Kim, Miriam Y.

Abstract Despite recent landmark advances in T-cell immunotherapy for the treatment of human cancer, metastatic solid tumors remain an intractable challenge. Macrophages are often the most abundant immune cell in the tumor microenvironment (TME), where they may convert into immunosuppressive (M2) tumor-associated macrophages (TAMs) and participate in disease progression. Currently, macrophage-orientated immunotherapeutic approaches under clinical development in oncology seek to reduce TAM infiltration (CSF-1 antagonists) or enhance TAM phagocytosis (CD47 antagonists). Transfer of autologous, activated, but nontargeted macrophages failed to demonstrate antitumor efficacy in past clinical trials. We hypothesized that genetically engineering human macrophages with CARs against tumor-associated antigens could redirect their phagocytic activity and lead to therapeutic efficacy with the potential for the induction of an antitumor T-cell response. We first demonstrate that CD3-zeta-based CARs are capable of inducing phagocytosis in human THP-1 macrophages, while truncated intracellular-domain deficient CARs were not. Targeted phagocytosis and clearance of CD19+, mesothelin +, and HER2+ cells by CARs targeted against each respective antigen was significantly superior to that by control untransduced (UTD) macrophages. We demonstrate that primary human macrophages, which are resistant to most viral vectors, are efficiently transduced by the chimeric fiber adenoviral vector Ad5f35 (~70% in 10 normal donors). Using Ad5f35, we engineered primary human macrophages with a CD3-zeta-based CAR against HER2. Anti-HER2 primary human CAR macrophages demonstrated targeted phagocytosis against HER2+ but not HER2- cell lines, with phagocytic activity dependent on both the CAR and antigen densities. Furthermore, CAR, but not UTD, macrophages led to potent dose-dependent killing of three distinct HER2-high cell lines in vitro. We sought to test the efficacy of anti-HER2 primary human macrophages in xenograft models of human HER2+ ovarian cancer. A single dose of CAR, but not UTD macrophages, led to tumor regression and improved overall survival in both intraperitoneal and disseminated models of disease. We show that macrophage transduction with Ad5f35, a double-stranded DNA virus, leads to a broad gene expression change, an interferon signaling signature, and phenotypic clustering toward classically activated M1 macrophages. CAR macrophages upregulated co-stimulatory ligand and antigen processing/presentation genes and led to enhanced T-cell stimulation in vitro and in vivo. Lastly, CAR, but not UTD, macrophages showed a broad resistance for M2 conversion in response to immunosuppressive cytokines. In conclusion, we show that primary human CAR macrophages are capable of targeted tumor phagocytosis, lead to improved overall survival in xenograft models, and demonstrate enhanced T-cell stimulation. CAR macrophages are a novel cell therapy platform for the treatment of human cancer. This abstract is also being presented as Poster B29. Citation Format: Michael Klichinsky, Marco Ruella, Olga Shestova, Andrew Best, Kristin Blouch, Xueqing M. Lu, Saad S. Kenderian, Miriam Y. Kim, Roddy O'Connor, Stephen Wallace, Miroslaw Kozlowski, Dylan M. Marchione, Maksim Shestov, Benjamin A. Garcia, Carl June, Saar Gill. Human chimeric antigen receptor (CAR) macrophages for cancer immunotherapy [abstract]. In: Proceedings of the AACR Special Conference on Tumor Immunology and Immunotherapy; 2018 Nov 27-30; Miami Beach, FL. Philadelphia (PA): AACR; Cancer Immunol Res 2020;8(4 Suppl):Abstract nr PR07.

