Enhancing IgA-mediated neutrophil cytotoxicity against neuroblastoma by CD47 blockade

Abstract

Background

Approximately half of the neuroblastoma patients develop high-risk neuroblastoma. Current treatment involves a multimodal strategy, including immunotherapy with dinutuximab (IgG ch14.18) targeting GD2. Despite achieving promising results, the recurrence rate remains high and poor survival persists. The therapeutic efficacy of dinutuximab is compromised by suboptimal activation of neutrophils and severe neuropathic pain, partially induced by complement activation.

Methods

To enhance neutrophil cytotoxicity, IgG ch14.18 was converted to the IgA isotype, resulting in potent neutrophil-mediated antibody-dependent cell-mediated cytotoxicity (ADCC), without complement activation. However, myeloid checkpoint molecules hamper neutrophil cytotoxicity, for example through CD47 that is overexpressed on neuroblastomas and orchestrates an immunosuppressive environment upon ligation to signal regulatory protein alpha (SIRPα) expressed on neutrophils. In this study, we combined IgA therapy with CD47 blockade.

Results

In vitro killing assays showed enhanced IgA-mediated ADCC by neutrophils targeting neuroblastoma cell lines and organoids in comparison to IgG. Notably, when combined with CD47 blockade, both IgG and IgA therapy were enhanced, though the combination with IgA resulted in the greatest improvement of ADCC. Furthermore, in a neuroblastoma xenograft model, we systemically blocked CD47 with a SIRPα fusion protein containing an ablated IgG1 Fc, and compared IgA therapy to IgG therapy. Only IgA therapy combined with CD47 blockade increased neutrophil influx to the tumor microenvironment. Moreover, the IgA combination strategy hampered tumor outgrowth most effectively and prolonged tumor-specific survival.

Conclusion

These promising results highlight the potential to enhance immunotherapy efficacy against high-risk neuroblastoma through improved neutrophil cytotoxicity by combining IgA therapy with CD47 blockade.

Correspondence to Dr Jeanette Leusen; [email protected]

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Prior research has demonstrated IgA therapy as a viable alternative to IgG therapy in the treatment of neuroblastoma. The evolving landscape of checkpoint molecules compelled us to investigate the role of the CD47/signal regulatory protein alpha (SIRPα) axis as a checkpoint mechanism impeding IgA functionality, with the goal of advancing IgA therapy against neuroblastoma.

WHAT THIS STUDY ADDS

• This study confirms that the CD47/SIRPα interaction impedes IgA therapy efficacy, emphasizing its inhibitory role. Furthermore, our findings show that disrupting this axis via CD47 blockade significantly improves IgA’s therapeutic potential.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study’s findings suggest the capability for combining CD47 blocking antibodies with IgA therapy, providing a promising avenue for enhanced targeted antibody immunotherapy. This approach has the potential to outperform current treatments such as dinutuximab, providing a new direction for clinical studies.

Introduction

Neuroblastoma is the most prevalent extracranial solid tumor in childhood, often diagnosed before the age of 5.¹ Neuroblastoma emerges from neural crest cells forming the sympathetic nervous system and typically develops in the adrenal medulla or paraspinal ganglia.² Neuroblastomas are highly heterogeneous, with some tumors regressing spontaneously, while up to 50% of neuroblastoma patients develop high-risk neuroblastoma, with a 5-year survival rate of only 45%.^{3 4}

The current standard treatment for high-risk neuroblastoma involves surgery, chemotherapy with hematopoietic stem-cell rescue, radiation therapy, and immunotherapy.⁵ In 2015, the US Food and Drug Administration (FDA) approved dinutuximab (IgG ch14.18) for immunotherapy treatment of these patients. Dinutuximab is an antibody directed against the end-terminal penta-oligosaccharide of GD2, a disialoganglioside expressed on the cell surface. GD2 is expressed at low levels in normal tissues of the central nervous system and peripheral nerves but is highly expressed in many tumors of neuroectodermal origin, including neuroblastoma.^{6 7} GD2 is overexpressed in nearly all neuroblastoma patients, making it an attractive target for immunotherapy.⁸

Dinutuximab activates NK cells, macrophages and granulocytes against the tumor by Fc gamma receptor- (FcγR) mediated effector functions, including antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cellular phagocytosis and trogocytosis.⁹ While immunotherapy has shown promise, the increase in event-free survival at 2 years has only been 20%.¹⁰ This could be in part explained by the severe side effects, such as neuropathic pain, which limits the dose administered.¹⁰ A single point mutation, K332A in the Fc region of dinutuximab disrupted C1q binding and reduced neuropathic pain in some patients, suggesting complement activation is at least partly associated with these side effects.^{11 12} The neuropathic pain and high relapse rate highlight the need for enhancement in current treatment strategies.^{13 14}

While often overlooked, granulocytes can be important effector cells in immunotherapy, as it is the most abundant leucocyte population in blood.¹⁵ Granulocytes, consisting mostly of neutrophils, are activated by dinutuximab via FcγRIIa and induce cytotoxicity against neuroblastomas.^16–18 However, neutrophils express to a much higher level, FcγRIIIb, which is a GPI-linked receptor, and therefore, will not activate neutrophils, but can act as a sink for IgG antibodies.¹⁹ To improve the activation of neutrophils, we previously converted IgG ch14.18 into the IgA isotype.²⁰ IgA ch14.18 demonstrated robust neutrophil-mediated ADCC through the activation of the Fc alpha receptor (FcαRI, CD89) and does not bind to FcγRs.²⁰ Additionally, IgA does not activate the complement system and therefore the isotype switch alleviated the neuropathic pain as demonstrated in mouse models.²⁰ Subsequently, IgA ch14.18 was further engineered into IgA3.0 ch14.18 to enhance its stability and manufacturability, making it more suitable for potential clinical applications.²¹

Neuroblastoma is regarded as an immunologically cold tumor, distinguished by limited lymphocyte infiltration and a suppressive tumor microenvironment (TME).^{22 23} Nevertheless, IgA ch14.18 showed encouraging results in extending survival in neuroblastoma tumor-bearing mice, but analysis of the TME showed very poor attraction of myeloid cells on IgA therapy, implying that the TME is immunological suppressed.²¹ Checkpoint molecules play a crucial role in inhibiting the immune response within the TME. CD47 is a myeloid checkpoint molecule, overexpressed on the surface of tumor cells, including neuroblastoma and interacts with signal regulatory protein alpha (SIRPα) on neutrophils, monocytes and macrophages. This interaction promotes immunosuppression and shields the tumor from immunosurveillance.^24–27 The CD47/SIRPα axis is commonly referred to as a “don’t eat me” signal in macrophages, as it suppresses phagocytosis of the tumor cells.^{26 28} Similarly, engagement of CD47 with SIRPα on neutrophils has been shown to suppress ADCC and trogocytosis.^{29 30} Neutrophils are regulated by various stimulatory signals and inhibitory checkpoint molecules, which together dictate the fate of the neutrophil. Blocking CD47 removes a suppressive signal, shifting the balance toward a more antitumor phenotype, allowing neutrophils to become more actively engaged in tumor cell killing. IgG ch14.18 in combination with CD47 blockade enhanced the cytotoxicity against neuroblastoma by neutrophils.²⁷ However, additional granulocyte-macrophage colony-stimulating factor (GM-CSF) was required to stimulate granulocytes and achieve an optimal antitumor response, possibly due to suboptimal activation of neutrophils by IgG ch14.18.^{16 27} Currently, the combination of IgA ch14.18 and CD47 blockade in neuroblastoma has not been studied.

We hypothesize that IgA therapy in combination with CD47 blockade could be a novel strategy to tackle high-risk neuroblastoma. First of all, IgA activates neutrophils more effectively and outperforms IgG in neutrophil-mediated cytotoxicity.²⁰ Second, the lack of a C1q-binding site in IgA prevents complement-associated neuropathic pain.²⁰ Finally, in various other cancer models, such as breast cancer, leukemia and lymphoma, the combination of IgA with CD47 blockade demonstrated superior efficacy compared with the combination of IgG with CD47 blockade.^30–32 In our study, we aimed to investigate the effect of CD47 blockade on the efficacy of IgA antibody therapy targeting GD2 in neuroblastoma.

