Progress, implications, and challenges in using humanized immune system mice in CAR‐T therapy—Application evaluation and improvement

Abstract

In recent years, humanized immune system (HIS) mice have been gradually used as models for preclinical research in pharmacotherapies and cell therapies with major breakthroughs in tumor and other fields, better mimicking the human immune system and the tumor immune microenvironment, compared to traditional immunodeficient mice. To better promote the application of HIS mice in preclinical research, we selectively summarize the current prevalent and breakthrough research and evaluation of chimeric antigen receptor (CAR) ‐T cells in various antiviral and antitumor treatments. By exploring its application in preclinical research, we find that it can better reflect the actual clinical patient condition, with the advantages of providing high‐efficiency detection indicators, even for progressive research and development. We believe that it has better clinical patient simulation and promotion for the updated design of CAR‐T cell therapy than directly transplanted immunodeficient mice. The characteristics of the main models are proposed to improve the use defects of the existing models by reducing the limitation of antihost reaction, combining multiple models, and unifying sources and organoid substitution. Strategy study of relapse and toxicity after CAR‐T treatment also provides more possibilities for application and development.

1. Summaries: current prevalent and breakthrough HIS mice applications of CAR‐T in various antiviral and anti‐tumor treatments.

2. HIS mice have better clinical patient simulation and promotion for the updated design of CAR‐T cell therapy than directly transplanted immunodeficient mice, but still have varies extents of defects.

3. Improvement and discussion: reducing the limitation of anti‐host reaction, combining multiple models, unifying sources and organoids substitution, etc.

4. Strategy Study of relapse and toxicity after CAR‐T treatment provided more possibilities.

1. INTRODUCTION

Humanized immune system (HIS) mice simulate the human immune system by implanting human hematopoietic stem cells (HSCs), lymphocytes, or immune tissues into immunodeficient mice. HIS mice were initially limited to studies on functional composition and response mechanisms of the human immune system, ¹ , ² , ³ in which human CD34⁺ HSCs were mostly used. HIS mice can also be used for the discovery and validation of novel viral transmission routes. ⁴ , ⁵ In recent years, HIS mice have been effective in pharmacotherapy efficacy evaluation and preclinical research and have been widely used in tumor immune research and development. In contrast, the human tumor transplantation model based only on immunodeficient mice lacks a human immune system and a tumor immune microenvironment, limiting the transformation research on immune mechanism and immunotherapy to some extent. For example, when the effects of CAR‐T cells on refractory glioblastoma were evaluated, the results differ from those of HIS mice, which is closer to the actual therapeutic efficacy of the human body. ⁶

The results of efficacy evaluation are quite different from those of clinical practice, whereas HIS mice of the immune system can attach the animal model to the human immune system. After their autoimmune function is reduced or even removed using immunodeficiency treatment like sublethal radiation, HIS is reconstructed by the transplantation of HSCs, peripheral blood mononuclear cells (PBMCs), immune cells, or immune organs. At present, HIS mice are mainly divided into three categories, the main difference being the different sources of constructing the human immune microenvironment. Models such as the humanized‐peripheral blood mononuclear cells (Hu‐PBMC), humanized‐hematopoietic stem cells (Hu‐HSC), and humanized‐bone marrow, liver, thymus (Hu‐BLT) are, respectively, constructed from human peripheral blood, human HSCs, and fetal liver or thymus. ⁷ , ⁸

CAR‐T therapy, as a tumor‐targeting device consisting of T lymphocytes and CAR tumor chimeric antigen receptors, has become a revolutionary immunotherapy for treating specific cancers. ⁹ It is necessary to consider enhancing the homology between CAR‐T cells and patients to exclude host‐specific rejection to estimate antitumor and antiviral drug development. The application of CAR‐T research based on HIS mice will have a more comprehensive and effective development value.