Abstract Acute myeloid leukemia (AML) is challenging to treat with antigen-specific immunotherapy due to the lack of leukemia-specific surface antigens that are absent from normal hematopoiesis. We propose a novel solution to this problem by removing CD33, an AML-associated antigen, from normal hematopoietic stem and progenitor cells (HSPCs) by gene editing, thus rendering them resistant to CD33-targeted therapy. This enables powerful CD33-directed immunotherapy such as chimeric antigen receptor T cells (CAR T cells) to be employed against AML without disrupting normal myeloid function. We developed an optimized protocol to generate CD33 KO HSPCs with CRISPR/Cas9 and obtained up to 90% CD33 KO in primary human CD34+ cells as confirmed by flow cytometry and DNA sequencing. To test whether CD33 KO HSPCs retained multilineage potential, we engrafted NSG mice with CD34+ cells treated with Cas9 and either a CD33-targeting gRNA or a control gRNA. CD33 KO HSPCs were able to engraft and differentiate into both lymphoid and myeloid lineages in the mice to the same degree as controls. Human myeloid cells in the peripheral blood (PB) of the CD33 KO-engrafted mice were deficient for CD33, consistent with the KO efficiency measured in vitro, while expression of other myeloid markers such as CD11b and CD14 were similar to controls (Figure 1A, B). To ensure that long-term repopulating stem cells were present in the CD33 KO HSPCs, we performed secondary transplants of bone marrow from mice after four months of primary engraftment. Persistent engraftment for up to three additional months was observed in both CD33 KO HSPCs and controls, proving that gene editing with CRISPR/Cas9 and the resulting CD33 KO does not affect the self-renewal capability of human hematopoietic stem cells. To determine whether CD33 loss would lead to any functional defects in the myeloid progeny, we investigated the properties of myeloid cells derived from CD33 KO HSPCs differentiated with SCF, GM-CSF, G-CSF, IL-3, IL-6 and EPO. Cell morphology and immunophenotype of the myeloid cells after in vitro differentiation were consistent with normal neutrophils and macrophages. Phagocytosis ability, as measured by uptake of E. coli particles, was identical in CD33 KO myeloid cells as compared to controls. CD33 KO myeloid cells were also able to secrete inflammatory cytokines (IL-1b, IL-8, IL-8, IL-10, IL-12, TNF) in response to LPS stimulation to the same degree as controls. RNA-seq showed a highly concordant gene expression profile of CD33 KO HSPCs and controls (R2=0.99), thus excluding any major impact of CD33 loss on downstream gene expression. Prior studies have shown that CAR T cells targeting CD33 (CART33) can eliminate AML in mouse models but also cause toxicity to normal HSPCs (Kenderian et al, Leukemia 2015). To test our hypothesis that CD33 KO HSPCs can evade CART33-mediated toxicity, we treated mice engrafted with either control or CD33 KO HSPCs with CART33. CD33-expressing cells were eliminated in both control and CD33 KO-engrafted mice, which led to the expected myeloablation in control mice; in contrast CD33 KO-engrafted mice continued to sustain human myeloid cells (Figure 2A). Bone marrow (BM) analysis showed significantly more human stem cells (CD34+38-) and myeloid progenitors (CD34+38+) in CD33 KO-engrafted mice compared to controls after CART33 treatment (Figure 2B). We then proceeded to engraft Molm14, an AML cell line that was engineered to express luciferase, into mice bearing control or CD33 KO hematopoiesis, and studied the effect of CART33 treatment in this model. As expected, AML was eliminated in both groups as measured by bioluminescent imaging (BLI) (Figure 2C), as were peripheral blood monocytes in the control group, while monocytes were still present in the CD33 KO-engrafted mice (Figure 2D). In conclusion, we show that CD33 is dispensable in hematopoietic differentiation, and that absence of CD33 from myeloid progeny does not cause any discernible functional changes. Gene editing to remove CD33 from normal HSPCs allows persistent myelopoiesis during CART33-mediated eradication of AML, thus paving the way to a novel system that maximizes efficacy while minimizing toxicity. We envision future clinical translation of this approach by administering CD33 KO HSPCs as an allogeneic stem cell transplant in combination with CART33 in patients with AML. Disclosures Kenderian: Novartis: Patents & Royalties, Research Funding. Ruella:Novartis: Patents & Royalties. Gill:Novartis: Patents & Royalties, Research Funding.

