ALBERT

Description: Background: T-cell therapy with CD19-CAR T cells has been very successful for the treatment of B-cell derived malignancies in humans. However, cytokine release syndrome, neurotoxicity, and development of CD19- escape variants have emerged as potential limitations. Developing CAR NK-cell based therapies might overcome some of these side effects since NK cells do not rapidly expand or persist long-term after adoptive transfer. However, CAR NK-cell therapies are less effective than CAR T-cell therapies in preclinical models. To overcome these limitations we have devised a strategy to genetically modify NK cells with CD22-CARs and CD19/CD3-bispecific T-cell engager (CD19-ENG) molecules. These NK cells should not only kill CD22+ B cells directly, but also redirect bystander T cells to malignant CD19+ B cells, enhancing antitumor effects and preventing immune escape. Methods: NK cells were generated using K562s expressing 41BBL and membrane bound IL15, and genetically modified with a retroviral vector encoding a CD22-CAR with a 41BB.ζ endodomain and/or a retroviral vector encoding CD19-ENG and mOrange separated by an IRES. To mimic immune escape, CD19 or CD22 knockout (ko) Ph+ leukemia cells (BV173) were generated by CRISPR/cas9. The effector function of genetically modified NK cells was evaluated using standard immunological assays. Results: After transduction 70-80% of NK cells expressed CD22-CARs, and ~50% expressed CD22-CARs and CD19-ENGs as judged by FACS analysis. We performed coculture and cytotoxicity assays using non-transduced (NT), CD22-CAR, CD19-ENG, and CD22-CAR/CD19-ENG NK cells as effectors and BV173 (CD19+/CD22+), BV173.koCD19, BV173.koCD22, Daudi (CD19+/CD22+), and KG1a (CD19-,CD22-) as targets. Cocultures were preformed +/- T cells and after 24 hours IFNγ and IL2 was determined by ELISA. In the absence of T cells, CD22-CAR and CD22-CAR/CD19-ENG NK cells only recognized CD22+ targets as judged by IFNγ production. Moreover, CD22-CAR/CD19-ENG and CD19-ENG NK cells efficiently redirected T cells to secrete IFNγ in the presence of CD19+/CD22- targets. No NK-cell population produced IL2, however CD22-CAR/CD19-ENG and CD19-ENG NK cells induced IL2 production of T cells in the presence of CD19+ targets. No significant cytokine production was observed in the absence of antigen (media, KG1a). Specificity of generated NK cells was confirmed in cytotoxicity assays. In vivo studies to confirm our in vitro findings are in progress. Conclusions: We have generated for the first time NK cells that kill B-cell malignancies through a CAR (CD22) and simultaneously redirect bystander T cells to a 2nd B-cell antigen (CD19) to enhance antitumor effects and prevent immune escape. Genetic modification of NK cells to enhance their antitumor activity and redirect bystander T cells may present a promising addition to current cell therapies for B-cell malignancies. Disclosures No relevant conflicts of interest to declare.

Description: There is a need to improve outcomes for patients with recurrent and/or refractory hematological malignancies. Immunotherapy holds the promise to meet this need, because it does not rely on the cytotoxic mechanism of conventional therapies. Among different forms of immunotherapy, redirecting T cells to hematological malignancies with bispecific antibodies (BsAbs) is an attractive strategy. BsAbs are an “off-the-shelf” product that is easily scalable in contrast to adoptive T-cell therapies. Among these, the bispecific T-cell engager blinatumomab has emerged as the most successful BsAb to date. It consists of 2 single-chain variable fragments specific for CD19 present on B-cell malignancies and CD3 expressed on almost all T cells. Blinatumomab has shown potent antitumor activity as a single agent, particularly for acute lymphoblastic leukemia, resulting in its US Food and Drug Administration approval. However, although successful in inducing remissions, these are normally short-lived, with median response durations of

