ALBERT

Description: Background: Natural killer T (NKT) cells are one of the primary effectors of the innate immune systems, and also have an important role to initiate and regulate adaptive immune responses. Previously, we reported that Vα24+ NKT cells proliferated more efficiently from granulocyte-colony stimulating factor (G-CSF)-mobilized peripheral blood mononuclear cells (PBMC) compared with steady-state blood cells. However, optimal culture conditions and characterization of ex vivo expanded NKT cells have not been fully evaluated. Object: We examined several different conditions including cell ratios, mediums, growth factors and incubation schedules to seek an optimal culture system for expansion of NKT cells. Methods: PBMC were collected from donors for hematopoietic stem cell transplantation before and after G-CSF administration, and were cultured in AIM-V medium supplemented with 10% auto-plasma, 100 ng/mL α-galactosylceramide (α-GalCer) and 100 U/mL recombinant human (rh) IL-2. IL-2 alone was repeatedly charged every 3 days to maintain its biological activity. After 12 days culture, we compared the expansion efficacy of Vα24+CD3+ NKT cells derived from PBMC with or without G-CSF. For depletion analysis, we used a magnetic cell sorting (MACS) system with labeling magnetic beads-conjugated monoclonal antibody against CD14, 56, 34, and TCR Vα24 chain. Results: The expansion fold of Vα24+CD3+ NKT cells were significantly higher with G-CSF (669 vs 182 fold, n=20). Among cell populations we tested, the proportion of CD14+CD16+ cells before cultures were associated with the efficacy of Vα24+CD3+ NKT cells expansion, and the proportion of CD34+, Vα24+, CD56+ and CD56+CD161+ cells were not. The magnitude of expansion of Vα24+CD3+ NKT cells was correlated with the percentage of CD14+ cells at the initiation of cultures. Proliferation of Vα24+CD3+ NKT cells was abrogated by the depletion of Vα24+cells, but notCD34+ cells. Depletion of CD56+ T cells induced higher expansion ratio of Vα24+CD3+ NKT, which was abolished when CD56+ and CD56− cells were cultured separately using a 3 μm pored-membrane filter. Furthermore, co-culture of enriched Vα24+ cells and purified CD56+ cells inhibited the proliferation of Vα24+CD3+ NKT cells. It was hypothesized that the repeated IL-2 supplementation resulted in enhancement of CD56+ cells (NK cells) to suppress the proliferation of Vα24+CD3+ NKT cells. We tested different administration schedule of IL-2 as follow: on day 0 only, day 0 & 3, day 0, 3 & 6, day 0, 3, 6 & 9 (that is every 3 days), and we found that Vα24+CD3+ NKT cells expanded most effectively when IL-2 was supplemented on day 0 only. In order to modify the number of CD14+ cells in culture system, we added back CD14+ cells to CD14− cells culture on day 0, 3, 6, 9 or every 3 days, but this did not result in significant enhancement of proliferation of Vα24+CD3+ NKT cells. Conclusions: For efficient ex vivo culture of Vα24+ NKT cells, the presence of Vα24+ NKT cells and CD14+ cells are critical. The NK cells may interfere the interaction between antigen presenting cells (APC) and NKT cells by hindering a function of antigen presentation or yielding direct cytotoxicity against APC. These findings are helpful to develop an efficient expansion system of NKT cells in feature adaptive immunotherapy.