This review summarizes and discusses the construction of three HIS mouse models and their application to CAR‐T treatment (Tables 1; Figure 1), pointing out a series of feasible measures and suggestions for improvement so that HIS mice can be better used in CAR‐T treatment‐related research.

TABLE 1.

Summary of HIS mouse models.

Model	Human tissue	Mouse strain	Construction strategy	Peripheral composition	Merit	Defect
Hu‐HSC	Mobilized peripheral/umbilical cord blood, fetal, liver, and bone marrow–derived human HSCs	NPG/BNDG/NCG/NOG/NSG/RGSF/ BRGS	1. Irradiation/busulfan 2. Intravenous/intrahepatic/intracardiac (new born); intravenous/intrafemoral/intraperitoneal (adult)	Lymphoid T, B, NK cells, myeloid monocytes and macrophage cells	1. No/weak GVHD 2. Real and accurate clinical patient simulation 3. Long‐term research	1. Long reconstruction period: 2–4 months 2. Lack of human thymus, weak function of T lymphocytes
Hu‐PBMC	CD34+ human PBMCs (T, B, NK, monocytes and dendritic cells)		Intravenous/intraperitoneal/intrasplenic	Human immune system dominated by mature activated T lymphocytes	1. Fast reconstruction period: 1–2 weeks 2. Short‐term research, especially requiring mature T cells	1. Immune system dominated by T cells but few other types of cells 2. Rapid response and proliferation of human T cells in mice, easily leading to strong GVHD
Hu‐BLT	Human fetal and liver and fetal liver/bone marrow–derived CD34+ HSCs (same donor)		1. Irradiation 2. Human fetal liver and thymus under renal capsule 3. HSC intravenous	Lymphatic T, B, NK, and dendritic cells, myeloid monocytes and macrophages, human HLA‐restricted lymphocyte immune system	1. Natural differentiated lymphocytes; GVHD probability lower than Hu‐PBMC probability 2. Functional expression of T‐cell masses in human thymus microenvironment 3. Effective adaptive immune response 4. Low GVHD incidence	1. Complicated and difficult modeling operation 2. Aborted fetal tissue with unavailable source and huge social and ethical disputes 3. GVHD probability higher than Hu‐HSC probability, mainly late‐stage chronic GVHD

Abbreviations: GVHD, graft‐versus‐host disease; HIS, humanized immune system; HLA, human leukocyte antigen; HSC, hematopoietic stem cell; Hu‐BLT, humanized‐bone marrow, liver, thymus; NK, natural killer cell; Hu‐HSC, humanized‐hematopoietic stem cells; PBMC, peripheral blood mononuclear cell.

FIGURE 1.

Summary of HIS (humanized immune system) model applications in CAR‐T. Created with BioRender.com.

2. HU‐HSC MODEL

Hu‐HSC models are transplanted with HSCs from mobilized peripheral blood, human umbilical cord blood, bone marrow, or fetal liver using their self‐renewal and differentiation ability to generate multilineage human immune cells, including lymphoid T, B, NK cells, and myeloid cells supported by some individuals (e.g., BRGSF‐HIS mice ¹⁰ ). The successful modeling of humanized immune mice relies on the favorable transfer of HSCs, which can achieve generalizations of human hematopoietic complexity by the selection of suitable strains and maximum control of human immune rejection, producing human‐like immune responses. Whether directly using natural genetically deficient mice or mice treated with gene editing, the resulting mutations or genetic defects can reduce their rejection of transplanted human HSCs. ¹¹ Hu‐HSC humanized mice are especially suitable for long‐term research in the fields of tumor immunity, ¹² infectious diseases (e.g., HIV/AIDS ¹³ ), reprogramming immunotherapy, ¹⁴ and so on. It is noteworthy that more accurate simulation of the internal immune environment of the human body has an excellent and outstanding clinical value for timely discovery of some specific treatments against human diseases, providing more convenience for the continuous development and innovation of treatment methods. Besides, the estimation of immune phenotype can provide reliable monitoring information for subsequent lymphocyte analysis in vivo, such as the anti‐inflammatory mechanism.

