Among those, spontaneous and genetically engineered immunodeficient strains can be administered in both primary human tumour tissue and cancer cell lines such that they grow to become one of the most widely used in vivo tools for target validation and efficacy.

Yet the immunodeficient nature of these host mouse models is also a hindrance to researching monoclonal antibody immuno-therapies that depend upon an intact immune system to combat tumour growth. This has driven the need for even more advanced models that combine a human immune system as well as human tumours in a living system. With the immense potential of immunotherapies, there is an ever greater dependence on the translational relevance of in vivo studies. While somewhat counterintuitive, the development of an even more severely immune-deficient rodent host has enabled human hematopoietic stem cell (HSC) engraftment to grow and repopulate the ‘empty’ host with human immune cells. When combined with human tumour tissue or cell line engraftment, a model for human immune cell function in the context of tumour growth is established.

A review of the current landscape of immunodeficient models, methods to overcome cross-species barriers and the value of combining human immune system mice with PDX (patient-derived xenograft) models illustrates how these models can greatly further immuno-oncology research.

Current state of immunodeficient models
Over the past 50 years, numerous immunodeficient mouse strains have been generated to model human immunity and disease. As a result of both increased understanding of basic immunological processes and advances in gene targeting technology, each successive generation of animals has incorporated additional defects that further disable the host immune function. Modern immunodeficient strains of mice have enabled greater success in human cell and tissue engraftment. Their commercial availability has also given the biopharmaceutical industry convenient access to the latest in vivo oncology models for pre-clinical studies.

The immune system is a highly complex network of cells and tissues that is tasked with the detection and elimination of foreign pathogens, toxins and malignant cells1. Early generations of immunodeficient mice, such as the athymic nude2 and scid strains3, carried single-gene mutations that disrupted components of adaptive immunity. The adaptive arm of the immune system mediates a highly-specific response through antibody-releasing B-cell (humoral) and T-cell effector mechanisms. While the Foxn1 mutation carried by the nude mouse leads to impaired T-cell development, the scid mutation promotes a more severe immunodeficiency, which is characterised by loss of both T- and Bcell function.

Despite the wide use of these strains in immunological and cancer research, the presence of innate immunity in these models prohibits major successes in human cell engraftment4. In contrast to the adaptive immune system, the innate arm serves as a non-specific defence mechanism that is rapidly mobilised at the site of infection or inflammatory damage2. Physical barriers (eg skin), cellular components (eg phagocytic cells, natural killer (NK) cells and granulocytes) and circulating proteins (eg hemolytic complement proteins) co-ordinate the innate immune response. A key development in the generation of immunodeficient mice has been the advent of the NOD-SCID strain5. In addition to impairing B- and T-cell function, introduction of the non-obese diabetic (NOD) background renders these animals highly immunodeficient by disrupting the innate immune system. NOD-SCID mice exhibit reduced function of NK cells, macrophages and antigen-preventing cells (APCs) and a lack of hemolytic complement. Positional genetics has also revealed that the NOD strain carries a variant allele of the SIRP-􀀀 gene6. SIRP-􀀀 (signal-regulatory protein alpha) is an inhibitory protein that interacts with CD47 expressed on the surface of macrophages to prevent phagocytosis. Consequently, the SIRP-􀀀 polymorphism leads to enhanced binding to human CD47 and reduced phagocytosis of engrafted cells.

Although a clear improvement over previous generations of immunodeficient mice, the NODSCID mouse has some disadvantages. NOD-SCID mice are difficult to engraft with human CD34+ HSCs, and the development of spontaneous murine lymphomas results in a reduced lifespan of the host mouse. While NK cell function is impaired, a low but detectable level of residual activity remains5. These drawbacks led to the generation of advanced NOD-SCID animals that carry mutations in the interleukin-2 receptor common gamma chain (Il2rg) gene (ie the CIEA NOG® mouse and the NSG™). Loss of the common gamma chain impairs signalling from at least five different interleukins (IL-2, IL-4, IL-7, IL-15 and IL-21). The loss of IL15 signalling is particularly important because it completely abolishes NK cell development4. As a result, these super-immunodeficient mice are far more suitable hosts for human HSC engraftment. They also lack spontaneous murine lymphomas due to the defect in IL-2 signalling, which extends their lifespan well beyond the original NOD-SCID strain and allows for longterm studies.