Description: [Background] Non-myeloablative regimens have been proven to allow engraftment following allogeneic stem cells transplantation with minimal procedure-related toxicity and lower costs. Cord blood has emerged as an appealing alternative source of hematopoietic stem cells for unrelated donor transplantation, but delayed engraftment and frequent transfusion were reported. No studies have formally evaluated the cost of reduced-intensity cord blood transplantation (RICBT). [Purpose] To evaluate the relationship among costs, baseline patient characteristics, and major complications of RICBT, we performed an economic analysis of data in a clinical trial of RICBT for hematologic diseases at a single institution. [Patients and Methods] Ninety-three patients with hematological diseases (median age, 55y; range, 17–79: median body weight, 53kg; range, 38–75) underwent RICBT from March 2002 to May 2004 in Toranomon Hospital. Mean follow-up period was 77 days (range, 13–863). Data on resource use, including hospitalizations, medical procedures, medications, and diagnostic tests, were abstracted from subjects’ clinical trial records. Resources were valued using the Japanese national insurance reimbursement system for inpatient costs at one hospital and average wholesale prices for medications. Monthly costs were calculated and stratified by treatment group and clinical phase. [Results] The median initial inpatient cost was $80,400 (range, 41,300–154,700). When baseline variables were considered, disease status was significant predictor of costs. When clinical events were considered, in-hospital death was associated with higher costs. The mean length of total inpatient days was 78 days (range, 31–222), and the mean length of inpatient days post transplant was 51 days (range, 15–131). The mean units of transfused RBC, Platelet, and FFP were 27u, 224u, and 27u, respectively. [Discussion] This study firstly demonstrates that the cost of RICBT was much higher as compared to previous RIST using peripheral blood or bone marrow. RICBT is an attractive therapy, however, economic problem lies before prevalence of RICBT. The increased numbers of transfusions and supportive care would have effects on costs. The association between mortality and higher costs suggest that prevention of clinical complication may have significant economic benefits. Interventions that decrease these complications may have favorable cost-benefit ratios, and will be the focus of future investigation.

Description: Erythropoietin (EPO) is required for the survival, proliferation and differentiation of erythroid progenitor cells. The scaffolding adaptor protein Grb2-associated binder-1 (Gab1) is tyrosine phosphorylated upon stimulation of EPO in several cell lines and erythroid progenitor cells, and interacts with signaling molecules such as SHP2 phosphatase and the p85 subunit of phosphatidylinositol 3-kinase (PI3K). However, biological functions of Gab1 in EPO receptor (EPOR)-mediated signaling has not yet been established. In this study, to explore the biological functions of Gab1 in vivo, Gab1-deficient F-36P human erythroleukemia cells were generated by means of transfection of the expression vector of siRNA against Gab1. WST-1 assay revealed that growth of Gab1-deficient F-36P cells was reduced to 61% and 77%, respectively, 5 days after incubation with lower concentrations of EPO (0.001 and 0.01 ng/ml), compared with that of mock-transfected F-36P (F-36P-mock) cells. In contrast, growth of Gab1-deficient F-36P cells at sufficient concentration of EPO (10 ng/ml) was similar to that of F-36P-mock cells. Analysis of apoptosis by flow cytometry using FITC-labeled annexin-V showed that the percentage of annexin-V-positive apoptotic cells in Gab1-deficient F-36P and F-36P-mock cells was increased to 19% and 34%, and 8% and 17%, respectively, 72 h after incubation with 0.01 and 0.001 ng/ml of EPO. These results indicate that Gab1 plays a crucial role in transducing EPOR-mediated survival signals. Next, we examined the molecular mechanism of EPOR-mediated signaling involved in survival of erythroid cells through Gab1. Western blot analysis showed that EPO-induced phosphorylation of threonine 202/ tyrosine 204 on Erk-1 and Erk-2 in Gab1-deficient F-36P but not in F-36P-mock cells was significantly suppressed. Interestingly, phosphorylation of serine 473 on Akt in Gab1-deficient F-36P cells in response to EPO was slightly suppressed in comparison with that in F-36P-mock cells. Therefore, Gab1-mediated survival signals appear to be mainly transmitted to downstream through activation of the Erk pathway, although the PI3K/Akt pathway may be involved in EPO-initiated survival signal transduction mediated by Gab1. Furthermore, EPO induced the association of SHP2 with EPOR in Gab1-deficient F-36P cells. Gab1 was associated with SHP2 in EPO-treated F-36P cells. In addition, Gab1 was constitutively associated with Grb2 in F-36P cells. Taken together, EPO induces the recruitment of Gab1 to EPOR through binding of Gab1 to SHP2, which is associated with EPOR. Because the guanine nucleotide exchange factor SOS1 is known to bind to the SH3 domain of Grb2, SOS1-Grb2 complex is recruited to vicinity of Ras at the plasma membrane to activate this GTP-binding protein through the interaction of Grb2 with Gab1, leading to activation of Erk.