Nevertheless, this model also has certain disadvantages. HSCs take a long time to reconstruct the immune system (2–4 months ¹² , ¹³ , ¹⁵ ). The HIS constructed by this model lacks the human thymus necessary for the development and maturation of human T cells; therefore, the function of T lymphocytes is relatively weak. Data show that if new‐born immunodeficient mice are used as HSC receptors, T cells are limited to express mouse major histocompatibility complex ¹⁶ (MHC) because of their maturation in mouse thymus, greatly resulting in a decrease in graft‐versus‐host disease (GVHD) and obvious limitations in human disease simulation, distribution, and quantity.

Hu‐HSC models have a wide application in exploring diverse treatments for CAR‐T. Based on the good clinical efficacy of CAR‐T in nonsolid tumors such as leukemia, in a treatment strategy against acute myeloid leukemia (AML), the comprehensive use of the Hu‐HSC model and the patient‐derived tumor xenotransplantation (PDX) model can simulate the clinical trial processes to explore the efficacy of CAR‐T targeting leukemic CD117⁺ cells. ¹⁷ The Hu‐HSC model was used to confirm the targeted elimination of CD117 CAR‐T cells in healthy individuals, which laid a pre‐dose basis for confirming the efficacy of CAR‐T in eliminating CD117^‐positive AML cells in subsequent tumor xenotransplantation models. When assessing CAR‐T cells for Epstein–Barr virus (EBV) infection and lymphoproliferative diseases，the Hu‐HSC model was used to determine the significant anti‐EBV effect of 7a1‐gp350 CAR‐T cells in prevention and treatment. The results of the bio‐layer interferometry analysis significantly revealed that inflammation in the spleen and plasma effectively reduced after the prophylactic administration of gp350CAR‐T cells, and its use in long‐term treatment could also inhibit EBV diffusion. ¹⁸ Meanwhile, in a breakthrough combination therapy research for pancreatic adenocarcinoma, ¹⁹ the use of the Hu‐HSC model helped overcome the difficulty of limited replication of oncolytic adenovirus‐infected rodents in vivo. The tumor microenvironment of the host in the Hu‐HSC model was changed by the stimulation of CAdTrio (an oncolytic adeno‐immunotherapy); therefore, HER2‐CART showed an unprecedented lasting antitumor effect, which could not be observed in the immunodeficient tumor model. This confirmed the clinical research value of CAdTrio and HER2‐CART combined immunotherapy for pancreatic tumors. With the continuous promotion of CAR‐T application to clinical practice, research on the recurrence and adverse reactions of many patients after treatment has also been carried out. In a research project specifically focused on HIS, a strategy to construct personalized Hu‐HSC mice that reflected the patient‐specific immunity and cancer pathological characteristics was pointed out. ²⁰

In addition to the aforementioned function like testing the efficacy of new CAR‐T cells, the Hu‐HSC disease model can be used to characterize the generation and function of cytokine lineage, memory CAR‐T cells, and regulatory T cells after CAR‐T treatment and to further explore the relevant mechanism of recurrence and complications after CAR‐T treatment.

3. HU‐PBMC MODEL

Hu‐PBMC models transplanted human PBMCs, including T, B, NK, monocytes, and dendritic cells, into adult immunodeficient mice to rapidly construct a human immune system dominated by mature and activated human T‐lymphocyte cells that produce strong effects through contact with mouse xenoantigen. The significant advantage of this model is its quick reconstruction, with a period as short as 1–2 weeks. ²¹ , ²² This procedure is also relatively simple and easy to operate. Therefore, for some studies in which human trials cannot be conducted due to ethical factors, the PBMCs of patients can be quickly and synchronously used to model, simulating the state of the immune system at different periods of clinical treatment. ²³ The model is suitable for tests that require short‐term studies of mature T cells, such as evaluating tumor immunity, GVHD, ²⁴ blood diseases, and viral infectious diseases ²⁵ , ²⁶ (e.g., HIV/AIDS).