Description: Background: CD123 is a potential immunotherapeutic target in AML due to its overexpression on leukemic stem cells that play an essential role in disease progression and relapse. Our previous study using T cells secreting CD123/CD3 bispecific engager molecules (CD123-ENG T cells) showed promising results in pediatric AML. Interleukin-15 (IL15) has emerged as a candidate immunomodulator as it enhances the cytolytic activity of CD8+ T-cells and induces long-lasting memory T cells. To improve the efficacy and persistence of CD123-ENG T-cells we developed IL-15 expressing CD123-ENG T cells. Here, we report characterization and efficacy of IL15 secreting CD123-ENG T cells in adult AML. Methods/Results: A cDNA encoding IL15 was cloned into retroviral vectors encoding CD123-ENG or CD19-ENG and the CD20 suicide gene separated by 2A sequences (CD20.2A.CD123-ENG.2A.IL15; CD20.2A.CD19-ENG.2A.IL15). ENG T cells were generated from human peripheral blood mononuclear cells (PBMCs) from normal donors by retroviral transduction and expanded in vitro. Non-transduced (NT) T cells and T cells expressing CD20 and CD123 (CD20.CD123-ENG T cells) served as controls. The transduction efficiency was between 62.81-95% (average 72%, n=3) and phenotypic analysis by flow cytometry showed reproducible CD4+, CD8+, central memory (CCR7+CD45RA-), effector memory (CCR7-CD45RA-), and naïve (CCR7+CD45RA+) T cells populations compared to NT cells. IL15 production of CD20.CD19-ENG.IL15 and CD20.CD123-ENG.IL15 T cells was confirmed by ELISA (84-154 pg/ml vs 32 and 44 pg/ml of NT and CD20.CD123-ENG T cells, p

Description: Background: CD123 is frequently expressed on hematologic malignancies including 96-98% of AML. CD123 has been a potential immunotherapeutic target in AML due to its association with leukemic stem cells that play an essential role in disease progression and relapse. Our previous study using T-cells secreting CD123/CD3-bispecific T-cell engagers (BiTEs) (CD123-ENG T-cells) showed a promising approach anti-AML activity, however T-cell persistence was limited. Interleukin-15 (IL15) has emerged as a candidate immunomodulator as it enhances T-cell expansion and persistence, and induces long-lasting memory T-cells. To improve the efficacy and persistence of CD123-ENG T-cells we developed IL15 expressing CD123-ENG T-cells. Here, we report on the characterization and efficacy of IL15-secreting CD123-ENG T cells in vitro and in vivo models of adult AML. Methods/Results: A cDNA encoding IL15 was cloned into retroviral vectors encoding CD123-ENG or CD19-ENG (CD123-ENG.IL15; CD19-ENG.IL15). ENG T-cells were generated from human peripheral blood mononuclear cells (PBMCs) from normal donors or T-cells from AML patients by retroviral transduction and in vitro expansion. Non-transduced (NT) T-cells and T-cells expressing CD123 (CD123-ENG T-cells) served as controls. IL15 production of CD19-ENG.IL15 and CD123-ENG.IL15 T cells was confirmed by ELISA (144-159 pg/ml vs 38 and 46 pg/ml of NT and CD123-ENG T cells, p

Description: Background: Pediatric and adult acute myeloid Leukemia (AML) is a disease with poor prognosis due to its high relapse rate and treatment related mortality. Adoptive immunotherapy has the potential to improve outcomes and CD123 present a promising immunotherapy target for AML. CD123-specific engager (ENG) T-cells are cells genetically modified to secrete bispecific T-cell engagers (BiTEs) that recognize CD3+ on T-cells and CD123+ on AML. We had previously shown that inclusion of an inducible MyD88/CD40 costimulatory molecule, which is activated by a chemical inducer of dimerization (CID), improved anti-tumor effect against CD123+ targets in vitro. We now wanted to determine, which component of the costimulatory molecule is critical for the improved effector function of CD123-ENG T cells and perform in vivo studies. Methods: We generated a panel of retroviral vectors encoding CD20 as a safety switch, CD123-ENG, and an inducible costimulatory molecule consisting of a myristoylation sequence, FKBP dimerizer domains and one of three signaling costimulatory domains (MyD88 (iM) ,CD40(iC) or MyD88/CD40(iMC)). A vector specific for CD19 (CD20.CD19-ENG.iMC) and non-transduced (NT) T-cells served as controls. We genetically modified T-cells using a retroviral transduction protocol and evaluated their effector function +/- CID. We assessed the effector function of transduced T-cells using flow cytometry, ELISA and luciferase based cytotoxicity assays. We evaluated antitumor activity and persistence in vivo in a xenograft AML model. Results: We successfully generated CD20.CD123-ENG.iM, CD20.CD123-ENG.iC and CD20.CD123-ENG.iMC T-cells. Transduction efficiency ranged from 45 to 70% (±5). The immunophenotype of CD123-ENG T-cells was predominandly CD8+ effector memory. CD20.CD123-ENG T-cells expressing inducible costimulatory molecules secreted higher levels of CD123-specific BiTEs, IL2 and IFNγ in comparison to CD20.CD123-ENG T-cells when cocultured with of MOLM-13 (CD123+) and CID (p