Description: Abstract 354 Adoptive cell therapy utilizes unique mechanisms of action to prevent the development of infections in immunocompromised patients and treat chemotherapy resistant malignancies. In adoptive cell therapy, the major effector cells appear to be CD8+ T cells, since they are armed with antigen-specific effector functions, i.e. cytotoxicity and cytokine secretion. However, the roles of antigen-specific CD4+ T cells in T cell immunity are also critical. In immunocompromised patients adoptively transferred with CMV-specific CD8+ T cells, long-term in vivo persistence was achieved only when CMV-specific CD4+ T cells were also present in vivo. Recently, adoptive transfer of a NY-ESO-1 specific CD4+ T cell clone was reported to induce a complete response in a patient with metastatic melanoma. These results suggest that adoptive cell therapy for infectious diseases and cancer can be improved by infusing both antigen-specific CD4+ helper T cells as well as CD8+ CTL. Unfortunately, however, few versatile systems are available for producing large numbers of antigen-specific human CD4+ T cells for the purpose of adoptive therapy. K562 is a human erythroleukemic cell line, which lacks the expression of HLA class I and II, invariant chain (Ii), and HLA-DM, but does express adhesion molecules such as ICAM-1 and LFA-3. Given this unique immunologic phenotype, K562 has served as a useful tool in clinical cancer immunotherapy trials. Previously, we reported the generation of a K562-based artificial APC (aAPC), which expresses HLA-A2, CD80, and CD83. aAPC/A2 can uniquely support the priming and prolonged expansion of large numbers of antigen-specific CD8+ CTL which display a central/effector memory phenotype, possess potent effector function, and can be maintained in vitro without any feeder cells or cloning. aAPC/A2 is equipped with constitutive proteasome and inducible immunoproteasome machinery and can naturally process and present CD8+ T cell peptides via transduced A2 molecules. We have successfully generated clinical grade aAPC/A2 under cGMP conditions and conducted a clinical trial where patient with advanced melanoma are infused with large numbers of MART1-specific CTL generated ex vivo using aAPC/A2, IL-2 and IL-15. Based on our experience with aAPC/A2 and CD8+ T cells, we have generated a series of novel aAPC (aAPC/DR1, DR3, DR4, DR7, DR10, DR11, DR13, and DR15) to stimulate HLA-DR-restricted antigen-specific CD4+ T cells. K562 has been engineered to express HLA-DRα and β chains as a single HLA allele in conjunction with Ii, HLA-DMα and β chains, CD80 and CD83. CD83 delivers a CD80-dependent T cell stimulatory signal that allows T cells to be long-lived. Following the transduction of Ii, CLIP (class II invariant chain-associated peptide) appeared on the cell surface of aAPC. Furthermore, CLIP expression on aAPC was almost completely abrogated by the introduction of HLA-DM. This result is in accordance with previous studies showing that HLA-DM catalyzes the removal of CLIP from DR thus enabling exogenous peptides to bind to empty DR molecules in late endosomes. In addition to its endogenous pinocytic activity, aAPC was made capable of Fcγ receptor-mediated endocytosis by transduction of CD64. Comparison of naïve aAPC and CD64-transduced aAPC confirmed that Fcγ receptor-mediated endocytosis is more efficient than pinocytosis to take up soluble protein and process and present DR-restricted peptides to CD4+ T cells. Using these standardized and renewable aAPC, we determined novel viral protein-derived DR-restricted CD4+ T cell epitopes and expanded large numbers of viral antigen-specific CD4+ T cells without growing bystander Foxp3+ regulatory T cells. Without any feeder cells or cloning, expansion of CD4+ T cells using aAPC and low dose IL-2 and IL-15 was sustainable up to 150 days. Immunophenotypic analysis using HLA-DR tetramers and specific mAbs revealed that expanded CD4+ T cells were CD45RA−, CD45RO+, CD62L+-, demonstrating a central/effector memory phenotype. Furthermore, intracellular cytokine analysis showed that expanded DR-restricted viral-specific CD4+ T cells secreted IL-2 and IFN-γ but much less IL-4, displaying a Th1-biased phenotype. Taken all together, these results suggest that K562-based aAPC may serve as a translatable platform to generate both antigen-specific CD4+ helper T cells and CD8+ CTL. Disclosures: No relevant conflicts of interest to declare.