The advantages of the Hu‐PBMC model have been fully applied in the research relevant to CAR‐T cell therapy. For its rapid modeling, the feasibility of clinical research with CD1α × CD3ε dual CAR‐T cells can be confirmed using the model, constructed by PBMCs of T‐cell acute lymphoblastic leukemia patients and synchronized with the characteristics of disease progression. ²⁷ The Hu‐PBMC model can be used to study the anti‐HIV activity of multi‐specific dual CAR‐T cells. ²⁸ One week after infection, the HIV infection rate could be monitored by the quantitation of luciferase activity in mouse spleen. The inhibition of HIV infection by multi‐specific CAR‐T cells was significantly stronger than that by single‐specific cells, which reasonably demonstrated the therapeutic advantages and application prospects of bimolecular CARs and multi‐specific modified T cells.

The model also has significant defects relative to Hu‐HSC mice. The constructed immune system considers T cells as the dominant cells, and the number of other types of cells is low, which may be due to the lack of relevant human cell regulatory factors. ²⁹ Human cytokines can be expressed in mice by repeated administration, introduction of viral vectors or plasmids encoding cytokines, and transgenic methods. ³⁰ , ³¹ Human T cells respond rapidly in mice to achieve rapid proliferation while creating a strong antihost response to mouse heterologous cells called GVHD. When GVHD occurs (~4–5 weeks after reconstitution ²¹ ), test mice represent evident signs of weight loss, arch back, diarrhea, hair removal, and peeling. ³² Therefore, the observational window period of Hu‐PBMC should be specially considered based on different strains in the experimental design to avoid GVHD fatality in the process of data acquisition and to prevent fatality from covering up due to disease symptoms and signs. Regarding CD19 CAR‐T cells that have been used in clinical practice to play a good role in leukemia treatment, combining the advantages of multiple models, Hu‐HSC, Hu‐PBMC, and PDX models, can provide efficient data support for CD8‐LV ³³ and CD4‐LV ³⁴ targeted delivery vectors, the key technologies for first selective generation of CD19 CAR‐T in vivo, breaking the previous restrictions on the production cycle and quantity of CAR‐T generated by activating and transmitting isolated cells before.

4. HU‐BLT MODEL

The lymphocyte immune system of the Hu‐BLT model restricted by human leukocyte antigen (HLA) is reconstructed by preimplantation of the human fetal liver and thymus and subsequent injection of fetal liver or bone marrow–derived CD34⁺ HSCs cells from the same donor. This is currently the most functional and perfect model for the reconstruction of the human immune system. Lymphocytes are primarily produced by natural differentiation, with a lower probability of producing GVHD than the Hu‐PBMC model. First, compared to the Hu‐HSC model, constructing the human thymic microenvironment composed of fetal thymus transplantation compensated for the limited expression of the T‐cell population. T cells mature within the human thymic epithelium of the same origin, and the maintained long‐term development in the thymus together with systemic development of T cells is conducive to the formation of an efficacious adaptive immune response. ³⁵ This model is well adapted for studies on adaptive immune responses such as HIV infection, ³⁶ inflammation in visceral organs, ³⁷ and delayed type hypersensitivity. ³⁸ , ³⁹

