Advancing Translational Oncology Through Humanized and Genetically Engineered Research Solutions

Background: Second-generation humanized NOG mice (HIS mice) enable evaluation of immune-oncology mechanisms within a physiologically relevant human immune microenvironment. However, translational interpretation of preclinical efficacy studies requires highly specific and reproducible quantification of human cytokines. To support robust immune monitoring, we co-evaluated a 41-plex humanized mouse cytokine/chemokine Luminex panel for performance across diverse human CD34+ donor sources and multiple HIS NOG model variants. Study Aims/Hypothesis: We hypothesized that the human-specific analytes in the MILLIPLEX® 41-plex humanized mouse multiplex assay would show (i) high specificity for human cytokines with minimal cross-reactivity to host murine analytes, (ii) reproducible quantification across donor-to-donor variation, and (iii) consistent detection sensitivity across advanced HIS NOG models supporting translational oncology applications. Results: Plasma samples were collected from HIS NOG cohorts reconstituted with ≥3 unrelated CD34+ donor pools and representing multiple second generation humanized platforms. Across all donors and models, human-restricted analytes yielded quantifiable signals above background. Murine-only control samples confirmed negligible cross-reactivity. Inter-assay and intra-assay CVs demonstrated technical reproducibility. Donor-specific cytokine patterns (e.g., Th1/Th2 balance, myeloid-associated chemokines) were preserved across models, indicating biological fidelity rather than assay drift. Baseline cytokine architecture correlated with reconstitution kinetics and human immune cell subset frequencies, supporting biological coherence.

Conclusions: These validation data demonstrate that the MILLIPLEX® 41-plex Luminex assay provides reproducible, donor-consistent, and human-specific cytokine quantification in second-generation HIS NOG mice. These data support the assay’s suitability for mechanistic and pharmacodynamic readouts in oncology studies requiring high-resolution human cytokine/chemokine profiling. Reliable multiplex cytokine measurement strengthens the translational bridge between HIS mouse studies and human immuno-oncology responses, enabling improved interpretation of therapeutic mechanisms, biomarkers, and treatment-induced immune modulation.

Abstract:

Humanized immune system (HIS) mouse platforms that support durable human NK and T cell function are essential for evaluating CAR-NK and CAR-T therapies. Residual murine Fe gamma receptors (FcyRs) can confound interpretation of lgG-based therapeutics and engineered cell products. We compared hlL-15 NOG and FcyR-null FcResolv™ hlL-15 NOG mice engrafted with human NK cells and observed equivalent ~12-week persistence and stable CD56.CD16• profiles, enabling extended engineered-cell studies. In CD34• HIS tumor models, anti-PD1 efficacy and pharmacodynamics were accurately detected only in FcResolv™ mice. FcyR-null HIS platforms improve predictive assessment of CAR-NK/CAR-T efficacy, trafficking, and off-target responses.

Abstract: 

Humanized immune system (HIS) mice, which can be engrafted with human cell line-derived or patient-derived tumors, have become essential tools for preclinical and IND-enabling development of cell therapies and biologics in oncology. However, selecting the appropriate HIS model remains challenging given the diversity of humanizable mouse strains and immune-engraftment protocols. We first compared immune profiles, clinical symptoms, body weight, and survival in severely immunodeficient mice engrafted with either cord blood-derived CD34⁺ hematopoietic stem cells (HSCs) or peripheral blood mononuclear cells (PBMCs). PBMC-engrafted mice demonstrated poor survival associated with early graft-versus-host disease (GvHD) and exhibited amplification of human T cells with a partially exhausted phenotype (Lag-3⁺, TIM-3⁺). In contrast, CD34-engrafted mice showed no health deterioration over 30 weeks and developed a complete human immune system comprising T, B, NK, and myeloid cells. These results place the CD34+ HSC engrafted HIS mice as overall superior model for drug assessment as it allows a longer, GvHD-free treatment window and a more complete immune system. We next implemented a universal, irradiation-free, chemoablation-based CD34⁺ HSC-engraftment protocol to compare the extent of immune humanization across several severely immunodeficient strains, including foundational models (NOG, NCG,BNDG, BRG, and next-generation strains (NOG-EXL, which overexpresses human GM-CSF and IL-3; and FcResolv NOG, which lacks murine Fcγ receptors). For each strain, we measured survival, overall humanization rate (percentage of human among total immune cells), and human immune-subset frequencies in blood. Finally, we demonstrate how hydrodynamic gene delivery of one or more human cytokines into HIS mice can boost certain immune populations when a transgenic strain overexpressing these cytokines is unavailable. Altogether, our findings provide a knowledge base for the selection of the humanized immune system mouse model with the most suitable immune reconstitution profile for assessing any drug candidate based on its mechanisms of action.

Abstract: 

Cytokine release syndrome (CRS) is a major risk during treatment with CAR-T cells, T-cell engagers, and, occasionally, checkpoint inhibitors. Because humanized immune-system mice are frequently used for the preclinical development of such therapies, we sought to provide a platform to derisk CRS before the clinical stage. We therefore characterized CRS induction, immune dynamics, and clinical manifestation in two different humanized mouse models. We humanized the immune system of initially immunodeficient mice by engrafting them either PBMCs or CD34+ hematopoietic stem cells. In the PBMC model, T cells were the only engrafted human immune population and exhibited a chronically induced, partially exhausted (TIM3+, LAG3+, PD1-) phenotype at baseline. At one week post-PBMC engraftment, circulating T-cells were at approximately at 1000 cells per mL, and OKT3 injection depleted them, thereby failing to induce CRS. At three weeks post-engraftment, however, circulating T-cell counts reached one million cells per mL and were no longer depleted by OKT3, which instead induced severe CRS and mortality within 24 hours. CRS induction in PBMC mice was accompanied by increases in circulating TNF-α, IFN-γ, and IL-2, as well as transient (≤6 hours) T-cell proliferation. In contrast, in the CD34 model, a more complete human immune system developed after engraftment, including myeloid, NK, dendritic, B, and non-exhausted T cells. This broader immune reconstitution enabled OKT3 to induce CRS with as few as 10,000 circulating T cells. Although clinical CRS symptoms were milder than those observed in PBMC-engrafted mice, the CRS cytokine signature in the CD34 model included not only the IFN-γ, TNF-α, and IL-2 surges observed in PBMC mice but also substantial increases in hallmark myeloid-derived CRS cytokines such as IL-6, CXCL10, and CCL2. Likewise, the cellular signature of CRS was more complete in CD34 mice, with induction of T-cell activation (CD69+, CD38+, HLA-DR+), exhaustion (PD1+, TIM3+), and proliferation (Ki67) markers, along with increased monocyte mobilization into the blood. In summary, our results demonstrate that the CD34 model displays mild-to-moderate clinical symptoms while more comprehensively recapitulating the hallmark cellular and molecular features of CRS.