In Vitro Evaluation of the Safety and Efficacy of Cibisatamab Using Adult Stem Cell-Derived Organoids and Colorectal Cancer Spheroids

Abstract

Objectives: Developing ex vivo models that replicate immune-tumor interactions with high fidelity is essential for advancing immunotherapy research, as traditional two-dimensional in vitro systems often lack the complexity required to fully represent these interactions. Methods: In this study, we establish a comprehensive 3D redirect lysis (3D-RDL) assay using colorectal cancer spheroids and adult stem cell-derived, healthy human organoids to evaluate the efficacy and safety profile of Cibisatamab, a bispecific antibody targeting carcinoembryonic antigens (CEAs) on cancer cells and CD3 on T cells. This model allows us to assess cytotoxic activity and immune responses, capturing variations in therapeutic response not observable in simpler systems. Our model integrates live imaging and cytotoxicity analyses to enable precise, real-time tracking of antibody effects on CEA-expressing tumor cells compared to healthy cells. Additionally, by standardizing effector-to-target cell ratios in each co-culture, we establish a reproducible workflow that enhances data accuracy and comparability across assays. Flow cytometry and Granzyme B release profiling further allow us to characterize immune cell activation, revealing distinct T cell activation markers and Granzyme B release patterns tied to Cibisatamab treatment. Results: Our results show that Cibisatamab effectively induces cell death in cancer spheroids with high CEA expression while being dose-dependent on target, off-tumor binding and killing on non-cancerous cells of healthy organoids with intermediate CEA levels. This highlights our model's potential to predict clinical immunotherapy outcomes, capturing complex responses like immune activation, therapeutic selectivity, and potential resistance mechanisms. Conclusions: These findings underscore the utility of our model as a reliable, physiologically relevant tool for screening new immunotherapies and advancing our understanding of tumor-immune dynamics.

Keywords: T cell bispecific (TCB) antibodies; adult stem cell-derived organoids; co-culture; preclinical safety and potency assessment.

Figure 1

The experimental design of the 3D-RDL assay and CEA target expression validation: (A) In the preparation of the 3D-RDL assay, single cells from cell lines (MKN-45 and DLD-1) and organoids (Rectum and Small intestine_2) are seeded in 96-well microwell plates. At day 2, hPBMCs are rested overnight in cRPMI medium. At day 3, hPBMCs (E:T, 5:1) and Cibisatamab (six doses; 0.001–100 µM) are added to the aggregated spheroids and organoids and incubated for 72 h. (B) CEA protein and RNA expression levels in healthy human organs. CEA is expressed in a variety of epithelial tissues such as the urogenital, respiratory, and gastrointestinal tracts. In the healthy human colon, the CEA protein is restricted to the apical side of the differentiated epithelial cells forming the luminal surface. Source: Human Protein Atlas. (C) CEACAM5 bulk RNA expression levels in healthy colon, rectal, and intestinal organoids (Doppl SA Biobank). (D) Relative CEA protein expression levels of control cell lines MKN-45 (CEA_high) and DLD-1 (CEA_low) and 2 different healthy organoid lines (Rectum and Small intestine_2) analyzed by flow cytometry.

Figure 2

The 3D-RDL results of co-cultured 3D grown human adenocarcinoma cell lines MKN-45 and DLD-1 in the presence of hPBMCs treated at different doses of Cibisatamab for 72 h. (A–D) Cytotoxicity of MKN-45 and DLD-1 cells in the presence of Cibisatamab and healthy hPBMCs (E:T,5:1) for 72 h in a 3D-RDL assay. Each well was composed of 31 microwells containing one organoid/spheroid each. (A) Lactate dehydrogenase (LDH) released in the supernatant was quantified, and raw absorbance values was displayed. Increasing concentrations of LDH reflect an increase in target cell death. (B) ATP content was measured through luminescence release upon complete cell lysis. Decreasing ATP concentrations show increase in target cell death. (C) Ethidium median intensity was evaluated with our high-throughput automated analysis pipeline (imaged based). Increased ethidium signals correlated with increased cell death. (D) Our image-based analysis pipeline enabled the measurement of the size of each organoid. A decrease in spheroid size generally correlated with a decrease in cell viability. Data represent mean ± SD from 3 different hPBMCs donors tested in duplicate. Ordinary 1-way ANOVA with Dunnett’s multiple-comparisons test was performed, and the mean of each column was compared to the mean of the reference column (Target only). * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. Only significant values are displayed (p < 0.05).

