iPSC-derived NK cells expressing high-affinity IgG Fc receptor fusion CD64/16A to mediate flexible, multi-tumor antigen targeting for lymphoma

Abstract

Introduction: NK cells can mediate tumor cell killing by natural cytotoxicity and by antibody-dependent cell-mediated cytotoxicity (ADCC), an anti-tumor mechanism mediated through the IgG Fc receptor CD16A (FcγRIIIA). CD16A polymorphisms conferring increased affinity for IgG positively correlate with clinical outcomes during monoclonal antibody therapy for lymphoma, linking increased binding affinity with increased therapeutic potential via ADCC. We have previously reported on the FcγR fusion CD64/16A consisting of the extracellular region of CD64 (FcγRI), a high-affinity Fc receptor normally expressed by myeloid cells, and the transmembrane/cytoplasmic regions of CD16A, to create a highly potent and novel activating fusion receptor. Here, we evaluate the therapeutic potential of engineered induced pluripotent stem cell (iPSC)-derived NK (iNK) cells expressing CD64/16A as an "off-the-shelf", antibody-armed cellular therapy product with multi-antigen targeting potential.

Methods: iNK cells were generated from iPSCs engineered to express CD64/16A and an interleukin (IL)-15/IL-15Rα fusion (IL-15RF) protein for cytokine independence. iNK cells and peripheral blood NK cells were expanded using irradiated K562-mbIL21-41BBL feeder cells to examine in in vitro and in vivo assays using the Raji lymphoma cell line. ADCC was evaluated in real-time by IncuCyte assays and using a xenograft mouse model with high circulating levels of human IgG.

Results: Our data show that CD64/16A expressing iNK cells can mediate potent anti-tumor activity against human B cell lymphoma. In particular, (i) under suboptimal conditions, including low antibody concentrations and low effector-to-target ratios, iNK-CD64/16A cells mediate ADCC, (ii) iNK-CD64/16A cells can be pre-loaded with tumor-targeting antibodies (arming) to elicit ADCC, (iii) armed iNK-CD64/16A cells can be repurposed with additional antibodies to target new tumor antigens, and (iv) cryopreserved, armed iNK-CD64/16A are capable of sustained ADCC in a tumor xenograft model under saturating levels of human IgG.

Discussion: iNK-CD64/16A cells allow for a flexible use of antibodies (antibody arming and antibody targeting), and an "off-the-shelf" platform for multi-antigen recognition to overcome limitations of adoptive cell therapies expressing fixed antigen receptors leading to cancer relapse due to antigen escape variants.

Keywords: ADCC - antibody-dependent cellular cytotoxicity; antibody; cancer; immunotherapy; natural killer (NK) cell.

Figure 1

Generation and function of iPSC-derived NK cells expressing CD64/16A. (A) Schematic of the genetic engineering of iPSCs and their differentiation to iNK-CD64/16A cells. (B) Flow cytometric phenotyping of CD16A and CD64 surface expression on unmodified iNK cells, iNK-CD64/16A cells, and healthly donor PBNK cells following a 2 week expansion with irradiated K562-mbIL21–41BBL feeder cells. Histograms are representative of several independent expansions. (C) Unmodified iNK cells, iNK-CD64/16A cells, and PBNK cells co-cultured with Raji cells at an E:T ratio of 8:1 with diluted rituximab, as indicated. (D) Unmodified iNK cells or iNK-CD64/16A cells were co-cultured with Raji cells at an E:T ratio of 8:1 in the presence of 0.01 μg/ml of rituximab with or without CD16A blocking mAb 3G8 F(ab′)₂ (5μg/ml). Percent cytotoxicity via ADCC at 18 hours is graphed (right), n = 3. (E) Unmodified iNK cells or iNK-CD64/16A cells co-cultured with Raji cells at the indicated E:T ratios in the presence of 5 μg/ml of rituximab. Percent cytotoxicity via ADCC at 48 hours is graphed (right), n = 3. Statistical significance was determined by ordinary one-way ANOVA with Tukey post hoc test. *p<0.05, ***p<0.001, ****p<0.0001, ns = not significant.

