ALBERT

Description: Abstract 2456 Poster Board II-433 Murine T cells exposed to rapamycin maintain flexibility towards Th1/Tc1 differentiation; the degree to which rapamycin might inhibit human Th1/Tc1 differentiation has not been fully evaluated. In the presence of rapamycin, T cell co-stimulation and polarization with IL-12 or IFN-α permitted human CD4+ and CD8+ T cell differentiation towards a Th1/Tc1 phenotype: by intracellular flow, median percentage expression of Foxp3, IFN-γ, and T-bet was 4%, 20%, and 72%, respectively. Phospho-flow cytometry revealed that such Th1/Tc1 cells expressed activated STAT1 and STAT4 in spite of mTOR blockade; STAT activation was abrogated by PI3 kinase inhibition. Rapamycin-resistant human Th1/Tc1 cells (Th1/Tc1.R cells): (1) had increased expression of the autophagy-related gene LC3BII by gene array and protein analysis; (2) preferentially expressed anti-apoptotic bcl-2 family members (reduced Bax, Bak; increased phospho-Bad); (3) maintained mitochondrial membrane potentials; and (4) had reduced apoptosis relative to control Th1/Tc1 cells not generated in rapamycin (p=0.04). The anti-apoptotic phenotype of Th1/Tc1.R cells was abrogated by co-incubation with the autophagy inhibitor, 3-methyl adenine. The in vivo effect of the Th1/Tc1.R cells was evaluated using two xenogeneic GVHD (x-GVHD) models. First, in an LPS-induced x-GVHD model, Th1/Tc1.R cells resulted in lethality in 75% recipients; soluble TNF-α receptor therapy with etanercept reduced the frequency of lethality to 15%. Second, using a non-LPS natural history model of x-GVHD, recipients of Th1/Tc1.R cells (relative to recipients of control Th1/Tc1 cells) had increased human T cell engraftment (day 30 post-BMT, p=0.001), increased human T cell cytokine levels, increased human T cell expression of the cytotoxic degranulation molecule CD107 (p=0.05), and increased human T cell infiltration of skin, gut, and liver. In this model, lethality due to x-GVHD was also increased in Th1/Tc1.R cell recipients (lethality increased from 20% to 70%, p=0.04). We conclude that rapamycin therefore does not impair human T cell capacity for type I differentiation. Rather, by promoting autophagy rapamycin permits stable expression of T-bet and generates an anti-apoptotic Th1/Tc1 effector phenotype, thereby yielding increased x-GVHD. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction: The technique of selectively depleting alloreacting T cells from the stem cell transplant (SCT) to prevent graft-versus-host disease (GVHD) while sparing T cells with useful reactivity against viruses and leukemia is commonly termed selective depletion (SD). This approach involves the ex vivo stimulation of donor T cells with host antigen presenting cells (APCs) followed by targeted elimination of the activated alloreactive donor T cell repertoire. SD has been achieved through immunotoxin or magnetic bead depletion targeting T cell activation markers such as CD25 or photodepletion with a rhodamine-like dye. These strategies, while having some success in clinical trials, are limited by poor depletion efficacy or off-target depletion. Here, we present an optimized, novel SD method employing ex vivo treatment of alloactivated T cells with pharmacological concentrations of the purine nucleotide adenosine that preserves quiescent lymphocytes, regulatory T cells (Treg), and T cells with antiviral and antileukemic specificities. Methods: Mature dendritic cell (DC) or irradiated peripheral blood mononuclear cell (PBMCs) "recipient" stimulators were co-cultured with HLA mismatched "donor" PBMC for up to 7 days. Pharmacologic grade adenosine was added at various time points in culture at doses ranging from 100uM-2mM. Depletion efficacy was analyzed by proliferation of the depleted product when challenged separately against stimulator, responder, and 3rd party APCs in 5 day CFSE-dilution assay to measure residual alloreactivity, background proliferation, and response to de novo antigens. Activity against viral antigens was assessed through 6 hour peptide challenge followed by flow-cytometry based intracellular cytokine staining. Leukemia-specific cells were generated from the allodepleted product through two weekly stimulations with autologous DCs transduced to express leukemia associated