Humanized Mouse Models: Types, Applications, and Key Limitations

Humanized mouse models are now indispensable tools in biomedical research, enabling scientists to study human immune function, evaluate immunotherapies, and test drugs in a living system that closely mimics human biology. By engrafting human cells, tissues, or hematopoietic stem cells into immunodeficient mouse hosts, these models bridge the gap between reductionist cell-culture assays and unpredictable first-in-human trials — making them central to modern translational medicine.

What Is a Humanized Mouse Model?

A humanized mouse model is an immunodeficient mouse that has been engrafted with human immune cells, tissues, or hematopoietic stem cells (HSCs) to reconstitute a partial or near-complete human immune system within the murine host. The fundamental rationale is straightforward: conventional mice carry mouse-specific cell-surface antigens, cytokine networks, and MHC (H-2) haplotypes that do not accurately reflect human HLA-restricted immune responses. This species mismatch makes them poor surrogates for predicting how a human patient will respond to a pathogen, a tumor antigen, or a biologic drug.

Three properties of the mouse host determine how well human cells engraft and persist:

• Absence of adaptive immunity — T and B cell deficiency prevents rejection of the human graft. The scid mutation (impaired V(D)J recombination) was the first such lesion exploited, yielding the C.B-17 SCID mouse in 1983.
• Reduced innate immunity — Residual NK cell activity in SCID mice limits engraftment. The NOD background attenuates NK function and reduces macrophage-mediated clearance.
• IL-2 receptor common gamma-chain (IL-2Rγ) deletion — Knockout of Il2rg abolishes NK cell development and blocks signaling for multiple cytokines (IL-2, IL-4, IL-7, IL-9, IL-15, IL-21), creating the permissive environment needed for robust human cell engraftment.

Combining all three lesions yields the NSG mouse — formally designated NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ — which remains the most widely used immunodeficient background today (Shultz et al., Nat Rev Immunol, 2012).

Types of Humanized Mouse Models

Five principal humanized mouse model strategies are in routine use. Each differs in the source of the human graft, the reconstitution timeline, the immune lineages present, and the experimental window before graft-versus-host disease (GvHD) terminates the study.

A note on PDX and CDX classification: PDX and CDX models are technically xenograft models rather than humanized mouse models in the immunological sense. They engraft human tumor material into an immunodeficient host but do not reconstitute a human immune system. This distinction matters when designing combination studies: testing a checkpoint inhibitor in a PDX model requires co-engraftment of PBMCs or HSCs to provide the immune effectors that the drug is meant to engage.

The hu-HSC model warrants additional detail. CD34⁺ HSCs can be sourced from umbilical cord blood (highest CD34⁺ frequency, no donor mobilization needed), G-CSF–mobilized peripheral blood (adult donors, scalable), or fetal liver (richest multilineage potential, increasingly restricted by regulation). Each source differs in engraftment efficiency and the relative output of T, B, NK, and myeloid lineages. Shultz et al. (2007) demonstrated robust multilineage human immune reconstitution in NSG mice following neonatal intrahepatic injection of cord-blood CD34⁺ cells, with stable engraftment persisting beyond 20 weeks (Shultz et al., J Immunol, 2007).

Applications of Humanized Mouse Models in Biomedical Research

Humanized mouse models have reshaped how researchers approach translational questions across oncology, infectious disease, and immune-mediated disorders.

Immunotherapy Development

Testing CAR-T cells, bispecific antibodies, and PD-1/PD-L1 checkpoint inhibitors requires a human immune compartment that can respond to human tumor antigens. The hu-PBL model (PBMC injection) provides rapid T cell reconstitution within 1–2 weeks, making it suitable for short-duration efficacy studies of adoptive cell therapies or checkpoint blockade. The hu-HSC model is preferred for studies requiring de-novo T cell differentiation or evaluating vaccines that depend on germinal center reactions.

