|Home | About | Journals | Submit | Contact Us | Français|
Update on humanized mouse models and their use in biomedical research.
The recent description of immunodeficient mice bearing a mutated IL-2 receptor gamma chain (IL2ry) facilitated greatly the engraftment and function of human hematolymphoid cells and other cells and tissues. These mice permit the development of human immune systems, including functional T and B cells, following engraftment of hematopoietic stem cells (HSC). The engrafted functional human immune systems are capable of T and B cell-dependent immune responses, antibody production, anti-viral responses, and allograft rejection. Immunodeficient IL2rynull mice also support heightened engraftment of primary human cancers and malignant progenitor cells, permitting in vivo investigation of pathogenesis and function. In addition, human-specific infectious agents for which animal models were previously unavailable can now be studied in vivo using these new generation humanized mice.
Immunodeficient mice bearing an IL2rynull mutated gene can be engrafted with functional human cells and tissues, including human immune systems, following engraftment with human hematolymphoid cells. These mice are now used as in vivo models to study human hematopoiesis, immunity, regeneration, stem cell function, cancer, and human-specific infectious agents without putting patients at risk.
Studies of human cell and tissue function have traditionally been limited to ex vivo analyses, non-invasive procedures, or clinical trials that are costly and severely limited due to ethical constraints. Small animal models of human cell and tissue function would overcome these limitations. A major breakthrough in the generation of humanized mice was the development of immunodeficient mice bearing a targeted IL2ry mutation. These mice permit functional in vivo studies of human cells and tissues [1–3]. This report summarizes recent progress in the use of humanized mice for the study of human diseases.
Human cell and tissue functions in murine hosts have been investigated since the first description of athymic (nude) mice . Humanization of immunodeficient mice advanced greatly with the discovery of the severe combined immune deficiency (scid) mutation  and the knockout of recombination activating genes 1 and 2 [6;7]. Additional manipulation of the mouse genome by knockout and transgenic technology led to increased engraftment and function of human cells and tissues . The recent description of immunodeficient mice bearing a mutated IL-2 receptor gamma chain (IL2rynull) facilitated greatly the in vivo engraftment and function of human cells. The history of humanized mice in biomedical research and the tremendous advances in this field resulting from the generation of immunodeficient mice with a mutated IL2ry gene has been reviewed recently [1–3]. The present review briefly summarizes the different models of humanized mice available for experiments, highlights recent advances in the model systems, and summarizes new findings that have emerged over the last year.
We use a simple definition of humanized mice as “mice engrafted with functional human cells or tissues or expressing human transgenes.” Depending on the experimental question, different models of immunodeficient humanized mice are utilized (Table 1). Hu-PBL-SCID mice are engrafted with peripheral blood mononuclear cells [PBMC, 8]. Hu-SRC-SCID mice are engrafted with human hematopoietic stem cells [HSC, 9]. SCID-hu mice are engrafted with human fetal liver and thymus . These models represent humanized mice engrafted with functional human immune systems. Immunocompetent mice expressing human transgenes also provide insights into human biology [11–13]. This review will focus on humanized mice engrafted with functional human cells and tissues.
Human cell and tissue function was enhanced greatly by the generation of immunodeficient IL2rynull mice [1–3], but there remain a number of limitations that are continuously being recognized and overcome. One of the first questions in the field arose from the bewildering number of available mouse strains and engraftment protocols. Recognizing this problem, standardized methodologies for establishment of humanized mice were recently published [14;15]. Optimal approaches for engraftment of newborn and adult NOD-scid IL2rynull mice with HSC (Hu-SRC-SCID) as well as engraftment of human PBMC (Hu-PBL-SCID) are detailed by Pearson et al . Optimal approaches for engraftment of BALB/c-Rag2null IL2rynull mice with human HSC (Hu-SRC-SCID) are described by Legrand et al . It is recommended that investigators working with humanized mouse models initially establish the models based on these standardized guidelines.
