Search

Relapse of leukemic cells that do not express the antigen targeted by chimeric antigen receptor (CAR) is still a risk. As is the potential for targeting hematopoietic stem cells (HSCs) that share the same antigen expression, off-tumor on-target toxicity. Further, CAR T cells that bind different epitopes of the same antigen can have different tumor-killing efficacies. Therefore, we screen murine single chain variable fragment (scFv) based for indirect affinity to identify a CAR that targets Acute myeloid leukemia (AML), while minimizing toxicities. Also, recent advances in CAR design have demonstrated that the requirement of two separate tumor antigens to be ligated by CARs can increase the specificity for tumor targets. So designing a CAR that only activates a T cell when it binds two separate AML antigens will allow T cells to enhance safety. Therefore, we set out to develop a affinity based multi-antigen CAR T cell therapy that targets well described antigens for AML, including CD33 and CD123. Mice were immunized with these antigens, spleens collected, and fused with myeloma cell lines. The antibodies of fused hybridomas were screened for binding and activation against antigens by high throughput flow cytometry. After screening, we derived multiple de novo CD33, and CD123 scFvs by sequencing. We incorporated CD33 and CD123 scFvs into standard mono-specific CARs utilizing a 41BB co-stimulatory domain to validate antigen-specificity. Gene transfer assessment of CAR T cells demonstrated about 50-80% transduction efficiency for CD33 and CD123 scFvs. There were no differences in CD4 and CD8 proportions in these CAR T cells. We next examined the CARs for their cytotoxic ability using a Real-Time Cell Analysis (RTCA) system. For the CD33 CARs, 2 (6A11-1 and 27A3-1) out of 5, and for the CD123 CAR, 2 (15A12-11 and 15 A12-12) out of 8, scFv sequences transduced into T cells were highly efficacious at killing target cells and generated significant amounts of cytokines such as IFN-g, TNF- a, and IL-6. CAR T cells with these same scFv sequences were able to proliferate better in response to targeted antigen. To find the best possible combination of CD33 and CD123 scFvs we double transduced T cells with four selected CD33 or CD123 scFvs each with only one co-stimulation domain either CD3z or 4-1BB in "AND" gate fashion. Clear differences in cytotoxic ability and cytokine production were observed. We selected 12 combination of CD33/123 CARs bi-specific CARs to evaluate in vitro efficacy, polyfunctionality, and safety. Finally, to find the best combination of CARs that would be less toxic to HSCs, we performed a Colony Forming Unit (CFU) assay with CD34+ bone marrow stem cells and found 5 bi-specific pairs that were less toxic to HSCs . Based on the CFU assay and PSI index, we were able to select the combination of CD33/123 scFvs that would target AML but minimize the killing of HSCs.

Introduction Acute myeloid leukemia (AML) is a highly aggressive form of leukemia with poor long-term survival. Our clinical development of a targeted radiation treatment for relapsed/refractory AML disease led to an impressive two-year overall survival (OS) rate of 41% (Stein et al., 2019) in contrast to Methods We first developed a TMI treatment methodology using a Precision X-RAD small animal irradiator. A whole body CT image was acquired, and three-dimensional (3D) dose calculations were performed using a Monte Carlo dose engine-based SmART-Plan treatment system (van Hoof et al, 2013; Downes et al., 2009; Faddegon et al., 2009). Before treatment, mice were further imaged to verify target anatomy prior to delivering precise radiation doses to the entire skeletal system and spleen while sparing vital organs (lungs, liver, gut). We used a Cbfb-MYH11/Mpl (CMM)-induced mice leukemia model. At 5-10% CMM-GFP cells in peripheral blood (~20-30% in bone marrow; high disease burden), the mice were treated with three doses of CSI (2.5mg/kg) on alternate days with or without 4 Gy radiation (TMI or total body irradiation [TBI]). Cy3 labeled CSI uptake studies were carried out 48h post RT by flow cytometry, and immune cell trafficking and activation studies were conducted by harvesting bone marrow and spleen cells on day 8 post RT. Results Dose volume histogram (DVH) and radiation dose painting show that the TMI treatment plan significantly reduced doses to critical organs (lungs, liver, gut) while maintianing the same dose to the skeleton and spleen, unlike TBI (same dose to entire body) (Figure 