Materials and methods

Antibodies

IgG and IgA3.0 (hereafter called IgA) antibodies targeting GD2 (ch14.18, dinutuximab) were produced by WuXi Biologics in CHO-K1 cells. To block CD47, we either used an engineered human SIRPα D1 domain with high affinity for both mouse and human CD47 fused to IgG1 L234A/L235A/P329G (LALAPG) as described in Chernyavska et al³³ or an IgG1 LALAPG anti-CD47 (Clone 2.3D11) produced and purified in-house. Antibodies were produced and purified as described previously.^{20 34} In short, antibodies were produced through transient transfection of Expi-CHO-S cells (Thermo Fisher Scientific). Subsequently, supernatant containing the antibody was affinity purified using a HiTrap Protein A column (Cytiva) for IgG antibodies or a HiTrap KappaSelect column (Cytiva) for IgA antibodies, coupled to a ÄKTA liquid chromatography system (Cytiva). The captured IgG was eluted and subsequently dialyzed against PBS, while IgA was additionally purified by size exclusion chromatography. The eluate was filtered using a 0.22 µm filter and the antibody concentration was measured by UV absorbance at 280 nm using the corresponding extinction coefficient (ε280).

Cell lines

The human neuroblastoma cell lines, IMR32, SKNFI, SKNAS, LAN-5, GIMEN, were obtained from the American Type Culture Collection. All cells were maintained in DMEM (Dulbecco’s Modified Eagle Medium, Thermo Fisher Scientific) supplemented with 10% fetal calf serum, 100 U/mL penicillin-streptomycin (Pen/Strep, Gibco, life technologies), and 2% MEM non-essential amino acids solution (Thermo Fisher Scientific), hereafter called complete DMEM. All cell lines were grown at 37°C in a humidified incubator containing 5% CO₂. ExpiCHO-S cells (Thermo Fisher Scientific) were cultured in ExpiCHO Expression Medium (Gibco) at 37°C, 8% CO2, shaking at 125 rpm. Cells were not cultured past 20 passages, and they were regularly tested for mycoplasma contamination by a Mycoalert mycoplasma detection kit (Lonza).

AMC691B organoids

Patient-derived organoids AMC691B were kindly provided by Professor Jan Molenaar and generated as described previously.³⁵ The organoids were cultured in neuroblastoma organoid medium (table 1), consisting of DMEM with low glucose and Glutamax, supplemented with 20% Ham’s F-12 Nutrient Mix, B-27 supplement minus vitamin A (50×), N-2 supplement (100×), 100 U/mL penicillin, 100 µg/mL streptomycin, 20 ng/mL animal-free recombinant human EGF, 40 ng/mL recombinant human FGF-basic, 200 ng/mL recombinant human IGF-I, 10 ng/mL recombinant human PDGF-AA and 10 ng/mL recombinant human PDGF-BB, at 37°C with 5% CO₂ and were subcultured once or twice per week depending on growth rate. AMC691B transduced with GFP-luciferase constructs was generated previously.³⁶

Table 1

Composition of the organoid medium

Dulbecco’s modified Eagle’s medium with low glucose and Glutamax	Thermo Fisher	Cat#21 885-108
Ham’s F-12 Nutrient Mix	Thermo Fisher	Cat#31765027
B-27 supplement minus vitamin A (50×)	Thermo Fisher	Cat#12587010
N-2 supplement (100×)	Thermo Fisher	Cat#17502048
Recombinant human EGF	Peprotech	Cat#AF-100-15
Recombinant human FGF-2	Peprotech	Cat#100-18B
Recombinant human IGF-1	R&D	Cat#100-11
Recombinant human PDGF-AA	Peprotech	Cat#100-13A
Recombinant human PDGF-BB	Peprotech	Cat#100-14B

ADCC assays

⁵¹Cr release ADCC assays were performed as previously described.³⁷ In brief, target cells were labeled with 100 µCi chromium-51 (PerkinElmer) per million cells for at least 2 hours at 37°C and 5% CO₂. Labeled cells were washed three times with medium, before being pretreated with 10 µg/mL of SIRPα fusion protein for 30 min at room temperature to block CD47. Human polymorphonuclear leucocytes (PMNs) and peripheral blood mononuclear cells (PBMCs) were isolated from peripheral blood from healthy donors at the UMC Utrecht by Ficoll density gradient centrifugation. After collecting the PBMC layer, the remaining pellet was subjected to erythrocyte lysis using RBC Lysis Buffer (Biolegend) and PMNs were collected. Whole leucocytes were obtained by lysing the red blood cells from peripheral blood, followed by resuspending the remaining leucocytes in the original volume. The effector to target (E:T) ratio used for PMNs and PBMCs was 40:1 and 100:1, respectively. For whole leucocytes, a volume of 50 µL was added to the target cells in a final volume of 200 µL. Antibodies were added at the concentrations specified in each experiment. After 4 hours of incubation at 37°C in a humidified incubator containing 5% CO₂, the plate was centrifuged, and the supernatant was transferred to a lumaplate (PerkinElmer) to be measured on a beta-gamma counter for radioactive scintillation (in cpm) (PerkinElmer). Specific lysis was calculated using the formula: ((experimental cpm–basal cpm)/(maximal cpm–basal cpm))×100. The maximum cpm was determined by treating target cells with 5% Triton X-100 (Sigma-Aldrich), and the baseline cpm was determined by chromium release from target cells in the absence of antibodies and effector cells.

A luciferase-based ADCC assay was performed to measure ADCC against AMC691B organoids. Three days prior to the assay, organoids were dissociated into single cells, and 5000 cells were seeded in organoid medium into a solid white flat bottom 96-well plate (Corning). PBMCs and PMNs were freshly isolated from healthy donor blood on the day of the assay as described above. Effector cells were added to the wells at an E:T ratio of 40:1 for PMNs and 100:1 for PBMCs (5%–20% of the PBMCs are NK cells). CD47 was blocked by adding SIRPα fusion protein at a final concentration of 10 µg/mL. Subsequently, the plate was incubated for 4 hours at 37°C in a humidified incubator containing 5% CO₂. Five min before luminescence measurement, IVISbrite D-luciferin (XenoLight, Perkin Elmer) was added to the wells at a final concentration of 150 µg/mL. Luminescence was measured on the Spectramax M3 (Molecular Devices) with a measurement interval of 1 s and a settling time of 0.1 s. Specific lysis was calculated as: 100−(((experimental luminescence−basal luminescence)/maximal luminescence)×100), where basal luminescence was obtained from medium-only controls and maximal luminescence was obtained from organoids-only controls.

Flow cytometry

For the FACS staining, 1,00,000 cells were seeded and stained with the indicated antibodies in FACS buffer (PBS, 0.01% bovine serum albumin, 0.01% sodium-azide) for 45 min on ice. Expression level of SIRPα (15–414, Biolegend) on various immune cell subsets within whole peripheral blood was determined using the following antibodies to determine different immune cell subsets: Anti-CD45 (HI30, Biolegend), anti-CD19 (SJ25C1, BD Biosciences), anti-CD3 (UCHT1, Biolegend), anti-CD14 (MΦP9, BD Biosciences), anti-CD56 (NCAM16.2, BD Biosciences) and anti-CD66b (G10F5, BD Biosciences).

Surface expression of GD2 on neuroblastoma cell lines was determined using an anti-GD2 antibody (14G2a, Biolegend) while CD47 expression was measured using anti-CD47 (CC2C6, Biolegend) antibody.

To quantify the number of CD47 molecules per cell, we performed a QIFIKIT (Agilent/Dako) analysis according to the manufacturer’s instructions. Cells were stained with a saturating concentration of 10 µg/mL unconjugated mouse IgG monoclonal antibody directed against CD47 (CC2C6, Biolegend). Measurements were performed using a BD FACS Canto II flow cytometer.

Trogocytosis

Neuroblastoma cells were labeled with 5 µM DiO (Molecular Probes) according to manufacturer’s instructions. Labeled cells were co-cultured with PMNs at an E:T ratio of 10:1 in complete DMEM for up to 4 hours with 10 µg/mL IgG or IgA (ch14.18) and 10 µg/mL IgG1 LALAPG SIRPα in indicated experiments at 37°C in a humidified incubator containing 5% CO₂. After co-culture, cells were washed with PBS, and stained with anti-CD66b-V450 (G10F5, BD Biosciences). The frequency of DiO⁺ neuroblastoma/target cells taken up by CD66b⁺ neutrophils was measured by flow cytometry using a BD FACS Canto II flow cytometer.

