Search

CD47 expressed on cells functions as a "don't eat me" signal by binding SIRPa on macrophages and blocking phagocytosis. Cancer cells including AML evade phagocytosis by overexpressing CD47, and high CD47 expression in AML has been associated with inferior outcomes. Here we studied the pre-clinical activity of the 5F9 (magrolimab), an antibody that blocks the CD47, combined with BCL-2 inhibitor venetoclax with azacitidine (VEN/AZA) in AML both in vitro and in vivo. We first tested the combined efficacy of VEN/AZA and 5F9 in vitro in phagocytosis assay, utilizing monocytes isolated from seven healthy donors by CD16 positive selection and inducing macrophage differentiation with recombinant human M-CSF (rhM-CSF, 500ng/ml) for 1 week. Three paired isogenic AML cell lines MOLM-13, VEN-resistant (VEN-RE) MOLM-13, and p53-mutant MOLM-13 harboring R248W mutation, were pre-treated with DMSO or combination of 10nM VEN and 1uM AZA for 24hrs. Leukemia cells were labeled with Calcein and co-cultured with macrophages for 1h at a ratio of 2:1 (T:M), followed by flow cytometry read-out, quantifying Calcein+ macrophages that engulfed leukemia cells. The active phagocytosis was calculated as the percentage of Calcein+CD206+ cells from total CD206+ cells (Figure 1A, experimental schema). The 5F9/VEN-AZA combination significantly increased phagocytosis not only in parental but also in VEN-RE and p53-mutant MOLM-13 cells when compared with 5F9 or VEN/AZA alone treatments (Figure 1B). The treatment of VEN or AZA induced AnnexinV in MOLM-13 cells which may function as a pro-phagocytic signal and contribute to the enhanced phagocytosis (Figure 1C). Further proteomics studies to identify additional "eat me" signals following VEN/AZA treatment are ongoing and will be presented. Next, we examined the therapeutic efficacy of 5F9 alone or combined with VEN/AZA in the AML patient-derived xenografts (PDX) models. NSG mice were engrafted with VEN-RE AML PDX #344 (no canonical AML mutations detected by targeted sequencing) and #574 (FLT3-ITD, DNMT3A, NPM1-mutated). After confirmed bone marrow (BM) engraftment by flow cytometry, mice were randomized to receive vehicle, 5F9 (10mg/kg/d, ip. twice a week, 3 weeks), VEN (50mg/kg/g, po. qd, 14 days) / AZA (2.5mg/kg/d, ip, qd, 7 days), or their combination. Circulating leukemia burden was monitored by hCD45 flow cytometry in peripheral blood (PB) (Figure 1D), and mice survival was followed. 5F9 reduced leukemia burden, delayed leukemia progression and significantly extended mice' survival in both models, with the median survival of 276.5 days of 5F9 group compared to 197 days of vehicle in #344, and 81.5 vs 52 days in #574 (Figure 1E ). On the contrary, VEN/AZA therapy had no or minimal impact on tumor burden or survival. Notably, the combination of 5F9 and VEN/AZA resulted in elimination of circulating blasts and significant extension of survival in both models, median of 385 days in #344 and 124.5 days in #574. In summary, the VEN/AZA combined with 5F9 increased phagocytosis in isogenic MOLM-13 cells in vitro regardless of VEN resistance or p53 status and prolonged the survival in vivo in VEN/AZA refractory AML PDX models. The mechanisms of increased phagocytosis by combination treatment and the induction of "eat-me signals" by VEN/AZA is currently under investigation and will be presented. This triplet combination is currently being tested in a clinical trial (NCT04435691, Daver et al., ASH 2021) with high preliminary efficacy. Figure 1 Figure 1. Disclosures Chao: Gilead Sciences, Inc.: Current Employment; TigaTx: Membership on an entity's Board of Directors or advisory committees; Stanford University: Patents & Royalties; Hepatx Inc: Consultancy, Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees; Iconovir Bio: Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees; Leukemia and Lymphoma Society: Membership on an entity's Board of Directors or advisory committees; Stanford University Medical School: Membership on an entity's Board of Directors or advisory committees; Foresite capital: Consultancy; Bioverge: Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees; Chimera Bioengineering: Current equity holder in publicly-traded company. Maute: Gilead Sciences: Ended employment in the past 24 months. Andreeff: Syndax: Consultancy; Karyopharm: Research Funding; ONO Pharmaceuticals: Research Funding; Senti-Bio: Consultancy; AstraZeneca: Research Funding; Daiichi-Sankyo: Consultancy, Research Funding; Breast Cancer Research Foundation: Research Funding; Medicxi: Consultancy; Aptose: Consultancy; Reata, Aptose, Eutropics, SentiBio; Chimerix, Oncolyze: Current holder of individual stocks in a privately-held company; Oxford Biomedica