Abstract Abstract 845 Background: Waldenstrom's macroglobulinemia (WM) is a relatively uncommon plasma cell disorder and most of previous literature is limited to small patient series. We have noted ethnic disparities in multiple myeloma, another plasma cell disorder. We undertook a large Surveillance Epidemiology and End Results (SEER) based analysis to describe outcome disparities in different subgroups of WM patients, with a focus on various ethnicities, so that therapeutic resources can be targeted more effectively. Methods: The SEER 17 Registry data (1973–2008) was utilized for patients with confirmed diagnosis of WM. To avoid bias of under representation of different ethnicities, analysis was restricted to patients with a diagnosis date of 1992 or later. Cases that received a diagnosis at death certificate or autopsy, no follow-up records, as well as lacking documentation on age at diagnosis, sex, or race/ethnicity were excluded. Cox proportional hazards models, adjusted for gender, age, race, year of diagnosis, marital status and stratified by SEER registries were used to evaluate associatio n between patient characteristics and survival. All statistical tests were two-sided and utilized the SAS software (v9.2) with a significance level of 0.05. Results: The final analysis included 2,840 WM patients (1,759 males; 62%, 1,081 females; 38%). The studied age-group cohorts included: 18–64 yrs (845; 30%), 65–74 yrs (795; 28%) and >75 yrs (1,200; 42%). Patients were stratified by race/ethnicity: White (2,471; 87%), African-American (AA) (102; 4%), Hispanic (133; 5%), Asian (129; 5%), and Native American (5; 0.1%). Patients were also stratified based on year of diagnosis (before or after 2002) to study the impact of certain novel agents (proteasome inhibitors, IMiDs) on WM treatment. There was a significant difference in the age at diagnosis of WM patients across different ethnicities, with AA the youngest (median 61.5 years) and Whites the oldest (median 73 years) subgroup (p<0.001). (Figure 1) Cause-specific mortality could not be evaluated due to variable documentation of reported causes of death for WM patients in the database. Survival analysis revealed that for all patients, females had better median overall survival (OS) than males (7.3 years vs. 6.2 years, HR 0.834; 95% CI 0.740, 0.939; p=0.003). Among the different age cohorts, patients with age ≥75 years had a significantly worse median OS (4.1 years) than those 65–74 year-old (7.3 years) or 18–64 year-old (10+ years) (p<0.001). Patients diagnosed after 2002 had a significantly better median OS (7.3 years) as compared to patients diagnosed previously (6.1 years) (p=0.002). Hispanics had the worst median OS (5 years) and Whites had the best median OS (6.8 years) across various ethnicities. In a multivariate model using gender, age, marital status, year of diagnosis and race, a significant interaction was noted between race, age and survival, with Hispanics having a significantly worse OS than Whites (HR 1.86; 95% CI 1.269, 2.726; p=0.003) (Figure 2). Conclusions: Studies of outcome disparities are important for evaluating disease characteristics and management needs of specific patient populations as well as optimal triaging of healthcare resources. We have performed the largest population-based analysis for WM including various ethnicities in the novel therapeutic agent era. We observed that AA patients had a significantly younger age at WM diagnosis. Older patients had worse OS across all ethnic subgroups and patients diagnosed after 2002 had a better OS suggesting an impact of novel therapeutic agents (e.g., proteasome inhibitors, IMiDs). Hispanic WM patients had the worst OS among all ethnicities. This survival difference was more significant and pronounced in patients >75 years. This suggests that the impact of race on survival is influenced by patient's age. These results will help in better understanding of various influences on disease biology and clinical behavior. Disclosures: No relevant conflicts of interest to declare.