The second question is the optimal recipient strain. HSC engraftment in different strains of immunodeficient mice following intrahepatic injection into newborns has been compared . NOD-scid IL2rynull mice and BALB/c-Rag2null IL2rynull l mice are equivalent in their generation of human B cells, dendritic cells (DC), and platelets whereas NOD-scid IL2rynull mice are superior in supporting human T cell development. Fetal liver and umbilical cord blood (UCB) HSC supported higher percentages of human engraftment than G-CSF-mobilized peripheral blood HSCs. Bone marrow HSC was not tested in this report. We have confirmed that UCB HSC-engrafted newborn NOD-scid IL2rynull mice are superior to BALB/c-Rag1null IL2rynull mice in their ability to support human T cell development, and further found that intrahepatic and intracardiac (i.v.) injections are equivalent (Brehm et al, submitted). The NOD vs. BALB/c support of T cell engraftment is not based on the scid vs. Rag1/2null mutations as NOD-scid IL2rynull mice and NOD-Rag1null IL2rynull mice are equivalent in their support of human HSC engraftment . These data suggest that investigators establishing humanized mice requiring a fully functional human immune system in the Hu-SRC-SCID model should consider basing their work on newborn engraftment of NOD-scid IL2rynull or NOD-Rag1null IL2rynull mice.
Another limitation being addressed is the species-specificity of a number of molecules. Examples include species-specific human cytokines. Transgenic expression of human IL15 enhanced human NK cell development and differentiation in HSC-engrafted BALB/c-Rag2null IL2rynull mice, and documented the critical role that IL15 trans-presentation has in regulating human NK cell homeostasis [18*]. A second cytokine, B Lymphocyte Stimulating factor (BLyS also termed BAFF) is important in B cell survival and differentiation [19;20]. Mouse BLyS cannot support human B cell survival, and administration of human recombinant BLyS to NOD-Rag1null Prf1null mice engrafted with human PBMC increased human B and, surprisingly, T cell engraftment [21*]. Generation of BLyS transgenic NOD-scid IL2rynull mice should enhance human B and T cell immune function in humanized mice and creation of these transgenic mice as well as other human-specific cytokine transgenic mice is underway .
Humanized mice have been used in the past year to investigate multiple types of human immune responses and to test potential therapeutics that modulate human immunity. The Hu-PBL-SCID model has an ~30 day window of analysis due to development of xenogeneic graft-versus-host disease [GVHD, 23]. Investigators have used this observation to establish an in vivo model of human immune-mediated GVHD. Using the Hu-PBL-SCID model based on NOD-scid mice, activation of human regulatory T cells by HIV-1 envelope protein gp120 delayed development of GVHD . A similar delay of GVHD was observed in the Hu-PBL-SCID model based on NOD-scid IL2rynull mice following treatment with soluble Fas ligand . King et al examined kinetics of engraftment and development of GVHD in the Hu-PBL-SCID model based on NOD-scid IL2rynull mice [26*]. They observed that most of the GVHD was directed against mouse MHC class I and II and mice deficient in MHC class I exhibited delayed GVHD. TNF inhibitors are used in the clinic to treat GVHD [27;28], and similarly, etanercept, a TNF inhibitor delayed the development of GVHD in this model system [26*].
Additional analyses of human T and B cells in Hu-SRC-SCID mice generated following engraftment of adult NOD-scid IL2rynull mice revealed that although human B cells develop, they are developmentally blocked , likely due to the inability of mouse BLyS to signal human B cells [21*]. The authors further suggested that human T cells selected in the thymus on mouse MHC class II is at least partially responsible for decreased human T cell immune responses.
One approach to enhance human T cell selection in the mouse thymus is to provide a human thymus autologous with the human HSC. This model, termed SCID-hu (Table 1), has been used extensively in the study of infectious agents (see below), and was recently used to establish a porcine islet xenograft rejection model  and an approach for induction of xenograft tolerance [31*]. Human HSC model systems supporting rejection of human allografts have to date not been reported. However, SCID-beige immunodeficient mice engrafted with human HSC generate multiple hematopoietic cells but not T cells [16;32]. Macrophages infiltrated human skin allografts in these mice, but produced little injury. However, when combined with adoptive transfer of autologous T cells to activate the infiltrating macrophages, the macrophages produced intimal expansion and calcification, reminiscent of atherogenesis or end-stage renal disease [33;34]. Regarding T cell-mediated allograft rejection, Racki et al demonstrated using the Hu-PBL-SCID model based on NOD-scid IL2rynull mice that either purified human CD4 or CD8 T cells can mediate human skin allograft rejection .
