The B7-H3-targeting mAb with a wildtype Fc portion B7-H3-WT, the B7-H3-SDIE and the corresponding control iso-SDIE were generated by chimerizing an anti-B7-H3 mAb (clone 8H8) and a control mAb (clone MOPC21), respectively, with the human immunoglobulin G1/K constant region (21). The mAbs were Fc-optimized with S239D/I332E modification according to previous descriptions (22). In brief, the mAb light and heavy chain plasmids were obtained using the EndoFree Plasmid Maxi kit from Qiagen (Hilden, Germany), following the manufacturer’s instructions. Antibodies were produced using the ExpiCHO cell system (Gibco, in Carlsbad, CA) following the manufacturer’s instructions. Antibodies were purified from media by protein A affinity chromatography (GE Healthcare, Chicago, IL) followed by preparative size exclusion chromatography (HiLoad 16/60 Superdex 200; GE Healthcare). To guarantee the quality and purity of the antibodies produced, we conducted analytical size exclusion chromatography (using Superdex 200 Increase 10/300 GL from GE Healthcare) and 4–12% gradient SDS-PAGE gels (Invitrogen, Carlsbad, CA) with gel filtration and Precision Plus standard from Bio-Rad (Hercules, CA), respectively.

Endotoxin levels of samples as determined by an EndoZyme assay (Biomerieux, Nuertingen, Germany) were<0.5 EU/mL.

Relative expression levels of B7-H3 mRNA were obtained for tumor and normal tissue samples from the Cancer Genome Atlas (TCGA) database and the GTEx project, using the Gene Expression Profiling Interactive Analysis (GEPIA) web server as previously described (23). Data sets for pancreatic adenocarcinoma (tumor = 179) and corresponding normal tissue (normal = 171) were downloaded from TCGA (http://www.oncolnc.org). The online web server GEPIA (http://gepia.cancer-pku.cn) was utilized for analysis of B7-H3 RNA levels depending on tumor stage (I-IV).

To determine B7-H3 mRNA, B7-H3 primers were QuantiTect Primer Assay Hs_CD276_1_SG (Qiagen), while RPL13 (Hs_RPL13_1_SG, Qiagen) served as housekeeping gene. For total RNA isolation, 1-2 x 10⁶ cells were used and the High Pure RNA Isolation Kit (Roche, Basel, Switzerland) was employed, followed by cDNA synthesis with FastGeneScriptase II (NIPPON Genetics Europe, Dueren, Germany) in accordance with the manufacturer’s instructions. Reverse transcription-polymerase chain reaction (RT-PCR) was conducted following previously described methods (24, 25). Quantitative PCR (qPCR) was performed utilizing the PerfeCTa SYBR Green FastMix (Quanta Biosciences, Beverly, MA) with a LightCycler 480 instrument from Roche.

All cell lines were purchased from ATCC (American Type Culture Collection) and cultured in Dulbeccos Modified Eagle Medium (DMEM) (Gibco). Cells were tested routinely for mycoplasma contamination every three months. Authenticity was determined on a regular basis by validating the respective immunophenotype described by the provider using flow cytometry. Healthy donor peripheral blood mononuclear cells (PBMC) were isolated using density gradient centrifugation (Biochrom, Berlin, Germany). PBMC were collected from healthy donors of various ages and genders, and randomly selected for each experiment. Cryopreserved cells were cultured in media overnight at 37°C prior to use in functional experiments. In every case, written consent was obtained in accordance with the Helsinki protocol. The study was conducted in adherence to the local ethics committee’s guidelines.

Fluorescence-conjugated B7-H3 mAb or isotype control (BioLegend, San Diego, CA) was used to analyze surface expression of B7-H3. To conduct dose titration and binding experiments, cells were stained with B7-H3-SDIE or iso-SDIE, followed by an anti-human PE conjugate (Jackson ImmunoResearch West Grove, PA). A murine B7-H3 hybridoma-derived antibody and the QIFIKIT (Dako, Hamburg, Germany) were employed to perform a quantitative analysis of immunofluorescence aimed to determine the number of B7-H3 molecules present on the cell surface, as previously described (22).

For staining of NK cell activation and degranulation, fluorescence-labeled CD3-APC/Fire, CD56-PECy, CD16-APC, CD25-BV711, CD69-PE (all from BioLegend), or CD107a-PE (BD Biosciences, Franklin Lakes, NJ) were used to stain the cells.

Target cell lysis was assessed via flow cytometry using a previously described method (26). In brief, pancreatic cancer cells were seeded in coculture with PBMC from healthy donors, with or without antibodies (1 μg/mL each), after loading with CellTrace™ Violet cell proliferation dye (Thermo Fisher Scientific, Waltham, MA) at 2.5 μM. To control for assay volumes, beads (Sigma, St. Louis, MO) were used. The percentage of viable target cells was determined by calculating the proportion of 7-AAD-negative cells after treatment relative to those in the control group and multiplying by 100.