Description: Background Extended-spectrum b-lactamase (ESBL)-producing bacteria are resistant to several types of antibiotics excluding carbapenems. Consequently, infections caused by ESBL-producing bacteria are difficult to treat. Especially for the immunocompromised including hematologic malignancies ESBL-producing bacteria are emergent pathogens. In addition, there are no evidence-based guidelines specifying the infection control for ESBL-producing bacterial outbreak. Along with the recent trend of wide spreading ESBL genes in the community, carriers ESBL-producing bacteria in the community become common. Methods An ESBL-producing Enterobacteriaceae (ESBL-E) outbreak observed between July 2012 and August 2012 in Kagawa University Hospital was surveyed using various molecular microbiology techniques. In this study, we attempted to identify the source of an outbreak of ESBL-producing bacteria in a medical oncology and immunology care unit. We used pulsed-field gel electrophoresis (PFGE) and PCR-based ESBL gene analysis as molecular typing methods. DNA extracted from ESBL-E was digested overnight with Xba I. Tenover criteria were used for the separation and identification of each band. ESBL-producing bacteria were screened by PCR/DNA for the detection of blaTEM, blaSHV, and blaCTX-M. Subsequently, ESBL genes with specific CTX-M subtypes (groups 1, 2, 8 and 9) were characterized using multiplex PCR amplification. The antibiotic susceptibility of strains isolated from clinical samples was assessed by microdilution methods using an IA20MIC mkII system (Koden Industry Co., Japan). Results In total, 21 isolates (12 K. pneumoniae and 9 E. coli) were obtained from clinical samples, including four control strains (two each of both bacteria), that were physically different from those obtained from other inpatient units in our hospital. PFGE for K. pneumoniae (digested by Xba I) produced similar patterns excluding one control strain. PCR classification of the ESBL gene for K. pneumoniae revealed that all strains other than the control strain carried CTX-M9. On another, ESBL genes derived from E. coli strains were varied in each strain with TEM, SMV, CTX-M-1, CTX-M-2, or CTX-M-9. This result was reconfirmed by direct DNA sequencing. Although the outbreak of K. pneumoniae was considered to be "clonal," PFGE and PCR classification of the ESBL genes for E. coli uncovered at least six different "non-clonal" strains possessing individual ESBL gene patterns. The pattern of antimicrobial susceptibility was more variable for K. pneumoniae than for E. coli. Conclusion Typing by PFGE and ESBL gene PCR analysis is practical for discriminating various organisms. In our cohort, two outbreaks were concomitantly spread with different transmission strategies, namely clonal and non-clonal, in the same unit. This might represent clinical evidence that transmissibility differs according to the type of strain. Our speculation is that the ability to acquire ESBL genes differs among bacteria, and the penetrance of resistance is strain-dependent. ESBL-E outbreaks can occur concurrently, as shown in this study. We observed that patient-to-patient transmission of ESBL-E occurred according to the properties of each individual strain and strain-specific maneuver could be optimal. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 2086 CD8+ T cells play a primary role in rejecting pathogens and tumors. Studies of CD8 null mice show that CD8 coreceptor is critical for the development of MHC class I-restricted CD8+ T cells in the thymus. However, while these mice possess low numbers of CTL with limited clonality, they are highly avid and contain acute and chronic infections. In humans, CD8 deficiency leads to no or only mild symptoms of immunodeficiency. These results suggest that the CD8 coreceptor is not absolutely necessary for the generation of antigen-specific CTL and that there exists a compensatory mechanism for the