UK: Research Funding; Novartis, Cancer UK; Leukemia & Lymphoma Society (LLS), German Research Council; NCI-RDCRN (Rare Disease Clin Network), CLL Foundation; Novartis: Membership on an entity's Board of Directors or advisory committees; Glycomimetics: Consultancy; Amgen: Research Funding. Daver: Hanmi: Research Funding; Bristol Myers Squibb: Consultancy, Research Funding; Gilead Sciences, Inc.: Consultancy, Research Funding; Daiichi Sankyo: Consultancy, Research Funding; Novimmune: Research Funding; Trillium: Consultancy, Research Funding; Glycomimetics: Research Funding; ImmunoGen: Consultancy, Research Funding; Sevier: Consultancy, Research Funding; Amgen: Consultancy, Research Funding; Genentech: Consultancy, Research Funding; FATE Therapeutics: Research Funding; Trovagene: Consultancy, Research Funding; Abbvie: Consultancy, Research Funding; Astellas: Consultancy, Research Funding; Pfizer: Consultancy, Research Funding; Novartis: Consultancy; Jazz Pharmaceuticals: Consultancy, Other: Data Monitoring Committee member; Dava Oncology (Arog): Consultancy; Celgene: Consultancy; Syndax: Consultancy; Shattuck Labs: Consultancy; Agios: Consultancy; Kite Pharmaceuticals: Consultancy; SOBI: Consultancy; STAR Therapeutics: Consultancy; Karyopharm: Research Funding; Newave: Research Funding. Konopleva: Rafael Pharmaceuticals: Other: grant support, Research Funding; Novartis: Other: research funding pending, Patents & Royalties: intellectual property rights; Ascentage: Other: grant support, Research Funding; Reata Pharmaceuticals: Current holder of stock options in a privately-held company, Patents & Royalties: intellectual property rights; Eli Lilly: Patents & Royalties: intellectual property rights, Research Funding; F. Hoffmann-La Roche: Consultancy, Honoraria, Other: grant support; KisoJi: Research Funding; Stemline Therapeutics: Research Funding; Ablynx: Other: grant support, Research Funding; AstraZeneca: Other: grant support, Research Funding; Cellectis: Other: grant support; Sanofi: Other: grant support, Research Funding; Calithera: Other: grant support, Research Funding; Genentech: Consultancy, Honoraria, Other: grant support, Research Funding; Agios: Other: grant support, Research Funding; Forty Seven: Other: grant support, Research Funding; AbbVie: Consultancy, Honoraria, Other: Grant Support, Research Funding.

Context: Magrolimab, an anti-CD47 antibody, induces tumor phagocytosis and eliminates leukemia stem cells. Azacitidine (AZA) synergizes with magrolimab by inducing prophagocytic “eat me” signals. Magrolimab + AZA is clinically effective in acute myeloid leukemia (AML) and myelodysplastic syndrome. Objective: To report Phase 1b data of magrolimab + AZA in untreated AML. Design: This is an open-label, nonrandomized, Phase 1b, interventional study. Interventions: Patients received a magrolimab priming/intrapatient dose-escalation regimen (1–30 mg/kg intravenous weekly followed by 30 mg/kg every 2 weeks in cycle 3 and beyond) and AZA (75 mg/m2 on days 1–7 on a 28-day cycle). Patients or Other Participants: Treatment-naive AML patients unfit for intensive chemotherapy. Main Outcomes Measures: Outcome measures included the percentage of patients with adverse events (AEs), objective response (OR), time to response, and duration of response. Results: Fifty-two patients were treated with magrolimab + AZA. Overall, 64% had poor-risk cytogenetics, and 65% had TP53 mutations. Magrolimab + AZA was well tolerated with a safety profile similar to AZA monotherapy. Treatment-related AEs (≥15% of patients) were anemia (31%), fatigue (19%), blood bilirubin increase (19%), neutropenia (19%), thrombocytopenia (17%), and nausea (15%). On-target anemia was generally transient and reversible, and no immune-related AEs associated with magrolimab were observed. Of 34 patients evaluable for efficacy, 22 (65%) achieved an OR, 15 (44%) achieved complete response (CR), 4 (12%) with CR with incomplete count recovery (CRi), 1 (3%) with partial response, 2 (6%) with morphological leukemia-free state (MLFS), 11 (32%) with stable disease (SD), and 1 (3%) with progressive disease (PD). Time to response was 2.04 months. In TP53-mutant patients, 15/21 (71%) achieved an OR, 10 (48%) achieved a CR, 4 (19%) with CRi, 1 (5%) with MLFS, 5 (24%) with SD, and 1 (5%) with PD. The median overall survival for TP53-mutant patients (n=34) and TP53–wild-type patients (n=16) was 12.9 and 18.9 months, respectively, with a median follow-up of 4 and 12 months, respectively. Conclusions: Magrolimab + AZA was well tolerated, with efficacy in both TP53-mutant and TP53–wild-type AML patients. A Phase 3 trial evaluating magrolimab + AZA in untreated TP53-mutant AML patients is planned (NCT04778397).