Additional human immune responses have been described in humanized mice. A model for asthma was used to identify a role for DC-derived CCL17 and CCL22 in attraction of Th2 cells and induction of airway inflammation . In a humanized mouse model of sepsis, human lymphocyte apoptosis and cytokine production recapitulated the findings in patients with septic shock . As new models are generated, investigation of multiple aspects of both immune and autoimmune responses of human immune systems will be possible.
One of the most prevalent uses of humanized mice is the study of human-specific infectious agents, particularly HIV [38;39]. Using the SCID-hu system, robust virus-specific immune responses following HIV infection were observed [40*]. Despite these robust responses, HIV viral load remained high and correlated with increased PD1 expression on human T cells, an observation also made in humans [41–43]. In a model based on NOD-scid Jak3null mice, a nucleoside reverse transcriptase inhibitor blocked HIV infection in Hu-PBL-SCID mice . Using the Hu-SRC-SCID model based on BALB/c-Rag1null IL2rynull mice, in vivo RNAi gene therapy against HIV-1 was investigated . Human HSC were transduced with a lentiviral vector expressing a shRNA against HIV-1 nef gene and engrafted into newborn BALB/c-Rag1null IL2rynull mice. Evidence was obtained that the mature human CD4 T cells recovered from the HSC-engrafted mice exhibited an inhibition of virus replication, confirming efficacy of the shRNA therapy.
In a key series of experiments using both the Hu-PBL-SCID and Hu-SRC-SCID models based on NOD-scid IL2rynull mice, Kumar et al validated a novel new drug for the prevention and treatment of HIV infection [46**]. They used a modified single chain antibody (scFv) to the human T cell marker CD7 to deliver siRNA in vivo against CCR5 and viral Vif and Tat genes to human CD4 T cells in humanized mice. They documented that HIV infection could be controlled in a prophylactic setting in both model systems when viral challenge was performed after initiation of siRNA treatment, as well as in a post-infection setting, where mice were engrafted with PBMC from an HIV-infected subject.
A number of reports have described Dengue virus humanized mouse models for which no animal model system previously existed. Using newborn HSC-engrafted NOD-scid IL2rynull mice infected with eight different viral strains representing the four genotypes of Dengue viruses, viremia, a thrombocytopenia, increase in body temperature and erythema were observed corresponding to clinical characteristics in Dengue-infected humans [47;48].
Another approach in addition to the SCID-hu model to enhance human T cell selection during development in the thymus is to transgenically express human HLA in the mouse recipient. In a report using Hu-SRC-SCID mice based on NOD-scid IL2rynull HLA-A2 transgenic mice engrafted with HLA-A2 HSC, Jaiswal et al documented the development of virus-specific HLA-A2-restricted human T cell responses to Dengue virus infection [49*]. This is one of the first two reports using HLA-transgenic mice to document a human T cell HLA-restricted immune response. The other report used Epstein Barr Virus (EBV) infection in a model also based on NOD-scid IL2rynull HLA-A2 transgenic mice engrafted with HLA-A2 HSC [50**]. HLA-A2-restricted cytotoxic and IFNy-producing human T cells against multiple EBV HLA-A2 epitopes were observed exhibiting similar patterns of reactivity to that detected in human EBV carriers. These two reports document that HLA expressed transgenically in mouse thymus can positively select developing T cells and lead to HLA-restricted immune responses in mice engrafted with human HSC. Development of additional HLA-transgenic immunodeficient mice is currently underway [2;39].