CRISPR/Cas9 of CD47

CD47 was knocked out using CRISPR/Cas9 technology. A single guide RNA (sgRNA, AGCAACAGCGCCGCUACCAG, IDT) targeting CD47 or a scrambled negative control (IDT) was used to guide the Cas9 nuclease to the target site. IMR32 cells were prepared for electroporation with the Neon Transfection System (Thermo Fisher) following the manufacturer’s instructions. Three pulses were delivered at 1200 V and 20 ms to introduce the sgRNA and Cas9 complex into the cells. Knockout efficiency was determined by flow cytometry, and the IMR32-CD47KO population was enriched by sorting the CD47 negative cells using the BD FACSAria II cell sorter.

Live cell imaging

One day prior to the live cell imaging experiment, SKNAS cells were seeded in an eight well µ-slide (Ibidi). Tumor cells were labeled with 1 µM CellTrace CFSE according to the manufacturer’s protocol (Thermo Fisher Scientific). Next, 10 µg/mL of IgA ch14.18 in the absence or presence of 10 µg/mL of IgG1 LALAPG SIRPα fusion protein and 1 µM TO-PRO-3 (Thermo Fisher Scientific) were added to the target cells. After 15 min, primary PMNs were added. Cells were kept in an enclosed incubation chamber at 37°C and 5% CO₂ on a Deltavision RT widefield microscope (GE Healthcare) equipped with an Olympus 40×/1.35 NA oil immersion objective and a Cascade II 1K EMCCD/E2V CCD-201 camera. Recordings were made at 25 s intervals for 2 hours. The fluorescence intensity of TO-PRO-3 was quantified using Imaris (Bitplane).

AMC691B organoids were seeded as single cells in an eight-well µ-slide, 3 days prior to the experiment. PMNs were freshly isolated and stained with 1 µM eFluor 450 dye (Thermo Fisher Scientific) according to manufacturer’s instructions. For the experiment, 10 µg/mL of IgA ch14.18 in the absence or presence of 10 µg/mL of IgG1 LALAPG SIRPα fusion protein was added to the organoids. Additionally, 1 µM propidium iodide nucleic acid stain (Thermo Fisher Scientific) was added to visualize cell death. Live cell imaging (xyzt) was performed within an incubation chamber maintained at 37°C and 5% CO2 using a Stellaris 5 confocal microscope (Leica) equipped with an HC PL APO 20×/0.75 dry objective (Leica) and power HyD S detectors (410–850 nm). Recordings were made at 32 s intervals for 4 hours. Image analysis was performed in LAS X (Leica) and Imaris (Bitplane).

Subcutaneous IMR32 mouse model

In this experiment, we used human FcαRI (CD89) transgenic (Tg) mice, previously generated at the UMC Utrecht and backcrossed on a SCID (NOD.CB17-Prkdcscid/scid/Rj) background.³⁸ All mice were bred and maintained at Janvier Labs in Paris, France, before being transported to the University of Utrecht’s Central Laboratory Animal Research Facility for the experiment. Here, mice were housed in a controlled environment with a 12:12 hours light-dark cycle with access to food and water ad libitum. We used male transgene-negative (nTg) littermates for the solvent (PBS) control, IgG treatment and CD47 blockade only treatment groups. These treatment groups do not receive IgA and thus do not require the presence of CD89. Importantly, no differences in terms of tumor outgrowth are reported between the CD89 Tg mice and the non-Tg littermates, allowing combined use in the same experiment.³⁹ We used male mice ranging in age from 10 to 32 weeks in this experiment. Mice were acclimatized for at least 1 week prior to the experiment and randomized based on weight and age. The treatment and analysis were both double-blind.

IMR32 cells were mixed in a 1:1 solution of PBS and high concentration matrigel (Corning). 2.5 million cells in 150 µL were injected subcutaneously and tumor growth was monitored using a caliper to measure length, width, and height until the tumor size reached 1500 mm³ (humane end-point). Starting from day 2 after tumor cell injection, the neutrophil-depleting agent, mouse anti-Ly6G (clone 1A8, mIgG2a, Absolute Antibody) was administered intraperitoneally three times a week at a dose of 100 µg per mouse.⁴⁰ Treatment was started from day 4 and was given intraperitoneally. IgG ch14.18 was administered once a week at a dosage of 5 mg/kg, IgA ch14.18 three times a week at 25 mg/kg. A higher IgA dose was administered to compensate for its shorter half-life. These doses were tested previously and showed comparable serum titers over time.^{20 21} Additionally, IgG1 LALAPG SIRPα was injected intraperitoneally every 9 days at 30 mg/kg. As a solvent control, mice received PBS.

Analysis of the TME

When the tumor reached a size of 1500 mm³, mice were euthanized, and their tumor was collected in ice-cold PBS. Additionally, blood samples were obtained using the cheek puncture method and collected in lithium-heparin tubes (Sarstedt) for further analysis. Following tumor collection, the tumors were cut into smaller pieces and digested using the mouse tumor dissociation kit from Miltenyi. Up to 1 g of tumor tissue was processed following the manufacturer’s protocol. The dissociation was carried out using the 37C_m_TDK_1 program on a gentleMACS Octo Dissociator (Miltenyi). After the dissociation process, the samples were passed through a 100 µm cell strainer to obtain a single-cell suspension. Subsequently, both the blood samples (online supplemental table 1) and 2.5 million tumor cells (online supplemental table 2) were stained for FACS analysis.

Data processing and statistical analyses

Flow cytometry data were analyzed using FlowJo software (TreeStar). Statistical analyses were performed using GraphPad Prism V.9.3.0 (GraphPad Software). Specific statistical tests performed for each experiment are indicated in the corresponding figure legends. Data are presented as mean±SD or SEM and a p<0.05 was considered statistically significant. Graphs and figures were generated using the aforementioned software, Adobe Illustrator and Biorender. Imaging data was analyzed using Imaris (Bitplane).

Results

Neuroblastoma cell lines express high levels of GD2 and CD47

The current standard treatment regimen for treating high-risk neuroblastoma with immunotherapy uses dinutuximab (IgG ch14.18). To evaluate the potential benefits of combining IgA ch14.18 with CD47 blockade therapy, we initially verified the coexpression of GD2 and CD47 in neuroblastoma patients. However, a complication factor is that GD2 is not a protein, and therefore, not encoded by a gene. Instead, B4GALNT1, encoding for GD2 synthase, plays a key role in GD2 synthesis. Additionally, ST8SIA1, responsible for GM3 synthase, catalyzes the initial steps in ganglioside synthesis. Our analysis of these genes revealed the coexpression of CD47 with both B4GALNT1 and ST8SIA1 in patient samples (online supplemental figure 1), suggesting a potential benefit of combining IgA ch14.18 with CD47 blockade. Subsequently, we assessed the expression levels of GD2 in a panel of commonly used neuroblastoma cell lines (figure 1A). Additionally, to investigate whether the CD47/SIRPα axis could play a role in neuroblastoma, we quantified CD47 expression levels using QIFIKIT analysis (figure 1B). Flow cytometry analysis revealed high expression levels of GD2 as compared with the GD2-negative GIMEN cells. Moreover, all neuroblastoma cell lines expressed high levels of CD47. Next, we assessed the expression of SIRPα, the receptor for CD47, in immune cell subsets in the blood. High RNA levels of SIRPA, encoding SIRPα, were observed in eosinophils and neutrophils, and to a lesser extent in various monocyte subsets and myeloid dendritic cells, according to data from the Human Protein Atlas (figure 1C). Moreover, SIRPα protein expression was analyzed in peripheral blood from healthy donors, revealing expression limited to monocytes and neutrophils (figure 1D), confirming that SIRPα expression is limited to the myeloid compartment. These findings suggest that targeting GD2 using IgA and simultaneously blocking CD47 may be an effective therapeutic strategy in neuroblastoma.

Figure 1. Expression levels of GD2 and CD47 in neuroblastoma cells and SIRP[alpha] in leucocytes. (A) The expression level of GD2 on various neuroblastoma cell lines was determined using flow cytometry. Data are shown as MFI+-SD. (B) The absolute expression level of CD47 was determined through QIFIKIT analysis using 10 [micro]g/mL anti-CD47 antibody (Clone: CC2C6). ABC for n=3 experiments is shown as mean+-SD. (C) RNA expression level of SIRPA across different immune cells, depicted in normalized transcript Per million (nTPM). Data are retrieved from the Human Protein Atlas ( www.proteinatlas.org ), accessed on April 29, 2024 (D) The expression level of SIRP[alpha] on leucocytes in blood from three healthy donors was evaluated. MFI values were normalized to the MFI value measured on T cells and are depicted as mean+-SD. MFI, mean fluorescence intensity; ND, not detected.