Immune checkpoint blockade has emerged as a promising cancer treatment paradigm. Unfortunately, there are still a large number of patients and malignancies that do not respond to therapy. A major barrier to validating biomarkers for the prediction and monitoring of responders to clinical checkpoint blockade has been the lack of imaging tools to accurately assess dynamic immune checkpoint expression. Here, we sought to optimize noninvasive immuno-PET imaging of human programmed death-ligand 1 (PD-L1) expression, in a preclinical model, using a small high-affinity engineered protein scaffold (HAC-PD1). Six HAC-PD1 radiotracer variants were developed and used in preclinical imaging and biodistribution studies to assess their ability to detect human PD-L1 expression in vivo. Radiotracer design modifications included chelate, glycosylation, and radiometal. HACA-PD1 was adopted as the naming convention for aglycosylated tracer variants. NOD scid γ-(NSG) mice were inoculated with subcutaneous tumors engineered to either be constitutively positive (CT26 hPD-L1) or be negative (ΔmPD-L1 CT26) for human PD-L1 expression. When the tumors had grown to an average size of 1 cm in diameter, mice were injected with 0.75–2.25 MBq (∼10 μg) of an engineered radiotracer variant and imaged. At 1 h after injection, organs were harvested for biodistribution. Of the practical immuno-PET tracer modifications considered, glycosylation was the most prominent design factor affecting tracer uptake, specificity, and clearance. In imaging studies, aglycosylated 64Cu-NOTA-HACA-PD1 most accurately visualized human PD-L1 expression in vivo. We reasoned that because of the scaffold’s small size (14 kDa), its pharmacokinetics may be suitable for labeling with the short-lived and widely clinically available radiometal 68Ga. At 1 h after injection, 68Ga-NOTA-HACA-PD1 and 68Ga-DOTA-HACA-PD1 exhibited promising target-to-background ratios in ex vivo biodistribution studies (12.3 and 15.2 tumor-to-muscle ratios, respectively). Notably, all HAC-PD1 radiotracer variants enabled much earlier detection of human PD-L1 expression (1 h after injection) than previously reported radiolabeled antibodies (>24 h after injection). This work provides a template for assessing immuno-PET tracer design parameters and supports the translation of small engineered protein radiotracers for imaging human immune checkpoints.

Intro: Relapsed/refractory diffuse large B cell lymphoma (r/r DLBCL) is associated with poor outcomes, and remains an area of unmet clinical need. Emerging data indicate that a key macrophage immune checkpoint termed CD47 can be effectively targeted in r/r DLBCL. CD47 is a potent "don't eat me" signal and impairs tumor cell phagocytosis by engaging the inhibitory macrophage receptor, SIRPα. Importantly, early clinical studies demonstrate that blocking CD47-SIRPα interactions with an anti-CD47 antibody (magrolimab) yields promising activity in r/r DLBCL patients, particularly when combined with rituximab, which further potentiates lymphoma cell phagocytosis via its Fc domain. Moving forward, it will be crucial to identify combinatorial strategies that improve the efficacy of anti-CD47 therapy in r/r DLBCL. Tumor-associated macrophages (TAMs) contribute to an immunosuppressive environment in cancer through secretion of cytokines, metabolites, and expression of immune checkpoints. Sustained PI3K-γ signaling is critical to maintain immune suppressive TAM functions, and PI3K-γ inhibition polarizes TAMs into a pro-inflammatory state that culminates in potent anti-tumor T cell responses, sensitizing cancers to PD-1 blockade therapy. We hypothesized that activating TAMs through PI3K-δ/γ inhibition would also augment their inherent phagocytic capacity and synergize with phagocytosis-targeting treatments such as CD47 blockade. Thus, we investigated the impact of the dual PI3K-δ/γ inhibitor, duvelisib, on expression of key phagocytic receptors and pathways in human macrophages, as well as its effect on enhancing lymphoma cell phagocytosis with the anti-CD47 antibody, magrolimab. Results: RNA sequencing of duvelisib-treated human monocyte-derived macrophages revealed that PI3K-δ/γ inhibition induced broad transcriptional changes suggestive of an activated macrophage state (Fig 1A), including upregulation of interferon response genes (IRF9) and costimulatory molecules (CD86), while decreasing expression of immune suppressive cytokines (IL10) and inhibitory ligands (PDL1). Moreover, duvelisib upregulated transcription of canonical pro-phagocytic receptors (LRP1, FCGR2A,CD44), lysosomal degradation components (LYZ, CTSS) and phosphatidylserine-bridging genes (MFG-E8, GAS6), while inhibiting metabolic pathways, most notably mTOR, each of which has been associated with increased and more efficient phagocytosis. Given these findings, we evaluated the effect of dual PI3K-δ/γ inhibition on macrophage phagocytosis of DLBCL cells in the context of CD47 blockade in vitro. Across a broad panel of human DLBCL cell lines