loss of CD8 expression. Common γ chain receptor cytokines (e.g. IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21) transmit STAT-mediated signaling in T cells. Importantly, it has been demonstrated that crosstalk between TCR and STAT signaling occurs in CD8+ T cells. For example, weak or partial TCR-initiated signaling via the Ras/MAPK pathway can be complemented by IL-2 induced STAT activation. This mechanism enables common γ chain receptor cytokines to modulate optimal and supplement suboptimal TCR signaling. Recently, we reported the generation of K562-based artificial APC (aAPC) expressing HLA-A2 as a single HLA allele in conjunction with CD80 and CD83 (wt A2-aAPC). This aAPC can prime naïve CD8+ T cells ex vivo and generate antigen-specific CD8+ CTL with a central memory~effector memory phenotype. It has been demonstrated that the D227K/T228A mutations of A2 molecules (mut A2) inhibit the CD8/MHC interaction without affecting TCR/pMHC interactions. In this human ex vivo study, using a modified aAPC which expresses mut A2 molecules (mut A2-aAPC), we tested our hypothesis that CD8 coreceptor-independent T cell stimulation in the presence of complementary adaptive cytokines preferentially stimulates high avidity antigen-specific CTL. When freshly isolated A2+ CD8+ T cells were restimulated against recall antigen with mut A2-aAPC, both antigen specificity measured by tetramer positivity and number of expanded antigen-specific CD8+ T cells were significantly lower compared with wt A2-aAPC. Similar results were obtained when naïve CD8+ T cells were primed ex vivo with wt-A2 aAPC in the presence of CD8 coligation and then restimulated with mut-A2 aAPC in the absence of CD8 binding. When naïve CD8+ T cells were initially primed with mut A2-aAPC in the absence of CD8 binding, subsequent restimulation using wt A2-aAPC in the presence of CD8 coligation was not able to induce the proliferation of antigen-specific CD8+ CTL. As expected, when naïve CD8+ T cells were both primed and restimulated with mut A2-aAPC in the complete lack of CD8 coligation, antigen-specific CD8+ CTL did not grow. However, adding IL-21 overcame this deficiency. When CD8-independent T cell priming and restimulation by mut A2-aAPC was supplemented with IL-21, antigen-specific CD8+ CTL expansion with high functional avidity occurred. Lack of CD8 binding to MHC results in partial TCR signaling because of the absence of CD8/Lck recruitment to the proximity of MHC/TCR. While the responses of IL-21R- Jurkat and its Lck-null mutant, J.CaM1.6, to IL-21 were minimal, both showed robust IL-21 responses when stably transduced with IL-21R. Intracellular staining revealed that IL-21 induced robust phosphorylation of STAT3 but not STAT1 upon stimulation in both IL-21R-transduced Jurkat and J.CaM1.6. When IL-21R-transduced Lck-null J.CaM1.6 cells were stimulated in the presence of IL-21, T cell responses were completely abrogated by STAT3 inhibition. In contrast, the MAPK inhibitor only partially blocked the T cell responses. The combination of suboptimal doses of the STAT3 inhibitor and the MAPK inhibitor completely nullified the T cell responses, indicating an additive effect of STAT3 and MAPK inhibition. These results suggest that STAT3 but not STAT1 is critically involved in IL-21 signaling that rescues the defective T cell responses caused by the lack of Lck. Taken all together, this study suggests that CD8 ligation is critical for the expansion of post-thymic peripheral antigen-specific CTL in humans. However, STAT3-mediated IL-21 signaling can complement partial TCR signaling caused by the lack of CD8 association and expand CTL with high functional avidity. Since high functional avidity CTL are optimal effectors for the clearance of pathogens and tumors in vivo, our findings may be important for generating high avidity CTL ex vivo for effective adoptive therapy. Disclosures: No relevant conflicts of interest to declare.