CD47 blockade enhanced ADCC by neutrophils

To investigate whether the CD47/SIRPα axis suppresses neutrophil function in IgA-mediated killing, we tested three different methods that disrupt this myeloid checkpoint: (1) an engineered human SIRPα domain 1 protein fused to an inactivated IgG1 Fc, hereafter SIRPα fusion protein, (2) an anti-CD47 antibody (Clone: 2.3D11), or (3) CD47 knockout (online supplemental figure 2) of the neuroblastoma cell line IMR32. ADCC induction by IgA in combination with either of these blocking strategies was tested in a ⁵¹Cr release assay. IgA ch14.18 alone induced a modest killing efficiency of up to 25% with whole leucocytes as effector cells, but this was significantly improved by disrupting the CD47 axis, which increased killing by more than twofold from 25% to over 50% (figure 2A). All three strategies disrupting the CD47/SIRPα axis showed comparable results, but we chose to continue with the SIRPα fusion protein due to its versatility in binding both human and mouse CD47, making it convenient for both in vitro and in vivo studies.

Figure 2. CD47 blockade enhanced ADCC capacity by neutrophils. (A) The specific lysis of IMR32 (-CD47KO) cells was evaluated in a 51 Cr release assay using whole leucocytes. IgA ch14.18 was added in titrated concentrations in the presence or absence of 10 [micro]g/mL CD47 blocking antibody (IgG1 LALAPG SIRP[alpha] or IgG1 LALAPG 2.3D11). (B) PBMC-mediated ADCC (E:T 100:1) of IMR32 cells was determined using IgG or IgA ch14.18 at different concentrations (0.001-10 [micro]g/mL) in the presence or absence of 10 [micro]g/mL IgG1 LALAPG SIRP[alpha]. (C) Specific lysis of neuroblastoma cell lines was evaluated in an ADCC assay using PMNs (E:T 40:1) and (D) whole leucocytes using IgA or IgG ch14.18 with or without preincubation with 10 [micro]g/mL IgG1 LALAPG SIRP[alpha]. (E) The fold increase in ADCC capacity on CD47 blockade. In all 51 Cr release ADCC assays, the lysis was evaluated after a 4-hour incubation. PMNs, PBMCs, and whole leucocytes were isolated from healthy donors’ blood. The mean+-SD specific lysis of a technical triplicate of a single donor is shown. Representative graphs of at least three independent experiments. ns>0.05, *p<0.05, **p<0.01, ***p<0.001, determined by two-way ANOVA followed by Tukey’s post hoc test. Ab, antibody; ADCC, antibody-dependent cell-mediated cytotoxicity; ANOVA, analysis of variance; PBMCs, peripheral blood mononuclear cells; PMNs, polymorphonuclear leucocytes; WLs, whole leucocytes.

When PBMCs were used as effector cells, we did not observe any improvement of either IgA or IgG ch14.18 on CD47 blockade (figure 2B), consistent with the fact that PBMC-mediated lysis is primarily driven by natural killer cells that lack SIRPα expression in our hands (figure 1C). In contrast, CD47 blockade enhanced both IgA and IgG ch14.18-induced lysis of all tested neuroblastoma tumor cells when PMNs were used as effector cells (figure 2C). IgA was more effective in inducing PMN-mediated ADCC than IgG. CD47 blockade improved both IgG-mediated and IgA-mediated lysis, but the combination with IgA resulted in the highest lysis. Using whole leucocytes as effector cells, we observed similar if not higher lysis mediated by IgA compared with IgG in the presence of the SIRPα fusion protein (figure 2D). In this assay, IgA-mediated lysis was primarily induced by neutrophils, and to a lesser degree monocytes, while IgG-mediated lysis was induced by NK cells, neutrophils and possibly monocytes. Moreover, the fold increase in lysis as a result of CD47 blockade was higher for IgA compared with IgG (figure 2E), showing that IgA can benefit more from CD47 checkpoint inhibition, likely due to the high SIRPα expression on neutrophils.

IgG-induced and IgA-induced trogocytosis of neuroblastoma cells

To characterize the tumor cell killing mechanism induced by the combination of GD2-targeting antibodies and CD47 blockade, we used DiO-labeled neuroblastoma tumor cells to identify the interaction between the tumor cells and PMNs. DiO, a lipophilic dye that incorporates into the cell membrane, allowed us to visualize the process of trogocytosis, whereby PMNs ingest pieces of the tumor cells (figure 3A). We observed that PMNs gradually became DiO-positive over time after incubation with the tumor cells, indicating the uptake of tumor cell fragments by trogocytosis (figure 3B). We observed increasing levels of trogocytosis over time in all experimental conditions. Surprisingly, IgG and IgA induced similar trogocytosis by PMNs against IMR32 cells. With SKNAS cells, IgA induced more trogocytosis compared with IgG. When combined with CD47, the trogocytosis capacity was increased for both IgG and IgA. This process ultimately led to tumor cell death, as shown in the ADCC assays (figure 2).

Figure 3. IgG and IgA-induced trogocytosis of neuroblastoma cells. (A) Neuroblastoma cells were labeled with DiO and co-cultured with IgG or IgA ch14.18 and PMNs (E:T 10:1) for up to 4 hours. PMNs were isolated from healthy donors’ blood. Trogocytosis was quantified by the frequency of uptake of DiO+staining by CD66b + neutrophils. (B) Trogocytosis of IMR32 (left) and SKNAS (right) cells using 10 [micro]g/mL IgG or IgA ch14.18 in the presence or absence of 10 [micro]g/mL IgG1 LALAPG SIRP[alpha] was measured for up to 4 hours. The mean+-SD of a technical triplicate of a single donor is presented. One representative graph represents at least n=3 independent experiments. The statistical analysis was performed using two-way ANOVA followed by Tukey’s post hoc test. ns>0.05, **p<0.01, ***p<0.001. ANOVA, analysis of variance; E:T, effector to target; Freq, frequency; PMNs, polymorphonuclear leucocytes.

IgA in combination with CD47 blockade induced swarming of neutrophils and killing of neuroblastoma cells

In the previous ⁵¹Cr release ADCC assays (figure 2C,E), SKNAS showed the highest resistance to IgA-induced lysis. Moreover, CD47 blockade on these cells resulted in the highest increase in IgA-mediated lysis by PMNs. Following that, we visualized the process of tumor cell killing by IgA-activated PMNs in a live imaging assay using widefield microscopy (figure 4A). CFSE-labeled SKNAS neuroblastoma cells (green) were treated with IgA and SIRPα fusion protein, which efficiently induced tumor cell death by PMNs as marked by red TO-PRO-3 staining (figure 4B, online supplemental video 1). Notably, the addition of CD47 blockade to IgA resulted in neutrophil swarming directed toward the tumor cells within a couple of minutes, characterized by increased migration (figure 4D). The combination treatment showed tumor cell killing after 30 min of incubation while IgA alone only induced minor cell death and CD47 block alone did not result in any cell death (figure 4C). These live cell imaging results emphasize the significance of disrupting the CD47/SIRPα axis, enabling strong activation of neutrophils by IgA. This is characterized by the swarming behavior, ultimately leading to the effective killing of tumor cells.

Figure 4. Live cell imaging of SKNAS cell killing by IgA. (A) Live cell imaging of SKNAS cells stained with CFSE (green) and co-cultured with PMNs (E:T 40:1) and 10 [micro]g/mL IgA ch14.18, with or without 10 [micro]g/mL of IgG1 LALAPG SIRP[alpha]. Dead cells were visualized using TO-PRO-3 (red). A 2-hour time-lapse was captured with a 25 s interval on a widefield microscope to visualize tumor killing. (B) Representative images of SKNAS cells at 0, 30, 60, and 120 min. White arrows denotes examples of neutrophil swarming. (C) Cell lysis was determined by quantifying the mean TO-PRO-3 fluorescence intensity. (D) The migration speed of PMNs was quantified using IMARIS analysis software. Statistical analysis was performed using a one-way ANOVA followed by Dunnett’s post hoc test, ns>0.05, ***p<0.001. ANOVA, analysis of variance; E:T, effector to target; PMNs, polymorphonuclear leucocytes.