and multiple monocyte donors, pre-treatment of macrophages with duvelisib significantly increased lymphoma cell phagocytosis by magrolimab (Fig 1B-C). This effect reflected a direct increase in the phagocytic capacity of treated macrophages, as lymphoma cells were not exposed to duvelisib. In separate experiments, we found that duvelisib rarely induced full apoptosis of DLBCL cells, but did promote the expression of "eat me" signals, such as calreticulin and phosphatidylserine on many of the cell lines examined (Fig 1D). Collectively, these results suggest that PI3K-δ/γ inhibition enhances the inherent phagocytic capacity of macrophages, while also sensitizing lymphoma cells for engulfment. Similarly, potent effects were not observed with PI3K-δ- or PI3K-γ-specific inhibitors, suggesting that combined PI3K-δ/γ inhibition is critical to mediating these beneficial effects. Finally, to determine whether duvelisib impacted the efficacy of magrolimab in vivo, we treated human DLBCL xenografts and observed that while duvelisib had minimal impact on lymphoma growth, the combination of duvelisib and magrolimab induced complete tumor rejection in a high proportion of mice, leading to increased survival compared to mice treated with magrolimab alone (Fig 1E). Conclusion: Collectively, our findings suggest that targeting the PI3K-δ/γ axis favorably induces key macrophage phagocytic pathways, while simultaneously upregulating important "eat me" signals on lymphoma cells. Together, these alterations lower the phagocytic threshold of TAMs and dramatically increase the efficacy of anti-CD47 therapy in DLBCL xenograft models. This work paves the way for early-phase studies of duvelisib and magrolimab in r/r DLBCL and other lymphomas. Disclosures Godfrey: Verastem: Research Funding; Merck: Research Funding; Gilead: Research Funding. Coma:Verastem Oncology, Inc: Current Employment, Current equity holder in publicly-traded company. Maute:Gilead Sciences, Inc.: Current Employment. Chao:Gilead Sciences: Current Employment. Pachter:Verastem Therapeutics: Current Employment. Kline:Kite/Gilead: Speakers Bureau; Merck: Research Funding; Verastem: Membership on an entity's Board of Directors or advisory committees, Research Funding; Karyopharm: Membership on an entity's Board of Directors or advisory committees; Seattle Genetics: Membership on an entity's Board of Directors or advisory committees.

7507 Background: Magrolimab (Hu5F9-G4) is an antibody blocking CD47, a macrophage immune checkpoint and don’t eat me signal on cancers. It induces tumor phagocytosis and eliminates leukemia stem cells. Azacitidine (AZA) synergizes with magrolimab by inducing eat me signals on leukemic cells, enhancing phagocytosis. We report Ph1b data including a potential MDS registration cohort. Methods: Magrolimab+AZA was given to untreated intermediate to very high risk IPSS-R MDS and intensive chemo unfit AML patients. A magrolimab priming/intrapatient dose escalation regimen (1-30 mg/kg QW, Q2W Cycle 3+) was used. AZA was dosed 75mg/m2 days 1-7. Efficacy was assessed by IWG 2006 (MDS) and ELN 2017 (AML) criteria. Results: 68 patients (39 MDS, 29 AML) with a median age of 72 were treated with magrolimab+AZA. 19% were intermediate cytogenetic risk with 68% poor risk (13% unknown). 27% were TP53 mutant. The combo was well-tolerated with safety similar to AZA alone. Common treatment-related AEs were anemia (38%), fatigue (21%), neutropenia (19%), thrombocytopenia (18%) and infusion reaction (16%). Treatment-related febrile neutropenia was 1.5%. Only 1 patient (1.5%) discontinued due to an AE. In RBC transfusion dependent patients, 58% of MDS and 64% of AML patients became transfusion independent. 30/33 (91%) efficacy evaluable MDS patients had an objective response (42% CR, 24% marrow CR (4/8 also with HI), 3% PR, 21% HI alone, 9% SD). MDS patient responses deepened on study, with a 56% CR rate in patients with ≥ 6 mo follow-up. In AML, 16/25 (64%) responded (40% CR, 16% CRi, 4% PR, 4% MFLS, 32% SD, 4% PD). In 12 TP53 mutant AML patients, 75% had a CR+CRi (42% CR, 33% CRi, 17% SD, 8% PD). Cytogenetic CR was seen in 35% and 50% of responding MDS and AML patients. 22% of MDS and 50% of AML patients with CR/CRi/marrow CR were MRD negative by flow cytometry. Median duration of response is not reached in either MDS or AML, including TP53 mutant AML, with a median follow-up of 5.8, 8.8 and 9.4 mos, respectively (range: 1.9 – 16.8 mos). 91% of MDS and 100% of AML responding patients are in response at 6 mos. The 6 mo overall survival estimate is 100% in MDS and 91% in TP53 mutant AML patients. Conclusions: Magrolimab is a macrophage targeting immunotherapy that with AZA is well tolerated with durable efficacy in MDS, AML, particularly TP53 mutant, a poor prognostic group. A potential registration single arm MDS cohort is ongoing (NCT03248479). ENHANCE, a randomized Ph3 MDS trial is planned. Additional patients/analyses will be reported. Funded by Forty Seven and CIRM. Clinical trial information: NCT03248479 .

