Bridging the Translation Gap in Cancer Immunotherapy

Immunotherapy is an effective way to fight cancer using the body’s own immune system. In vivo modeling is the cornerstone of immunotherapy development, but many models lack the accuracy needed to translate therapies into the clinic.

In this Innovation Spotlight, Brian Soper, senior scientific engagement manager at The Jackson Laboratory, discusses how the latest immuno-oncology platforms provide a better model of the tumor microenvironment, enabling a more clinically relevant understanding of a therapeutic’s interactions and efficacy.

Why do so many promising immuno-oncology therapies struggle to translate from preclinical studies to clinical success?

Many new therapies under development show extraordinary activity when tested in vitro under defined conditions where human cellular components are included using the researchers’ best estimate of the conditions found in vivo. However, the true in vivo condition includes many variables that cannot easily be modeled in a culture dish. These variables include other cell types that play an immunosuppressive role counteracting the mechanism of action of the therapy and complexities of the microenvironment at the tumor site as well as systemic biology that influences the bioavailability of the therapy. Furthermore, robust immune cell activation when tumor burden is high causes the release of cytokines that can lead to systemic toxicities that, in the most severe cases, include immune effector cell-associated neurotoxicity syndrome. Both bioavailability and toxicities can vary significantly from patient to patient.

How are humanized mouse models helping close that gap?

First, we need to define clearly what is meant when we say “humanized mouse model”. Here, we are referring to highly immunodeficient mice that often are further genetically modified to express human hematopoietic growth factors that support development of human immune cell populations following the engraftment of cord blood-derived hematopoietic stem cells (HSCs) or enable maintenance of mature immune cells following engraftment of peripheral blood mononuclear cells (PBMC).

The Jackson Laboratory has recently released a new platform (S15-DKO) that expresses multiple human growth factors and is deficient in expression of mouse MHC class I and II to allow support of a wider range of human immune cells following PBMC engraftment while also greatly diminishing xenogeneic graft-versus-host disease. Once the human immune system is established, the mice are further humanized by co-engraftment of a human tumor, either a blood cancer or solid tumor. These platforms enable better modeling of the complexities of the human tumor microenvironment and provide the full systemic model needed to understand the therapies’ absorption, distribution, metabolism, and elimination in addition to therapeutic efficacy. Preclinical studies can be designed to include HSC- or PBMC-engrafted mice from multiple donors to assist in capturing donor-to-donor differences in therapeutic efficacy as well as safety.

How does the Onco-Hu^® platform help researchers generate clinically relevant efficacy and safety data?

The Onco-Hu^® platform provides a more complete system, modeling the complexities of human-to-human cellular interactions that are not easily predicted and thus not created in the culture dish. This includes a holistic evaluation of the consequences of associated toxicities to understand if they are transitory or lead to more significant organ and tissue damage. Study design can include multiple donors to capture human diversity typical of a clinical study. The data sets generated include tumor efficacy and durability of response, cytokine release for understanding mechanism of action and safety, and clinical assessments of individual recipients. These can include assessments of blood chemistry as well as tissues for biochemical and histological analysis, all in support of investigational new drug (IND) filing.

What types of human immune cells can be studied using JAX humanized mouse models?

Immune humanized mice enable analysis of a wide range of human immune cells and their function. The choice of host along with the engraftment source (HSC vs PBMC) will determine what immune cells are present in the system. Human T cells are a strong focus area and have been modulated using checkpoint inhibitors, bispecific T-cell engagers, bispecifics, and trispecifics. These T cells are now being modified in vivo with viral or lipid nanoparticles to deliver genetic payloads for CAR T generation. Similar methods are being used for NK and γδ T cells. Other approaches include methods to block immunosuppression by myeloid suppressor cells and the adoptive transfer of ex vivo manipulated cells, including CAR T cells, NK cells, macrophages, or dendritic cells.

Can you share a real-world example of how JAX humanized mouse models helped researchers make better decisions during drug development?

One of the most impressive examples is a series of papers published by Roche. In 2018, they published a paper using HSC humanized mice to develop a dosing strategy for a new CD20 T cell bispecific (glofitamab, aka Columvi) to treat B cell lymphoma.¹ Treatment with glofitamab alone showed significant tumor efficacy, but this was coupled with strong T cell activation and high cytokine release, raising concerns of safety in the clinic. A prior clinically approved drug developed by Roche called obinutuzumab (aka Gazyva) demonstrated partial efficacy in the absence of high cytokine release.² The team devised a stepwise dosing protocol where immune and tumor humanized mice were first treated with obinutuzumab to debulk the tumor and then treated with glofitamab to enable a more complete tumor response in the absence of high cytokine release. The treatment method modeled in the humanized mice supported IND filings for clinical trials that led to clinical approval in 2023.^3,4

Given the success of this dosing strategy for a “signal one” approach to activate T cells, the team continued the use of humanized mice to evaluate the inclusion of a new bispecific targeting CD19 and CD28 to enable a “signal two” T cell signal to further perpetuate T cell activity and allow a deeper and even more durable response.⁵ Building on the prior dosing strategy, they first dosed with obinutuzumab to debulk the tumor, then treated with glofitamab, followed by treatment with the new CD19-CD28 bispecific. The efficacy data generated in the humanized mice supported initiation of a new clinical trial that is currently underway.

Looking ahead, what advances in preclinical modeling do you think will have the greatest impact on the future of immuno-oncology research?

As with any small animal model of human disease, the humanized platforms have limitations. While neutrophils can be established in certain host platforms, the overall numbers and proportion of these important cells are low and do not accurately model what is found in the human population. Similarly, while many of these platforms develop or support B cells, the ability of these cells to undergo robust class switching and IgG production is limited.

The new S15-DKO is an important advancement in that it enables support of mature human memory B cells capable of IgG production, raising the potential for improved modeling of human autoimmune disease using patient samples. The importance of B cell functionality is also reflected in the need by the scientific community for a platform capable of predicting large molecule treatment-induced immunogenicity, another leading cause of therapeutic drug failure in the clinic.

The Jackson Laboratory has a solid track record as a leading innovator in the creation of new humanized mouse platforms with improved human immune function. Extensive research and development are currently focused on improving B cell maturation and function. Having a platform with full B cell functionality will have a very significant impact on immuno-oncology both in terms of immunogenicity and cancer vaccine development.