NOD-scid IL2rynull mice engrafted as newborns with HSC were used to document a novel approach for enhancing immune responses following immunization. Targeting EBV antigen to human DC in vivo stimulated human T cell responses to EBV and induced anti-EBV antibody responses . Similarly, NOD-scid β2mnull mice engrafted with human HSC and autologous mature T cells and then infected with live attenuated trivalent influenza virus generated human T cell responses to influenza . The authors proposed this as a model for investigating antigen-presenting cells in immune responses as the response was completely dependent on reconstitution of the human myeloid compartment.
Finally, the first humanized mouse model for the study of Plasmodium falciparum was recently reported [53*]. NOD-scid IL2rynull mice were injected repeatedly with human red blood cells, which could then support productive infection with P. falciparum. Therapeutic efficacy of a number of anti-malarial agents was tested, permitting determination of the protective ED90 of the drugs against infection.
Overall, infectious disease studies in humanized mice are providing important pre-clinical model systems for the investigation of the pathogenesis of human-specific infectious agents, evaluation of therapeutics, and platforms to understand mechanisms of human immune responses following vaccination.
Immunodeficient IL2rynull mice permit engraftment with a number of primary human tumors . Comparing growth of human melanoma lung metastases in NOD-scid, NOD-scid β2mnull, and NOD-scid IL2rynull mice  the absolute NK deficiency in NOD-scid IL2rynull mice appeared to be a major factor in its enhanced support of primary tumor growth . The enhanced engraftment of NOD-scid IL2rynull mice was confirmed in a report showing that human acute leukemia cells generate a faster and more efficient disease as compared to that observed in NOD-scid and NOD-scid β2mnull mice . Human immune responses via antibody-dependent cellular cytotoxicity against primary adult T cell leukemia/lymphomas, Hodgkin lymphoma, and cutaneous T cell lymphoma in NOD-scid IL2rynull mice could be potentiated by defucosylated anti-CCR4 antibody [56;57*], suggesting a novel approach to enhance anti-tumor immunity. Confirming the utility of humanized mice for studies of cancer, it was shown that primary lung tumors transplanted into NOD-scid IL2rynull mice recapitulated in vivo tumor characteristics, including maintenance of stroma and passenger leukocytes that could, when activated with IL12, become tumor effector CTLs .
Cellular therapy for treatment of multiple human diseases, particularly diabetes, is a promising approach. Successful human islet transplantation , has expanded into a number of clinical trials (www.clinicaltrials.gov). Because sufficient islets are not available for the tremendous need, focus has turned to development of beta cells from stem cells (http://www.betacell.org/). However, animal models for testing the safety and efficacy of this and other forms of human cellular therapy, such as regulatory T cells  or embryonic stem cells , are needed and humanized mice are being developed to address these critical needs. For example, immunodeficient NOD-Rag1null Prf1null Ins2Akita mice that spontaneously develop a non-autoimmune hyperglycemia [62–65] can be engrafted with human islets . This model can be used to test the in vivo function of human beta stem cells. Backcrossing of the Ins2Akita mutation to the NOD-Rag1null IL2rγnull strain will permit human beta stem/progenitor cells to be transplanted into mice bearing a functional human immune system [17;67], similar to the situation that occurs during transplantation in the clinic.
Humanized mice as pre-clinical models for the in vivo study of human cells and tissues have been under development for over 30 years. With the recent generation of immunodeficient IL2rγnull mice, the ability of humanized mice to serve as preclinical models is becoming a reality, but not yet ideal. Additional modifications of the model systems and genetic manipulation of the host continue to improve the ability of humanized mice to more accurately recapitulate the in vivo function of human cells and tissues. Novel insights into human disease are now possible due to the availability of humanized mice wherein human cells and tissues can be studied in vivo without putting patients at risk.
This work was supported by National Institutes of Health (NIH) Grants AI46629, DK53006, HL077642, an institutional Diabetes Endocrinology Research Center (DERC) grant DK32520, a Cancer Center Core grant CA34196, the Beta Cell Biology Consortium, grants from the Juvenile Diabetes Foundation, International, the Brehm Foundation, and the Helmsley Foundation. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
Funding National Institutes of Health, the Juvenile Diabetes Foundation, International, The Brehm Foundation, and the Helmsley Foundation.
The authors have no conflict of interest to report.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.