electroporation with CD7 gRNA, TRAC gRNA, and Cas9 mRNA. On day three, T cells were transduced with lentivirus particles encoding either CD7 CAR or CAR CD19 control and allowed to expand for a further 6 days. Transduction efficiency and ablation of CD7 and TRAC were confirmed by flow cytometry. Multiplex CRISPR/Cas9 gene-editing resulted in the simultaneous bi-allelic deletion of both CD7 and TRAC in 72.8%±1.92 of cells, as determined by both non-homologous end joining (NHEJ) and FACS analyses. To prevent alloreactivity, CD3+ CAR-T were removed from the product by magnetic depletion. Of particular importance is that by using two distinct methods for assessing "off-target" nuclease activity across the entire human T cell genome (Guide-seq and probe capture), we could only detect one gene, an intronic modification of RMB33, that was inappropriately targeted using this approach. No obvious genomic rearrangements were detected by these analyses. UCART7 effectively expanded and killed T-ALL cell lines (CCRF-CEM, MOLT3, and HSB2) and human primary T-ALL blasts in vitro. Next, we tested the capacity of UCART7 to kill primary T-ALL in vivo without xenogeneic GvHD. Considerable expansion of alloreactive T cells, severe GvHD (mean clinical GvHD score = 5.66), and a robust graft vs. leukemia effect were observed in recipients of WT T cells. In contrast, GvHD was completely absent, T cells were undetectable, and considerable tumor burden was observed in mice receiving TRACΔ T cells. Mice receiving UCART7, however, had no GvHD and, unlike UCART19 controls, effectively cleared primary human T-ALL in NSG mice. Fratricide-resistant and allo-tolerant 'off-the-shelf' UCART7 signifies a novel strategy for treatment of relapsed and refractory T-ALL and non-Hodgkin's T cell lymphomas without a requirement for autologous T cells and represents the first clinically feasible adoptive T cell therapy for T cell malignancies. Disclosures No relevant conflicts of interest to declare.

Regulatory T cells (Tregs) suppress graft-versus-host disease (GVHD) while preserving a beneficial graft-versus-leukemia (GVL) effect. Thus, their use in allogeneic stem cell transplantation (SCT) provides a promising strategy to treat GVHD. However, 3 obstacles prevent their routine use in human clinical trials: (1) low circulating number of Tregs in peripheral blood, (2) loss of suppressor function after in vitro expansion, and (3) lack of Treg-specific surface markers necessary for efficient purification. FOXP3 is exclusively expressed in Tregs and forced expression in CD4+CD25− T cells can convert these non-Tregs into Tregs with functional suppressor function. Here, we show that the FDA-approved hypomethylating agents, decitabine (Dec) and azacitidine (AzaC), induce FOXP3 expression in CD4+CD25− T cells both in vitro and in vivo. Their suppressor function is dependent on direct contact, partially dependent on perforin 1 (Prf1), but independent of granzyme B (GzmB), and surprisingly, Foxp3. Independence of Foxp3 suggests that genes responsible for the suppressor function are also regulated by DNA methylation. We have identified 48 candidate genes for future studies. Finally, AzaC treatment of mice that received a transplant of major histocompatibility complex mismatched allogeneic bone marrow and T cells mitigates GVHD while preserving GVL by peripheral conversion of alloreactive effector T cells into FOXP3+ Tregs and epigenetic modulation of genes downstream of Foxp3 required for the suppressor function of Tregs.

The CXCR4–SDF-1 axis plays a central role in the trafficking and retention of normal and malignant stem cells in the bone marrow (BM) microenvironment. Here, we used a mouse model of acute promyelocytic leukemia (APL) and a small molecule competitive antagonist of CXCR4, AMD3100, to examine the interaction of mouse APL cells with the BM microenvironment. APL cells from a murine cathepsin G-PML-RARα knockin mouse were genetically modified with firefly luciferase (APLluc) to allow tracking by bioluminescence imaging. Coculture of APLluc cells with M2-10B4 stromal cells protected the leukemia cells from chemotherapy-induced apoptosis in vitro. Upon injection into syngeneic recipients, APLluc cells rapidly migrated to the BM followed by egress to the spleen then to the peripheral blood with death due to leukostasis by day 15. Administration of AMD3100 to leukemic mice induced a 1.6-fold increase in total leukocytes and a 9-fold increase of circulating APL blast counts, which peak at 3 hours and return to baseline by 12 hours. Treatment of leukemic mice with chemotherapy plus AMD3100 resulted in decreased tumor burden and improved overall survival compared with mice treated with chemotherapy alone. These studies provide a proof-of-principle for directing therapy to the critical tethers that promote AML-niche interactions.