Developed at the Central Institute for Experimental Animals (CIEA) in Japan and first published in 2002, the NOD-SCID-based strain carrying an Il2rg mutation was called NOD-Shiscid- Il2rgamma(null) or NOG7. The CIEA NOG® mouse, now distributed exclusively for CIEA by Taconic Biosciences, carries a truncating mutation in the cytoplasmic tail of Il2rg. This genetic modification disrupts key intracellular signalling pathways required for the development and function of multiple immune cell populations. A similar strain, the NSG™, was developed independently by the Jackson Laboratory and was published three years later, in 20058. The NSG™ mouse carries a complete null mutation of the Il2rg gene, which prevents both cytokine ligand binding and signal transduction. Given the superior level of human cell and tissue engraftment achieved through the Il2rg mutation, the NOG and NSG strains are now considered as the optimal hosts for the generation of human immune system engrafted mouse models4,9.

Many methodologies have been established that facilitate the generation of human immune system engrafted mice4,10. Each methodology has advantages, disadvantages and intended uses depending on the area of research. Although initially developed for simpler immunodeficient models, the protocols for human immune system engraftment are now commonly applied to super-immunodeficient strains, including NOG and NSG. For studies of mature immune cell populations, peripheral blood mononuclear cells (PBMCs) can be transplanted into adult mice through intravenous, intraperitoneal, or intrasplenic routes and used to evaluate function of memory and effector T-cells, APCs and NK cells11. The resulting model is a short-term one best suited for studies related to vaccine and T-cell immune responses and graft versus host disease (GvHD).

Methods have also been established to enable the development of a multi-lineage immune system in mice. A technically challenging but powerful model is the bone marrow, liver, thymus (BLT) mouse12. Generation of this model involves engraftment of human liver and thymus tissue under the kidney capsule followed by transplantation of autologous liver-derived CD34+ HSCs. The resultant hematopoietic-lymphoid system permits development of T- and B-cells, monocytes, macrophages and dendritic cells and has been extensively used to study microbial pathogenesis, vaccine response and CD4 T-helper-cell biology.

Probably the most widely-used method for humanising models is through tail-vein injection of CD34+ HSCs into pre-conditioned (irradiated) juvenile mice13. Because of the multi-lineage capacity of HSCs, this system has been extensively used to investigate normal hematopoiesis. Multiple cell lineages are detected 12-16 weeks post-engraftment and remain stable in the blood, bone marrow, spleen and thymus. CD4+ and CD8+ T-cells successfully differentiate on the NOG background4, but human myeloid lineages, differentiating from the common myeloid progenitor, are poorly represented, and the frequency and function of the lymphoid- derived NK cell lineage are also sub-optimal.

The aforementioned methods of obtaining human immune system engrafted models have been applied to the CIEA/Taconic NOG platform, including the huPBMC-NOG mouse, the BLTNOG mouse, and the CD34+ engrafted huNOG mouse. Advanced humanised mice that support the development of additional cell lineages, such as the recently-released huNOG-EXL model, have also been generated. The Jackson Laboratory, for instance, offers humanised models similar to those offered by CIEA/Taconic, including NSG-background humanised mice that have been engrafted with human PBMCs (huPBMC) or CD34+ HSCs (huCD34).

Advances in human immune system models
An emerging arena for the application of humanised immune system mice is the field of immunooncology. The interplay of the immune system with tumour cells is now widely recognised as an important feature of cancer that can be exploited therapeutically. However, these interactions are complex and involve multiple immune cell types, some of which are poorly developed in current human immune system engrafted mice. In order to realise the full potential of these models, a system is required that permits more comprehensive downstream lineage commitment following HSC engraftment. The restricted development of certain cell lineages, especially those of the myeloid family, is due in part to cross-species differences that lead to the poor binding of host-mouse cytokines by human progenitor cell cytokine receptors4,11. To overcome this limitation, CIEA has recently modified and improved upon the core NOG model by introducing human transgenes that encode multiple cytokines critical to myeloid differentiation (ie GM-CSF and IL-3)14. This second-generation NOG strain, called NOG-EXL (extended lineage), exhibits an increase in total CD45+ leukocyte number, a number that can also be detected earlier than in non-transgenic NOG mice (as early as six weeks).

The presence of human cytokines in the huNOG-EXL mouse supports increased differentiation of the human myeloid lineage. Moreover, further analysis of the granulocyte sub-lineage of huNOG-EXL mice revealed efficient development of basophils, eosinophils, immature neutrophils and mast cells. Differentiation of the monocyte sub-lineage is also improved in the huNOG-EXL model, as a greater frequency of macrophages are evident in non-lymphoid tissues including the lungs, liver and the spleen compared to core NOG mice engrafted in the same way. The functional capacity of these expanded lineages was demonstrated by challenging huNOG-EXL mice with several allergens, a process that resulted in a strong mast-cell-mediated response14.