Enhanced IgA-mediated killing of neuroblastoma organoids by CD47 blockade

In addition to ADCC assays using cell lines, we evaluated the efficacy of our approach in a more complex three-dimensional (3D) model representative of neuroblastoma tumors by performing killing assays with patient-derived organoids (AMC691B) (figure 5A). Organoids provide a 3D tumor model mimicking the bulkiness of solid tumors. AMC691B organoids have comparable GD2, but lower CD47 expression than IMR32 cells (figure 5B). IgG-mediated cytotoxicity against AMC691B organoids was low when PMNs were used as effector cells, whereas IgA was more effective, achieving up to 40% lysis at the highest antibody concentration (figure 5C). When PBMCs were used as effector cells, IgG killing was more effective, resulting in comparable lysis to IgA in a PMN ADCC. Consistent with our findings using cell lines, CD47 blockade enhanced antibody-mediated killing by PMNs of AMC691B organoids, but no improvement was observed with antibody-mediated killing by PBMCs. Moreover, among all tested conditions, the combination of IgA and CD47 blockade was the most effective in inducing tumor cell killing by PMNs, achieving lysis as high as 70%. Neutrophil cytotoxicity against the organoid was further assessed with live cell imaging (figure 5D, online supplemental video 2). Addition of IgA rapidly induced killing of the organoid, however, efficient killing only lasted for 20 min (figure 5E). When the CD47 checkpoint inhibitor was added, IgA cytotoxicity was more persistent and effective killing was observed for 90 min. Overall, the combination strategy induced more lysis of the tumor cells, which was accompanied by increased swarming of neutrophils, as also observed with SKNAS cells. Our observations that IgA ch14.18 with CD47 blockade can enhance immune responses even against 3D tumor structures further support the therapeutic potential of this combination strategy.

Figure 5. Enhanced IgA-mediated killing of neuroblastoma organoids by CD47 blockade. (A) Representative image of neuroblastoma organoid line AMC691B 3 days after seeding. (B) Flow cytometric analysis of GD2 and CD47 expression levels on AMC691B compared with IMR32 cells. Expression was shown as MFI value+-SD. (C) Specific lysis of neuroblastoma organoids determined in a 4-hour luciferase-based ADCC assay using PMNs (E:T 40:1) or PBMCs (E:T 100:1) from healthy donors. A concentration range of IgA or IgG ch14.18 was added in the presence or absence of 10 [micro]g/mL SIRP[alpha] fusion protein. The mean+-SD specific lysis of a technical triplicate of a single donor is shown. One representative graph represents at least n=3 independent experiments. Significance levels are indicated as ns>0.05, ***p<0.001, determined by two-way ANOVA followed by Tukey’s post hoc test. (D) Live cell imaging of AMC691B-GFP + . Organoids were co-cultured with eFluor 450 (blue) stained PMNs (E:T 40:1). IgA-mediated lysis was recorded after addition of 10 [micro]g/mL IgA ch14.18 and 10 [micro]g/mL SIRP[alpha] fusion protein. Cell lysis was visualized using propidium iodide (red). A 4-hour time-lapse was captured with a 32 s interval on the Stellaris five confocal microscope. Representative images at 0, 1, 2, 3 and 4 hours. (E) Cell lysis was determined as means of total propidium iodide intensity. Fold changes from the intensity measured at t=0 hour are calculated and shown. Ab, antibody; ADCC, antibody-dependent cell-mediated cytotoxicity; ANOVA, analysis of variance; E:T, effector to target; MFI, mean fluorescence intensity; PBMCs, peripheral blood mononuclear cells; PMNs, polymorphonuclear leucocytes.

Checkpoint inhibition prolonged survival of IMR32 tumor-bearing mice

To assess whether the enhanced tumor cell killing on CD47 blockade could be translated into the in vivo setting, we performed a mouse experiment using CD89 transgenic SCID mice inoculated with IMR32 tumor cells. Previously, we could delay tumor outgrowth in this tumor model with a dose of 60 mg/kg of IgA ch14.18.²¹ In the present experiment, we administered a suboptimal dose of 25 mg/kg IgA ch14.18, which previously failed to improve survival in tumor-bearing mice, and again it showed no observable differences in tumor outgrowth when compared with the solvent group (figure 6A). In comparison, 5 mg/kg IgG slightly delayed tumor outgrowth in these mice. To explore the therapeutic potential of CD47 blockade in mice, we intraperitoneally administered the SIRPα fusion protein at a dose of 30 mg/kg. Combining 25 mg/kg IgA with CD47 blockade resulted in significant suppression of tumor outgrowth, which corresponded to significantly prolonged survival of the tumor-bearing mice (figure 6B). Remarkably, by day 60, 50% of the mice in the IgA and CD47 blockade combination group were still alive, compared with survival rates of 0% and 27% in the IgG and IgG+CD47 blockade treatment groups, respectively. To assess the contribution of neutrophils to the IgA-mediated antitumor response, we depleted neutrophils in the IgA and CD47 blockade combination treatment group with intraperitoneally injections of 100 µg mouse anti-Ly6G thrice a week (online supplemental figure 3B). The therapeutic effect was completely lost when neutrophils were depleted, and tumor outgrowth matched that of the solvent control group, highlighting the essential role of neutrophils for IgA-mediated therapy.

Figure 6. Combination of IgA and CD47 blockade suppressed subcutaneous IMR32 tumor outgrowth in SCID mice. (A) Male SCID mice were subcutaneously injected with 2.5 million IMR32 tumor cells on day 0. Neutrophil depletion was started on day 2, and treatment began on day 4, following the treatment scheme outlined in the methods section. Tumor progression was monitored over time by measuring subcutaneous tumor dimensions with a caliper. Mice were euthanized once tumors reached a volume of 1500 mm 3 . Tumor growth data are presented up to the day when the first mice reached the humane end-point, which occurred on day 39. Tumor sizes are shown as mean +-SEM of 9-13 mice per treatment group. *p<0.05, **p<0.01, ***p<0.001, using two-way ANOVA. Asterisks indicated significance compared with solvent control group. (B) Kaplan-Meier survival plot of IMR32 tumor-bearing mice. **p=0.008, ***p<0.001 was determined with the log-rank (Mantel-Cox) test. Asterisks indicated significance compared with solvent control group. (C) Immune cell composition in blood was obtained through a cheek puncture on days 16 and 33, and subsequently analyzed using flow cytometry. Data from both days are pooled and shown as mean +-SD of 4-6 mice per group. No significant differences were measured as analyzed with one-way ANOVA. (D) Immune cell composition within the tumor was assessed using flow cytometry at the humane end-point. Data are shown as mean +-SD of 3-4 mice per treatment group. *p<0.05, **p<0.01, ***p<0.001, using one-way ANOVA. Asterisks indicated significance compared with solvent control group. ANOVA, analysis of variance; s.c., subcutaneous.

Given the myeloid checkpoint inhibitor’s high affinity for both human and mouse CD47, it effectively replicates the antibody sink challenge currently encountered by clinically applied anti-CD47 antibodies like magrolimab. Therefore, our model provides a realistic representation of the clinical setting. Despite the high affinity for CD47, administration of the SIRPα fusion protein alone did not show any therapeutic effect due to the Fc dead tail. Moreover, in an earlier pilot study, administration of the SIRPα fusion protein did not affect the red blood cell count in mice (online supplemental figure 3A), which is vital since red blood cell-related toxicities are a major concern with CD47 checkpoint inhibitors.

To examine the effect of our treatment strategies on the immune cell composition, we used flow cytometry to analyze both blood and the tumor (figure 6C,D, online supplemental figure 4). There were no significant differences in immune cell composition in the blood on the different treatment regimens.

In the tumor, IgA treatment slightly elevated the number of myeloid cells, although not significant. More notably, IgA in combination with CD47 blockade triggered a surge in the myeloid cell count in the tumor (figure 6D). Neutrophil, monocyte, eosinophil and macrophage counts were all significantly elevated, with increases ranging from 7-fold to 20-fold compared with PBS control. The majority of myeloid cells were neutrophils, which accounted for more than half of the total myeloid cell population. Additionally, the evaluation of SIRPα expression levels on these immune cells across the different treatment groups showed no significant differences (online supplemental figure 5). Our data imply that by disrupting the CD47/SIRPα axis, the immunosuppressive environment is relieved, allowing IgA to effectively activate and recruit neutrophils and other myeloid cells to the tumor.