Figure 3

T cell activation after 72 h of co-culture of hPBMCs and MKN-45 cells in 3D-RDL assay format. (A) CD8+ and (B) CD4+ T cell response in the presence of healthy hPBMCs (E:T, 5:1), MKN-45 cells and increasing doses of CEA TCB antibodies. (C) CD8+ and CD4+ co-expression in the presence of healthy hPBMCs (E:T, 5:1), MKN-45 cells, and increasing doses of Cibisatamab. T cell activation upon Cibisatamab treatment was measured by flow cytometry while measuring the expression of CD25 (green) and CD69 (orange). Data represent non-linear regression analysis of the mean ± SD from 3 different hPBMCs donors tested in duplicate.

Figure 4

The 3D-RDL results of co-cultured rectal and intestinal healthy organoids in the presence of hPBMCs treated at different doses of Cibisatamab for 72 h. (A–C) Cytotoxicity of healthy rectal and intestinal organoids in the presence of Cibisatamab and healthy hPBMCs (E:T, 5:1) for 72 h in a RDL assay. Each well was composed of 31 microwells containing one organoid/spheroid each. (A) Ethidium median intensity was evaluated with our high-throughput automated analysis pipeline (imaged based). Increased ethidium signals correlated with increased cell death. (B) Our image-based analysis pipeline enabled the measurement of the size of each organoid. Decreases in spheroid size generally correlated with decreases in cell viability. (C) ATP content was measured through luminescence release upon complete cell lysis. Decreasing ATP concentrations showed increases in target cell death. (D) Representative brightfield images of rectal organoids cultured in microwells in different conditions after 48 h: untreated rectal organoids (target cells alone); rectal organoids treated with Triton-X 0.8% (+Triton-X 0.8%); rectal organoids co-incubated with hPBMCs; rectal organoids co-incubated with hPBMCs and 100 nM of Cibisatamab. Scale bars: 800 μm, 400 μm, and 200 μm. (E) The cytotoxicity of cancer spheroids and healthy organoids in the presence of Cibisatamab and healthy hPBMCs (E:T, 5:1) for 72 h in a 3D-RDL assay. Data represent non-linear regression analysis of the mean ± SD from 3 different hPBMCs donors tested in duplicates. Specific lysis was calculated as follows: ((Sample—Target Only)/(Triton-X 0.8%—Target Only)). Data represent mean ± SD from 3 different hPBMCs donors tested in duplicate. Ordinary 1-way ANOVA with Dunnett’s multiple-comparisons test was performed, and the mean of each column was compared to the mean of the reference column (Target only). ** p < 0.01, and **** p < 0.0001. Only significant values are displayed (p < 0.05).

Figure 5

Granzyme B quantification after 72 h of co-culture of rectal/intestinal healthy organoids and 3D grown MKN-45/DLD-1 spheroids in the presence of hPBMCs treated at different doses of CEA TCB. (A,B) The quantification of Granzyme B released in the supernatant after 72 h RDL assay by flow cytometry. Data represent non-linear regression analysis of the mean ± SD from 3 different hPBMCs donors tested in duplicate. Ordinary 1-way ANOVA with Dunnett’s multiple-comparisons test was performed, and the mean of each column compared to the mean of MKN-45 treated at 100 nM. ** p < 0.01 and **** p < 0.0001. Only significant values are displayed (p < 0.05).