e15098 Background: Magrolimab (Hu5F9-G4) is a monoclonal IgG4 antibody that binds to CD47 receptor and blocks its interaction with SIRPA, thereby enabling tumor cell phagocytosis and activation of an anti-tumor T-cell response. Magrolimab has shown encouraging efficacy and tolerable safety profiles in treatment-naïve myelodysplastic syndromes/acute myeloid leukemia (MDS/AML) patients on an every week (QW) maintenance dosing schedule. Reduced dosing frequency (e.g. every other week – Q2W) is desired to enhance patient convenience and quality of life. Methods: Interim pharmacokinetic (PK) and CD47 receptor occupancy (RO) data from study 5F9005 (Clinical Trial: NCT03248479) were used for this analysis. A previously developed population PK (PopPK) model of magrolimab was verified with observed PK data from study 5F9005. The PK model was linked to peripheral blood and bone marrow blast CD47 RO. The model-predicted RO was validated with observed RO data. Simulations of PK and RO were conducted to justify the proposed reduced dosing frequency of Q2W starting in Cycle 3 onward. Simulations were compared with preliminary observed RO data from the ongoing 5F9005 study after switching to Q2W dosing. Results: The results showed that observed magrolimab PK in study 5F9005 was consistent with model predictions. Model predicted RO was aligned with observed RO data. PK and RO simulations based on the model indicated > 99% RO in peripheral blood and ~90% RO in bone marrow are expected after switching to Q2W dosing regimen from Cycle 3 onward. Preliminary observed RO data confirmed these predictions. Conclusions: Magrolimab PK in AML/MDS is consistent with other patient populations including solid tumors and lymphomas. Model-based simulations and preliminary RO data support Q2W dosing of magrolimab from Cycle 3 onward as optimal RO is expected with the proposed dosing regimen. Clinical trial information: NCT03248479 .

114 Background: Magrolimab (M, Hu5F9-G4) is an antibody targeting CD47, a macrophage “don’t eat me” signal that demonstrates preclinical synergy with cetuximab (C) in refractory KRAS wild type (KRASwt) and KRAS mutant (KRASm) colorectal (CRC) tumors. Methods: Phase (Ph) 1 doses of M+C were escalated in solid tumor patients (pts) and Ph 2 efficacy was explored in previously treated KRASwt and KRASm CRC patients. Day 1 priming with 1 mg/kg of M was used to mitigate on-target anemia followed by maintenance doses ranging from 10 to 45 mg/kg in combination with C. Ph 2 pts were treated with 30 or 45 mg/kg of M and 400/250 mg/m2 of C. Results: In 78 treated pts (32 Ph 1 and 46 Ph 2), the median age was 59 years (range 19-82), and median prior therapies was 5 (range 1-14). No maximum tolerated dose was reached. Treatment-related adverse events (TRAEs) of any Grade (G) included dermatitis acneiform 36%, dry skin 33%, fatigue 32%, infusion reactions 31%, headache 30%, diarrhea 23%, nausea 23%, chills 23%, and anemia 22%. There were no fatal TRAEs and 3/78 (4%) discontinued M treatment due to any adverse events. In the combined Ph 1+2 study, 2 of 30 evaluable KRASwt CRC pts had confirmed PRs for 7.0 and 12.5 months (mo), for a 6.7% objective response rate (ORR). Both had prior C treatment. The median progression-free survival (mPFS) and median overall survival (mOS) was 3.6 mo (95%CI 1.8-5.4) and 10.1 mo (95%CI: 6.9-14.4), respectively. In 40 evaluable KRASm pts, there were no responses but 45% had stable disease (SD) and the mPFS and mOS were 1.9 mo (95%CI: 1.8-3.5) and 10.4 mo (95%CI: 5.7-16.4), respectively. In 28 KRASm pts who were TAS102/regorafenib naïve, preliminary mOS was 12.4 mo (95%CI: 5.9-not reached) which is longer than that reported for historical controls. Tumor biopsies showed treatment-related increases in macrophage immune cell infiltrates in SD pts, and baseline T cell infiltration was associated with longer OS. Pharmacokinetic profiles will be presented. Conclusions: M+C is a novel, well-tolerated combination immunotherapeutic treatment regimen. Responses were observed in two previously treated CRC pts and survival is encouraging in KRASm pts. Funded by Forty Seven and California Institute for Regenerative Medicine. Clinical trial information: NCT02953782.