1A-D). In TMI-treated mice, the CSI-Cy3 uptake was significantly higher in CMM cells than in the CMM cells of mice treated with CSI alone, indicating improved delivery post RT (Figure 1E, F). The trafficking of T helper cells (CD4+) and cytotoxic T lymphocytes (CD8+; CTLs) as well as effector (CD62L-CD44+) and effector memory T cells (CD62L+CD44+) is significantly augmented in the bone marrow of TMI+CSI treated mice compared to levels in mice treated with TBI+CSI or CSI alone (Figure 1G, H). An increased total number of IFNγ secreting CTLs, and a higher CD8:Treg (CD4+CD25+FOXP3+) ratio indicate enhanced anti-tumerogenic activity in TMI+CSI treated mice over that of mice treated with TBI+CSI or CSI alone (Figure 1I, J). Similarly, myeloid cell trafficking and activation was augmented in TMI-treated bone marrow (data not shown). The benefit of augmented immune modulation in TMI+CSI combinatorial therapy is reflected in a significantly increased survival (~39 median survival days) over untreated mice and those treated with CSI or TMI alone ( ~8-9 days median survival) (Figure 1K). Conclusion This is the first report of a novel radio-immunotherapy using a systemic targeted precise RT (TMI) in combination with STAT3 down-regulation (CSI) in AML. As hypothesized, low-dose targeted RT in combination with blocking of STAT3 signaling improved immune cell trafficking and activation, thereby enhancing the efficacy of this combinatorial therapy in high-disease burden. Further, newly developed low-dose TMI shows enhanced immune modulation over conventional TBI, suggesting the benefits of localized targeted RT in hematological malignancies. Disclosures Hui: Janssen Research & Development, LLC: Consultancy, Honoraria.

CD19 targeted 2nd generation chimeric antigen receptor T (CAR T) cells have been successful against relapsed and/or refractory B cell malignancies. The pending FDA-approval of 2 separate CD19 targeted CAR T products highlight the need to understand the biology behind this novel therapy. CAR design includes a single-chain variable fragment, which encodes antigen-binding, fused to a transmembrane domain, co-stimulatory domain, and CD3ζ activation domain. The two CARs likely to be approved as standard of care include a 41BB or CD28 co-stimulatory domain. CD28 is a critical co-stimulatory receptor required for full T cell activation and persistence, while 4-1BB is a member of the tumor necrosis factor receptor family and also a critical T cell co-stimulatory factor. Early evaluation of the co-stimulatory domains role in CAR design confirmed that they are required to enhance T cell function, but lacked insight regarding their mechanism for this enhancement. Furthermore, clinical outcomes suggest that the co-stimulatory domains in CARs support different T cell functions in patients. For example, while overall outcomes are similar between 41BB (19BBz) and CD28-containing CARs (1928z), 19BBz CAR T cells can persist for years in patients, but functional 1928z CAR T cells rarely persist longer than a month. Recent studies are providing insight to these differences and have demonstrated that 4-1BB-containing CARs reduce T cell exhaustion, enhance persistence, and increase central memory differentiation and mitochondrial biogenesis, while CD28-containing CARs support robust T cell activation and exhaustion, and are associated with effector-like differentiation. However, these studies have been performed mostly in vitro or in immune deficient mice, which limits their ability to model complex immune biology. Therefore, we evaluated murine CD19-targeting CARs with a 4-1BB (m19BBz) or CD28- (m1928z) co-stimulatory domain in relevant animal models of immunity. We directly compared m19BBz and m1928z CAR T cell immune phenotype, cytotoxicity, cytokine production, gene expression, intracellular signaling, and in vivo persistence, expansion, and B cell acute lymphoblastic leukemia (B-ALL) eradication. In vitro assays revealed that m1928z CAR T cells had enhanced cytotoxicity and cytokine production compared to m19BBz CAR T cells. Also, evaluation of m1928z and m19BBz CAR T cells displayed similar immune phenotypes, but markedly different gene expression with m1928z CAR T cells upregulating genes related to effector function and exhaustion, while m19BBz CAR upregulated genes critical for NFkB regulation, T cell quiescence and memory. In vivo, both m1928z and m19BBz CAR T cells supported equivalent protection against B-ALL. Similar to patients, in our mouse models there are