antigens WT1 and PRAME. Results: Adenosine reduced alloreactive T lymphocytes in HLA-mismatched allogeneic co-cultures to background control frequencies or lower (Figure 1). Optimum allodepletion was achieved with addition of 2mM adenosine on days 1, 2, and 5 after establishing co-culture. Allodepleted lymphocyte products comprised 37±3% of initial cell numbers with 90±1% viability (n=18). Similarly, this method reproducibly achieved SD in four haploidentical donor-recipient pairs, reducing CFSE-tracked proliferation in challenge against haplo APCs below the background of proliferation in challenge against autologous APCs. Adenosine depletion equally affected CD4 and CD8 T cell subsets while sparing NK and B cell populations. Analysis of naïve, effector memory (EM), central memory (CM), and terminal effector (EMRA) T cell compartments revealed no significant differences in depletion of a particular subset in the allodepleted T cell products (n=6). Critically, CD4+CD25+FOXP3+ Treg populations (n=6) and activity against viral antigens (cytomegalovirus, Epstein-Barr virus, and adenovirus) were maintained and in some cases enriched (n=6). Significant activity against leukemia associated antigens (PRAME and WT1) was achieved in allodepleted products from 3 donors. Conclusion: This novel SD technique employing adenosine as a pharmacological agent satisfies the requirements of allodepletion with preservation of viral and leukemia-specific immune responses and thus presents a potentially economical method to deplete alloreacting T cells in SCT products for clinical applications. Disclosures No relevant conflicts of interest to declare.

Description: We have recently shown that ex vivo manufactured, rapamycin-resistant donor Th2 cells (Th2.R cells) prevent the rejection of allogeneic bone marrow grafts through a process that involves IL-4-mediated polarization of host T cells towards a Th2 phenotype. Because donor T cell therapy lacks feasibility in the setting of cardiac allograft transplantation using cadaveric donors, we developed an approach to prevent rejection that incorporates host Th2 cell therapy prior to organ transplantation. The aims of the study were to: (1) develop a method for the ex vivo manufacture of rat Th2.R cells and inject such syngeneic Th2.R cells just prior to class I and class II disparate cardiac allografting; (2) determine whether such adoptive host Th2.R cell transfer polarizes post-transplant immunity towards a Th2 phenotype; and (3) evaluate in a preliminary manner whether Th2.R cell transfer reduces graft rejection in a model that uses sub-optimal cyclosporine therapy. Recipient-type (Black Norway, BN) CD4+ T cells were co-stimulated with anti-rat anti-CD3 and –CD28 antibodies in the presence of rapamycin, rrIL4, rhuIL-2, and rhIL7 (3-day culture interval). The Th2.R cell cytokine phenotype was determined by intra-cellular flow cytometry and multi-analyte testing of supernatants obtained after repeat co-stimulation. Th2.R cells were injected (i.v.; 1 × 107 cells) 2–3 hours prior to cardiac allografting (donor strain, Dark Agouti; DA). Cardiac function was evaluated clinically by cardiac palpation through day 28 post-grafting. At day 28, histology studies and immune function studies were performed, including assessment of host-anti-donor alloreactivity. Engraftment controls (cohort #1) and rejection controls (cohort #2) received cyclosporine (CSA) through the full 28 days or only through day 18, respectively; the experimental cohort (#3) received host Th2.R cells followed by the short-course of CSA. Supernatants from the Th2.R cell product contained 1500 pg/ml IL-4 and only 50 pg/ml of IFN-γ; the frequency of IL-4+ cells and IFN-γ+ cells was 10% and 1%, respectively. Clinical rejection was observed in cohort #2 (3/3 subjects) whereas all subjects in cohorts #1 and #3 had full cardiac function through day 28 (cohort #1- 2/3 subjects; cohort #3- 3/3 subjects). Relative to rejection controls, Th2.R cell recipients had a reduced frequency of intra-cardiac CD8+ T cells (%CD8+ T cells, 9.2 ± 6.2 vs. 1.1 ± 0.3; p