Description: Background: Acute myeloid leukemia (AML) represents a treatment challenge due to its poor prognosis and high associated morbidity and mortality rates. Thus, new therapies are needed. CD123-ENG T-cells are cells genetically modified to secrete bispecific antibodies that recognize CD3 (T-cells) and CD123 (AML). Preclinical studies show that these cells recruit bystander T-cells to kill CD123+ blasts in vitro and in vivo. However, CD123-ENGs are unable to maintain sequential killing capability of CD123+ targets. To overcome this limitation, we have devised an approach that provides inducible Myd88/CD40 costimulation activated by a chemical inducer of dimerization (CID). This strategy should increase the persistence, expansion and anti-AML activity of CD123-ENG T-cells. Methods: We generated a retroviral vector encoding a CD20 safety switch, CD123-ENG, and the inducible costimulatory molecule MyD88.CD40 (iMC) linked by 2A sequences (CD20.2A.CD123-ENG.2A.iMC). We used a vector specific for CD19 (CD20.2A.CD19-ENG.2A.iMC) and non-transduced (NT) cells as controls for non-specific effects of the iMC construct. We genetically modified T-cells using a retroviral transduction protocol and evaluated their effector function +/- CID. We assessed the effector function of transduced T-cells using standard immunological assays and a flow cytometry based sequential killing assay. Results: We successfully generated CD20.CD123-ENG.iMC T-cells, which maintained a transduction efficiency above 50% throughout our study period. We performed coculture and cytotoxicity assays using NT, CD20.CD123-ENG, CD20.CD19-ENG.iMC and CD20.CD123-ENG.iMC T-cells +/-CID as effectors and MOLM13 (CD123+), Kg1a (CD123+) and K562 (CD123-) as targets. Cocultures were performed +/- CID. CD20.CD123-ENG.iMC T-cells maintained CD123 antigen specificity, as evidenced by cytotoxicity and cytokine assays. CD20.CD123-ENG.iMC T-cells + CID secreted increased IL-2 and IFN-γ in the presence of CD123+ targets (KG1a and MOLM13) when compared to baseline and to CD20.CD123-ENG T-cells. In addition, CD20.CD123-ENG.iMC T-cells + CID displayed enhanced sequential killing capabilities at a 1:1 ratio, compared to CD20.CD123-ENG T-cells. Conclusion: CD20.CD123-ENG.iMC T-cells recognize and kill CD123+ AML blasts in an antigen dependent manner. In addition, control ENG.iMC T-cells have no antitumor activity, indicating that activation of Myd88/CD40 does not induce nonspecific AML blast killing. CD20.CD123-ENG.iMC T-cells have improved effector function in the presence of CID as judged by increased cytokine production and their ability to sequentially kill CD123+ target cells in vitro. Based on our promising in vitro experiments, we have initiated in vivo studies in AML xenograft models, which are currently in progress. Disclosures Leen: Marker: Equity Ownership.