To explore the broader application of combined therapy beyond neuroblastoma, we also tested our combination strategy in an epidermoid carcinoma mouse model. Consistent with the observations in the neuroblastoma model, combining IgA EGFR with CD47 blockade enhanced the tumor response, as demonstrated by the hampered tumor growth and improved survival of tumor-bearing mice (online supplemental figure 6). The combination treatment strategy resulted in a remarkable fourfold enhancement in the survival rates of A431 tumor-bearing mice. This further supports the potential of our combined therapy approach across different cancer types.

Discussion

This study demonstrates a superior therapy combining IgA ch14.18 with CD47 blockade compared with the original IgG ch14.18 therapy in a preclinical neuroblastoma mouse model. The isotype switch of dinutuximab from IgG to IgA resulted in improved neutrophil recruitment and activation of ADCC. Despite effective neutrophil activation by IgA, neuroblastoma, like many other cancers, overexpresses checkpoint molecules as an evasion strategy. All of the neuroblastoma cell lines included here, which were originally derived from high-risk patients, showed CD47 overexpression. This is consistent with the CD47 expression observed in neuroblastoma patient samples.^{27 41} Additionally, SIRPα protein levels were higher in colon cancer tissue compared with adjacent healthy tissue, although the specific immune cells were not identified.⁴² Another study showed increased SIRPA gene expression in tumor-associated macrophages in colorectal cancer tissues, with a notable but not statistically significant increase in neutrophils.⁴³ The protein expression of SIRPα on tumor-associated neutrophils in neuroblastoma remains unidentified, as the antitumor role of neutrophils has only recently gained interest as a significant immune cell population. However, investigating this in future studies would certainly be valuable.

The interaction of CD47 on the tumor cells with SIRPα on neutrophils reduced the ADCC capacity as shown in chromium release assays. Notably, SKNAS and SKNFI, which had lower GD2 levels, benefited the most from the combination of IgA therapy and CD47 blockade. Presumably, the relatively weaker IgA-mediated neutrophil activation due to the lower antigen abundance made these cells more sensitive to CD47/SIRPα resistance. This emphasizes the importance of checkpoint inhibition, particularly in neuroblastoma patients with low GD2 expression.^{44 45}

Our findings demonstrated that disrupting the CD47/SIRPα axis enhanced neutrophil-mediated ADCC by both IgG and IgA. However, ADCC by PBMCs was not improved, probably due to the lack of SIRPα expression on NK cells, the most important effector cell of IgG antibody therapy in the PBMC fraction. Both IgG and IgA ch14.18 induced significant trogocytosis, which was further improved by CD47 blockade. Surprisingly, despite the similar frequency of trogocytosis induced by IgG and IgA, it did not translate to comparable ADCC by neutrophils, especially evident when targeting IMR32 cells. Specifically, IgG-mediated lysis of IMR32 cells by neutrophils was lower than that of IgA. These observations suggest that IgG-activated neutrophils like IgA-activated neutrophils take small bites of the cell membrane, however, with insufficient strength to disrupt the membrane’s integrity and induce cell death. This could be explained by weaker ITAM signaling as a result of the 1:1 stoichiometry observed with IgG vs the 1:2 stoichiometry observed with IgA. Consequently, IgA activates four ITAMs as opposed to the two ITAMs activated by IgG.¹⁹ Furthermore, when trogocytosis does not result in induced cell death, the cell membrane may be repaired via repair mechanisms and antibody-bound antigen is shaved from the cell surface, as seen with CD20.^{46 47} It would not be surprising if GD2 experienced something comparable, eventually leading to relapse in patients.^{45 48}

The combination of IgA ch14.18 and CD47 blockade significantly increased neutrophil cytotoxicity against neuroblastoma cell lines and organoids, which was underscored by data from our in vivo study. In our subcutaneous IMR32 tumor model, combining IgA therapy with CD47 blockade hampered tumor outgrowth, contributing to a significant improved survival rate of tumor-bearing mice. The therapeutic response was primarily attributed to the presence of neutrophils, as the depletion of neutrophils abolished the observed therapeutic effect. The combination of IgA therapy and CD47 blockade could essentially maintain a positive feedback loop in which robust neutrophil activation triggers the release of cytokines and chemokines, which attract more immune cells directed against the tumor cells.

We used an in-house produced CD47 blocking protein consisting of an engineered SIRPα D1 domain fused to IgG1 with LALAPG mutations to ablate Fc functions.^{33 49} The SIRPα fusion is able to target human CD47 on the IMR32 tumor as well as the ubiquitously expressed mouse CD47 on healthy murine cells. Since CD47 is widely expressed in healthy cells, including red blood cells, CD47 antibodies are rapidly captured, requiring a high dose to achieve effective concentrations at the tumor site, which is known as the CD47 “sink” effect.⁵⁰ As a result, a relatively high dose of 30 mg/kg was required as determined by Kauder et al.⁵¹ It is worth noting that, despite the high dose given to the mice, there was no decrease in red cell count, implying that including an Fc dead tail in the CD47 targeting strategy mitigated potential side effects such as anemia. Furthermore, CD47 block alone did not induce apparent differences in leucocyte composition in blood and within the tumor. To improve clinical application, the problems caused by the CD47 “sink” effect could be circumvented by using novel targeting strategies, such as bispecific antibodies or QPCTL inhibitors.^{24 52}

Although the addition of dinutuximab to the treatment protocol for high-risk neuroblastoma patients has improved 5-year survival rates, recurrences remain a challenge.¹³ Dinutuximab has been shown to promote macrophage-induced phagocytosis as well as neutrophil-mediated ADCC.^{16 20 21 53 54} However, the efficacy of IgG in activating myeloid cells remains suboptimal, leaving room for improvement. As a result, combination strategies involving CD47 blockade were investigated, leading to the initiation of a phase 1 clinical trial (NCT04751383) examining the use of dinutuximab and the CD47 monoclonal antibody, magrolimab, for pediatric neuroblastoma and osteosarcoma patients.^{26 27} Unfortunately, due to unexpected toxicities, this trial had to be suspended. The specific clinical implications of this outcome are unknown. Alternatively, using IgA in combination with CD47 blockade was found to be more effective in our study in extending tumor-specific survival than using IgG alone or IgG in combination with CD47 blockade. This suggests that IgA therapy in combination with CD47 checkpoint inhibition may be a more effective treatment option than IgG.

Additionally, infiltration of myeloid cells is often associated with restricted lymphocyte responses. However, CD47 blockade indirectly modulates the adaptive immune system by enhancing cross-priming of cytotoxic T cells by myeloid cells.⁵⁵ Disruption of the CD47/SIRPα axis promotes tumor cell engulfment by dendritic cells and macrophages, followed by antigen presentation.^{51 56 57} Moreover, there is emerging evidence suggesting that neutrophils might engage in antigen presentation on taking up antigens through trogocytosis.^{55 58} In essence, targeting the CD47/SIRPα axis serves as a bridge connecting the innate and adaptive immune system, facilitating the infiltration of immune cells into the immunologically “cold” neuroblastoma tumors. Furthermore, CD47 can directly hinder T cell activation by interacting with the checkpoint molecule thrombospondin-1 on T cells.⁵⁹ However, due to the limitations of our immunocompromised mouse model, we were unable to investigate the effect of CD47 blockade on the adaptive immune system. Exploring the effect of CD47 blockade on the adaptive immune response in immunocompetent mouse models would be a valuable direction for future research.

In conclusion, our study showed a promising interplay between IgA-mediated therapy and CD47 blockade, providing improved immune responses against neuroblastoma. Our results demonstrated that a combined strategy significantly improved immune cell infiltration in an otherwise immune-suppressed TME. In vivo, strong induction and activation of neutrophils by IgA restrained tumor growth and prolonged survival. The combination of IgA ch14.18 and CD47 blockade emerges as a potent therapeutic option for high-risk neuroblastoma patients.

We would like to thank the flow cytometry and imaging facility at the CTI, UMCU and Corlinda ten Brink from the Cell Microscopy Core, Department of Cell Biology, for the assistance and services. We thank the MDD at the UMCU for their services and the GDL laboratory for the excellent care of the laboratory animals.

Data availability statement

Data are available on reasonable request. The data generated in this study are available on request from the corresponding author.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study involves human participants and was approved by UMC Utrecht 07-125. Participants gave informed consent to participate in the study before taking part. All animal experiments followed international guidelines and were approved by the national Central Authority for Scientific Procedures on Animals (CCD) and the local experimental Animal Welfare Organization (AVD115002016410).