Because of the better HIS simulation effect of the Hu‐BLT model overcoming the disadvantages of the first two models, its application in CAR‐T relevant research has also increased in recent years. The ability to produce mature T cells differentiated in human thymic epithelium made it possible for T cells derived from the Hu‐BLT model to be directly used to construct CAR‐T in vitro. In a research on the development of 4‐1BB and CD28 dual CAR‐T cells that played an anti‐HIV role, the CAR‐T generated by the Hu‐BLT‐derived and human‐derived T cells showed the equivalent of in vitro expansion dynamics, CAR surface expression level, and cytokine expression level. ⁴⁰ No significant difference in the CAR‐T functional characteristics between the two CAR‐Ts indicated that Hu‐BLT has a more significant clinical correlation when simulating autologous CAR‐T treatment and is suitable for the repeated development of high‐performance 4‐1BB and CD28 dual CAR‐T cells. Additionally, due to the advantages of the model, multilineage human immune cell functional lymphohematopoietic system and low GVHD incidence, T cells derived from the Hu‐BLT model can be used to construct CD19 CAR‐T acting on the Hu‐HSC B‐cell acute lymphoblastic leukemia (B‐ALL) disease model. ⁴¹ The results show that the therapeutic effect of CD19 CAR‐T in HIS mice similar to that in clinical leukemia patients further proves Hu‐BLT has more significant clinical relevance when simulating autogenous CAR‐T therapy and is suitable for evaluating the efficacy of and exploring the brand‐new CAR‐T therapy. Simultaneously, due to the negative correlation between normal B cells and anti‐CD19 CAR‐T cells, the level of human B cells in mice can also serve as a potential indicator of CAR‐T cell.

Although the Hu‐BLT model has the best simulation effect of the human immune system compared with the first two models, it still has great limitations and promotion space in practical construction and application. The surgical operation in the BLT modeling process is complex and difficult. Human tissue needs to be obtained from naturally aborted fetuses, which is socially and ethically controversial. ⁴² These two points make it difficult for such models to achieve batch modeling, limiting its application in the development of immune drugs. Relating to the substantial technology of in vivo CAR‐T generation proven to be effective as described previously, ³³ , ³⁴ it is considered that if the data results obtained through Hu‐BLT model research and confirmation will be more conducive to extrapolation and application to clinical patients, but more tremendous construction cost and difficulty in obtaining human materials will be faced.

In addition, the probability of GVHD in BLT mice is higher than that in HSC mice, mainly exhibiting late‐chronic GVHD, ⁴³ and the incidence of GVHD is different in BLT models based on different types of immunodeficient mice. Of these, NSG (NOD/SCID γc^−/−) mice have a high incidence of GVHD, whereas C57BL/6 Rag^−/−γ c^−/−CD47^−/− triple‐knockout (TKO) BLT mice and C57BL/6 Rag2^−/−γc^−/− double‐knockout BLT mice do not develop clinical symptoms of GVHD, and TKO‐BLT mice can also achieve good reconstruction effect of human cells similar to NSG‐BLT mice. ⁴⁴ For meeting the basic requirements, we must select immunodeficient mice and construction strategies with lower incidence of GVHD. In particular, mice with gene editing, transgene, and knockout at special sites meeting different experimental requirements produce the best effect. To avoid xenogeneic immune response and preserve the thymic microenvironment required for long‐term reconstruction and development of lymphocytes, we emphasize the selection of CD34⁺ HSCs homologous to fetal thymus and liver tissues; however, it has been reported that the human immune system can be constructed more easily with a large number of allogeneic adult HSC CD34⁺ cells with cryopreserved fetal thymus. ⁴⁵ The different treatments of human HSC cells and fetal thymus in two experiments show that the time to reach the same level of human reconstruction is reduced by improving adult HSC CD34⁺ cell number and using frozen depletion of fetal thymus primary T cells.

5. IMPROVEMENT

5.1. Reduce heterogeneity

Innovative research regarding the existing advantages and disadvantages of the HIS model is still ongoing on humanized mice. Regarding the differences between the reconstructed immune environment and the human immune environment, data confirm that humanized mice still differ from the real human body in the microenvironment of the drug effect, and heterogeneity is an important reason why the therapeutic effect of some theoretical active molecules cannot be truly confirmed by mouse models. One literature study suggested that using a TKO‐BLT‐humanized mouse model revealed an abnormal interaction between mouse cells and IFNα14 that can activate NK cells in HIV, making it a limited antiviral activity different from that in the human environment. ⁴⁶ Heterogeneity is the fundamental cause of immune rejection such as GVHD and host‐versus‐graft reaction. Exploring how to effectively reduce heterogeneity is significant for improving the life span of HIS models and expanding their scope of application.