Introduction Magrolimab (Hu5F9-G4, 5F9) is a first-in-class IgG4 antibody targeting CD47, a macrophage immune checkpoint and "don't eat me" signal expressed on cancer cells. Blockade of CD47 leads to phagocytosis of tumor cells. Magrolimab synergizes with rituximab to eliminate CD20-positive lymphoma by enhancing antibody-dependent cellular phagocytosis. Magrolimab +rituximab demonstrated encouraging safety and efficacy in a Phase (Ph)1b dose escalation cohort in patients with relapsed/refractory (r/r) DLBCL and FL that were rituximab-refractory (Advani et al., NEJM 2018). CD24 is an additional "don't eat me" signal, and has been proposed as a potential target for immunotherapy (Barkal et al. Nature 2019). Here we describe immunohistochemical analysis of CD47 and CD24 in primary patient biopsies from an ongoing follow-up Phase 2 trial of Non-Hodgkin's lymphoma (including DLBCL and indolent lymphoma) patients treated with magrolimab+rituximab. Results By immunohistochemistry, 54 out of 54 patients assessed were positive for expression of CD47 at screening. High levels of CD47 expression were maintained during treatment with clinically-efficacious doses of magrolimab. Therapeutic response did not correlate with CD47 H-score expression levels. At screening, patients presented with highly variable levels of CD24 expression, with 40 of 54 samples showing positive staining in >30% of cells. We observed no significant correlation between CD24 expression by H-score and response to therapy (complete response + partial response) in either DLBCL or indolent lymphoma. Conclusions CD47 shows consistently high expression in primary biopsies from Non-Hodgkin's lymphoma patients and remains persistently high during treatment. In our Phase 2 trial, therapeutic response did not correlate with CD47 expression levels, suggesting that the degree of target expression is not a primary driver of magrolimab efficacy in Non-Hodgkin's lymphoma. Although we observe wide variation in CD24 expression within this cohort, its levels are not predictive of outcome for this disease indication. We are evaluating the expression of additional biomarkers to identify those Non-Hodgkin's lymphoma patients most likely to benefit from magrolimab+rituximab combination therapy. Disclosures Maute: Forty Seven Inc.: Employment, Equity Ownership, Patents & Royalties. Chen:Forty Seven Inc.: Consultancy, Equity Ownership. Marjon:Forty Seven Inc.: Employment, Equity Ownership. Duan:Forty Seven Inc.: Employment, Equity Ownership. Choi:Forty Seven Inc.: Employment, Equity Ownership. Chao:Forty Seven, Inc.: Employment, Equity Ownership, Patents & Royalties. Takimoto:Forty Seven, Inc.: Employment, Equity Ownership, Patents & Royalties. Agoram:Forty Seven Inc.: Employment, Equity Ownership. Volkmer:Forty Seven, Inc.: Employment, Equity Ownership, Patents & Royalties.

Small-cell lung cancer (SCLC) is a highly aggressive subtype of lung cancer with limited treatment options. CD47 is a cellsurface molecule that promotes immune evasion by engaging signal-regulatory protein alpha (SIRPα), which serves as an inhibitory receptor on macrophages. Here, we found that CD47 is highly expressed on the surface of human SCLC cells; therefore, we investigated CD47-blocking immunotherapies as a potential approach for SCLC treatment. Disruption of the interaction of CD47 with SIRPα using anti-CD47 antibodies induced macrophage-mediated phagocytosis of human SCLC patient cells in culture. In a murine model, administration of CD47-blocking antibodies or targeted inactivation of the Cd47 gene markedly inhibited SCLC tumor growth. Furthermore, using comprehensive antibody arrays, we identified several possible therapeutic targets on the surface of SCLC cells. Antibodies to these targets, including CD56/neural cell adhesion molecule (NCAM), promoted phagocytosis in human SCLC cell lines that was enhanced when combined with CD47-blocking therapies. In light of recent clinical trials for CD47-blocking therapies in cancer treatment, these findings identify disruption of the CD47/SIRPα axis as a potential immunotherapeutic strategy for SCLC. This approach could enable personalized immunotherapeutic regimens in patients with SCLC and other cancers.

