In recent decades, chimeric antigen receptor (CAR)–engineered immune effector cells have demonstrated promising antileukemic activity. Nevertheless, their efficacy remains unsatisfactory on solid cancers, plausibly due to the influence of tumor microenvironments (TME). In a novel mouse cancer model with a humanized immune system, tumor-infiltrating immunosuppressive leukocytes and exhausted programmed death protein-1 (PD-1) ^high T cells were found, which better mimic patient TME, allowing the screening and assessment of immune therapeutics. Particularly, membrane-bound programmed death ligand 1 (PD-L1) level was elevated on a tumor cell surface, which serves as an attractive target for natural killer (NK) cell–mediated therapy. Hematopoietic stem cell–derived CAR-NK (CAR pNK) cells targeting the PD-L1 showed enhanced in vitro and in vivo anti-solid tumor function. The CAR pNK cells and nivolumab resulted in a synergistic anti-solid tumor response. Together, our study highlights a robust platform to develop and evaluate the antitumor efficacy and safety of previously unexplored therapeutic regimens.

Immune checkpoint blockade (ICB) monotherapy shows early promise for the treatment of nasopharyngeal carcinoma (NPC) in patients. Nevertheless, limited representative NPC models hamper preclinical studies to evaluate the efficacy of novel ICB and combination regimens. In the present study, we engrafted NPC biopsies in non-obese diabetic-severe combined immunodeficiency interleukin-2 receptor gamma chain-null (NSG) mice and established humanized mouse NPC-patient-derived xenograft (NPC-PDX) model successfully. Epstein–Barr virus was detected in the NPC in both NSG and humanized mice as revealed by Epstein–Barr virus-encoded small RNA (EBER) in situ hybridization (ISH) and immunohistochemical (IHC) staining. In the NPC-bearing humanized mice, the percentage of tumor-infiltrating CD8⁺ cytotoxic T cells was lowered, and the T cells expressed higher levels of various inhibitory receptors, such as programmed cell death protein 1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) than those in blood. The mice were then treated with nivolumab and ipilimumab, and the anti-tumor efficacy of combination immunotherapy was examined. In line with paired clinical data, the NPC-PDX did not respond to the treatment in terms of tumor burden, whilst an immunomodulatory response was elicited in the humanized mice. From our results, human proinflammatory cytokines, such as interferon-gamma (IFN-γ) and interleukin-6 (IL-6) were significantly upregulated in plasma. After treatment, there was a decrease in CD4/CD8 ratio in the NPC-PDX, which also simulated the modulation of intratumoral CD4/CD8 profile from the corresponding donor. In addition, tumor-infiltrating T cells were re-activated and secreted more IFN-γ towards ex vivo stimulation, suggesting that other factors, including soluble mediators and metabolic milieu in tumor microenvironment may counteract the effect of ICB treatment and contribute to the tumor progression in the mice. Taken together, we have established and characterized a novel humanized mouse NPC-PDX model, which plausibly serves as a robust platform to test for the efficacy of immunotherapy and may predict clinical outcomes in NPC patients.

Cancer therapeutics have undergone immense research over the past decade. While chemotherapies remain the mainstay treatments for many cancers, the advent of new molecular techniques has opened doors for more targeted modalities towards cancer cells. Although immune checkpoint inhibitors (ICIs) have demonstrated therapeutic efficacy in treating cancer, adverse side effects related to excessive inflammation are often reported. There is a lack of clinically relevant animal models to probe the human immune response towards ICI-based interventions. Humanized mouse models have emerged as valuable tools for pre-clinical research to evaluate the efficacy and safety of immunotherapy. This review focuses on the establishment of humanized mouse models, highlighting the challenges and recent advances in these models for targeted drug discovery and the validation of therapeutic strategies in cancer treatment. Furthermore, the potential of these models in the process of uncovering novel disease mechanisms is discussed.

The paired box 6 (PAX6) transcription factor is crucial for normal pancreatic islet development and function. Heterozygous mutations of PAX6 are associated with impaired insulin secretion and early-onset diabetes mellitus in humans. However, the molecular mechanism of PAX6 in controlling insulin secretion in human beta cells and its pathophysiological role in type 2 diabetes (T2D) remain ambiguous. We investigated the molecular pathway of PAX6 in the regulation of insulin secretion and the potential therapeutic value of PAX6 in T2D by using human pancreatic beta cell line EndoC-βH1, the db/db mouse model, and primary human pancreatic islets. Through loss- and gain-of-function approaches, we uncovered a mechanism by which PAX6 modulates glucose-stimulated insulin secretion (GSIS) through a cAMP response element–binding protein (CREB)/Munc18-1/2 pathway. Moreover, under diabetic conditions, beta cells and pancreatic islets displayed dampened PAX6/CREB/Munc18-1/2 pathway activity and impaired GSIS, which were reversed by PAX6 replenishment. Adeno-associated virus–mediated PAX6 overexpression in db/db mouse pancreatic beta cells led to a sustained amelioration of glycemic perturbation in vivo but did not affect insulin resistance. Our study highlights the pathophysiological role of PAX6 in T2D-associated beta cell dysfunction in humans and suggests the potential of PAX6 gene transfer in preserving and restoring beta cell function.