Enhancing the inclusiveness and tolerance of mouse residual phagocytes for heterologous human hematopoietic cells using transgene is a viable renewal direction. Gene modifications in human SIRPα of mouse signaling regulatory proteins have been reported to produce relatively desirable results, ⁴⁷ increasing implantation and maintenance of heterologous human hematopoietic cells, which serve as a manifestation of rejection resistance. For widely distributed and difficult to control cytokines, such a way of human gene knock‐in to replace the corresponding mouse genes can eliminate mouse cytokines fundamentally at the gene level, improving the stability and long‐term validity of human‐related cytokine expression, ⁴⁸ even if this method is more time consuming and costly, with respect to direct injection of recombinant protein together with fluid DNA injection.

Given that heterologous cytokines have more unrestricted and widely diverse therapeutic potential, more consideration should be given to improve the stability of heterologous cell construction, apart from limiting the same source, such as the fetal thymus frozen‐elimination method for heterogeneous HSC modeling in the previous BLT model. ⁴⁵ Gene targeting design involving multiple human cell regulatory factors ⁴⁹ and MHC modification ⁵⁰ , ⁵¹ , ⁵² based on NSG mice can produce mice with a higher immune deficiency level so as to relatively lessen GVHD occurrence of various kinds of HIS mice.

Further exploring the subtle differences in the overall microenvironment after immune cell maturation and production of various cytokines will also be a favorable method to reduce the abnormal interaction between humanized and test factors. Such heterogeneous interactions require the development of a longer‐term trial procedure, confirmed using more combined drug tests.

5.2. The application of HIS mice in the study of adverse toxicity after CAR‐T treatment

With the development of CAR‐T therapy, the toxic complications caused by CAR‐T have gradually been widely valued and explored, including cytokine release syndrome (CRS) and neurotoxicity. The clinical application of CAR‐T can contribute to relapse and posttreatment toxic attack due to many possible factors, whose mechanism is relatively complex. Tian's group pointed out that the lack of cross‐communication between the mouse microenvironment and murine cytokines and human cells was the principal element limiting the myeloid, NK differentiation level, and related research mediated by them. ⁵³ Using immunodeficient transgenic mice moderately expressing human cytokines was an excellent strategy to improve such models, and persistent overexpression led to abnormal immune function, but suggested that it can improve the similarity of disease occurrence related to human abnormal immune regulation to a certain extent, which was conducive to the study of complications after CAR‐T treatment.

Concerning an abnormal increase in cytokine levels, although some cytokines with cross‐regulatory mechanisms have not been studied and defined in HIS mice, HIS models can still be used to preliminarily study the concurrent mechanisms of two toxic reactions by combining relevant clinical cases and the basic theory of information regulation. Regarding the fact that the interleukin‐6 (IL‐6) receptor antagonist used clinically for a long period can slow down CRS but cannot effectually inhibit neurotoxicity, a Hu‐HSC mice‐assisting study has confirmed that interleukin‐1 released by circulating mononuclear cells was the upstream factor inducing IL‐6 secretion, and its antagonism can simultaneously produce a synergistic elimination effect against CRS and neurotoxicity. ⁵⁴ Simultaneously, in connection with the experimental results of in vitro experiments and HIS model, granulocyte‐macrophage colony‐stimulating factor as a new successful antagonistic target ⁵⁵ proved not only to achieve a complete and potent elimination of CAR‐T complications but also to enhance and maintain the efficacy of CAR‐T against leukemia to a certain extent.