functional differences between the mouse CD19-targeted CAR T cells. At 1 week post-infusion m19BBz CAR T cells are present in the blood of mice at significantly greater levels than m1928z CAR T cells. Furthermore, m19BBz CAR T cells enhance proliferation and/or anti-apoptosis protein expression to enhance B cell killing, which is evidenced by our observation that irradiation significantly weakens the in vivo efficacy of m19BBz but not m1928z CAR T cells. Our results suggest that B cell killing by m1928z CAR T cells is not impacted by irradiation because of their efficacious cytotoxicity of B cells. In contrast, m19BBz CAR T cells have enhanced viability and anti-apoptosis protein expression, which allows them to compensate for reduced effector function. We investigated potential mechanisms for the enhanced viability and anti-apoptosis of m19BBz CAR T cells and determined that NFkB signaling is upregulated much greater by m19BBz than m1928z. We have observed this difference in both a reporter cell line and primary mouse T cells. We are now dissecting what cellular components mediate increased NFkB signaling by the m19BBz CAR. Our animal models recapitulate equivalent anti-leukemia efficacy of CD19-targeted CAR T cells regardless of co-stimulatory domain, but underscore that anti-leukemia killing is mediated by different methods depending on the co-stimulatory domain. Our work sheds light on how 4-1BB mechanistically regulates and impacts CAR T function and has implications for future CAR design and evaluation. Disclosures No relevant conflicts of interest to declare.

Mixed hematopoietic chimerism permits durable tolerance of T, B and NK cells to xenoantigens in a rat to mouse bone marrow transplant model. However, it is unclear whether tolerance of human NK cells to pig xenoantigens can be induced by mixed hematopoietic chimerism. We assessed the tolerance of human NK cells towards pig cells in a humanized mouse model with established pig and human mixed xenogeneic chimerism. Pig and human mixed chimeras (MCs) were generated by injection of pig bone marrow cells to irradiated pig cytokine (IL3, GMCSF and SCF) transgenic NOD-scid common gamma chain knockout (NSG) mice followed by injection of human fetal liver CD34+ cells 3 day later. In the control group, only human CD34+ cells were transplanted. 12 weeks post-transplant, hydrodynamic injection of plasmid encoding human Flt3L followed by injection of three rounds of recombinant IL15/IL-15 receptor alpha Fc complex was given to promote human NK cell reconstitution. The control non–mixed chimeric group (Non-MC) received the same treatment without pig cells. 12 days following induction of human NK cell reconstitution, human NK cells from both MC and Non-MC mice were isolated from the spleen and their cytotoxic responses in vitro to pig cells were determined in a chromium release assay. In addition, the presence of pig cells in various tissues of the chimeric mice was studied. While human NK cells were usually undetectable in peripheral blood prior to the injection of human Flt3L plasmid and IL-15/IL-15 receptor alpha-Fc complex, they were detected by 5 days post-injection of IL-15 protein. 12 days post-induction of human NK cell reconstitution, pig cells remained detectable in peripheral blood, spleen, bone marrow and liver in the chimeric mice together with human NK cells. The co-existence of human NK cells and pig cells suggested that human NK cells in MCs might be tolerant to pig cells. Consistent with this notion, cytotoxicity assays showed that human NK cells from MCs demonstrated decreased killing of pig PBMC blasts compared to NK cells from Non-MC mice. Importantly, killing of K562 cells by NK cells from MCs mice was also decreased compared to that of Non-MC mice, suggesting that human NK cell tolerance to pig cells induced by mixed chimerism was associated with global hyporesponsiveness, as we have previously observed in a rat-to-mouse bone marrow transplantation model. Moreover, higher percentages of CD56highCD16low and lower percentages of CD56lowCD16high human NK cell subsets were observed in bone marrow of chimeric mice than in non-chimeric mice, indicating that the development of human NK cells in bone marrow might be altered by the presence of pig cells. In summary, our data suggest that mixed xenogeneic chimerism may induce tolerance of human NK cells towards porcine cells, but the tolerance may be associated with global hyporesponsiveness. Disclosures No relevant conflicts of interest to declare.