Inhibition of osteoclast (OC) activity has been associated with decreased tumor growth in bone in animal models. Increased recognition of factors that promote osteoclastic bone resorption in cancer patients led us to investigate whether increased OC activation could enhance tumor growth in bone. Granulocyte colony-stimulating factor (G-CSF) is used to treat chemotherapy-induced neutropenia, but is also associated with increased markers of OC activity and decreased bone mineral density (BMD). We used G-CSF as a tool to investigate the impact of increased OC activity on tumor growth in 2 murine osteolytic tumor models. An 8-day course of G-CSF alone (without chemotherapy) significantly decreased BMD and increased OC perimeter along bone in mice. Mice administered G-CSF alone demonstrated significantly increased tumor growth in bone as quantitated by in vivo bioluminescence imaging and histologic bone marrow tumor analysis. Short-term administration of AMD3100, a CXCR4 inhibitor that mobilizes neutrophils with little effect on bone resorption, did not lead to increased tumor burden. However, OC-defective osteoprotegerin transgenic (OPGTg) mice and bisphosphonate-treated mice were resistant to the effects of G-CSF administration upon bone tumor growth. These data demonstrate a G-CSF–induced stimulation of tumor growth in bone that is OC dependent.

Background: Relapse and drug resistance remains the major cause of therapy failure in Acute Myeloid Leukemia (AML). Inhibitors of apoptosis proteins (IAPs) are a family of proteins that suppress proapoptotic signaling and represent attractive targets to overcome chemotherapy resistance. Debio 1143 (a.k.a. AT-406) is an orally-active IAP antagonist able to promote apoptosis alone and/or in combination with chemotherapy. We recently reported the results of a Phase 1 study of Debio 1143 in combination with daunorubicin and cytarabine (araC) in poor-risk AML patients (DiPersio et al. J Clin Oncol 32:5s, 2014, a7029). Although Debio 1143 showed good tolerability and could be safely combined with chemotherapy, 62% (8/13) of patients with complete remissions relapsed within the study period. Here, we used a preclinical AML model to begin examining the effects of BM stroma on the efficacy of Debio 1143 combined with araC or doxorubicin (dox). Methods: We generated murine acute promyelocytic leukemia (APL) cell lines that are resistant to araC and dox both in vitro and in vivo by chronically exposing APL cells to increasing concentrations of drug in vitro. This APL model was previously generated by knocking in the human PML-RARa cDNA into the cathepsin G locus (Westervelt et al. Blood, 2003). We used WST-1 cell proliferation assays and bioluminescence imaging (BLI) to evaluate the efficacy of Debio 1143 to induce apoptosis alone and in combination with araC or dox in vitro in chemotherapy sensitive and resistant APL cells in the presence and absence of M2-10B4 stromal cells. Dox or araC was provided for (a) a 6h period initiated 24h before Debio 1143 (i.e., with an intervening 18h washout of the chemotherapy), (b) simultaneously with Debio 1143, or (c) 6h after Debio 1143. To assess Debio 1143 in vivo, we injected APL cells into syngeneic mice. In our first experiment, mice (n=5/cohort) were treated with PBS, araC (500 mg/kg, s.c.), dox (50 mg/kg, s.c.), Debio 1143 (100 mg/kg, p.o.) or their combination on days 5, 6, 12, and 13 after APL injection. In our second experiment, mice (n=8/cohort) were treated with PBS, araC (200 mg/kg, s.c.), dox (10 mg/kg, s.c.), Debio 1143 (100 mg/kg, p.o.) or their combination on days 3 to 7 and 10 to 16 after APL injection. In both experiments, Debio 1143 was given 2 hours prior to chemotherapy. Results: Figure 1shows that coculture of APL cells with stromal cells protects them from araC-induced killing (Figure 1A) and that our araC-resistant cells are indeed resistant to araC (Figure 1B). Treatment with Debio 1143 alone has no effect on parental or araC-resistant APL survival in the presence or absence of stroma cells. However, Debio 1143 in combination with araC killed the parental APL cells and overcame the stroma cell-mediated chemotherapy protection (Figure 1A). Although combination therapy partially inhibited araC-resistant APL cell proliferation in the absence of stroma, it had no effect when the cells were cultured in the presence of stroma (Figure 1B). We obtained similar results with dox and dox-resistant cells (data not shown). This in vitro data with the parental APL cell line was confirmed in vivo. Combination therapy with chemotherapy (araC or dox) and Debio 1143 for 4 days (EXP#1) or 10 days (EXP#2) significantly prolonged the overall survival of mice by 4 to 7 days compared to single agent therapy (Debio 1143, araC, or dox) in this aggressive tumor model (Table 1). Preliminary in vitro data suggests that pretreatment of parental or araC-resistant APL cells with chemotherapy (araC or dox) enhances their killing by Debio 1143 compared to simultaneous treatment. Conclusions: Debio 1143 combined with chemotherapy inhibits parental and chemotherapy-resistant APL growth in vitro and in vivo. Future mechanistic studies are required to determine how Debio 1143 can overcome stroma-mediated resistance. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures Rettig: Debiopharm International SA: Research Funding. Vuagniaux:Debiopharm International SA: Employment. Zanna:Debiopharm International SA: Employment. DiPersio:Debiopharm International SA: Research Funding.