The Jackson Laboratory also offers NSG mouse strains genetically modified to express various human cytokines. These include mice that express human stem cell factor (hSCF)15,16 and a multiallelic strain that combines hSCF expression with IL-3 and GM-CSF human growth factors (NSG™- 3GS, as well as NSG™-SGM3)17. As with the NOG-EXL model, both transgenic strains show improved myeloid lineage commitment, however the circulating concentration of cytokines in the blood is very different and can be potentially confounding for various long-term applications. The huNOG-EXL host mouse expresses between 0.03 and 0.08ng/ml of human GM-CSF/ and IL3 cytokines compared to 2-4ng/ml in the NSG™- SGM3. Although a direct comparison of these transgenic NOG and NSG™ models has not been carried out, the huNOG-EXL’s lack of hSCF, which is an established mediator of mast cell differentiation, does not impair development of this sub-lineage. This is likely due to sufficient cross-reactivity of host-mouse SCF with human Kit receptors expressed on HSCs in huNOG-EXL mice14.

Human immune system models for immuno-oncology applications
The ability of tumour cells to evade immune detection is a hallmark of cancer and plays an important role in disease progression and treatment resistance18. Stimulating anti-tumour immunity as a therapeutic strategy is therefore being actively pursued throughout the biopharmaceutical industry. To maximise the potential of translational success in new immunotherapies, a robust pre-clinical in vivo model is needed. A currently popular system involves transplantation of murine tumour cell lines into immunocompetent hosts. These syngeneic models are technically simple, have low associated costs and can be used to examine tumours with low and high immunogenicity. However, pre-clinical efficacy testing is challenging because mouse cell molecular targets are not always recognised by humanised monoclonal antibodies. Moreover, rapid tumour growth observed in these models can lead to an exhausted T-cell phenotype resulting in an artificial reflection of human immunity19. As a result, the generation of novel mouse models that harbour tumours and immune cell populations of human origin would greatly benefit the field of immuno-oncology.

Over the past several years, PDX models have been extensively used in cancer research20. Unlike the traditional xenograft model in which conventional cancer cell lines are transplanted into nude mice, the PDX platform utilises fresh patientderived materials that are passaged in vivo into highly immunodeficient animals. While the NODSCID strain is suitable for some tumour types, higher engraftment rates are commonly observed using more advanced NOG and NSG superimmunodeficient models21. When used at low passage (<6-10), PDX models have been shown to be more effective in representing parental tumours at the genetic and histological level20. Importantly, these models also maintain tumour heterogeneity and exhibit similar response rates to standard treatment regimens found in the clinic21. However, the application of PDX models to the rapidly growing field of immuno-oncology poses a significant challenge because the mechanism of action of immunotherapies requires a functional human immune response. Finding strategies that meet this challenge is an active area of investigation.

Although the idea of harnessing the power of the immune system to fight cancer is not new22, clinical success has driven increasing momentum. A major focus has been the development of immune checkpoint inhibitors. Immune checkpoints are critical signalling pathways that mediate self-tolerance and regulate the duration and intensity of immune response23. Two major checkpoint pathways that have been exploited therapeutically are Cytotoxic T-cell lymphocyte antigen 4 (CTLA4) and programmed cell death protein 1 (PD1). CTLA4 is a cell-surface receptor that is expressed exclusively on T-cells and that serves to dampen Tcell activation through inhibition of the CD28 costimulatory molecule24. Although the exact mechanism behind the expression remains unclear, CTLA4 and CD28 share the CD80 and CD86 ligands expressed on antigen-presenting cells (APCs). Because CTLA4 has a higher ligand binding affinity, it may out-compete CD28 and attenuate T-cell activation23. For therapeutic purposes, a number of monoclonal antibodies have been produced that target CTLA4 and PD1 to enhance anti-tumour immunity. The most well-known CTLA4 checkpoint inhibitor is Ipilimumab (Yervoy), the first drug of its class to gain FDA approval25.

