Abstract 3372: Predictive in vitro assays and PBMC donor banking: reducing variability in humanized mouse studies

Abstract Humanized mouse models are an essential tool for evaluating immunotherapies in vivo. The use of peripheral blood mononuclear cells (PBMCs) from healthy donors is common as they provide a functional human immune compartment capable of mediating responses such as T-cell activation and tumor clearance. However, donor variability introduces a significant challenge: PBMCs from different individuals differ in their immune cell composition, and alloreactivity. To address this, we established a PBMC donor bank. Each of the currently six donors are characterized for major phenotypic markers by flow cytometry, as well as HLA-type. In addition to the PBMCs itself we created a bank of expanded T cells as source for in vitro and in vivo assays. The phenotypic characterization of the PBMC revealed differences in all major cell types (T, B, NK, NK T cells and monocytes). However, all data were in physiological range. To deeper analyze the immune cells from different donors, we executed in vitro co-culture assays with a representative tumor cell line MDA-MB-231 to screen for killing potential. Solitumab, a CD3/EPCAM bispecific antibody was tested in the presence of expanded T cells from all donors in five independent runs against MDAMB-231. Efficacy was determined by CTG, oneGlo and flow cytometry in parallel to identify the optimal assay read-out. All three assay read-outs displayed acceptable run to run variability. However, flow cytometry was best suited to discriminate between cell viability of immune and tumor cells. The killing activity of the expanded T cells in vitro was predictive for the onset of GvHD in PBMC engrafted immune compromised NSG mice in vivo. Not only induced donors with higher killing efficiency in vitro an earlier onset of GvHD (day 18 post engraftment for donor 6 vs day 40 for donor 2) it also impacted overall survival (29 days vs 36 days, n.s. Kruskall-Wallis). The engraftment of human PBMC in vivo was characterized by analyzing human immune cells in peripheral blood of engrafted mice over the course of the experiment, GvHD scoring and body weight. Each donor displayed a specific pharmacokinetic of engraftment and different immune cell subtyping. These data are donor immanent, as they could be reproduced using the same donor in multiple independent experiments. The ability to induce antitumoral activity for immune modulating compounds was proven in different experimental set-ups for all six donors. A side-by-side comparison testing Solitumab against MDAMB-231 in vivo is currently under way. This approach enables the reduction of variability in tumor growth and treatment response for PBMC based in vivo studies. By banking PBMCs from multiple donors, researchers can pre-screen immune functionality using in vitro T cell killing assays before committing to in vivo studies. This ensures that only PBMC donors with the desired immune characteristics are selected for humanization in vivo thereby supporting the 3Rs. Citation Format: Matthias Bleisch, Philipp Meyer, Ina Rohleff, Eva Oswald, Anna Edinger, Julia B. Schueler. Predictive in vitro assays and PBMC donor banking: reducing variability in humanized mouse studies abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3372.