Figure 2

iNK-CD64/16A can be armed with mAbs to mediate potent ADCC. (A) iNK-CD64/16A cells were armed with biotinylated rituximab at the indicated concentrations, washed, and stained with a streptavidin fluorophore. Unarmed cells (0 μg/ml) were also stained with streptavidin fluorophore. Cell staining levels were analyzed by flow cytometry. (B) iNK-CD64/16A, iNK control cells, or healthy donor PBNK cells were armed with biotinylated rituximab (5 μg/ml), washed, stained with streptavidin fluorophore, and surface stained with CD64 antibody. Cells were gated on CD3- CD56+. Cell staining was analyzed by flow cytometry. (C) iNK-CD64/16A, iNK control cells or healthy donor PBNK cells were armed with rituximab (5 μg/ml), washed, and co-cultured with Raji-NLG cells at an E:T ratio of 8:1. Cytotoxicity via ADCC was assessed by live cell imaging. (D) iNK-CD64/16A cells were armed with the indicated mAbs, washed, and co-cultured with Raji-NLG cells at an E:T of 8:1. Cytotoxicity was assessed by live cell imaging. The percentage of live target cells remaining are displayed relative to targets alone at time 0. Viable Raji cells at 46 hours is graphed (right), n = 3. Statistical significance was determined by ordinary one-way ANOVA with Tukey post hoc test, ****p<0.0001.

Figure 3

Armed iNK-CD64/16A cells retain tumor-targeting antibody. (A) ADCC of Raji cells co-cultured with rituximab-armed (5 μg/ml) or unarmed iNK-CD64/16A at E:T ratio of 10:1 (left, Round 1) or left in the incubator at 37°C for 48 hours, washed, counted, and re-plated for co-culture with Raji cells at E:T ratio of 8:1 for 24 hours (right, Round 2). Staining for rituximab retention was performed following the first 48 hours (middle). (B) Rituximab-armed (5 μg/ml) or unarmed iNK cells co-cultured with Raji target cells at the indicated E:T ratios for 24 hours, effector cells were transferred with no adjustment for E:T for an additional two rounds of co-culture. Raji viability at 24 hours for rounds 1 and 2, and 48 hours for round 3 is graphed, n = 3. (C) Co-culture of Raji target cells with rituximab-armed (5 μg/ml) iNK-CD64/16A cells at an initial E:T ratio of 0.25:1 over two rounds of killing with no adjustment of E:T for round 2. Statistical significance was determined by paired two-tailed Student’s t-tests. *p<0.05, ***p<0.001, ****p<0.0001.

Figure 4

Armed iNK-CD64/16A cells target multiple tumor antigens. (A) CD19 and CD20 expressing Raji cells (red) and CD19 knockout (green) Raji cells were mixed in a 1:1 ratio and co-cultured with unarmed or armed iNK-CD64/16A cells for 48 hours at E:T ratio of 8:1. Loncastuximab armed + rituximab had rituximab added to the co-culture at 5 μg/ml. (B) Viable Raji cells at 48 hours, n = 3. Statistical significance was determined by ordinary one-way ANOVA with Tukey post hoc test. ***p<0.001, ****p<0.0001, ns = not significant.

Figure 5

iNK-CD64/16A cells are functional and mediate ADCC in presence of serum IgG. (A) iNK-CD64/16A cells were unarmed, armed with 5 μg/ml of rituximab (Armed iNK-CD64/16A-ritux) or had 5 μg/ml or rituximab added to the co-culture in media (iNK-CD64/16A-ritux added) containing 10% FBS or human AB serum against Raji target cells at an E:T ratio of 8:1 for 24 hours. Viable Raji cells at 24 hours are graphed (right), n = 3. (B) CD107a expression following co-culture of unarmed, rituximab-armed, or obinutuzumab-armed iNK-CD64/16A cells with Raji targets at an E:T ratio of 1:1 in 75% whole blood for 5 hours. Statistical significance was determined by ordinary one-way ANOVA with Tukey post hoc test. n = 3. ****p<0.0001, ns = not significant.

Figure 6

iNK-CD64/16A armed with rituximab mediates ADCC in vivo in human IgG xenograft mouse model. (A) Treatment schematic. NSG mice were i.v. injected with Raji-Luc cells (1x10⁵); i.p. administered GAMMAGARD (20 mg/mouse); i.v. administered iNK-CD64/16A cells (1x10⁷); and tumor burden was evaluated by BLI, as indicated. (B) Bioluminescence images through the first 3 weeks of the experiment. Tumor bearing mice were randomized into three groups: Raji alone (n=6); iNK-CD64/16A (n=6), which received unarmed iNK-CD64/16A cells; and iNK-CD64/16A-rituximab (n=5), which received iNK-CD64/16A cells armed with 5 μg/ml of rituximab. All cells were cryopreserved, thawed, washed, and then administered. GAMMAGARD was administered to all mice. (C) Graphical representation of cumulative BLI data. Statistical significance was determined by ordinary one-way ANOVA with Tukey post hoc test. (D) Kaplan-Meier curves for the in vivo Raji experiment. Statistical significance was determined by log-rank (Mantel-Cox) test. *p<0.05, **p<0.01.