Immunotherapies represent the next generation of treatments for cancer. The recent clinical successes of immunotherapies have highlighted the potential to achieve stunning efficacy against otherwise intractable cancers. One such clinically validated immunotherapy target is the T-cell surface receptor PD-1 (programmed cell death protein 1), which inhibits T-cell cytotoxicity upon ligation with PD-L1 (programmed death-ligand 1) on a target cell. Antibody therapies targeting PD-1/PD-L1 can be strikingly effective cancer treatments, as demonstrated in the clinic by the recently approved pembrolizumab and nivolumab. However, antibody therapies have significant drawbacks as agents against this pathway. Due to their large size (~150 kDa), antibodies have limited tissue penetrance. Furthermore, due to the presence of the prophagocytic Fc domain, antibodies bound to PD-1 or PD-L1 on the T-cell cause counterproductive lymphocyte depletion. We hypothesized that a small engineered protein capable of blocking PD-1:PD-L1 signaling could avoid these issues and function as a superior immunotherapy. To accomplish this goal, we used yeast display-based directed evolution to affinity-mature a soluble form of the PD-1 ectodomain to increase its affinity for human PD-L1. After several rounds of selection from two generations of mutant libraries, we identified a high-affinity PD-1 variant. By surface plasmon resonance, high-affinity PD-1 binds to human PD-L1 with 110 pM affinity and a half-life of binding of >40 minutes, a dramatic improvement from wild-type PD-1. Direct in vivo binding assays demonstrate that this small protein penetrates solid tissues much more effectively than an anti-PD-L1 monoclonal antibody. Unlike anti-PD-L1 antibodies, high-affinity PD-1 does not lead to the depletion of effector T-cells. Consistent with these advantages, in syngeneic CT26 tumor models, high affinity PD-1 was effective in treating both small (~50 mm3) and large tumors (>150 mm3), whereas the activity of anti-PD-L1 antibodies was completely abrogated against large tumors. Furthermore, radiolabeling of high-affinity PD-1 enabled us to utilize our reagent as a non-invasive tracer for PET imaging. Taken together, our data demonstrates the benefits of using small, non-antibody therapeutics as cancer immunotherapy and immune diagnostic agents. Citation Format: Sydney R. Gordon, Roy Maute, Aaron Mayer, Melissa McCracken, Arutselvan Natarajan, Nan Guo, Richard Kimura, Jonathan M. Tsai, Aashish Manglik, Andrew Kruse, Sanjiv Gambhir, Irving L. Weissman, Aaron M. Ring. Engineered PD-1 variants as immunotherapies for cancer. [abstract]. In: Proceedings of the CRI-CIMT-EATI-AACR Inaugural International Cancer Immunotherapy Conference: Translating Science into Survival; September 16-19, 2015; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2016;4(1 Suppl):Abstract nr B101.

CD47 allows cancer cells to evade the immune system by signaling through SIRPa, an inhibitory receptor on macrophages. Therapies that block CD47 convert tumor-promoting macrophages to a tumoricidal state within the tumor microenvironment. We recently developed next-generation CD47 antagonists by engineering the extracellular domain of SIRPa. As single-domain polypeptides, these “high-affinity SIRPa variants” have an affinity for human CD47 (KD) as low as 11.1 pM, approximately 50,000-fold improved over wild-type SIRPa. By themselves, the high-affinity SIRPa variants are inert and therefore non-toxic in mouse and primate studies. However, when combined with tumor-specific antibodies, the high-affinity SIRPa variants act as immunotherapeutic adjuvants to antibody therapies by maximizing the ability of macrophages to destroy cancer cells. In our current study, we hypothesized these novel CD47-blocking agents could be applied to the treatment of small cell lung cancer (SCLC), a cancer with poor prognosis for which no clinically-approved antibodies or immunotherapies exist. We examined a panel of human SCLC samples and found all samples tested expressed high levels of CD47 on their surface. Using purified macrophages in vitro, we found that CD47-blocking therapies were able to induce macrophage phagocytosis of SCLC cell lines and primary patient samples. As a proof-of-concept, treatment of mice bearing primary SCLC tumors with CD47-blocking antibodies was able to inhibit tumor growth and significantly prolong survival. To identify novel SCLC antigens that can be targeted in combination with the high-affinity SIRPa variants, we screened SCLC samples by high-throughput flow cytometry using LEGENDScreen comprehensive antibody arrays. We identified several new and established therapeutic targets on the surface of SCLC cells, including CD99, CD56, CD166, CD326, and CD164. We identified antibodies to these antigens that could elicit macrophage phagocytosis in vitro, validating these antigens as targets for immune-based therapies. The ability of these antibodies to induce phagocytosis was dramatically enhanced when combined with the high-affinity SIRPa variants. Future studies will test these immunotherapeutic combinations in vivo against SCLC samples to develop novel therapeutic combinations for patients. We propose this strategy as a universal method to identify new tumor antigens and overcome macrophage immunosuppression within the tumor microenvironment. Citation Format: Kipp Weiskopf, Peter J. Schnorr, Nadine Jahchan, Aaron M. Ring, Roy L. Maute, Anne K. Volkmer, Jens-Peter Volkmer, Kenan C. Garcia, Julien Sage, Irving L. Weissman. Overcoming macrophage immunosuppression in small cell lung cancer with high-affinity SIRPa variants. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3629. doi:10.1158/1538-7445.AM2014-3629