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is the only curative treatment for patients with relapsed/refractory leukemia, and marrow failure states such as myelodysplasia and aplastic anemia. However, allo-HSCT is complicated by allogeneic donor T cell-mediated graft-versus-host disease (GvHD) which can be life-threatening especially in recipients of unrelated or HLA-mismatched hematopoietic stem cell products. These same alloreactive donor T cells also mediate a beneficial graft-versus-leukemia (GvL) effect. We have recently reported that interferon gamma receptor deficient (IFNγR-/-) allogeneic donor T cells induce significantly less GvHD in both a MHC fully-mismatched (B6 (H-2b) → Balb/c (H-2d)) (75% vs 0% overall survival) and a minor-mismatched (B6 (H-2b) → B6x129 (H-2b)) allo-HSCT models (100% vs 10% overall survival) compared to WT T cells (Choi et al Blood 2012). In addition, IFNγR-/- donor T cells maintain a beneficial GvL effect, which has been examined in both systemic leukemia and solid tumor models using luciferase-expressing A20 cells derived from Balb/c. We also found that IFNγR-/- T cells migrate primarily to the spleen while WT T cells to GI tract and peripheral lymph nodes (LNs) using bioluminescence imaging (BLI), suggesting that altered T cell trafficking of IFNγR-/- T cells to GvHD target organs might be the major reason for the reduced GvHD. We further demonstrated that the IFNγR-mediated signaling (via JAK1/2 - STAT pathway) in alloreactive donor T cells is required for expression of CXCR3 which has been implicated in trafficking of T cells to areas of inflammation and target organs, commonly known to be the sites of GvHD. Here, we examine if inhibition of IFNγR signaling using a small molecule inhibitor can recapitulate the reduced GVHD with potent anti-leukemia effects similarly to that seen with IFNγR-/- T cells. We find that INCB018424, an inhibitor of JAK1/JAK2 which mediate IFNγR signaling, blocks CXCR3 expression in vitro. Most importantly, in vivo administration of INCB018424 (100 ug, s.c., twice a day, day 1-31) after allo-HSCT alters T cell trafficking and significantly reduces GvHD (70% vs. 0% overall survival, n=10/group, p=0.0012). We also find that INCB018424 preserves the beneficial GvL effect, which has been examined in both systemic leukemia and solid tumor models using luciferase-expressing A20 cells derived from Balb/c (B6 to Balb/c model) and APL cells from B6x129 (B6 to B6x129 model). Of note is that INCB018424, when given after transplant, had no significant effect on neutrophil or platelet recovery compared to animals receiving placebo. Thus, the IFNγR signaling pathway represents a promising therapeutic target for future efforts to mitigate GvHD while maintaining GvL after allo-HSCT. Moreover, this pathway could be targeted and exploited in other diseases besides GvHD such as those from organ transplantation, chronic inflammatory diseases and autoimmune diseases. Disclosures: No relevant conflicts of interest to declare.