Abstract Chimeric antigen receptor T cell (CART) therapy has developed as a powerful and potentially curative therapy in hematological malignancies over the last few years. CD19 directed CART cells have resulted in impressive complete response rates of ~90% in acute lymphoblastic leukemia that are durable for the majority of patients. However, the overall response rates in other malignancies such as chronic lymphocytic leukemia are around 50%. This could be partially related to CART exhaustion and dysfunction induced by leukemia cells. In this study, we aim to evaluate the role of inhibitory receptors/pathways in inducing CART cell dysfunction and exhaustion in hematological malignancies. As a tumor model, we used an acute myeloid leukemia (AML) cell line (MOLM14) and primary AML samples and treated them with CD33 or CD123 directed CART cells (second generation CARs using 4-1BB and CD3z signaling domains and a lentiviral vector). Incubation of primary AML samples or MOLM14 cell line with CD33 or CD123 directed CARTs resulted in a significant up-regulation of PD-L1 on tumor cells after 24 hours of incubation (0% on day 0 vs 80% on day 1, P <0.001), and up-regulation of PD-1 and TIM-3 on T cells 3-7 days post co-culture (8% of T cells expressed PD-1 on day 0 vs 43% on day 3, P=0.03 and 13% of T cells expressed TIM-3 on day 0 vs 71% on day 3, p=0.001, Figure 1). For in vivo experiments, we used NSG (NOD-SCID-g-/-) mice and engrafted them with the MOLM14 cell line. Treatment of these AML xenografts with suboptimal doses of CD33 or CD123 CARTs resulted in initial anti-tumor responses, followed by disease relapses in 40-60% of the mice (Figure 2A). T cells were isolated from the bone marrow of these mice and analyzed for differential expression of inhibitory receptors. There was a significantly increased up-regulation of TIM-3 receptors on T cells isolated from mice with relapsed disease compared with T cells isolated from mice in remission after CART cell therapy (Figure 2B). Next, we investigated the role of adding checkpoint blockade to improve T cell function ex vivo after CART cell therapy. Marrows of mice that relapsed after CART cell therapy contained both residual CART cells and leukemia and were used to model the administration of checkpoint blockade in the setting of CART cell exhaustion. Cells were cultured with PD-1 or TIM-3 blocking antibodies or the combination of both (10 ug/ml for 72 hours). CART cell effector functions such as cytokine production and Ki-67 proliferation marker improved in the presence of checkpoint antagonists especially when both PD-1 and TIM-3 blocking antibodies were combined (figure 2C). Finally, we tested the combination of checkpoint blockade with CARTs in AML xenografts. In this approach, we treated MOLM14 xenografts with suboptimal doses of CD33 or CD123 directed CARTs or with control untransduced T cells (UTD), with or without checkpoint blocking antibodies. NSG mice bearing MOLM14 AML xenografts were established. Engraftment was confirmed by bioluminescent imaging. The tumor bearing mice were then treated with suboptimal doses (0.25-0.5 x106 total T cells I.V) of CD33 or CD123 directed CARTs or with control untransduced T cells (UTD). Mice also received PD-1 blockade, TIM-3 blockade or the combination of both on days 3, 6, 9 and 12 post T cell therapy, with the rationale for early checkpoint blockade being based on our in vitro observations of early upregulation of inhibitory ligands on AML after exposure to CART cells. Mice were then followed with serial imaging to assess disease burden. The addition of checkpoint antagonists to untransduced T cells, in particular anti-TIM3, did not lead to an anti-leukemic effect. However, the addition of PD-1 or TIM-3 blockade to CART cell therapy resulted in a synergistic anti-tumor activity as shown in Figure 3. The durable complete response rate was: 45% for treatment with CART123 alone, 80% for treatment with CART123+PD-1 blockade, 100% for treatment with CART123+TIM-3 blockade, and 80% for treatment with CART123+ both PD-1 and TIM-3 blockade). Our preclinical results indicate that PD-1 and TIM-3 pathways are involved in CART exhaustion and dysfunction in AML. Combination of checkpoint inhibitors with CART cells may lead to enhanced efficacy in AML and other hematological malignancies. Current studies are investigating mechanisms of synergy and the role of these combinations in other hematological malignancies. Figure 1. Figure 1. Figure 2. Figure 2. Figure 3. Figure 3. Disclosures Kenderian: Novartis: Patents & Royalties, Research Funding. Ruella:Novartis: Patents & Royalties, Research Funding. Porter:Novartis: Patents & Royalties, Research Funding. June:University of Pennsylvania: Patents & Royalties: financial interests due to intellectual property and patents in the field of cell and gene therapy. Conflicts of interest are managed in accordance with University of Pennsylvania policy and oversight; Novartis: Research Funding. Gill:Novartis: Patents & Royalties, Research Funding.