Description: Background: T cells from a stem cell source are inevitably contaminated, and over 5.0×104/kg T cells are thought to induce graft-versus-host disease (GVHD) in HLA-mismatched or haplo-identical stem cell transplantations (SCTs) [4]. To suppress GVHD reactions, a procedure for T-cell depletion (TCD) was developed over the past several decades, especially for HLA-mismatched and haplo-identical SCTs, which are at high risk for GVHD. To reduce the incidence of GVHD, a potentially effective agent is anti-thymocyte globulin (ATG), which is generally administered at a dose of ≥ 5-10 mg/kg. Based on data regarding the use of ATG for the treatment of aplastic anemia, we hypothesized that ATG might accommodate engraftment and inhibit GVHD. We attempted to use a lower dose of ATG to decrease non-relapse mortality (NRM) in Japanese patients undergoing an HLA-matched SCT. Patients and method: We treated patients with hematological diseases who underwent an allogeneic SCT after March 2010 without or with 2.5 mg/kg ATG. The inclusion criteria for underlying disease included both hematological malignancies and bone marrow failures. All consecutive patients transplanted from an allogeneic related or unrelated donor were included. Cord blood transplantations were omitted from this analysis. The patients who underwent an SCT before February 2010 (n=20) were examined as the control group without ATG treatment. ATG was administered 1 day prior to the transplantation day at 2.5 mg/kg with 500 mg/body methylpredonisolone as a preconditioning procedure. GVHD prophylaxis, tacrolimus 0.03 mg/kg and short-term MTX (10-7-7 mg/m2) was adapted for both the ATG group and the control group. Results: Thirty-nine (21 male, 18 female) recipients were recruited (median age 49 yrs, range 19-64 yrs). Their underlying diseases were acute myeloid leukemia (n=14), acute lymphoblastic leukemia (n=10), myelodysplastic syndrome (n=5), lymphoma (n=7), and myeloma, aplastic anemia, and other malignancy (n=1 each). Preparation regimens were myeloablative for 17 patients (14 cyclophosphamide [CY]/total body irradiation [TBI], two busulfan [BU]/CY, and another) and non-myeloablative for the other 22 patients (14 fludarabine/melphalan [Flu/Mel] and eight Flu/BU). All but one patient achieved engraftment, and one secondary graft failure was observed. The overall incidences of acute and chronic GVHD were 63.2% and 15.8% for the ATG-treated patients (40.0% and 25.0% for the control cohort), respectively. Acute GVHD (grades II to IV and III to IV) in the recipients who received ATG occurred in 21.1% and 0.0% (control cohort, 10.0% and 5.0%), respectively. The estimated probability of overall survival (OS) 2.5 yrs after transplantation was 77.8% for the ATG group (controls, 57.1%). The relapse rate 2.5 yrs after transplantation was 21.1% and 20.0% in the ATG and control groups, respectively. The NRM rate was decreased after ATG treatment: 25.0% vs. 10.5% (not significant). The causes of mortality with or without ATG were recurrent diseases (n=1 and 2), infection (n=1 and 0), and adverse events caused by transplant-related complication (n=1 and 5), respectively. No deaths due to acute or chronic GVHD occurred. Discussion: Low-dose ATG could suppress the incidence of severe acute GVHD and chronic GVHD without increasing the NRM, although our study design did not have enough power to make a conclusion about the efficacy of low-dose ATG. However, partial T-cell depletion may be effective for HLA-matched SCT recipients. Our results show that ATG at 2.5 mg/kg can be used safely for the Japanese transplant population of HLA-matched donors. Low-dose ATG is a potential treatment to partially disempower T cells from a stem cell source, which are inevitably contaminated. Recent developments in the prophylaxis for GVHD, such as selective cytotoxic T-cell depletion by using a post-transplant CY regimen, are promising strategies to fully suppress T cells as the GVHD enhancer. Previous studies revealed the clinical efficacy of GVHD prophylaxis but did not clarify the significance of its survival benefit. Likewise, our present findings indicated a lack of survival benefit by ATG treatment in this small study. However, the low-dose ATG contributed to a reduction of severe GVHD. Although early mortality after transplantation is decreasing, late-onset comorbidity including chronic GVHD remains a significant problem. