An obstacle to studying human cancer is the limited availability of models with human stroma and immune cells. This is particularly relevant for head and neck squamous cell carcinomas (HNSCC) given the pivotal role the immune system plays in their development. An unmet need is an investigation of how cancer stem cells (CSCs) modulate immunity, because it is critical to understand how CSCs harness oncogenic signaling to evade immune surveillance. To address these outstanding questions, we will use two unique tools: 1) CSCs derived from a panel of patient HNSCCs propagated in mice, and 2) a humanized mouse (HM) model that enables studying tumor-stromal interactions. We have defined CSCs as ALDH+CD44high in multiple patient cases, and documented that SOX2 is responsible for key CSC features such as growth, invasion, and resistance to therapy (1). Further, we observed an association between SOX2 and tumor expression of the ligand PD-L1, which activates PD-1 in T cells leading to immune escape. Additionally, SOX2 modulates SOX9, which regulates the other PD-1 ligand, PD-L2. After treatment of HNSCC HM mice with a PD-1 inhibitor, we documented resistance associated with an increase in PD-L2; thus, we propose that alternate ligands are used by HNSCC as a resistance mechanism. We initially developed a mismatched HM (mHM; tumor and hematopoietic stem cells [HSCs] from different sources), and now have generated an autologous HM (aHM; tumor, HSCs and mesenchymal stem cells [MSCs] from the same patient). Tumors grown in HM more closely resemble the originator tumor than those grown in non-HM mice, and the drift in gene expression caused by prior passaging was partially reversed. Signaling in key immune and stroma pathways was more prominent and closely resembled the originator patient in HM vs. non-HM models. The immune cells mounted effective tumor-specific immune responses, mediated by human immune cells including T cells. Two salient and under- studied issues limit the wider application of HM: tumor-immune mismatch and T cell education. Tumor and immune matching can affect the faithfulness of immune response, and that can only be appreciated by comparing mHM and aHM. HM are subject to xenogeneic education of human T cells by mouse thymic cells, resulting from the lack of a functional human thymus in such models. The goals of this proposal are to: 1) study the regulation by SOX2 and SOX9 of PD-1 ligands; 2) understand how SOX2 and SOX9 affect CSC properties; 3) generate and characterize aHM and mHM from 10 HNSCC patients; 4) test if SOX2 and SOX9 are involved in resistance to PD-1 inhibitors by testing them on HM; and finally, we will 5) generate thymic epithelium from the same patient's HSCs that will result in HM with a fully autologous HNSCC, thymus and immune system, thus enabling immune cell education in a strictly human context. This project will lead to a deeper and more mechanistic understanding of the interplay between HNSCC CSCs and the immune/stroma systems, and ultimately, improve precision medicine.

Public Health Relevance

Little progress has been made in understanding the cancer-host interaction because human cancer is studied either using cultured cancer cells or on mice transplanted with human tumors where cancer cells interact with host cells of a different species in the absence of an immune system. To address these limitations, we are proposing to study head and neck cancer and its therapy in a humanized mouse model, where a patient's tumor will be grown in mice that have that same patient's immune system, and that same patient's thymus where human immune cells mature. Our goal is to use a humanized mouse model of head and neck cancer to understand how a small population of so called ?cancer stem cells? generates a tumor mass without being destroyed by the immune system, and translate this information into better treatments for our patients.

National Institute of Health (NIH)
National Cancer Institute (NCI)
Research Project (R01)
Project #
Application #
Study Section
Basic Mechanisms of Cancer Therapeutics Study Section (BMCT)
Program Officer
Hargrave, Sara Louise
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Colorado Denver
Internal Medicine/Medicine
Schools of Medicine
United States
Zip Code
Keysar, Stephen B; Eagles, Justin R; Miller, Bettina et al. (2018) Salivary Gland Cancer Patient-Derived Xenografts Enable Characterization of Cancer Stem Cells and New Gene Events Associated with Tumor Progression. Clin Cancer Res 24:2935-2943
Dionne, Lai Kuan; Peterman, Eric; Schiel, John et al. (2017) FYCO1 regulates accumulation of post-mitotic midbodies by mediating LC3-dependent midbody degradation. J Cell Sci 130:4051-4062
Keysar, Stephen B; Le, Phuong N; Miller, Bettina et al. (2017) Regulation of Head and Neck Squamous Cancer Stem Cells by PI3K and SOX2. J Natl Cancer Inst 109:
Morton, J J; Bird, G; Keysar, S B et al. (2016) XactMice: humanizing mouse bone marrow enables microenvironment reconstitution in a patient-derived xenograft model of head and neck cancer. Oncogene 35:290-300
Morton, J Jason; Bird, Gregory; Refaeli, Yosef et al. (2016) Humanized Mouse Xenograft Models: Narrowing the Tumor-Microenvironment Gap. Cancer Res 76:6153-6158
Gan, Gregory N; Altunbas, Cem; Morton, John J et al. (2016) Radiation dose uncertainty and correction for a mouse orthotopic and xenograft irradiation model. Int J Radiat Biol 92:50-6
Le, Phuong N; McDermott, Jessica D; Jimeno, Antonio (2015) Targeting the Wnt pathway in human cancers: therapeutic targeting with a focus on OMP-54F28. Pharmacol Ther 146:1-11
Gan, Gregory N; Eagles, Justin; Keysar, Stephen B et al. (2014) Hedgehog signaling drives radioresistance and stroma-driven tumor repopulation in head and neck squamous cancers. Cancer Res 74:7024-36
Jimeno, Antonio; Weiss, Glen J; Miller Jr, Wilson H et al. (2013) Phase I study of the Hedgehog pathway inhibitor IPI-926 in adult patients with solid tumors. Clin Cancer Res 19:2766-74
Keysar, Stephen B; Astling, David P; Anderson, Ryan T et al. (2013) A patient tumor transplant model of squamous cell cancer identifies PI3K inhibitors as candidate therapeutics in defined molecular bins. Mol Oncol 7:776-90

Showing the most recent 10 out of 14 publications