The cause of CAR‐T toxicity in vivo can be understood by focusing on the fundamental principles of CAR‐T maintaining activity in vivo. A research application involving the Hu‐HSC model mentioned in a conference abstract demonstrated that long‐term treatment of CAR‐T could lead to the injury of myeloid cells in vivo. ⁵⁶ Because normal myeloid cells are indispensable for CAR‐T to maintain its antimyeloid malignant activity, damage to myeloid cells can in turn mediate and affect the loss of the corresponding CAR‐T activity, indicating the challenges when extending CAR‐T cell therapy to myeloid malignancies. Similarly, the application of the Hu‐PBMC model also confirmed that human CD19 CAR‐T cells can lead to depletion of B cells in vivo and posttransplantation phenomena as CRS. ⁵⁷ Due to the limitations of Hu‐PBMC (Table 1), the prediction of toxicity after CAR‐T treatment is limited by xenotransplantation reaction and GVHD symptom concealment. The aforementioned characteristics of the Hu‐HSC model can be studied to successfully represent cytokine lineage after CAR‐T treatment, ²⁰ combine the Hu‐PBMC model with the Hu‐HSC model, and remedy its defects to explore the potential mechanism of CRS and neurotoxicity in patients after CAR‐T cell therapy.

5.3. Remedying the disadvantages of the HIS model in CAR‐T application

Considering the advantages of the aforementioned three humanized mouse models in the main research fields of CAR‐T cell therapy, the following methods can be considered in CAR‐T‐related efficacy assessment to make up for model defects.

1. Two or even multiple complementary models were combined for parallel control tests to contrast the results. This approach has a precedent in model applications. ⁵⁸ A good preclinical model beneficial to treatment research on monoclonal antibodies against CD28 was identified by a strong contrast between the two models, Hu‐BLT and Hu‐HSC. This combined scheme can translate the defects of the model itself into the value of reinforcing the contrast effects. screening of CAR‐T treatments can be accomplished more efficiently using the combination of two complementary models.

2. Regarding the fetal tissue limitation required by the Hu‐BLT model, the liver and thymus organoids formed by the differentiation of embryonic stem cells or induced pluripotent stem cells stimulated by various kinds of cytokines can help replace fetal tissue; thus, the difficulty and ethical restriction in model construction can be relatively reduced. In a perspective article, it planned to use 3D organoids, Hu‐HSC model, and reduce HIS model heterologous antagonism in numerous aspects through human MHC introduction and mouse MHC inhibition to construct HIS mice with prostate organoids and HSCs of the same origin implanted in patients. This assumption combines the merits and demerits of different models, enhancing synchronization, authenticity, stability, and comprehensiveness of the preclinical efficacy evaluation of CAR‐T antitumor efficacy, and contributes to the development related to individualized, specific, and precise medical treatment in the field of cancer and even other fields. ⁵⁹

3. A variety of related cells and factors not only for CAR‐T design but also for HIS model construction should be used from the same source for direct heterologous graft or graft after culture expression. At present, relevant active molecules are derived from the human body, so specific immune gene knockout mice are still mainly employed to reduce xenogeneic immune rejection. If the relevant active molecules have to be used from allograft sources due to objective restrictions like the source individual's own defects, the main approach to reduce allogeneic rejection is combining antigen–antibody binding blockers ⁶⁰ and antibodies against multiple homologous lymphocytes. ⁶¹ Therefore, we put forward an idea: under the premise that techniques of interspecific nuclear transplantation have been preliminarily certified across a variety of mammals, ¹⁴ , ⁶² , ⁶³ especially the demonstration of the developmental potential of embryonic stem cells produced by somatic heterogeneous transplantation in humans, ⁶⁴ we consider whether it is possible to reduce heterologous rejection and ensure the normal expression of human immune factors that construct the immune microenvironment by nuclear transfer that uses mature human HSCs as donors and mouse oocytes as receptors, hematopoietic differentiating in vitro. There is no relevant evidence to confirm this hypothesis, and its feasibility requires to be confirmed by more trials in the future. Based on cellular reprogramming induced by cell fusion with stem cells, the fusion of mouse and human HSCs or human HSCs and mouse embryonic stem cells ⁶⁵ , ⁶⁶ may produce the desired results.