The second major immune checkpoint, PD1, serves both to regulate the inflammatory immune response in peripheral tissues during infection and to limit autoimmunity23. PD1 is a cell-surface receptor expressed on T-cells but also found in a broader range of immune cell populations, including B-cells and NK cells. Binding of PD1 to its ligands (PDL1 or PDL2) inhibits intracellular kinase signalling leading to reduced T-cell activity26. Deregulation of the PDL1 or PDL2 expression pathway (eg, through up-regulation of PD1 ligand expression on tumour cells) can lead to escape from recognition in the tumour microenvironment and thus represents a key immune resistance mechanism27. The anti-PD1 monoclonal antibodies Pembrolizumab (Keytruda) and Nivolumumab (Opdivo) have recently obtained US FDA (Food & Drug Administration) approval for use in treating non-small cell lung cancer (NSCLC) and renal cancer28.

Despite these advances, questions remain about the potential effectiveness of combining human immune system engrafted mice with PDX models in immuno-oncology research. While PDX tumour growth in immunodeficient animals has been well characterised, it was, until recently, unknown whether mice repopulated with functional human immune cells can tolerate the tumour grafts. To address this question, Champions Oncology implanted tumour fragments derived from NSCLC patients into huNOG (CD34+ humanised) and NOG (immunodeficient) mice. The results revealed similar tumour growth kinetics in both models. In a second experiment, the company investigated the utility of the huNOG model in immunotherapies. Tumour-bearing huNOG mice were treated with Ipilimumab, the approved CTLA4 checkpoint inhibitor, and tumour size was subsequently monitored. Remarkably, more extensive tumour regression was observed in the treated cohort than in the untreated control animals. In addition to the dramatic anti-tumour effect, a measurable increase in T-cell proliferation was detected in Ipilimumabtreated mice.

Taken together, these initial studies indicate that the presence of a human immune system does not lead to rejection of tumours in PDX models and that an appropriate immune response can be elicited following administration of an immune checkpoint inhibitor. Future studies that involve testing the efficacy of new checkpoint inhibitors and combination therapies (eg standard-of-care chemotherapeutics combined with other classes of immunotherapeutic agents) will no doubt provide critical insight into the value of these models as a pre-clinical tool in immuno-oncology research.

The tumour microenvironment is a complex tissue containing heterogeneous cell populations that regulate tumour progression29. Tumour-infiltrating myeloid-derived suppressor cells (MDSCs) are a mixed population of immature myeloid progenitors, including granulocyte and monocyte precursors, that potently inhibit T-cell activation30. Given the immunosuppressive function of MDSCs, it may be a worthwhile anti-cancer strategy to target that cellular subset, especially in combination with T-cell activating immunotherapies31.

The huNOG-EXL model, which permits myeloid lineage commitment, is well-suited to such studies. In preliminary work at CIEA, huNOGEXL mice were engrafted with a human cancer cell line, allowed to grow for 14 days, then homogenised and assayed by flow cytometry for the presence of tumour-infiltrating suppressive myeloid cells. The utility of this model was supported when CD11b+ MDSCs and tumour-specific IL-4Ra+ tissue associated macrophages (TAMs) were readily detected at increased levels in tumours but not in normal splenic tissue. In view of these early results, the huNOG-EXL mouse may provide a much-needed model for examining the potential of targeting immunosuppressive cells in the tumour microenvironment and thereby inducing antitumour immunity. In addition, this model could be used in combination with advanced technologies such as high-throughput proteomics to identify new immune checkpoint targets for immunotherapy drug development.

Much progress has been made in the generation of highly immunodeficient mice, including xenograft models and models that enable efficient humanisation of immune systems. With recent breakthroughs in immuno-oncology, the combined use of these models has great potential for the discovery and development of new immunotherapies. By improving the predictability of clinical outcomes, the time and cost associated with late-stage drug development failures can be reduced. The ultimate goal of these models, however, is to identify therapeutic agents that can be swiftly translated into effective clinical treatments for patients living with cancer.

Parkin, J and Cohen, B. An overview of the immune system. Lancet 357, 1777-89 (2001).

2 Flanagan, SP. ‘Nude’, a new hairless gene with pleiotropic effects in the mouse. Genet. Res. 8, 295–309 (1966).

3 Bosma, GC, Custer, RP and Bosma, MJ. A severe combined immunodeficiency mutation in the mouse. Nature 301, 527- 30 (1983).

4 Ito, R, Takahashi, T, Katano, I and Ito, M. Current advances in humanized mouse models. Cell. Mol. Immunol. 9, 208-14 (2012).

5 Shultz, LD et al. Multiple defects in innate and adaptive immunologic function in NOD/LtSz-scid mice. J. Immunol. 154, 180-91 (1995).

6 Takenaka, K et al. Polymorphism in Sirpa modulates engraftment of human hematopoietic stem cells. Nat. Immunol. 8, 1313-23 (2007).

