Humanized Mouse Models for Immuno-Oncology Studies
Immunocompetent mouse models are used for many immuno-oncology applications but have limitations regarding specific human immunological mechanisms. These limitations can be partially overcome by using humanized mice.
To use humanized mice efficiently, and to be sure that you’re using the right model to answer your specific questions, the presence and potency of the different human immune cell populations in each model need to be fully understood.
Humanized Mice Generated with Human PBMCs
Humanized mice are developed by injecting PBMCs into immunodeficient mice, such as NSG™/NOG® and BRG mice. PBMC engraftment supports short-term studies, creating transient, partially-reconstituted human immunity.
Specifically, a high level of functional and educated T cell populations are present, with the donor immune repertoire transfer retaining the immune memory and antigen specificity from the donor.
Uniform T cell activation is observed, dependent upon human PBMC xeno-reactivity with foreign host major histocompatibility (MHC) class I and class II. This leads to rapid GvHD occurrence in a few weeks, limiting length of studies. In order to delay the onset of GvHD and enable longer term studies, knockout mice for MHC class I (MHC-KbDb -/-) and/or MHC class II (H2-Ab1-/-) have been successfully used.
Other Hematopoietic Lineages
Engraftment of other important hematopoietic lineages is not supported in this model. This is due to the rapid expansion of T cells, and the lack of cross reactivity of important cytokines for myeloid or lymphoid differentiation, maturation, or activation e.g. IL‐3, IL‐15, GM‐CSF, and macrophage colony stimulating factor (M‐CSF).
For this reason, the majority of myeloid derived cells (i.e. macrophages and DC) are absent and NK cells are present at only a low level.
Also, although the donor humoral repertoire is transferred, PBMC humanized mice show a low level of human B cells with no de novo immune response. This low B cell level is due to a decrease in the proportion of B cells over time, reflecting the strong expansion of T cells. Nevertheless, the absolute number of B cells increases slightly over time and some antibody production has been observed.
Overall, PBMC humanized models can provide significant insight into human tumor immunology and treatment efficacy, targeting T cell function (e.g. agents such as BiTEs, immune checkpoint inhibitors/agonists, CAR-T cell therapy).
Humanized Mice Generated through Hematopoietic Stem Cell Engraftment
Alternatively, using human CD34+ hematopoietic stem cells (HSCs) provides a longer term, stable model of human immune system engraftment. Humanized mice are generated through inoculating human hematopoietic cells (from cord blood stem cells) into immunocompromised mice (e.g. NOG, NSG). For immunotherapies targeting immune cells of myeloid lineages, transgenic immunocompromised mice expressing GM-CSF and IL-3 are more suitable models.
HSC-humanized mice cannot support robust adaptive immune responses. Poor B cell response is seen, due to incomplete B cell maturation with low levels of IgM and limited class-switching to IgG. To overcome this limitation, hIL-6 knock-in mice have been successfully used. These models show improved T cell engraftment and serum IgG production, with IgG-switched memory B cells displaying a diverse antibody repertoire and high specificity against immunized antigens.
This limitation has also been worked around by expression of human GM-CSF and IL-4 in humanized mice. This improves class-switching, producing significant levels of antigen-specific IgG following immunization.
Immunocompromised mice routinely used in CD34+ models do not express HLA molecules on thymic epithelial cells. Due to this limitation T cells are educated and selected on mouse MHC (H2 antigen). This results in T cells unable to recognize antigens in an HLA-restricted manner. To overcome this problem, HLA class I (HLA-A2.1tg or HLA-A2/HHD) and II (HLA-DR1tg or HLA-DR4tg) transgenes were added into NSG mice allowing the development of human T cell repertoires and responses.
Alternatively, BLT (bone marrow, liver, and thymus) mice can also overcome this limitation. The hBLT model is generated by co-xenotransplantation of human fetal liver and thymic tissue under the murine renal capsule, as well as intravenous autologous CD34+ HSC injection. This model shows superior engraftment of hematopoietic lineages with HLA restricted T cells, and It is the only model that leads to the generation of a human mucosal immune system. However, this model is difficult to establish due to the surgical procedure required, as well as time-consuming to develop.
NK Cells and Myeloid Compartment
Due to poor interspecies cytokine cross-reactivity, NK cells are largely absent in HSC-humanized mice, and have impaired functionality. With only 65% homology between mouse and human IL-15, supplementation of hILl-15 is essential for development and functionality of human NK cells. Transgenic hIL-15 immunodeficient mice have been established showing significantly higher levels of human NK cells in NSG-hIL15 compared to NSG mice. Also, higher proportions of human NK cells express granzyme A, B, and perforin compared to NK cells in NSG mice.
The myeloid compartment is also poorly reconstituted in NSG humanized mice. Human myelopoiesis is not efficient in these models, due to the lack of species cross-reactivity of nonhematopoietic cell-derived growth factors including CSF-1, GM-CSF, IL-3, and erythropoietin.
Super immunodeficient murine backgrounds such as NSG-SGM3 and NOG-EXL support greater human myeloid cell reconstitution than the NSG and NOG. HuNSG-SGM3 expresses human SCF, GM-CSF, IL-3, and has higher engraftment of monocytes, macrophages, and dendritic cells (DC) compared with the huNSG model. The huNOG-EXL mouse expresses human GM-CSF and IL-3, with higher levels of myeloid cell differentiation following human HSC engraftment compared with huNOG.
Regarding DC, it has been published that FLT3-ligand treatment of humanized mice results in higher DC frequency and functionality.
In summary, the hCD34+ model is used more when you need to assess agents which require multi-lineage immunity. As the human immune system is more fully reconstituted, these models are also useful to investigate more specific mechanistic questions surrounding the stimulation or suppression of the immune response. With the wide therapeutic window, hCD34+ mice are also great for long term studies such as assessing memory response and long-term antitumor effects.
Hopefully this guide has provided a brief introduction to the two commonly used humanized model approaches, setting out the advantages and limitations of PBMC and HSC-humanized models, and where you might start in choosing the right model for your studies based on immune cells needed.