Abstract Chemo-refractory acute myeloid leukemia (AML) is associated with poor prognosis and treatment options are extremely limited. Most of these patients are ineligible for allogeneic stem cell transplantation. Chemo-refractory AML is thought to arise due to selection pressure of resistant clones from prior use of chemotherapy or in some cases pre-exist due to properties of the leukemic stem cells (LSC). CLEC12A (also known as CLL1) has previously been described as being selectively over expressed in LSCs. Successful modalities to target CLEC12A and eradicate the LSC would overcome chemo-refractoriness in AML and would represent a vertical advance in the field. In this study, we confirm that CLEC12A is heterogenously expressed on AML blasts and over-expressed on AML LSC. We also show that CLEC12A is overexpressed on bone marrows from patients with AML that fail to achieve a complete remission after induction chemotherapy, suggesting that it could be a marker for residual disease that is refractory to chemotherapy. We then separated AML blasts into CLEC12A positive or negative cells by magnetic sorting. CLEC12A positive blasts selected from AML patients were more resistant to chemotherapy compared to CLEC12A negative blasts (20% killing of CLEC12A positive AML cells versus 43% of CLEC12A negative AML cells when cultured with cytarabine 10 µg/ml, P=0.01). This finding was confirmed by using the AML MOLM14 cell line engineered to overexpress CLEC12A. CLEC12Ahigh MOLM14 cells were more resistant to chemotherapy compared to wild type MOLM14 cells (P=0.003). We then evaluated CLEC12A resistance to chemotherapy in a patient derived AML xenograft model. We found a relative increase in CLEC12A positive cells post Ara-C induction chemotherapy in AML xenograft models (Figure 1). The observation that CLEC12A positive cells are more resistant to chemotherapy provided a solid rationale to target CLEC12A with chimeric antigen receptor T (CART) cells. We therefore developed a second generation CLEC12A directed CAR construct using CD3z and 41BB costimulatory domains and generated CLEC12A CART cells by lentiviral transduction with this construct. Upon incubation with primary AML samples or AML cell lines, CLEC12A CART cells resulted in modest effector functions, due to the heterogeneity of CLEC12A expression on AML blasts. However when CLEC12A overexpressed MOLM14 cell line or CLEC12Apos selected leukemic cells were used as targets, CLEC12A-CART cells resulted in potent cytotoxicity, proliferation and cytokine production, indicating that CLEC12A-CART cells are more specific for LSC. To test the in vivo anti-leukemic activity of CLEC12A CARTs, we used primary human AML blasts xenografted into NSG-S mice (NOD-SCID-γc-/-, additionally transgenic for human stem cell factor, IL3 and GM-CSF). Treatment with CLEC12A CART (single dose, 1x105 total T cells via tail vein injection) resulted in modest activity against AML when employed as monotherapy. To investigate the potential role of CLEC12A CART cells in eradication of MRD and LSC, mice were treated first with chemotherapy (cytarabine 60 mg/kg intraperitoneal injection daily for 5 days) followed by a single dose (1x105 total T cells via tail vein injection) of either CLEC12A CARTs or control untransduced T cells (UTD). Treatment with CLEC12A CART cells resulted in eradication of leukemia and prolonged survival in these mice (overall survival at 200 days of 100% after CLEC12A CARTs compared to 20% after UTD, p=0.01, Figure 2). In conclusion, our preclinical studies reveal that CLEC12A positive cells in leukemia are resistant to chemotherapy and can be successfully targeted with CART cells. CLEC12A CART cells can potentially be employed as a consolidation regimen after induction chemotherapy to eradicate LSC and MRD in AML. Disclosures Kenderian: Novartis: Patents & Royalties, Research Funding. Ruella:novartis: Patents & Royalties: Novartis, Research Funding. Singh:Novartis: Employment. Richardson:Novartis: Employment, Patents & Royalties, Research Funding. June:Tmunity: Equity Ownership, Other: Founder, stockholder ; Immune Design: Consultancy, Equity Ownership; Novartis: Honoraria, Patents & Royalties: Immunology, Research Funding; University of Pennsylvania: Patents & Royalties; Celldex: Consultancy, Equity Ownership; Johnson & Johnson: Research Funding; Pfizer: Honoraria. Gill:Novartis: Patents & Royalties, Research Funding.