7 Ito, M et al. NOD/SCID/gamma(c)(null) mouse: an excellent recipient mouse model for engraftment of human cells. Blood 100, 3175-82 (2002).

8 Ishikawa, F et al. Development of functional human blood and immune systems in NOD/SCID/IL2 receptor {gamma} chain(null) mice. Blood 106, 1565–73 (2005).

9 Shultz, LD, Brehm, MA, Garcia-Martinez, JV and Greiner, DL. Humanized mice for immune system investigation: progress, promise and challenges. Nat. Rev. Immunol. 12, 786-98 (2012).

10 Shultz, LD, Ishikawa, F and Greiner, DL. Humanized mice in translational biomedical research. Nat. Rev. Immunol. 7, 118-30 (2007).

11 Mosier, DE, Gulizia, RJ, Baird, SM and Wilson, DB. Transfer of a functional human immune system to mice with severe combined immunodeficiency. Nature 335, 256-9 (1988).

12 Melkus, MW et al. Humanized mice mount specific adaptive and innate immune responses to EBV and TSST-1. Nat. Med. 12, 1316-22 (2006).

13 Lapidot, T et al. Cytokine stimulation of multilineage hematopoiesis from immature human cells engrafted in SCID mice. Science 255, 1137-41 (1992).

14 Ito, R et al. Establishment of a human allergy model using human IL-3/GM-CSFtransgenic NOG mice. J. Immunol. 191, 2890-9 (2013).

15 Brehm, MA et al. Engraftment of human HSCs in nonirradiated newborn NODscid IL2r? null mice is enhanced by transgenic expression of membranebound human SCF. Blood 119, 2778-88 (2012).

16 Takagi, S et al. Membranebound human SCF/KL promotes in vivo human hematopoietic engraftment and myeloid differentiation. Blood 119, 2768-77 (2012).

17 Wunderlich, M et al. AML xenograft efficiency is significantly improved in NOD/SCID-IL2RG mice constitutively expressing human SCF, GM-CSF and IL-3. Leukemia 24, 1785-8 (2010).

18 Hanahan, D and Weinberg, RA. Hallmarks of cancer: the next generation. Cell 144, 646–74 (2011).

19 Dranoff, G. Experimental mouse tumour models: what can be learnt about human cancer immunology? Nat. Rev. Immunol. 12, 61-6 (2012).

20 Siolas, D and Hannon, GJ. Patient-derived tumor xenografts: transforming clinical samples into mouse models. Cancer Res. 73, 5315- 9 (2013).

21 Hidalgo, M et al. Patientderived xenograft models: an emerging platform for translational cancer research. Cancer Discov. 4, 998-1013 (2014).

22 Parish, CR. Cancer immunotherapy: the past, the present and the future. Immunol. Cell Biol. 81, 106-13 (2003).

23 Pardoll, DM. The blockade of immune checkpoints in cancer immunotherapy. Nat. Rev. Cancer 12, 252-64 (2012).

24 Schwartz, RH. Costimulation of T lymphocytes: the role of CD28, CTLA-4, and B7/BB1 in interleukin-2 production and immunotherapy. Cell 71, 1065- 1068 (1992).

25 Sondak, VK, Smalley, KSM, Kudchadkar, R, Grippon, S and Kirkpatrick, P. Ipilimumab. Nat. Rev. Drug Discov. 10, 411-2 (2011).

26 Freeman, GJ. Engagement of the PD-1 Immunoinhibitory Receptor by a Novel B7 Family Member Leads to Negative Regulation of Lymphocyte Activation. J. Exp. Med. 192, 1027-1034 (2000).

27 Dong, H et al. Tumorassociated B7-H1 promotes Tcell apoptosis: a potential mechanism of immune evasion. Nat. Med. 8, 793-800 (2002).

28 Webster, RM. The immune checkpoint inhibitors: where are we now? Nat. Rev. Drug Discov. 13, 883-4 (2014).

29 Gajewski, TF, Schreiber, H and Fu, Y-X. Innate and adaptive immune cells in the tumor microenvironment. Nat. Immunol. 14, 1014-22 (2013).

30 Talmadge, JE and Gabrilovich, DI. History of myeloid-derived suppressor cells. Nat. Rev. Cancer 13, 739-52 (2013).

31 Mahoney, KM, Rennert, PD and Freeman, GJ. Combination cancer immunotherapy and new immunomodulatory targets. Nat. Rev. Drug Discov. 14, 561-584 (2015).