Autologous human tumor-immune organoid model for patient-specific prediction of response to immunotherapy

27 January 2026

Despite important investment in biomedical research, the most common malignant adult brain tumor - glioblastoma (GBM) - remains largely refractory to existing treatments.

Tumors can evade or suppress immunity through a variety of mechanisms, including antigen heterogeneity and  immunosuppressive signaling. Owing in large part to inter-species differences in the immune system and the tumor microenvironment (extracellular matrix composition, stromal cells behaviour, cytokine signaling, angiogenesis), preclinical animal models are poorly predictive of responses to immunotherapy in cancer patients and are ill suited to investigate the mechanisms by which tumors evade immunity. In addition, animal models cannot capture the genetic, microenvironmental and antigenic heterogeneity of human tumors, leading to therapies that show highly variable responses in patients and hindering opportunity to advance personalized medicine.

To advance effective therapies for GBM patients, it is therefore critically important to employ patient-specific GBM models that recapitulate human-relevant tumor-immune interactions. With this goal in mind, Baisiwala et al. from the University of California developed its immune-human organoid tumor transplantation (iHOTT) model - an autologous co-culture platform that integrates patient-derived tumor cells and matched peripheral blood mononuclear cells (PBMC) within human cortical organoids.

Where in vitro cancer platforms often represent simplified immune models containing select immune cells, such as monocytes and T cells, Baisiwala and colleagues made use of whole patient-derived PBMCs that represent the full functional pool of immune cells, including naïve T cells, effector memory T cells, NK, monocytes, and B cells. Unlike tumor-infiltrating lymphocytes, PBMCs have not been exposed to persistent antigen stimulation, immunosuppressives cytokines and metabolic stress that results in chronic T cell exhaustion, characterized by impaired proliferation and cytotoxicity. As such, PBMCs are more responsive to antigens and better suited for studying early tumor recognition and clonal expansion. These PMBCs were matched to patient-derived tissues, thus preserving autologous immune–tumor matching (same HLA, same antigen interaction), natural antigen presentation (tumor antigen-presenting dendritic cells and monocytes), immune infiltration and patient-specific responses to tumor antigens and immunotherapy.

iHOTT preserves patient-derived glioblastoma cell heterogeneity, enabling more accurate modeling of antigen‑specific immune responses, tumor evolution, therapeutic resistance, and patient‑specific treatment responses. Human cortical organoids, into which glioblastoma cells were transplanted, contain human ECM proteins, scaffolding, and cells, capturing the interactions between the tumor and its microenvironment.  In contrast to mouse models of GBM, that often consist of inbred strains with clonal tumor lines, the iHOTT model recapitulates the genetic drivers, clonal neoantigens, TCR repertoire, HLA genotype, immune microenvironment, and evolutionary history that are unique to each patient, paving the way for precision immunotherapy.

Treatment of iHOTT with pembrolizumab, a PD-1 blocking antibody, mirrored immune cell distributions, cell-cell interactions and T-cell responses observed in patients, supporting the physiological relevance of this platform. iHOTT appears to be particularly well suited for testing T-cell-mediated and CAR-T-based immunotherapies, using flow cytometric, cytokine, and transcriptomic profiling.

Future adaptations to this tumor immune organoid model for preclinical research will include incorporating myeloid-driven immunosuppressive signals and conditioning PBMCs ex vivo to mimic exhausted phenotypes.

What Puts It on the Frontier

• Engineering of a physiologically-relevant immune-human organoid tumor transplantation platform that faithfully recapitulates patient-specific multicellular tumor-immune interactions

• The obtained platform preserves autologous immune–tumor matching by integrating patient-derived tumor cells and matched peripheral blood mononuclear cells that represent the full functional pool of patient immune cells

• The transplanted patient-specific glioblastoma cells capture the genetic drivers, clonal neoantigens, TCR repertoire, HLA genotype, immune microenvironment, and evolutionary history within a controlled human neural microenvironment

Impact Snapshot

• Enhanced predictivity of preclinical in vitro models of cancer, particularly for testing T-cell-mediated and CAR-T-based immunotherapies

• Prediction of patient-specific responses to immunotherapy, enabling personalized medicine

Reference

Baisiwala S, Fazzari E, Li M, et al. A human tumor-immune organoid model of glioblastoma. Cell Reports, Volume 45, Issue 1116790, January 27, 2026. https://doi.org/10.1016/j.celrep.2025.116790