Abstract Chimeric antigen receptor T (CART) cell therapy results in impressively high remission rates in B cell neoplasms but is limited by the development of cytokine release syndrome (CRS). CRS is characterized by the development of high-grade fevers, hypotension, fluid overload and respiratory compromise, coincides with T cell expansion and is associated with marked elevation of interleukin-6, interferon-ɣ and other inflammatory cytokines. Severe CRS is seen in 25-80% of patients treated with CD19 directed CART cell therapy and mortality has been reported. While the use of the anti-IL6 receptor antibody tocilizumab with or without steroids can usually reverse this syndrome, there is concern that the early introduction of immunosuppressive medications could impair the anti-tumor activity and therefore most investigators currently reserve tocilizumab as therapy for severe (grade 3-4) CRS. Ruxolitinib is a JAK/STAT pathway inhibitor that is FDA approved for myelofibrosis and polycythemia vera and has resulted in a significant reduction of inflammatory cytokines in clinical studies. In preclinical models, treatment with ruxolitinib ameliorates cytokine production and manifestations of hemophagocytic lymphohistiocytosis. Therefore, there is a compelling rationale to investigate and develop ruxolitinib as a modality to prevent CRS after CART cell therapy. To our knowledge, there are no relevant models for CRS after human CART therapy, thus severely limiting the development of CRS prevention modalities that would in turn enhance the clinical feasibility of CART therapy. In this study, we established a novel acute myeloid leukemia (AML) xenograft model to study the development of CRS. Here, NSG-S (Non Obese Diabetic, SCID ɣ -/- mice that are additionally transgenic for human stem cell factor, IL-3 and GM-CSF) are engrafted with blasts from AML patients and treated with 1x106 CD123-directed CART cells (CART123), ten-fold higher than we previously used in our primary AML xenograft models. These animals develop an illness characterized by progressive weight loss, generalized weakness, emaciation, hunched bodies, withdrawal and poor motor response. This illness starts within one week of CART cell injection, correlates with T cell expansion and rapidly evolves and results in the death of the animals in 7-14 days (Figure 1A). Serum from these mice seven days after CART123 shows an extreme elevation of human IL-6, Interferon-γ, tumor necrosis factor-α, and other inflammatory cytokines (Figure 1B), resembling human CRS after CART cell therapy. NSG-S mice bearing primary AML were treated with CART123 and randomized to receive different doses of ruxolitinib (30, 60, or 90 mg/kg) or vehicle by oral gavage twice a day. Treatment started on the day of CART123 injection and continued for a week. Mice treated with ruxolitinib 60 or 90 mg/kg exhibited less severe clinical illness as manifested by attenuated weight loss when compared with mice treated with CART123 alone (Figure 2A). With the exception of ruxolitinib-only treated mice, all other groups exhibited an equivalent anti-leukemic effect (Figure 2B). We therefore established ruxolitinib 60 mg/kg for further experimental use. Most importantly, ruxolitinib treatment attenuated inflammatory cytokines (figure 2C) and led to long-term survival (figure 2D). Here we have described for the first time a clinically relevant animal model of human CRS and demonstrated that the JAK/STAT inhibitor ruxolitinib can prevent the development of severe CRS without impairing the anti-tumor effect of CART cells. These findings provide a useful platform for the future study of CRS prevention and treatment modalities. These experiments indicate that ruxolitinib could also be combined with CART cell therapy for the prevention of CRS in patients identified to be at high risk for the development of CRS. Disclosures Kenderian: Novartis: Patents & Royalties, Research Funding. Ruella:novartis: Patents & Royalties: Novartis, Research Funding. Lacey:Novartis: Research Funding. Melenhorst:Novartis: Patents & Royalties, Research Funding. June:Novartis: Honoraria, Patents & Royalties: Immunology, Research Funding; Pfizer: Honoraria; Johnson & Johnson: Research Funding; University of Pennsylvania: Patents & Royalties; Tmunity: Equity Ownership, Other: Founder, stockholder ; Celldex: Consultancy, Equity Ownership; Immune Design: Consultancy, Equity Ownership. Gill:Novartis: Patents & Royalties, Resear\ch Funding.