HIV and Infectious Disease Research

HIV-1 productively infects human CD4⁺ T cells and macrophages but does not replicate in murine cells, making humanized mice the only small-animal model for studying HIV pathogenesis in vivo. BLT and hu-HSC mice support systemic HIV-1 infection with CD4⁺ T cell depletion and viral reservoir establishment comparable to human disease. This system enabled pivotal preclinical testing of CCR5 antagonists for HIV entry inhibition (Denton et al., J Virol, 2008).

Hematological Malignancy Models

Primary human AML blasts engraft poorly in syngeneic immunocompetent mice but grow efficiently in SCID and NSG hosts. The landmark study by Lapidot et al. (1994) was the first to demonstrate that a rare CD34⁺CD38⁻ population within primary human AML could initiate leukemia in SCID mice, establishing the concept of the leukemia-initiating cell or leukemic stem cell (Lapidot et al., Nature, 1994). NSG mice dramatically improved engraftment of primary AML, ALL, and myeloma, enabling patient-specific drug sensitivity testing.

Oncology Drug Development and PDX Models

PDX models preserve the histological architecture, mutational landscape, and drug-response heterogeneity of the original patient tumor far better than established cell lines. Large PDX biobanks now enable co-clinical trials where mouse and patient receive the same treatment protocol, facilitating real-time biomarker discovery. However, the absence of an immune microenvironment in standard PDX limits their utility for immuno-oncology programs; this is addressed by co-engrafting human PBMCs or HSCs alongside the tumor graft.

Key Limitations of Humanized Mouse Models

Despite their transformative impact, no humanized mouse model fully recapitulates human immunophysiology. Researchers should be aware of the following constraints:

• Incomplete myeloid reconstitution. Monocytes, neutrophils, and tissue macrophages engraft poorly in conventional NSG mice because key myeloid survival and differentiation factors (M-CSF, GM-CSF, IL-3) are not cross-reactive across species. This blunts the model's utility for studying innate immune responses to infection or tumor-associated macrophage biology.
• Absent lymphoid architecture. HSC-humanized mice do not develop normal lymph node organization or germinal centers at physiological frequencies, limiting the modeling of affinity maturation and long-lived B cell memory.
• GvHD in PBMC models. Human T cells alloreact against murine MHC, causing lethal xenogeneic GvHD typically within 4–6 weeks of PBMC injection, compressing the experimental window.
• Species cross-reactivity gaps. Many murine cytokines do not efficiently signal through human receptors and vice versa, creating an artificial cytokine environment that may skew immune cell differentiation.
• HLA restriction. T cells educated in the mouse thymus (hu-HSC model) are selected on murine MHC, producing HLA-unrestricted T cells that may respond differently than human-thymus–educated cells in the BLT model.

Next-generation strains are addressing these gaps. The MISTRG mouse (NSG background with human M-CSF, IL-3, GM-CSF, TPO, and SIRPα knock-in) dramatically improves myeloid reconstitution and NK cell function (Rongvaux et al., Nat Biotechnol, 2014). MISTRG6 adds human IL-6, further improving B cell responses. DRAG mice carry human HLA-DR4 and HLA-A2, enabling MHC-matched T cell education without BLT surgery.

Products for Humanized Mouse Research from BioHippo

Supporting every stage of your humanized mouse experiment — from engraftment to readout — BioHippo offers a curated portfolio of reagents validated for human immune cell research.

• CD34⁺ HSC monitoring: Quantify engraftment efficiency with the Human CD34 ELISA Kit (CUSABIO, CSB-E14091h) or track stem cell surface expression with Anti-Human CD34 Antibody (QBEND-10), APC by flow cytometry.
• Human vs. mouse immune cell discrimination: The Anti-Human CD45 Antibody (BC8), PerCP allows precise gating of human leukocytes against residual murine CD45⁺ cells in peripheral blood and tissue digests — essential for reconstitution assessment by flow cytometry.
• Tumor xenograft cell lines: BioHippo's Cells collection includes AML cell lines (KG-1, KG-1a) and PDX-derived breast cancer cell lines (MC-BR-BTY-0006, MC-BR-BTY-0019) suitable for CDX and immune-oncology co-engraftment models.
• Cytokine profiling: Monitor human cytokine secretion post-engraftment with human-specific ELISA kits for IL-32, IL-34, CD40L, and other immune mediators — all available in the BioHippo ELISA catalog.