Footnote

GvT and JL contributed equally.

Contributors Conceptualization: CC, MS, GvT and JL. Methodology: CC, MS, MN and MJ. Formal Analysis: CC and MS Investigation: CC, MS, MN, MJ, EP and FvdH. Resources: FvdH and JW. Writing—original draft: CC; Writing—review and editing: CC, MS, MN, MJ, JW, GvT and JL. Supervision: GvT and JL. Funding acquisition: JL. Guarantor: JL.

Funding CC and MJ are funded by grant #11944 from The Dutch Cancer Society (KWF). MS and MN are funded by Villa Joep (project 17 IgA and anti-GD2).

Competing interests CC, MS, MN, MJ, EP, FvdH, JW and GvT have nothing to disclose. JL is the scientific founder and a shareholder of TigaTx.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ Smith MA, Seibel NL, Altekruse SF, et al. Outcomes for children and adolescents with cancer: challenges for the twenty-first century. J Clin Oncol 2010; 28: 2625–34. doi:10.1200/JCO.2009.27.0421

² Aygun N. Biological and genetic features of neuroblastoma and their clinical importance. Curr Pediatr Rev 2018; 14: 73–90. doi:10.2174/1573396314666180129101627

³ Cheung N-KV, Dyer MA. Neuroblastoma: developmental biology, cancer Genomics, and Immunotherapy. Nat Rev Cancer 2013; 13: 397–411. doi:10.1038/nrc3526

⁴ 2019 neuroblastoma - childhood: statistics. 2020.

⁵ Chung C, Boterberg T, Lucas J, et al. Neuroblastoma. Pediatr Blood Cancer 2021; 68: e28473. doi:10.1002/pbc.28473

⁶ Sait S, Modak S. Anti-Gd2 Immunotherapy for Neuroblastomas. Expert Rev Anticancer Ther 2017; 17: 889–904. doi:10.1080/14737140.2017.1364995

⁷ Ladenstein R, Pötschger U, Valteau-Couanet D, et al. Interleukin 2 with anti-Gd2 antibody Ch14.18/CHO (Dinutuximab beta) in patients with high-risk neuroblastoma (HR-Nbl1/SIOPEN): a Multicentre, randomised, phase 3 trial. Lancet Oncol 2018; 19: 1617–29. doi:10.1016/S1470-2045(18)30578-3

⁸ Paret C, Ustjanzew A, Ersali S, et al. Gd2 expression in Medulloblastoma and neuroblastoma for personalized Immunotherapy: A matter of subtype. Cancers (Basel) 2022; 14: 6051. doi:10.3390/cancers14246051

⁹ Morandi F, Sabatini F, Podestà M, et al. Immunotherapeutic strategies for neuroblastoma: present, past and future. Vaccines (Basel) 2021; 9: 43. doi:10.3390/vaccines9010043

¹⁰ Yu AL, Gilman AL, Ozkaynak MF, et al. Anti-Gd2 antibody with GM-CSF, Interleukin-2, and Isotretinoin for neuroblastoma. N Engl J Med 2010; 363: 1324–34. doi:10.1056/NEJMoa0911123

¹¹ Navid F, Sondel PM, Barfield R, et al. Phase I trial of a novel anti-Gd2 Monoclonal antibody, Hu14.18K322A, designed to decrease toxicity in children with refractory or recurrent neuroblastoma. J Clin Oncol 2014; 32: 1445–52. doi:10.1200/JCO.2013.50.4423

¹² Sorkin LS, Otto M, Baldwin WM, et al. Anti-GD(2) with an FC point Mutation reduces complement fixation and decreases antibody-induced allodynia. Pain 2010; 149: 135–42. doi:10.1016/j.pain.2010.01.024

¹³ Maris JM. Recent advances in neuroblastoma. N Engl J Med 2010; 362: 2202–11. doi:10.1056/NEJMra0804577

¹⁴ London WB, Castel V, Monclair T, et al. Clinical and biologic features predictive of survival after relapse of neuroblastoma: A report from the International neuroblastoma risk group project. J Clin Oncol 2011; 29: 3286–92. doi:10.1200/JCO.2010.34.3392

¹⁵ Rosales C. Neutrophil: A cell with many roles in inflammation or several cell types? front Physiol. Front Physiol 2018; 9: 113. doi:10.3389/fphys.2018.00113

¹⁶ Chen RL, Reynolds CP, Seeger RC. Neutrophils are cytotoxic and growth-inhibiting for neuroblastoma cells with an anti-Gd2 antibody but, without cytotoxicity, can be Growth- stimulating. Cancer Immunol Immunother 2000; 48: 603–12. doi:10.1007/s002620050008

¹⁷ Cheung N-KV, Sowers R, Vickers AJ, et al. Fcgr2A polymorphism is correlated with clinical outcome after Immunotherapy of neuroblastoma with anti-Gd2 antibody and granulocyte macrophage colony-stimulating factor. J Clin Oncol 2006; 24: 2885–90. doi:10.1200/JCO.2005.04.6011

¹⁸ Metelitsa LS, Gillies SD, Super M, et al. Antidisialoganglioside/granulocyte macrophage–colony-stimulating factor fusion protein facilitates neutrophil antibody-dependent cellular cytotoxicity and depends on FcγRII (Cd32) and Mac-1 (Cd11B/Cd18) for enhanced Effector cell adhesion and Azurophil granule Exocytosis. Blood 2002; 99: 4166–73. doi:10.1182/blood.v99.11.4166

¹⁹ Brandsma AM, Bondza S, Evers M, et al. Potent FC receptor signaling by IgA leads to superior killing of cancer cells by neutrophils compared to IgG. Front Immunol 2019; 10: 704. doi:10.3389/fimmu.2019.00704

²⁰ Evers M, Stip M, Keller K, et al. Anti-Gd2 IgA kills tumors by neutrophils without antibody-associated pain in the Preclinical treatment of high-risk neuroblastoma. J Immunother Cancer 2021; 9: e003163. doi:10.1136/jitc-2021-003163

²¹ Stip MC, Evers M, Nederend M, et al. Iga antibody Immunotherapy targeting Gd2 is effective in Preclinical neuroblastoma models. J Immunother Cancer 2023; 11: e006948. doi:10.1136/jitc-2023-006948

²² Voeller J, Erbe AK, Slowinski J, et al. Combined innate and adaptive Immunotherapy overcomes resistance of Immunologically cold Syngeneic murine neuroblastoma to Checkpoint inhibition. J Immunother Cancer 2019; 7: 344. doi:10.1186/s40425-019-0823-6

²³ Masih KE, Wei JS, Milewski D, et al. Exploring and targeting the tumor immune Microenvironment of neuroblastoma. J Cell Immunol 2021; 3: 305–16. doi:10.33696/immunology.3.111

²⁴ Chan C, Lustig M, Baumann N, et al. Targeting myeloid Checkpoint molecules in combination with antibody therapy: A novel anti-cancer strategy with IgA antibodies Front Immunol 2022; 13: 932155. doi:10.3389/fimmu.2022.932155

²⁵ Jaiswal S, Jamieson CHM, Pang WW, et al. Cd47 is upregulated on circulating hematopoietic stem cells and leukemia cells to avoid Phagocytosis. Cell 2009; 138: 271–85. doi:10.1016/j.cell.2009.05.046

²⁶ Theruvath J, Menard M, Smith BAH, et al. Anti-Gd2 Synergizes with Cd47 blockade to mediate tumor eradication. Nat Med 2022; 28: 333–44. doi:10.1038/s41591-021-01625-x

²⁷ Martínez-Sanz P, Hoogendijk AJ, Verkuijlen PJJH, et al. Cd47-SIRPα Checkpoint inhibition enhances neutrophil-mediated killing of Dinutuximab-opsonized neuroblastoma cells. Cancers (Basel) 2021; 13: 4261. doi:10.3390/cancers13174261

²⁸ Willingham SB, Volkmer J-P, Gentles AJ, et al. The Cd47-signal regulatory protein alpha (Sirpa) interaction is a therapeutic target for human solid tumors. Proc Natl Acad Sci U S A 2012; 109: 6662–7. doi:10.1073/pnas.1121623109

²⁹ Matlung HL, Babes L, Zhao XW, et al. Neutrophils kill antibody-opsonized cancer cells by Trogoptosis. Cell Rep 2018; 23: 3946–59. doi:10.1016/j.celrep.2018.05.082