Background Exportin 1 (XPO1) also called CRM1, is a widely expressed nuclear export protein, transporting a variety of molecules including tumor suppressor proteins and cell cycle regulators. Targeted inhibition of XPO1 is a new strategy to restore multiple cell death pathways in various malignant diseases. SINEs are novel, orally available, small molecule Selective Inhibitors of Nuclear Export (SINE) that specifically bind to XPO1 and inhibit its function. Methods We used WST-1 cell proliferation assays, flow cytometry, and bioluminescence imaging to evaluate the efficacy of multiple SINEs to induce apoptosis alone and in combination with cytarabine (AraC) or doxorubicin in vitro in chemotherapy sensitive and resistant murine acute promyelocytic leukemia (APL) cells. This murine model of APL was previously generated by knocking in the human PML-RARa cDNA into the 5’ regulatory sequence of the cathepsin G locus (Westervelt et al. Blood, 2003). The abnormal co-expression of the myeloid surface antigen Gr1 and the early hematopoietic markers CD34 and CD117 identify leukemic blasts. These Gr1+CD34+CD117+ APL cells partially retain the ability to terminally differentiate toward mature granulocytes (mimicking more traditional AML models) and can be adoptively transferred to secondary recipients, which develop a rapidly fatal leukemia within 3 weeks after tumor inoculation. To assess the safety and efficacy of SINEs in vivo, we injected cryopreserved APL cells intravenously via the tail vein into unconditioned genetically compatible C57BL/6 recipients and treated leukemic and non-leukemic mice (n=15/cohort) with 15 mg/kg of the oral clinical staged SINE KPT-330 (currently in Phase 1 studies in patients with solid tumors and hematological malignancies) alone or in combination with 200 mg/kg cytarabine every other day for a total of 2 weeks. Peripheral blood was obtained weekly from mice for complete blood counts and flow cytometry to screen for development of APL. Results The first generation SINE, KPT214, inhibited the proliferation of murine APL cell lines in a dose and time dependent manner with IC50 values ranging from of 95 nM to 750 nM. IC50 values decreased 2.4-fold (KPT-185) and 3.5-fold (KPT-249) with subsequent generations of the SINEs. Consistent with the WST-1 results, Annexin V/7-aminoactinomycin D flow cytometry showed a significant increase of APL apoptosis within 6 hours of KPT-249 application. Minimal toxicity against normal murine lymphocytes was observed with SINEs even up to doses of 500 nM. Additional WST-1 assays using AraC-resistant and doxorubicin-resistant APL cell lines demonstrated cell death of both chemotherapy-resistant cell lines at levels comparable to the parental chemosensitive APL cell lines. Combination therapy with low dose KPT-330 and AraC showed additive effects on inhibition of cell proliferation in vitro. This additive effect of KPT-330 and chemotherapy on APL killing was maintained in vivo. As shown in Figure 1, treatment with AraC or KPT-330 alone significantly prolonged the survival of leukemic mice from a median survival of 24 days (APL + vehicle) to 33 days or 39 days, respectively (P < 0.0001). Encouragingly, combination therapy with AraC + KPT-330 further prolonged survival compared to monotherapy (P < 0.0001), with some mice being cured of the disease. Similar in vivo studies with the AraC-resistant and doxorubicin-resistant APL cells are just being initiated. Conclusions Our data suggests that the addition of a CRM1 inhibitor to a chemotherapy regimen offers a promising avenue for treatment of AML. Disclosures: Shacham: Karyopharm Therapeutics Inc.: Employment, Equity Ownership, Membership on an entity’s Board of Directors or advisory committees, Patents & Royalties. Kauffman:Karyopharm Therapeutics Inc.: Employment, Equity Ownership, Membership on an entity’s Board of Directors or advisory committees, Patents & Royalties. McCauley:Karyopharm Therapeutics, Inc: Employment, Equity Ownership.