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 4284 While adoptive T cell therapy is a promising treatment modality for cancer, the optimal approach to generate T cell grafts ex vivo is currently unknown. CD4+ T cells help generate effective immune responses by sustaining CD8+ T cell proliferation, preventing exhaustion, and establishing long-lived functional memory. Incorporation of CD4+ T cell help to expand CD8+ T cells may provide a novel strategy to generate CTL grafts for adoptive therapy. In mouse models, common γ-chain receptor cytokines and CD40/CD40L can mediate CD4+ T cell help. However, CD4+ T cell help in humans has yet to be fully defined. We therefore developed an in vitro model for human CD4+ T cell help, which utilizes a novel artificial APC, aAPC/mOKT3. K562-based aAPC/mOKT3 expresses a membranous form of anti-CD3 mAb, CD54, CD58, CD80, and CD83 and stimulates CD3+ T cells regardless of HLA haplotype or antigen specificity. Using aAPC/mOKT3, we stimulated CD8+ T cells in the presence or absence of CD4+ T cells and found that CD8+ T cells expanded better when coincubated with CD4+ T cells, suggesting the presence of CD4+ T cell help. Coculture experiments using transwell plates suggested that the observed CD4+ T cell help of CD8+ T cell expansion involved both soluble factors and cell-cell contact. To identify molecules mediating the observed CD4+ T cell help, supernatants of CD4+/CD8+ T cell mixed and separate cultures were measured for a panel of soluble factors. IL-2 and IL-21 were detected at lower levels in mixed cultures, consistent with more consumption or less production of these cytokines. Blockade of either IL-2 or IL-21 in CD4+/CD8+ T cell mixed cultures resulted in a reduction of CD8+ T cell expansion, indicating that, for both cytokines, more consumption rather than less production occurred and that IL-2 and IL-21 may serve as mediators of CD4+ T cell help. However, the addition of IL-21 to CD8+ T cells stimulated with aAPC/mOKT3 in the presence of IL-2 did not improve CD8+ T cell expansion, suggesting that IL-2 plus IL-21 cannot solely replace CD4+ T cell help. We found that the presence of CD4+ T cells upregulated the expression of IL-21R on CD8+ T cells. When we introduced IL-21R on CD8+ T cells and stimulated with aAPC/mOKT3 in the presence of IL-2 and IL-21, CD8+ T cell proliferation was restored. These results suggest that CD4+ T cells help CD8+ T cells proliferate ex vivo by secreting both IL-2/IL-21 and upregulating IL-21R. When peripheral CD3+ T cells from normal donors were stimulated with aAPC/mOKT3, the number of both CD4+ and CD8+ T cells increased. However, in contrast to other pan T cell expansion systems, aAPC/mOKT3 preferentially expanded CD8+ T cells. No obvious skewing in the Vβ usage of both CD4+ and CD8+ T cell populations was revealed by TCR Vβ repertoire analysis, supporting “unbiased” T cell expansion by aAPC/mOKT3. Moreover, HLA-restricted antigen-specific CD8+ CTL with high functional avidity could be generated from CD3+ T cells initially expanded for 4 weeks using aAPC/mOKT3. Using aAPC/mOKT3, tumor-infiltrating lymphocytes (TIL) were successfully expanded without adding soluble mAb or allogeneic feeder cells. As in peripheral T cell cultures, CD8+ T cells predominantly expanded in all cultures, including those that initially contained a minimal percentage of CD8+ T cells. Importantly, Foxp3+ Treg cells did not proliferate. Expanded T cells highly expressed CD27 and CD28, which are associated with T cell survival and persistence in vivo. They also secreted high levels of IFN-γ and IL-2, lower amounts of IL-4, and no IL-10. These results demonstrate that the aAPC/mOKT3-based system can expand functional CD8+ TIL in the presence of autologous CD4+ T cells. In conclusion, we have determined that CD4+ T cell-dependent CD8+ T cell expansion required both soluble factors secreted by and cell contact with CD4+ T cells. Among the soluble factors secreted by CD4+ T cells, IL-2 and IL-21 were necessary. Furthermore, upregulation of IL-21R on CD8+ T cells by CD4+ T cells was critical for an optimized response to IL-21. Thus, in humans, CD4+ T cells help CD8+ T cells proliferate by secreting IL-2/IL-21 and upregulating IL-21R. Our aAPC enabled expansion of CD8+ TIL in the presence of CD4+ T cell help without using soluble mAb or allogeneic feeder cells. Taken together, these results demonstrate the indispensable role of CD4+ T cell help on expanding CD8+ T cells and suggest a novel strategy to generate anti-tumor T cells ex vivo for adoptive therapy. Disclosures: No relevant conflicts of interest to declare.