The exclusion of dead cells from flow cytometric analysis was achieved using 7-AAD (BioLegend) staining (1:200) or LIVE/DEAD™ Fixable Aqua (Thermo Fisher Scientific). The BD FACS Canto II or BD Fortessa (BD Biosciences) were used for sample analysis. Data analysis was carried out with FlowJo software (FlowJo LCC, Ashland, OR).

To assess the activation and degranulation of NK cells in healthy donor PBMC, coculture experiments were conducted with 20,000 pancreatic cancer cells and PBMC at an effector-to-target (E:T) ratio of 2.5:1 in the absence or presence of treatment (1 μg/mL). For degranulation analysis, Brefeldin A (GolgiPlug, BD Biosciences) was added to the coculture and cells were harvested after 4 hours. Subsequently, cells were stained for CD107a expression and analyzed using flow cytometry. After 24 hours, the cells were collected and labeled to evaluate CD69 and CD25 expression by flow cytometric analysis. NK cells were identified as CD3^- CD56⁺ cells within PBMC.

For the analysis of cytokine secretion, NK cells or PBMC from healthy donors were cultured with or without 20,000 pancreatic cancer cells at an E:T ratio of 2.5:1 in the presence or absence of the proposed mAbs at a concentration of 1 µg/mL. After a 4-hour incubation, the supernatants from the coculture were analyzed for the secretion of IFN-γ and TNF using Legendplex assays (BioLegend) following the manufacturer’s recommendations.

The cytotoxicity of PBMC against pancreatic cancer cells was evaluated using BATDA Europium assays (PerkinElmer, Waltham, MA) over a 2-hour duration, as previously described (27). The percentage of specific lysis was calculated using the formula:

To conduct long-term cytotoxicity analyses, we employed the xCELLigence RTCA system (Roche Applied Science, Penzberg, Germany) to perform a real-time assay at 30-minute intervals for a 120-hour observation period. Pancreatic cancer cells (5,000 cells/well) were seeded in 96-well plates and cocultured with PBMC from healthy donors at an E:T ratio of 10:1; this was done with and without the specified mAbs (1 µg/mL each).

Statistical analyses were conducted using GraphPad Prism software (version 9) (GraphPad, Boston, MA). Mean ± standard deviation of replicates or individual values were used to present data. Continuous variables were analyzed using Student’s t-test, Mann-Whitney U test, one-way ANOVA, and Friedman’s test. In the case of significant differences observed by ANOVA, group-wise comparison was performed using Tukey’s multiple comparison test. In the case of significant differences found by Friedman’s test, Dunn’s multiple comparison test was employed. All statistical tests were considered statistically significant when p was below 0.05.

B7-H3 expression has been reported in multiple solid tumors, including pancreatic cancer (28, 29). As a first step, we analyzed B7-H3 mRNA expression using TCGA datasets consisting of 179 pancreatic adenocarcinomas and 171 normal pancreatic tissue samples. Our analysis reveals that B7-H3 mRNA expression is substantially higher in pancreatic cancer tissues than in normal tissues ( Figure 1A ). A comparative analysis of B7-H3 mRNA levels in the TCGA data set of pancreatic adenocarcinomas, categorized by disease stage (I-IV), showed no significant differences in B7-H3 expression levels ( Figure 1B ). Next we studied B7-H3 expression mRNA levels in the pancreatic cancer cell lines AsPC-1, Capan-1, Capan-2, MIA PaCa-2, and PANC-1 (30), revealing differing B7-H3 expression levels across the examined cell lines ( Figure 1C ). A key determinant for an effective immunotherapeutic target antigen is its expression level on the cell surface. Our analysis, conducted by flow cytometry, revealed that B7-H3 is expressed on the cell surface of all pancreatic cancer cell lines tested, albeit to varying extent ( Figure 1D ). The range of B7-H3 surface molecules varied from 18,030 (Capan-1) to 39,840 molecules/cell (PANC-1) ( Figure 1E ). These results led us to select AsPC-1, MIA PaCa-2, and PANC-1 for functional assessments implementing our B7-H3 targeting Fc-optimized B7-H3-SDIE mAb. Dose titration experiments (ranging from 0.0005-10 µg/mL) showed that binding of B7-H3-SDIE saturates at a concentration of 1 µg/mL ( Figure 1F ). This concentration was used for subsequent functional analyses. Binding of an antibody to its target molecule frequently results in dose-dependent modulation of target antigen expression levels. This, in turn, may impair therapeutic efficacy of mAb treatment (31). To assess the induction of this phenomenon known as antigen shift, we examined B7-H3 expression after incubating pancreatic cancer cells with B7-H3-SDIE (range 0.0003-10 µg/mL) for 24 and 72 h. We observed only a slight decrease (maximum of 25% reduction) in B7-H3 surface expression ( Figures 1G, H ).

