Analysts at Technavio have predicted the global humanized mouse model market to grow at a compound annual growth rate (CAGR) of almost 10% during 2017–2021, attributing this growth to the increasing number of applications in fields including oncology, immunology, neuroscience, and toxicology. In oncology, the emerging success of immunotherapies such as antibody-based checkpoint inhibitors is driving demand for accurate, more predictive preclinical animal models. Both syngeneic and humanized models have valuable roles in immuno-oncology drug discovery and development, and it is important to understand the advantages, limitations, and best uses of each. GEN spoke to several researchers—who participated in a recent Cambridge Healthtech conference on preclinical animal models in oncology—on advances in engineered animal models used to study novel cancer immunotherapies.
Paula Miliani de Marval, Ph.D., research associate director at Charles River Laboratories, has seen exponential growth in the demand for syngeneic mouse models, which are immunocompetent animals implanted with tumors of mouse origin. Compared with xenograft models created using immunodeficient mice, syngeneic mice provide a full picture of how the mouse immune system affects and interacts with a tumor.
As the focus of cancer therapy has been shifting from directly treating the tumor to harnessing the immune system to attack the tumor, Charles River has followed this trend—by developing super-immunocompromised mice (Figure 1). While these models have some limitations, with not all human immune cells being fully represented, “they can be used effectively to study how the immune system responds to immune checkpoint inhibition,” says Dr. Miliani de Marval, to do proof-of-concept studies, immune target validation, and determine if the expected responses occur.
Illustrating the value of these models in immuno-oncology, Dr. Miliani de Marval described a recent study of combination checkpoint inhibitors to treat solid tumors. The study involved a NCG Il2-γ knockout mouse model—created by sequential CRISPR/Cas9 editing of the Prkdc and Il2rg loci in the NOD/Nju mouse—and human RKO colorectal carcinoma xenograft, which has high PD-L1 expression. The model was used to evaluate the efficacy of the monoclonal antibody-based drugs pembrolizumab (anti-PD-1) and ipilimumab (anti-CTLA-4) when administered alone or in combination. The antibodies each demonstrated efficacy as monotherapies, but, surprisingly, combination therapy did not have any additive effect. The results were replicated in a triple-negative breast cancer model.
Ongoing research aims to achieve a broader range and normal balance of immune-cell representation in transgenic humanized models. The goal is to recreate accurately the tumor and immune microenvironment in much the same way it is in syngeneic mouse models.
Engineering Large-Animal Models
Surrogen, a wholly-owned subsidiary of Recombinetics, develops swine models of human diseases using gene-editing technologies including TALEN- and CRISPR-based techniques. Large animals are often used in preclinical studies to assess the safety of drug candidates. Pigs, in particular, have demonstrated effectiveness as efficient biomedical models for evaluating drug efficacy, according to Adrienne Watson, Ph.D., senior research scientist at Surrogen.
One of the main cancer types being targeted at Surrogen is brain tumors, for which the first line of treatment is usually surgery to remove as much of the tumor as possible. The brain and other organs of pigs are anatomically (in size) and physiologically more similar to humans than are those of smaller animal models such as mice. Using pigs makes it possible to perform the types of surgeries and imaging studies that cancer patients might typically undergo. Furthermore, swine share vast genetic and metabolic similarities to humans, making nearly any genetic disease feasible to model in this animal.
“The pig can be more predictive of how a drug will act in patients and accelerate the timeline of bringing a drug through preclinical development and to the clinic,” says Dr. Watson.
In developing a new model animal, Surrogen reviews extensive genetic datasets and scientific literature to identify the exact mutations associated with a specific type of human cancer in large groups of patients. The company then engineers those exact mutations into the swine genome. Using genome-editing tools, targeted mutations are made in fetal swine fibroblasts. Piglets carrying a particular mutation—for example, associated with neurofibromatosis type 1 (NF1)—are then produced through somatic cell nuclear transfer.
These models allow researchers to study the development, progression, and metastatic behavior of tumors, beginning at an animal’s birth. The models are used to identify predictive blood biomarkers, characterize different types of tumors, search for therapeutic targets, and study the safety and effectiveness of experimental treatments.
For NF1, Surrogen has shown that the appearance and progression of the disease in humans, which typically develops before puberty, is mirrored in the model pigs. Figure 2 shows a young pig with hyper-pigmented café-au-lait spots on its skin, which are characteristic of NF1 and under which deep nerve tumors often develop in both humans and pigs.