Abstract 2971 Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is the only curative treatment for patients with relapsed/refractory leukemia, and marrow failure states such as myelodysplasia and aplastic anemia. However, allo-HSCT is complicated by allogeneic donor T cell-mediated graft-versus-host disease (GvHD) which can be life-threatening especially in recipients of unrelated or HLA-mismatched hematopoietic stem cell products. These same alloreactive donor T cells also mediate a beneficial graft-versus-leukemia (GvL) effect. Thus, the clinical goal in allo-HSCT is to minimize GvHD while maintaining GvL. Recent studies have suggested that this might be achieved by infusing regulatory T cells (Tregs) which in some preclinical models suppress GvHD-causing alloreactive donor T cells but have only limited effects on GvL-promoting alloreactive donor T cells. Unfortunately, Tregs exist in low frequency in the peripheral blood, are costly to purify and expand, and after expansion are difficult to isolate due to the lack of cell surface markers, all of which prevent their routine use in the clinic. Thus, alternative therapeutic approaches that do not require Tregs are needed. Using a MHC-mismatched GvHD model, B6 (H-2b) → Balb/c (H-2d), we demonstrated that infusion of IFN γR deficient allogeneic donor T cells induce significantly less GvHD, compared to WT T cells, determined by survival (74% vs. 0 % in overall survival; p =0.0004), weight and percentages of B220+ B cells (12.4% vs. 3.8%; p =0.0205), CD3+ T cells (14.3% vs. 4.3%; p =0.0025) in blood. Of note was that the IFN γR deficient donor T cells maintained a beneficial GvL effect, which was examined in both a systemic leukemia and a solid tumor model using luciferase-expressing A20 cells derived from Balb/c. We found that IFN γR deficient donor T cells responded normally to allogeneic antigens as measured by in vitro mixed lymphocyte reaction analyses, and express similar levels of granzyme B, compared to WT T cells. However, IFN γR deficient T cells trafficked predominantly to the spleen while WT T cells trafficked to gastrointestinal tract and peripheral lymph nodes, which are major GvHD target organs, based on in vivo bioluminescence imaging. All of these findings suggest that the reduced GvHD was not due to reduced function, altered subsets or relative deficiency of allogeneic donor T cells but from modification of in vivo trafficking of IFN γR deficient donor T cells compared to WT T cells. We further demonstrated that the IFN γR-mediated signaling in alloreactive donor T cells was required for expression of CXCR3 which has been implicated in trafficking of T cells to areas of inflammation and target organs, commonly known to be the sites of GvHD. CXCR3−/− T cells demonstrated a reduction in GvHD while maintenance of the same robust GvL effect using the same MHC mismatched transplant model. Thus, the IFN γR-CXCR3 axis represents a promising therapeutic target for future efforts to mitigate GvHD while maintaining GvL after allo-HSCT. Current studies are focused on 1) whether forced expression of CXCR3 rescues the GvHD-inducing potential of IFN γR deficient donor T cells and 2) if inhibition of IFN γR signaling (IFN γR, JAK1 and/or JAK2, CXCR3 and STAT1) using both neutralizing antibodies and small molecule inhibitors can recapitulate the anti-GvHD and pro-GvL effects seen in IFN γR−/− and CXCR3−/− T cells. Disclosures: No relevant conflicts of interest to declare.

Abstract 2537 Introduction: Regulatory T cells (Tregs) have been shown to mitigate graft-versus-host disease (GvHD) while preserving the beneficial graft-versus-leukemia (GvL) effect in animal models of allogeneic bone marrow transplantation (BMT). However, three major obstacles prevent their use in human clinical trials: the low numbers of Tregs, loss of suppressor activity following in vitro expansion, and the lack of Treg-specific markers to purify expanded Tregs. The locus of the Foxp3 gene, the master regulator of Tregs, is unmethylated and expressed only in Tregs. We have recently reported that the hypomethylating agent azacitidine (AzaC) induces FOXP3 expression in non-Tregs, converting them into Tregs in vitro and in vivo when administered after allogeneic BMT completely mitigating GvHD without abrogating GvL (Choi, et al Blood 2010). Three possible mechanisms for these effects include: 1) AzaC induces FOXP3+ Tregs, which in turn mitigate GvHD without abrogating GvL by regulating alloreactive donor T cells, 2) AzaC directly suppresses the proliferation of alloreactive donor T cells reducing GvHD, 3) AzaC alters donor T cell trafficking to GvHD target organs to prevent GvHD without altering interaction of donor T cells with recipient leukemia or trafficking of leukemic cells. Methods: Balb/c (CD45.2+, H-2Kd) were lethally irradiated one day prior to injection of T cell-depleted BM cells isolated from B6 (CD45.1+, H-2Kb) and luciferase-expressing A20 leukemia cells derived from Balb/c. Allogeneic donor T cells isolated from B6 (CD45.2+, H-2Kb) were given 11 days after BMT. AzaC (2 mg/kg) was administrated subcutaneously every other day (4 doses total) starting 4 days after T cell injection. In vivo bioluminescence imaging (BLI) was performed to assess leukemia cell localization. For T cell proliferation/trafficking analyses, Balb/c were lethally irradiated one day prior to injection of T cell-depleted BM cells isolated from B6 (CD45.1+). Allogeneic donor T cells isolated from B6 (CD45.2+) were transduced with Click Beetle Red luciferase and were given 11 days after BMT, followed by AzaC treatment as described above. BLI was performed to track the donor T cells. Results: While neither T cell