Important advances have been achieved in the development of immunodeficient mouse strains for humanization, with a number of different options that can be used according to the research interest1. Provided here is a general protocol for the humanization of NS γ-chainnull mice and genetically similar strains to be employed in three different models for studying HIV infection. In the first experimental approach, irradiated newborn mice are injected with human CD34+ HSCs, which can be derived from cord blood, fetal liver, or mobilized peripheral blood3,21. Appropriate irradiation of NS γ-chainnull mice is a critical step, as it eliminates the mouse bone marrow and other progenitor cells, allowing efficient reconstitution of human cell populations. However, some reports have evidenced reconstitution of human cells in different mouse strains, without irradiation27. In this regard, proper doses of irradiation must be provided, since NS γ-chainnull mice are radiosensitive, and high γ-irradiation could induce thymic lymphomagenesis21,28.
Other critical steps and factors that could affect the level of engraftment include the route of injection (intrahepatic, intravenous, intracardiac), mice age, percentage of purity of CD34+ HSCs, and operator expertise29. In the second and third approaches based on hu-PBL-NS γ-chainnull mouse models, some critical factors include the route of injection (intraperitoneal, intravenous, intrasplenic), mice age, and number of human cells injected, which can influence the final level of engraftment. Regarding this latter factor, several studies have used 5–10 x 106 PBMCs for engraftment22,23,30, whereas the present protocol suggests the use of 3.5 x 106 PBMCs. Of note, this number of cells is sufficient for the reconstitution of T-cells and for HIV replication, both in the acute and reactivation models, and also delays the development of GVHD23. Nonetheless, investigators should optimize the humanization conditions according to the research objectives. Moreover, it is important to validate the HIV strain used for infection of huNS γ-chainnull mice. Here, the R5 tropic HIV-1 BaL strain is used, which yields high levels of viral replication in huNS γ-chainnull mice. Other reporter strains, such as those containing luciferase or fluorescent proteins, are also suitable for single-cell analysis of HIV-infected cells31.
Overall, three major limitations are evidenced in huNS γ-chainnull mouse models following engraftment with CD34+ HSC. First, due to the absence of a human thymic environment, T-cells are educated in the context of murine MHC molecules, restraining subsequent antigen-specific stimulation via their T-cell receptors. This issue limits the use of NS γ-chainnull mouse models for studying the HIV-specific T-cell response. Nonetheless, this limitation can be overcome by the use of BLT mice or NS γ-chainnull mice with transgenic expression of HLA molecules16,17. Second, typically there is poor reconstitution of myeloid populations in NS γ-chainnull mouse models, limiting the study of these subsets that have relevance in the context of antigen-presentation and pathogenesis of HIV infection14,15. In this case, the use of mouse strains with transgenic expression of hematopoietic factors is recommended8,15,32.
Third, there is a 1) poor development of lymphoid follicle structures in the secondary lymphoid tissues and 2) lack of tertiary lymphoid tissues, which is related to the low levels of innate immune cells (i.e., dendritic cells in huNS γ-chainnull mice) that are critical for the development of follicles33. This issue is associated with a poor humoral response in huNS γ-chainnull mouse models34. Nonetheless, some reports have evidenced the development of follicle-like structures in huNS γ-chainnull mice4, whereas spleen- and lymph node-confined follicular T-cells (expressing the follicle-homing chemokine receptor CXCR5) are detected in huNS γ-chainnull mice and related strains15). Again, the use of 1) BLT mice or 2) mouse strains with transgenic expression of hematopoietic factors and/or with expression of HLA molecules can improve the reconstitution of myeloid populations, development of organized secondary and tertiary lymphoid structures, and effective T-cell and B-cell responses8,35,36.
Similar to the limitations of CD34+ HSC-humanized NS γ-chainnull mouse models, there is a lack of antigen-specific T-cell and humoral responses, absence of myeloid populations, and organized lymphoid structures in hu-PBL-NS γ-chainnull mice. In addition, an important limitation of the hu-PBL-NS γ-chainnull mouse model (acute and reactivation models of HIV infection) is the short window for monitoring, since these mice develop xenogeneic GVHD23. The development of GVHD could also induce undesired phenotypic and functional changes of immune populations, inherent of the pathogenic process23,37. Nonetheless, this model has the advantage of being simpler and more accessible, particularly considering that human PBMCs are more easily acquired from healthy or HIV-infected donors38. In addition, the injection of primary cells directly from patients is useful for the study of cell- or pathogen-intrinsic conditions of the donor, such as viral drug resistance mutations or donor-specific immune alterations. Of note, for the reactivation model, in vitro assays with HIV reactivation agents can be performed to corroborate the response of PBMCs before injection into mice39. Another limitation for some institutions is that this work requires BSL2+ facilities to handle HIV-infected animals due to regulations.
The huNS γ-chainnull mouse models have some advantages in comparison with other animal models for studying HIV infection, such as nonhuman primates infected with simian immunodeficiency virus. For instance, huNS γ-chainnull mice allow the creation of gene knockout or transgenic strains, which permit the evaluation of specific gene targets. Additionally, the use of primary human cells in huNS γ-chainnull mice avoids possible species-specific restrictions, such as the case of interferon-stimulated genes in nonhuman primates, which can influence the antiviral response and course of infection40. Thus, the kinetics of infection are highly consistent between huNS γ-chainnull mice. Finally, huNS γ-chainnull mouse models are less expensive, do not require complex core facilities, and are more accessible.
In summary, CD34+ HSC-humanized and hu-PBL-NS γ-chainnull mouse models offer a variety of possibilities for the study of chronic, acute, and reactivation events in HIV infection. With the recognition and overcoming of the aforementioned limitations of these models, the use of NS γ-chainnull mice may be a powerful tool for virological, immunological, and drug pre-clinical studies, as well as for genome editing and cell-based immunotherapies.