Description: Background: The outcome for patients with high risk acute myeloid leukemia (AML) remains poor. Thus new targeted therapies are needed and immunotherapies have the potential to fulfill this need. Adoptive transfer of tumor-specific T cells is one promising approach; however infused T cells do not redirect the large reservoir of resident T cells to tumors. To overcome this limitation we have recently developed a new approach to render T cells specific for tumor cells, which relies on genetically modifying T cells with a secretable, bispecific T cell engager (ENG-T cells). Secretion of bispecific protein should activate infused cells as well as bystander T cells against tumor. Consistent and prolonged synthesis of engagers by T cells should be superior to the intermittent, direct infusion of the recombinant bispecific antibody, not only because these recombinant proteins have short half-lives but also because they do not accumulate at tumor sites. The goal of this project was to generate T cells secreting IL3Rα (CD123) and CD3 bispecific T cell engagers (CD123-ENG T cells) and to evaluate their effector function in vitro and in vivo. Methods: CD123-ENG T cells were generated by transducing T cells with a retroviral vector encoding a CD123-specific T cell engager consisting of an scFv recognizing CD123 linked to an scFv recognizing CD3. The retroviral vector was also fashioned to include an mOrange gene downstream of an IRES element. The effector function of CD123-ENG T cells was evaluated in vitro and in a xenograft model. Results: Transduction of CD3/CD28-activated T cells resulted in mOrange expression in transduced T cells (median transduction efficiency 78%, range 49-92%) The presence of CD123-ENG molecules on the cell surface of both transduced and non-transduced T cells was demonstrated by FACS analysis using an F(ab) antibody that recognizes the CD123 scFv. Coculture of CD123+ AML cells (MV-4-11, MOLM-1, KG1a) and K562 cells genetically modified to express CD123 (K562-CD123) with engager T cells resulted in robust T-cell activation as judged by IFNγ and IL2 secretion. In contrast CD123-negative cells (K562) did not activate T cells. Likewise, control engager T cells (targeting an irrelevant antigen) were not activated when cultured with CD123+ cells. Antigen-dependent recognition was confirmed with cytotoxicity assays, in which engager T cells specifically killed CD123+ AML cells at an effector:target ratios ranging from 40:1-5:1 (p

Description: Background: CD123 is frequently expressed in hematologic malignancies including AML. CD123 has been a potential immunotherapeutic target in AML due to its association with leukemic stem cells that play an essential role in disease progression and relapse. Our previous study using T-cells secreting CD123/CD3-bispecific T-cell engagers (BiTEs) (CD123-ENG T-cells) has shown activity in preclinical studies, recognizing and killing acute myeloid leukemia (AML) blasts in vitro and in vivo. CD123-ENG T-cells secrete bispecific molecules that recognize CD3 (T-cells) and CD123 (AML blasts), and are able to direct transduced T-cells and recruit bystander T-cells to kill CD123-positive blasts. Venetoclax is a BCL-2 inhibitor that can restore functional apoptosis signaling in AML cells, and has been FDA approved for the treatment of AML patients in combination with hypomethylating agents. To improve the efficacy of CD123-ENG T-cells we explored efficacy in AML by combining targeted immunotherapy (CD123-ENG T cells) with targeted inhibition of anti-apoptotic BCL-2 (venetoclax) in vitro and in vivo models of AML. Methods : CD123-ENG T-cells were generated by retroviral transduction and in vitro expansion. Non-transduced (NT) T-cells served as control. In vitro, GFP+ MOLM-13 AML cells were pretreated with venetoclax (0, 10µM, and 20µM) for 24 hours prior to co-culture with CD123-ENG or NT T-cells at an effector/target ratio of 1:10. After 16 hours, MOLM-13 AML cells were analyzed by flow cytometry and quantitated using counting beads; cytotoxicity was calculated relative to untreated MOLM-13 control. The anti-AML activity of the combination was further evaluated in a MOLM-13-luciferase xenograft AML mouse model. Leukemia progression was assessed by bioluminescence imaging. The frequency of MOLM13 AML and human T cells in periphera blod (PB) was determined by flow cytometry. Results: In vitro, we demonstrated that pretreatment of Molm13 AML cells with venetoclax enhanced the cytolytic activity of CD123-ENG T-cells compared to NT- or no T-cell controls. Interestingly, venetoclax sensitized Molm13 to CD123-ENG T-cell killing in a dose-dependent manner (Fig.1; 50%/31% killing by CD123-ENG T-cells versus 27%/14% of killing by NT T cells post pretreatment with 10µM or 20µM ventoclax, p