Description: Abstract 3580 Poster Board III-517 Immune cell expression of programmed death ligand-1 (PD-L1) represents a particularly important molecular mechanism responsible for control of auto- and allo-immunity mediated by effector memory T cells expressing PD1 receptor. As such, we have reasoned that an immuno-gene therapy approach that enables T cell expression of PD-L1 will represent a novel method of immune regulation. Advantageous features of this proposed therapy include a capacity to: (1) enforce long-term, stable expression of PD-L1; (2) build-in an independent surface marker to allow specific transduced cell enrichment; (3) utilize cellular delivery vehicles comprised of highly functional T cells that persist in vivo after adoptive transfer; and (4) incorporate an enhanced cell fate control or ‘suicide’ gene to permit in vivo control of the immuno-gene therapy. Given these considerations, we developed a recombinant lentiviral vector (LV) incorporating an EF1-α promoter that first encodes the cDNA for a fusion protein consisting of human CD19 (truncated, non-signaling) combined with mutated human TMPK that efficiently activates AZT as a pro-drug (Sato et al; Mol Therapy, 2007); then, after an IRES element, the vector encodes full-length human PD-L1. LV was made after transfection of 293T cells and then concentrated and titered. Initial experiments used Jurkat cells to optimize virus infection and to confirm co-expression of CD19 and PD-L1 by flow cytometry. In previous work, we have demonstrated that ex vivo T cell expansion in rapamycin induces an anti-apoptotic phenotype that permits enhanced in vivo T cell persistence in murine models and human-into-mouse xenogeneic transplant models. As such, we established the goal of infecting primary human CD4+ T cells manufactured using ex vivo co-stimulation (anti-CD3, anti-CD28), Th1-type polarization (inclusion of IFN-α), and exposure to high-dose rapamycin (1 μM); using a 6-day culture system and subsequent anti-CD19 column purification, 〉90% of resultant transduced T cells expressed PD-L1. Next, we utilized a xenogeneic transplantation model (Rag2−/−γc−/− hosts) to assess in vivo persistence of the gene-modified T cells and transgene expression (10,000 T cells transferred i.v. into each host). In vivo experiment #1 demonstrated that recipients of gene-modified T cells had increased numbers of human T cells in the spleen that co-expressed CD19 and PD-L1 relative to recipients of non-transduced but identically expanded human T cells (harvested at day 5 after adoptive transfer; 38,000 cells/spleen vs. 1000 cells/spleen, p=0.02). Such in vivo harvested T cells were secondarily co-stimulated ex vivo and propagated for an additional 5 days: co-expression of CD19 and PD-L1 persisted in ∼ 50% of T cells harvested from the gene-modified T cell cohort, and T cell numbers were maintained ex vivo (yield of CD19+PD-L1+ cells, 28,600 vs. 1500; p=0.0001). In vivo experiment #2 confirmed and extended these results. At day 21 after adoptive transfer, recipients of gene-modified T cells had increased numbers of human T cells that co-expressed CD19 and PD-L1 relative to recipients of non-transduced but identically expanded human T cells in both the spleen (2800 cells/spleen vs. 390 cells/spleen, p=0.01; n=10 per cohort) and bone marrow (71,600 cells/marrow vs. 6500 cells/marrow, p=0.0001; n=10 per cohort). Such in vivo harvested T cells at day 21 after adoptive transfer were secondarily co-stimulated ex vivo and propagated for an additional 6 days: co-expression of CD19 and PD-L1 persisted in ∼ 50% of T cells harvested from the gene-modified T cell cohort, and T cell numbers were maintained ex vivo (yield of CD19+PD-L1+ cells harvested from spleen, 71,200 vs. 1800, p=0.0008; yield of CD19+PD-L1+ cells harvested from marrow, 226,000 vs. 1400, p=0.0001). Because the rapamycin-resistant T cell vehicle utilized in these experiments manifests an anti-apoptotic phenotype that confers long-term engraftment potential, it is likely that the demonstrated durability in transgene expression relates both to the efficiency of the LV method utilized and to a T cell pro-survival function. In conclusion, the LV-mediated transfer of this novel combination of CD19/TMPK fusion protein and PD-L1 results in stable transgene expression in primary human T cells in vitro and in vivo, thereby opening an avenue to assess PD-L1 mediated immuno-gene therapy under cell fate control. Disclosures: No relevant conflicts of interest to declare.