Abstract Chimeric antigen receptor (CAR) T cell immunotherapy has demonstrated profound results in hematologic malignancies but clinical efficacy in the solid tumor setting has not been observed. Barriers to T cell entry and function may partially explain this observation. As solid tumors actively recruit myeloid cells, we hypothesized that macrophages have the potential to be a powerful cellular immunotherapeutic agent in this setting if properly activated and redirected. We here describe the development of CAR macrophages (CARMA), demonstrating the feasibility, mechanism, and efficacy of this platform. To examine the function of CARs in macrophages, first generation anti-CD19, anti-mesothelin, or anti-HER2 CARs with a CD3ζ intracellular domain were introduced into the THP1 macrophage model. In vitro function was assessed via quantitative phagocytosis and luciferase-based specific killing assays. CARMA selectively phagocytosed and cleared cognate antigen-bearing tumor cells. To demonstrate the requirement for CAR-mediated intracellular signaling for activity, a CD3ζ-null CAR construct was tested in vitro. The deletion of CD3ζ significantly reduced the phagocytic and killing capacity (p&lt;0.01) of CARMA. We identified Ad5f35, a chimeric adenovirus, as a novel and highly efficient viral vector for the transduction of normal donor and cancer patient macrophages (&gt;70% CAR expression). Ad5f35 transduction polarized human macrophages toward a durable immunostimulatory M1 phenotype and rendered CARMA resistant to subversion toward the immunosuppressive M2 phenotype, as defined by surface markers and metabolomics. CARMA enhanced the proliferative capacity of CD8+ T cells in phytohemagglutinin activation assays and secreted factors that activated by-stander macrophages. Primary human anti-HER2 CARMA demonstrated targeted phagocytosis and killing of HER2 expressing ovarian and breast cancer cell lines, and exhibited a six-fold higher luciferase-based killing capacity of SKOV3 cells compared to trastuzumab in vitro (p=0.002). Anti-HER2 CARMA was evaluated in vivo in an intraperitoneal (IP) SKOV3 ovarian cancer xenograft model. Mice that received IP CARMA had a decrease in tumor burden of approximately two orders of magnitude and had a 30-day survival benefit relative to untreated or control macrophage treated mice (p=0.018). In a systemically disseminated SKOV3 model, a single dose of IV CARMA led to a durable anti-tumor response (38-fold reduction relative to control on day 31 post-treatment; p=0.016) Lastly, we demonstrated that the blockade of the anti-phagocytic CD47/SIRPα axis enhanced the phagocytic capacity of CARMA. In summary, we here demonstrate that human macrophages engineered with a CAR exhibit targeted anti-tumor function in both in vitro and in vivo preclinical models. This novel cellular immunotherapeutic approach has a clear translational potential for the treatment of solid tumors. Citation Format: Michael Klichinsky, Marco Ruella, Olga Shestova, Saad S. Kenderian, Miriam Y. Kim, Roddy O'Connor, John Scholler, Carl June, Saar Gill. Chimeric antigen receptor macrophages (CARMA) for adoptive cellular immunotherapy of solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4575. doi:10.1158/1538-7445.AM2017-4575