Next, we investigated the effects of B7-H3-SDIE on NK cells in the absence of tumor cells. To this end, isolated NK cells or whole PBMC were incubated with B7-H3-SDIE or corresponding isotype control. B7-H3-SDIE did not induce NK cell activation ( Figures 2A, B ) or secretion of the immune effector cytokines IFN-γ and TNF ( Figure 2C ) in the absence of target cells. Furthermore, we did not observe any changes of total PBMC numbers, confirming that B7-H3-SDIE does not induce unspecific cell lysis ( Figure 2D ).

As a next step, we investigated if and how our mAb B7-H3-SDIE activates immune effector cells to attack pancreatic cancer cells. For this purpose, we used PBMC from healthy donors as effector cells. Cocultures of PBMC and pancreatic cancer cells (AsPC-1, MIA PaCa-2, and PANC-1) in the presence or absence of B7-H3-SDIE or control mAb iso-SDIE were analyzed for the expression of the activation markers CD69 and CD25 on NK cells upon counterstaining for CD56, CD3 and CD16 after 24 h by flow cytometry. A comparison between the Fc-optimized B7-H3-SDIE and a Fc wildtype mAb (B7-H3-WT) revealed that B7-H3-SDIE binds more strongly to CD16 on NK cells, resulting in increased activation capacity ( Supplementary Figures S1A-C ). Both CD16⁺ and CD16^- NK cells showed greater activation when exposed to B7-H3-SDIE compared to B7-H3-WT or control treatments ( Supplementary Figures S1D, E ). Following treatment with B7-H3-SDIE, the proportion of CD16⁺ NK cells decreased significantly compared to treatment with B7-H3-WT ( Supplementary Figure S1F ). Following treatment with B7-H3-SDIE, the activation markers CD69 and CD25 were significantly upregulated on NK cells. The control mAb iso-SDIE with nonrelevant target specificity had no significant effect ( Figures 3A, B ). B7-H3-SDIE induced activation of various NK cell subsets (CD56^brightCD16^-, CD56^dimCD16^- and CD56^dimCD16⁺) with most pronounced effects observed with the CD56^dimCD16⁺ subpopulation ( Supplementary Figure S2 ). This subpopulation is crucial for treatment with ADCC inducing mAbs, as it is not only the major circulating subset, but also the primary ADCC inducing subset (32, 33). Flow cytometric analysis further showed increased expression of the degranulation marker CD107a on NK cells in cocultures of PBMC with pancreatic carcinoma cells when treated with B7-H3-SDIE. In contrast, the Fc control mAb iso-SDIE had no significant effect ( Figure 3C ). In line, degranulation was induced in all analyzed NK cell subsets (CD56^brightCD16^-, CD56^dimCD16^- and CD56^dimCD16⁺) ( Supplementary Figure S3 ). Activated NK cells produce cytokines, such as IFN-γ and TNF, that can effectively stimulate an immune response to fight tumors. These cytokines can directly impact target cells or activate other immune cells to attack target cells (34). Legendplex assays used to analyze cytokine secretion into culture supernatants of PBMC and pancreatic cancer cell lines demonstrated a significant increase in the release of both IFN-γ and TNF after treatment with B7-H3-SDIE, whereas the isotype control again had no relevant effect ( Figures 3D, E ).

Finally, to study the ability of B7-H3-SDIE to induce tumor cell lysis, we co-cultured the pancreatic cancer cell lines AsPC-1, MIA PaCa-2, and PANC-1 with PBMC from healthy donors. We found that B7-H3-SDIE substantially induced lysis of target cells in short-term cytotoxicity assays for all the cell lines examined, whereas the presence of control mAb exhibited no discernible impact on target cell lysis ( Figure 4A ). Analyses using long-term flow cytometry-based lysis assays over 72 h further demonstrated the marked efficacy of B7-H3-SDIE against pancreatic cancer cells ( Figure 4B ). The potential of B7-H3-SDIE to induce target cell lysis was additionally confirmed by analyzing tumor cell lysis over a 120-hour period using xCELLigence assays, despite the varying morphological features, growth rates, and levels of B7-H3 surface expression across the different pancreatic cancer cell lines ( Figure 4C ). To summarize, our findings have demonstrated that B7-H3-SDIE exerts a significant effect on the activation, degranulation, and cytokine release of NK cells, ultimately resulting in the lysis of tumor cells.