Description: Rapamycin-generated donor Th2 cells attenuate established acute murine GVHD (Foley et al, JI, 2005) and are dependent in part upon IL-4 and IL-10 secretion (ASBMT Meeting, 2007). That is, Th2.rapa cell recipients (Th2 infusion, d 14 post-BMT) had increased survival relative to GVHD controls (post-BMT survival, median days; 33.7±0.4 vs. 24.8±1.2; p=0.0002) whereas recipients of IL-4 or IL-10 knockout Th2.rapa cells did not have increased survival (28.9±0.3 and 24.6±0.2 days, respectively; p=NS). These data indicate that Th2.rapa cells operate through a Th2-type mechanism rather than a Treg cell mechanism; in addition, we found that Th2.rapa cells expressed low levels of the Treg cell transcription factor, Foxp3 (

Description: Abstract 501 Introduction: Previously, we found that ex vivo usage of rapamycin and a type I polarizing cytokine could be utilized to generate murine Th1/Tc1 cells that persisted in vivo and mediated increased GVHD. The type I cytokine profile and increased in vivo persistence of this population may prove useful in autologous transplant settings. In recent experiments using culture flasks and polarization with IFN-α, we found that human T cells similarly can adopt a Th1/Tc1 phenotype in rapamycin; furthermore, such human T cells have increased in vivo persistence when transferred into immune-deficient murine hosts. In this project, our goal was to evaluate whether human Th1/Tc1 cell generation in rapamycin varied depending on use of culture flasks or more clinically relevant, closed system polyolefin culture bags. Methods: Human lymphocytes were collected by steady-state apheresis and subsequent counter flow centrifugal elutriation. CD3+ T cells were purified and co-stimulated using anti-CD3/anti-CD28 tosyl-activated P450 magnetic beads. T cells were cultured under four conditions: recombinant human (rh) IL-2 alone; rhIL-2 plus rapamycin; rhIL-2 plus rhIFN-α; or rhIL-2 plus rhIFN-α plus rapamycin. After 6 days of culture in flasks or bags, cells were harvested and co-stimulated; cytokine production was measured by Luminex assay, and cell surface markers were assessed using flow cytometry. Cells were further characterized using an in vivo xenogeneic cytokine storm model of GVHD (x-GVHD). Specifically, immune deficient Rag2/γc knockout mice were irradiated, injected with ex vivo expanded T cells, and challenged with lipopolysaccharide (LPS) injection on day 6 after adoptive transfer to induce lethal levels of TNF-α. Results: Relative to T cells expanded in only IL-2, further addition of IFN-α yielded T cells with increased secretion of IFN-γ (33.5 vs. 5.7 ng/ml, p=0.01) and reduced secretion of IL-4 (0.6 vs.18.7 pg/ml, p=0.02). Inclusion of both IFN-α and rapamycin also yielded cells with preferential secretion of IFN-γ relative to IL-4; IFN-γ and IL-4 secretion values did not vary significantly between flask and bag conditions. However, relative to flask expanded T cells, Th1/Tc1 cells expanded in bags had increased co-expression of T central memory molecules CD62L and CCR7 (median percentage of co-expression increased from 20% to 40%, p