Advances in gene-editing technology are making it easier for companies such as Surrogen to progress from making single mutations in a genome to creating animal models with multiple mutations, which is especially important for studying cancer. Surrogen is also creating increasingly humanized swine models by introducing entire human genes. These models facilitate studies of monoclonal antibody-based immunotherapies. Also in development are inducible, organ-specific swine models, such as animals in which researchers can activate oncogenes known to induce pancreatic cancer only in pancreatic cells, or turn off tumor suppressor genes only in certain areas of the brain to induce glioblastoma.
Targeting T Cells and Natural Killer Cells
Preclinical immuno-oncology research in animal models requires human immune cells that are fully functional. “You need to be able to study human proteins on appropriate immune-cell targets in a preclinical setting” and to widen the possibility for detecting both on-target and off-target effects, says Azusa Tanaka, Ph.D., product manager at Taconic Biosciences. Taconic uses CRISPR/Cas9, knockout/knockin, and transgenic technologies to create genetically engineered mice and rats, and offers a super-immunodeficient model portfolio that includes the CIEA NOG mouse. The NOG portfolio includes genetically humanized mice expressing human cytokines, and cell-humanized NOG mice such as hematopoietic stem cell (HSC) human immune system (HIS)–engrafted mouse models (Figure 3).
A key challenge is to understand the limitations of each technology and to select the best model for a particular application, according to Dr. Tanaka. “Use a syngeneic model first to understand the mechanism of action, the target, and the cells involved in an experimental therapy,” she advises. “Once you have identified the specific cell target (e.g., T cells, myeloid cells), then select a humanized model that allows reconstitution of that particular cell type to study a protein of interest. For example, targeting a myeloid cell requires a transgenic mouse that expresses human GM-CSF and IL-3 cytokines, and studying multiple myeloma may call for a model that expresses human IL-6. [Taconic’s] HIS-engrafted models include the HSC-engrafted mice that transgenically express human GM-CSF and hIL-3, called huNOG-EXL™.”
Among the advantages of using mice as models for drug discovery and development is the ability to create syngeneic and humanized animals, and their relatively small size and fast reproduction rate, making it easier and less costly to generate larger populations to perform high-throughput testing.
Dr. Azusa describes the advantages of using humanized mice to study immunotherapeutic approaches for cancer that has metastasized to the bone. Bone metastasis is especially difficult to target and treat, due in part to the development of drug resistance, and because HSCs reside in the bone marrow. HSC-engrafted (HIS) mice can mimic the coexistence of HSCs and cancer metastasis in the bone, allow researchers to study the crosstalk between human immune-cell populations, HSCs in the bone marrow, and the tumor and bone microenvironment, and enable studies of immuno-oncology drugs.
While T cells have largely been the focus of immuno-oncology drug discovery to date, researchers are increasingly looking closely at natural killer (NK) cells as targets and as the basis for cell and gene therapy. Human NK cells require human IL-15 for survival, and Taconic provides a human transgenic super-immunodeficient mouse that expresses human IL-15 to promote the expansion of human NK cells.
Matching Immune and Tumor Cells
“There are a lot of different versions of humanized mouse models, which can make it confusing for consumers,” says Marcus W. Bosenberg, M.D., Ph.D., associate professor at Yale University. One issue to consider is how the human immune system is introduced into the mouse.
Dr. Bosenberg studies melanoma and typically uses a more primitive mouse model, the NSG™ immunodeficient mouse strain from The Jackson Laboratory, into which he remakes the human immune system using lymphocytes removed from the same tumor that he transplants into the mouse. As the tumor-infiltrating lymphocytes (TILs) and the tumor are derived from the same patient, the interaction between immune cells and the tumor will more closely mirror what occurs in the patient. Similarly, the response to an immunotherapeutic drug—such as an immune-checkpoint inhibitor—in mice with matched TILs and tumor cells is anticipated to be more predictive of the response in humans.
Another advantage of using TILs, instead of HSCs derived from umbilical cord blood or some other source to rebuild the immune system of immunodeficient mouse models, is a reduced risk of graft-versus-host disease, which can limit the lifespan of the animals.
Most of Dr. Bosenberg’s research relies on syngeneic mouse models, which “have been the workhorse for industry because they are relatively fast, cheap, and easy to use,” he says. At present, though, only approximately 5% of cancer drugs that make it into Phase I testing are approved for commercial use. The hope is that “humanized mice will improve the predictive value of preclinical testing,” says Dr. Bosenberg. In the meantime, researchers are also exploring faster, less costly alternatives that will allow them to avoid the use of animal models. How predictive these complex, in vitro three-dimensional tissue models—such as spheroids—are, also remains to be determined.