Abstract 703 In order to gain insights into the role of microRNA (miRNAs) in mature B cell function and lymphomagenesis, we performed a comprehensive miRNA expression profiling of normal mature B cells and of germinal center (GC)-derived lymphomas. miRNA expression profiles were generated using a commercial array platform designed to interrogate approximately 700 miRNAs. Normal GC B cells were isolated from tonsil tissue of 5 pediatric healthy donors and subjected to miRNA profiling in parallel with tumor specimens obtained from 10 Burkitt Lymphoma (BL), 16 Follicular Lymphoma (FL) and 20 Diffuse Large B Cell Lymphoma (DLBCL) patients. Each tumor type displayed a distinct miRNA profile and appeared to be clearly separated from the normal counterpart. Interestingly, a set of miRNAs was expressed in normal GC cells, but not in lymphoma samples, suggesting that structural and/or functional loss of miRNAs occur during lymphomagenesis. Among these, miR-28 was found to be up-regulated in GC B cells, while it was completely silenced in BL and significantly reduced in a large fraction of DLBCL and FL. Quantitative RT-PCR analysis confirmed miR-28 reduced expression in these GC-derived lymphoma subtypes including both primary biopsies and cell lines. MiR-28 is an intragenic miRNA encoded by the LPP gene locus, located on chromosome 3q28. Deletions affecting miR-28 and LPP were previously reported in FL (Schwaenen et al. Genes Chromosomes Cancer 2009; 48:39-54) and similarly we identified LPP deletions in about 9% of DLBCL investigated for copy number alterations by SNP arrays. Further FISH analyses performed in cell lines lacking miR-28 expression (12 DLBCL and 27 BL) failed to identify chromosomal aberrations in the LPP locus, suggesting that mechanisms other than genetic losses, possibly of epigenetic nature, are involved in miR-28 silencing in lymphomas. In order to investigate the effects of miR-28 in lymphoma cells, we generated stable lymphoma cell lines displaying inducible expression of miR-28. Re-expression of miR-28 in lymphoma cells led to a retarded growth due to a combination of G1-cell-cycle arrest and increased apoptosis. Furthermore, lymphoma cells expressing miR-28 lost their clonogenic properties as shown by their inability to form colonies in soft agar. Taken together these results suggest a tumor suppressor function for miR-28 in lymphoma cells. Toward the identification of the direct miR-28 target genes, we implemented computational target prediction methods with gene expression profiling data obtained from the miR-28 engineered cell lines. Approximately two thousand miR-28 direct candidate targets were computationally predicted, 88 of which displayed transcriptional down-regulation upon miR-28 induction, suggesting that miR-28 may affect the stability of the target transcript. The candidate targets included genes involved in the control of cell proliferation and apoptosis and in cell signaling, consistent with the phenotypic changes induced by miR-28 expression. In 4 out of 5 tested candidate targets, 3′-UTR reporter gene assay confirmed the direct effect of miR-28 on the target gene. Finally, we have generated a conditional, GC B cell-specific miR-28 knock-out mouse model, which will provide critical insights on the physiologic as well as the tumor suppressor role of miR-28 in vivo. Disclosures: No relevant conflicts of interest to declare.

18 Background: Magrolimab (M, Hu5F9-G4) is an antibody targeting CD47, a “don’t eat me” signal for macrophages that enhances ovarian cancer cell phagocytosis in preclinical models in combination with the PD-L1 inhibitor avelumab. CD47 blockade can also enhance cross-priming of T cells. Methods: In Part 1 (P1) M+A doses were escalated in ST patients (pts) while Part 2 (P2) enrolled platinum-resistant or refractory OC pts. All received a 1 mg/kg Day 1 priming dose of M to mitigate on-target anemia due to macrophage-mediated extravascular hemolysis followed by 30 or 45 mg/kg maintenance doses in P1 or 45 mg/kg in P2, in combination with 800 mg of A Q2 wks. Results: In the 34 total pts (13 in P1 and 21 in P2), median age was 66 years (range 47-88), and median # of prior therapies was 5 (range 1-10). In P1, no dose limiting toxicities occurred. In P1+P2, no Grade (G) 4 or 5 treatment-related adverse events (TRAEs) occurred. TRAEs of any G in > 20% pts included headache 62%, fatigue 47%, infusion related reaction 44%, pyrexia 38%, chills 35%, nausea 35%, anemia 24%, and vomiting 21%. In P1, a patient (pt) with metastatic papillary adenocarcinoma of the finger on 45 mg/kg of M had a confirmed partial response (PR) lasting for 4 months before progressing. In P2, 1 pt had an unconfirmed PR but progressed per RECIST v1.1 on the next scan achieving a best response of stable disease (SD). In the 18 OC P1+P2 pts with at least one response evaluation, 56% had SD and 44% had progression. Analysis of tumor biopsies for immune cells infiltration and CD47, PD-L1, and other marker expression is ongoing. In the 13 OC tumors analyzed to date, only 2 were PD-L1 positive. One PD-L1+ tumor had PD-L1 expression only on infiltrating immune cells. The other PD-L1+ had tumor cell expression and was the only OC pt with tumor shrinkage. This supports the potential importance of PD-L1 expression for this combination. Pharmacokinetic profiles will be presented. Conclusions: M+A is a novel, well-tolerated combination treatment regimen with 1 observed PR in a ST pt and a 56% SD rate in OC pts. Tumor shrinkage was noted in the only OC pt with tumor cell expression of PD-L1 which warrants further evaluation in PD-L1+ OC pts. Clinical trial information: NCT03558139.