4. Compared to Hu‐PBMCs directly injected into mature T cells and Hu‐BLT providing a developmental environment via fetal thymus, weak T‐cell function is a distinct disadvantage of Hu‐HSC. CAR‐T applications mentioned earlier show that limited research on the long‐term side effects of CAR‐T related to nonhematopoietic tissue infiltration resulted from short‐term maintenance of T‐cell liveness. ¹⁷ Insufficient proliferation leading to sustained stimulating failure ¹⁸ and impossible exploration of CAdTrio inducing antigen diffusion and then producing antigen‐specific cytotoxic T lymphocytes due to lack of a thymic development environment ¹⁹ also confirmed its limitations for further development. What we desired can be achieved by combining Hu‐PBMC, synchronously injecting human PBMC with CD34⁺, or simulating the thymic microenvironment with numerous cytokines required for T cell development. T cells redirected for universal cytokine‐mediated killing (TRUCK), the technology of fourth‐generation CAR‐T, ⁶⁷ provides an effective and feasible strategy. By targeting the design of the CAR domain with transgenic protein information sets, the regulatory factors related to T cell proliferation and functional differentiation can be temporally and spatially enriched, thus compensating for the disadvantages of Hu‐HSC and expanding its further exploration of the value and hidden problems in CAR‐T treatment. Given this, Hu‐HSC is expected to provide more favorable guarantees and value supports for the safety and effectiveness of CAR‐T clinical treatment in the future.

6. CONCLUSIONS

This review summarized the basic characteristics of three main models and enumerated the practical cases of their application in CAR‐T therapy development and complication mechanism research, attaching great importance to the vital impact on HIS models in improving the clinical relevance and authenticity of CAR‐T research in antitumor and antivirus aspects. In brief, with the continuous improvement and optimization of the HIS mouse model, its application will further promote CAR‐T to explore wider therapeutic fields in the future, improving the safety and effectiveness of clinical patient application via studying the negative effects of its application.

The development of animal models that more accurately mimics the natural immune response in the human body will further increase research in biological medicine in the future and promote the development of multiple effectual vaccines and therapies. Beginning with modeling the human immune response to common pathogens, we can study the similarities and differences between the HIS mouse models and the human immune system with respect to immune mechanisms. ⁶⁵ , ⁶⁸ Exploring the immune advantage model requires tractable data such as human pathological metabolism data, humanized mouse kinetics, degree of specificity, and characterization of viral infection. Using these more basic simulation data, the HIS model can be developed in more feasible directions, while keeping the potential of stable and reliable preclinical assessment models, which is worth promoting to CAR‐T and more therapeutic fields. Fundamental medical research findings translated more successfully into clinical disease research may greatly reduce the risk of clinical human experiments.

AUTHOR CONTRIBUTIONS

Hanwei Yue summarized and analyzed the main information on model construction and was a major contributor to write this manuscript. Lin Bai guided the project, reviewed the manuscript, critically revised the key intellectual content, and was ultimately responsible for the decision to submit for publication. Both authors read and approved the final manuscript.

FUNDING INFORMATION

CAMS Innovation Fund for Medical Sciences (No. 2021‐I2M‐1‐035) and National Key Research and Development Project (No. 2022YFA1103803).

CONFLICT OF INTEREST STATEMENT

Lin Bai is an editorial board member of AMEM and coauthor of this article. To minimize bias, she was excluded from all editorial decision making related to the acceptance of this article for publication.

ETHICAL APPROVAL AND CONSENT TO PARTICIPATE

Not applicable.

Yue H, Bai L. Progress, implications, and challenges in using humanized immune system mice in CAR‐T therapy—Application evaluation and improvement. Anim Models Exp Med. 2024;7:3‐11. doi: 10.1002/ame2.12353

DATA AVAILABILITY STATEMENT

Not applicable.