For custom humanized mouse model services or to discuss reagent suitability for your specific engraftment protocol, contact us at support@biohippo.com or visit our request-a-quote page.

Frequently Asked Questions

What is a humanized mouse model?

A humanized mouse model is an immunodeficient mouse engrafted with human immune cells, hematopoietic stem cells, or tissues so that it carries a functional human immune compartment. The humanized mouse allows researchers to study human-specific immune responses, test immunotherapies, and model human infectious diseases in a living organism — experiments that are not feasible in conventional mice because murine immune cells lack the human surface antigens, HLA haplotypes, and cytokine sensitivities required for authentic human immune biology.

How are humanized mouse models made?

The most common approach is to irradiate an NSG mouse (to eliminate residual innate immune activity) and inject human CD34⁺ hematopoietic stem cells — sourced from cord blood, mobilized peripheral blood, or fetal liver — intravenously or intrahepatically. Over 12–20 weeks, the HSCs home to the bone marrow and generate human T cells, B cells, NK cells, and dendritic cells. Alternatively, adult human PBMCs can be injected intraperitoneally (hu-PBL model) for rapid T cell reconstitution within 1–2 weeks, though at the cost of a shorter experimental window due to xenogeneic GvHD. The BLT model adds surgical implantation of human fetal thymus and liver fragments under the renal capsule, enabling HLA-restricted T cell education.

What is an NSG mouse?

The NSG mouse — formally NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ — is the current gold-standard immunodeficient background for humanized mouse research. It combines the NOD background (reduced NK cell function, impaired macrophage activity), the scid mutation in Prkdc (abolishes functional T and B cells by blocking V(D)J recombination), and a targeted null mutation in the IL-2 receptor gamma chain gene Il2rg (eliminates NK cells and blocks multiple cytokine-receptor signaling pathways). Together these three lesions create a profoundly immunodeficient host that accepts human cell grafts with high efficiency and low residual immune rejection.

What are the main limitations of humanized mouse models?

The principal limitations are: (1) incomplete myeloid reconstitution — monocytes and macrophages engraft poorly because murine cytokines such as M-CSF poorly cross-react with human receptors; (2) GvHD in PBMC models limits experiments to approximately 4 weeks; (3) absent or disorganized lymph node architecture impairs germinal center reactions and affinity maturation; (4) T cells educated in the mouse thymus are not HLA-restricted; and (5) species differences in pharmacokinetics, metabolic enzymes, and cytokine networks can complicate translation of drug PK/PD data to humans. Next-generation strains such as MISTRG (humanized M-CSF, IL-3, GM-CSF knock-in) partially overcome myeloid reconstitution deficiencies.

Which humanized mouse model is best for immunotherapy research?

The choice depends on the therapy class and study duration. For CAR-T cell and checkpoint inhibitor studies requiring a rapid readout (under 4 weeks), the hu-PBL model (PBMC injection into NSG mice) provides the fastest human T cell reconstitution. For longer-term studies involving de-novo lymphocyte differentiation, vaccine responses, or multilineage immune reconstitution, the hu-HSC model using CD34⁺ cord blood cells in NSG is preferred. If HLA-restricted antigen presentation is critical — for example, testing peptide vaccines or TCR-engineered T cells — the BLT model or emerging HLA knock-in strains (DRAG mice) should be considered. For tumor-immune co-engraftment studies combining a PDX with an immune compartment, hu-PBL or hu-HSC engraftment is layered onto the xenograft model.