or leukemia cell trafficking was affected by the AzaC treatment, proliferation of donor T cells was significantly reduced compared to mice treated with PBS. The observed reduced T cell proliferation is not likely due to the direct effect of AzaC on T cells since the AzaC treatment preserved GvL activity comparable with the PBS control group. In addition, T cells isolated from both AzaC and PBS groups were equally reactive against third party antigen presenting cells, based on mixed lymphocyte reactions and cytotoxic T lymphocyte killing assays. These data along with our previous report demonstrating that the AzaC treatment increases Tregs in vivo strongly suggest that the therapeutic effect of AzaC on GvHD and GvL are mediated by the AzaC-induced Tregs which preferentially target alloreactive T cells while preferentially sparing anti-tumor T cells. Currently, secondary transplantation of Treg-depleted/replete T cells isolated from AzaC/PBS-treated recipient mice is underway to further confirm that donor T cells in the AzaC-treated mice are fully functional and that alloresponses of donor T cells are regulated by AzaC-induced Tregs. Conclusions: In vivo administration of AzaC after donor T cell infusion mitigates GvHD while preserving GvL via peripheral conversion of alloreactive donor T cells to FOXP3+ Tregs that preferentially inhibit alloreactive T cells while sparing anti-tumor T cells. These data provides the foundation for future clinical trials using epigenetic therapy aimed at mitigating GvHD without abrogating GvL and overcoming HLA barriers. Disclosures: No relevant conflicts of interest to declare.

Hematopoietic stem cells (HSC) reside in the bone marrow (BM) and interact with stroma cells and extracellular matrix. CXCR4/SDF-1 axis regulates the trafficking of HSC to and from the BM. We utilized a PML-RARα knock-in mouse model of human acute promyelocytic leukemia (APL) to study APL interaction with the normal BM. We have previously shown there is a rapid mobilization of APL cells from the BM into peripheral blood (PB) after administration of AMD3100, a competitive inhibitor of CXCR4. We hypothesize that we can sensitize these tumor cells to chemotherapy by interrupting the interaction between APL and the BM stroma. We transduced banked APL cells with a dual function reporter gene that encodes a fusion protein comprised of Click Beetle Red luciferase, a bioluminescence imaging (BLI) optical reporter gene, and EGFP for ex vivo cell sorting (Luc/EGFP). Upon iv injection into genetically compatible recipients (F1 129/B6 mice), APL rapidly migrated to the BM with increased BLI signal in the femurs, spine, ribs, and skull, at 4 days after injection, followed by spleen infiltration and death due to leukostasis by day 15. 129/B6 F1 mice (n=28) were injected iv with 106 APL cells. By day 12 all mice had ±5% APL cells in PB. 8 mice received AraC (500mg/kg/sq) on days 12 and 13, and another 8 mice received AraC+AMD (5mg/kg/sq) 1 hour before and 3 hours after each AraC injection. 6 mice received only AMD and 6 control mice were observed. Total body BLI signal, WBC, and blasts per μl of blood on days 19 and 23 were higher in AraC versus AraC+AMD (p

GVHD is the major cause of morbidity and mortality following allogeneic hematopoietic stem cell transplantation. We developed a consistent and sensitive model of human GVHD by retroorbitally (ro) injecting huT cells into sublethally irradiated NOD/SCID-β2mnull mice (β2). Using this model, 59% (20/34) mice conditioned with 250cGy and injected ro with 107 huT cells developed lethal XGVHD and exhibited huT cell engraftment, weight loss, and infiltration in multiple target tissues. β2 mice injected with huT through the lateral tail vein (iv) had only small and transient huT cell engraftment and failed to develop lethal XGVHD. To evaluate why the ro versus iv route affected huT cell engraftment and the development of lethal XGVHD, we performed in vivo BLI and sequential [18F]FHBG microPET imaging studies. We used huT cells transduced with a click beetle red luciferase/enhanced green fluorescent protein (CBRluc/egfp) BLI reporter gene (huT*) and/or a chimeric CD34-thymidine kinase fusion suicide/PET gene. We injected 107 huT* cells either iv or ro into 250cGy irradiated β2 mice. Mice injected iv (n=5) failed to develop lethal XGVHD, whereas 60% of the mice injected ro (n=5) developed lethal XGVHD. During the first week, mice injected ro presented early expansion of the huT cells in the retroorbital cavity and local regional draining lymph nodes followed by widespread dissemination throughout the whole body, with increased percentage of huT in the blood (p10-fold increase in the BLI signal during the first 2 weeks. This homing pattern and strong BLI signal of the ro-injected huT* cells remained until death from GVHD. The BLI signal of the iv-injected huT* cells increased between days 21 and 45. However, this late expansion was not associated with the development of XGVHD. We also generated huT cells that co-expressed the suicide gene CD34-TK75 and CBRluc/egfp (huT**). We observed the same results after we injected 0.25×107 huT** cells into 250cGy irradiated and anti-CD122 mAb-treated NOD/SCID mice followed with BLI and PET. These studies demonstrated the ability to image trafficking of the transduced T cells in vivo with BLI and PET. In summary, all mice that received huT iv showed initial clearing of the huT cells in the lungs and a significant delay in their in vivo expansion abrogating their potential to cause XGVHD. In contrast, the development of lethal XGVHD in immunodeficient mice was dependent upon the initial retention and early expansion of huT in the retroorbital cavity.