³⁰ Treffers LW, Ten Broeke T, Rösner T, et al. Iga-mediated killing of tumor cells by neutrophils is enhanced by Cd47–SIRPα Checkpoint inhibition. Cancer Immunol Res 2020; 8: 120–30. doi:10.1158/2326-6066.CIR-19-0144

³¹ Baumann N, Arndt C, Petersen J, et al. Myeloid Checkpoint blockade improves killing of T-acute Lymphoblastic leukemia cells by an Iga2 variant of Daratumumab. Front Immunol 2022; 13: 949140. doi:10.3389/fimmu.2022.949140

³² Evers M, Rösner T, Dünkel A, et al. The selection of variable regions affects Effector mechanisms of IgA antibodies against Cd20. Blood Adv 2021; 5: 3807–20. doi:10.1182/bloodadvances.2021004598

³³ Chernyavska M, Hermans CKJC, Chan C, et al. Evaluation of Immunotherapies improving macrophage anti-tumor response using a Microfluidic model. Organs-on-a-Chip 2022; 4. doi:10.1016/j.ooc.2022.100019

³⁴ Meyer S, Nederend M, Jansen JHM, et al. Improved in vivo anti-tumor effects of IgA-Her2 antibodies through half-life extension and serum exposure enhancement by Fcrn targeting. mAbs 2016; 8: 87–98. doi:10.1080/19420862.2015.1106658

³⁵ Bate-Eya LT, Ebus ME, Koster J, et al. Newly-derived neuroblastoma cell lines propagated in serum-free media recapitulate the genotype and phenotype of primary neuroblastoma tumours. Eur J Cancer 2014; 50: 628–37. doi:10.1016/j.ejca.2013.11.015

³⁶ M. Kholosy W, Derieppe M, van den Ham F, et al. Neuroblastoma and DIPG Organoid Coculture system for personalized assessment of novel anticancer Immunotherapies. JPM 2021; 11: 869. doi:10.3390/jpm11090869

³⁷ Brandsma AM, Ten Broeke T, Nederend M, et al. Simultaneous targeting of FC rs and FC RI enhances tumor cell killing. Cancer Immunol Res 2015; 3: 1316–24. doi:10.1158/2326-6066.CIR-15-0099-T

³⁸ van Egmond M, van Vuuren AJ, Morton HC, et al. Human immunoglobulin A receptor (FcαRI, Cd89) function in transgenic mice requires both FCR Γ chain and Cr3 (Cd11B/Cd18). Blood 1999; 93: 4387–94.

³⁹ Stip MC, Jansen JHM, Nederend M, et al. Characterization of human FC alpha receptor transgenic mice: comparison of Cd89 expression and antibody-dependent tumor killing between Mouse strains. Cancer Immunol Immunother 2023; 72: 3063–77. doi:10.1007/s00262-023-03478-4

⁴⁰ Olofsen PA, Stip MC, Jansen JHM, et al. Effective, long-term, neutrophil depletion using a Murinized anti-Ly-6G 1A8 antibody. Cells 2022; 11: 3406. doi:10.3390/cells11213406

⁴¹ Aktas S, Ozdemir APE, Serinan EO, et al. “Don’t eat me” signals of neuroblastoma by CD47 for immune escape: A novel prognostic biomarker. International Cell Death Research Congress; 2018: 1538. doi:10.3390/proceedings2251538 Available: https://www.mdpi.com/2504-3900/2/25/1538/htm

⁴² Ji K, Zhang Y, Jiang S, et al. SIRPα blockade improves the antitumor immunity of radiotherapy in colorectal cancer. Cell Death Discov 2023; 9: 180. doi:10.1038/s41420-023-01472-4

⁴³ Huang C, Wang X, Wang Y, et al. Sirpα on tumor-associated myeloid cells restrains antitumor immunity in colorectal cancer independent of its interaction with Cd47. Nat Cancer 2024; 5: 500–16. doi:10.1038/s43018-023-00691-z

⁴⁴ Schumacher-Kuckelkorn R, Hero B, Ernestus K, et al. Lacking Immunocytological Gd2 expression in neuroblastoma: report of 3 cases. Pediatr Blood Cancer 2005; 45: 195–201. doi:10.1002/pbc.20301

⁴⁵ Schumacher-Kuckelkorn R, Volland R, Gradehandt A, et al. Lack of Immunocytological Gd2 expression on neuroblastoma cells in bone marrow at diagnosis, during treatment, and at recurrence. Pediatr Blood Cancer 2017; 64: 46–56. doi:10.1002/pbc.26184

⁴⁶ van Rees DJ, Bouti P, Klein B, et al. Original research: cancer cells resist antibody-mediated destruction by neutrophils through activation of the Exocyst complex. J Immunother Cancer 2022; 10: e004820. doi:10.1136/jitc-2022-004820

⁴⁷ Beum PV, Peek EM, Lindorfer MA, et al. Loss of Cd20 and bound Cd20 antibody from opsonized B cells occurs more rapidly because of Trogocytosis mediated by FC receptor-expressing Effector cells than direct Internalization by the B cells. J Immunol 2011; 187: 3438–47. doi:10.4049/jimmunol.1101189

⁴⁸ Tibbetts R, Yeo KK, Muthugounder S, et al. Anti-Disialoganglioside antibody Internalization by neuroblastoma cells as a mechanism of Immunotherapy resistance. Cancer Immunol Immunother 2022; 71: 153–64. doi:10.1007/s00262-021-02963-y

⁴⁹ Weiskopf K, Ring AM, Ho CCM, et al. Engineered SIRPα variants as Immunotherapeutic Adjuvants to anticancer antibodies. Science 2013; 341: 88–91. doi:10.1126/science.1238856

⁵⁰ Bouwstra R, van Meerten T, Bremer E. Cd47-SIRPα blocking-based Immunotherapy: current and prospective therapeutic strategies. Clin Transl Med 2022; 12: e943. doi:10.1002/ctm2.943

⁵¹ Kauder SE, Kuo TC, Harrabi O, et al. Alx148 blocks Cd47 and enhances innate and adaptive antitumor immunity with a favorable safety profile. PLoS One 2018; 13: e0201832. doi:10.1371/journal.pone.0201832

⁵² Logtenberg MEW, Jansen JHM, Raaben M, et al. Glutaminyl cyclase is an enzymatic modifier of the Cd47- SIRPα axis and a target for cancer Immunotherapy. Nat Med 2019; 25: 612–9. doi:10.1038/s41591-019-0356-z

⁵³ Michon J, Moutel S, Barbet J, et al. In vitro killing of neuroblastoma cells by neutrophils derived from granulocyte colony-stimulating factor-treated cancer patients using an anti-Disialoganglioside/anti-FC gamma RI Bispecific antibody. Blood 1995; 86: 1124–30. Available: https://dx.doi.org/10.1182/blood.V86.3.1124.1124

⁵⁴ Chan GCF, Chan CM. Anti-Gd2 directed Immunotherapy for high-risk and metastatic neuroblastoma. Biomolecules 2022; 12: 358. doi:10.3390/biom12030358

⁵⁵ van Duijn A, Van der Burg SH, Scheeren FA. Cd47/SIRPα axis: bridging innate and adaptive immunity. J Immunother Cancer 2022; 10: e004589. doi:10.1136/jitc-2022-004589

⁵⁶ Liu X, Pu Y, Cron K, et al. Cd47 blockade triggers T cell-mediated destruction of Immunogenic tumors. Nat Med 2015; 21: 1209–15. doi:10.1038/nm.3931

⁵⁷ Xu L, Wang S, Li J, et al. Cd47/SIRPα blocking enhances Cd19/Cd3-Bispecific T cell Engager antibody-mediated lysis of B cell malignancies. Biochem Biophys Res Commun 2019; 509: 739–45. doi:10.1016/j.bbrc.2018.12.175

⁵⁸ Lin A, Loré K. Granulocytes: new members of the antigen-presenting cell family. Front Immunol 2017; 8: 1781. doi:10.3389/fimmu.2017.01781

⁵⁹ Schwartz AL, Nath PR, Allgauer M, et al. Antisense targeting of CD47 enhances human cytotoxic T-cell activity and increases survival of mice bearing B16 melanoma when combined with anti-CTLA4 and tumor irradiation. Cancer Immunol Immunother 2019; 68: 1805–17. doi:10.1007/s00262-019-02397-7