We and others have previously demonstrated that early administration of the prodrug ganciclovir (GCV) protects mice from lethal GVHD after BMT with allogeneic naive (uncultured) thymidine kinase (TK)-expressing transgenic (Tg) T cells. However, the efficacy of GCV control of GVHD following BMT with ex-vivo activated, TK transduced and selected (Td) donor T cells has not been thoroughly characterized because of the significantly reduced GVHD-inducing potential of cultured cells compared to naive T cells. In this study we used the intracellular fluorescent dye CFSE and in vivo bioluminescence imaging (BLI) to monitor the proliferative kinetics, GVHD-inducing potential and GCV sensitivity of Td and naive T cells after allogeneic and syngeneic BMT. Murine T cells were transduced with either a CD34/TK chimeric fusion suicide gene or a click beetle red luciferase-EGFP (CBR/EGFP) dual function reporter gene. High efficiency (>50%) Td of T cells was accomplished 24 h after activation with anti-CD3/anti-CD28 coated beads and gene-modified cells were purified to > 98% using a MoFlo cell sorter or CD34 immunomagnetic selection 48h post-infection. Naive B6 CD34/TK+ Tg T cells were isolated the day of BMT. Purified T cells were then injected (4e6 total T cells), along with T cell depleted B6 (CD45.1+) BM, into lethally irradiated allogeneic (BALB/c, CD45.2+) or syngeneic (B6, CD45.2+) recipients. The CBR/EGFP BLI signal increased nearly 3 orders of magnitude between days 1 and 8 post-BMT, remained steady for a week, and then only gradually declined over the next month (only a 3-fold decrease between days 14 to 42 post-BMT). Consistent with this BLI data, we found, using CFSE, that the naive and Td CD34/TK+ T cell groups exhibited similar division kinetics during the first week after BMT. At 3 days post-BMT in the allogeneic recipients, CD4 and CD8 cells from both the naive and CD34-TK+ donor T cells were detected in the 8 divisions discernible by CFSE, with approximately equal percentages (5%–15%) of cells in each division cycle. By the next day, however, virtually all of the CD4 and CD8 cells from both donor T cell groups had divided more than 7 times. In contrast, less than 10% of the CD4 or CD8 cells of either donor T cell group had undergone more than 7 divisions in the syngeneic recipients. Most importantly, untreated animals (No GCV) that received either naive or Td T cells died of GVHD, with a median survival of 25 days and 26 days, respectively. To evaluate the ability of GCV to prevent this GVHD, we administered a 7 d course of GCV beginning 1, 4, or 10 days after BMT. Treatment with GCV from days 1–7 after BMT significantly decreased the BLI signal intensity and protected the allogeneic recipients from lethal GVHD, irrespective of the donor T cell source. We observed similar protection from lethal GVHD with significantly reduced weight gain in animals that received Td T cells and treatment with GCV from days 4–10. In contrast, lethal GVHD was evident in day 4–10 GCV treated mice that received naive T cells and in all recipients (irrespective of donor T cell source) that were treated with GCV from days 10–16 after BMT. These observations may have important implications in future clinical trials using TK-expressing T cells to control GVHD after BMT.