Human Immune System (HIS) and Patient-Derived Xenograft (PDX) models

Post Title

Human Immune System (HIS) and Patient-Derived Xenograft (PDX) models

Champions Oncology and others in the field working with patient derived xenografts (PDX) are developing and providing translationally relevant preclinical models to test therapeutic efficacy of IO agents as monotherapy or rational combinations in humanized mice engrafted with functional human immune system.

Boosting a patient’s immune system to attack and reject a tumor has been hailed as a major breakthrough in cancer treatment. The activated and educated immune system against a tumor can reduce the burden in some patients. However, it isn’t always effective. Moreover, it is associated with the potential to result in severe nonspecific immune activation, thereby necessitating the halting treatment or even contributing to patient death.

Current available preclinical models for testing immune-oncology (IO) agents include syngeneic mouse models. Due to the presence of a functional immune system, they have the potential to afford an understanding of the mechanism of action of IO agents as well as identify interactions with other immune compartments and the tumor itself, but these models are not always reliable predictive of clinical outcomes.

Human Immune System (HIS) Preclinical Mouse Model – Humanized Mice

At Champions, these are developed by adoptive transfer of cord-blood derived CD34 cells in immunodeficient mice strains. Lack of a functional murine immune system allows for CD34 cells to be engrafted in the spleen and bone marrow, followed by downstream lineage development of lymphoid and myeloid cells in 12-16 weeks. Immune cell development is measured by the presence of huCD45 in periphery at 12 weeks; if the levels are above 25% the mice are considered humanized mice.

Key leukocyte surface markers in humanized mice

PDX tumors are subcutaneously implanted in the flanks of these mice; when tumors are palpable the mice are divided into groups and therapeutic response against IO agents is evaluated in terms of tumor growth inhibition (TGI) and mechanism of action. The NOG-EXL mice strain allows for superior engraftment; i.e., specific and functional immune lineage development that can be targeted and studied due to transgene expression of hu GM-CSF and huIL-3. These mice exhibit higher engraftment, and downstream myeloid lineage development (1).

Humanization in human IL-2 NOG mice allows for development of functional NK cells and evaluation of therapeutic agent targeting NK cells (2). Complete HLA match with PDX is not possible in this platform; therefore, in order to reduce graft vs. host disease, partial HLA haplotype matching is done.

CD34 cells are more resistant to graft vs. host disease and hence where haplotype HLA match is not possible HLA mismatch studies can successfully be performed. The lack of human thymus is one of the major caveats in this platform and, as a result, the human T cells undergo education in mouse thymus and are restricted to H2-MHC (3). However, the haplotype HIS PDX platform does allow for:

1) Evaluation of IO therapies in PDX derived from IO therapy responsive and non-responsive patients and patients who have acquired resistance to IO therapy providing a relevant translational platform to develop better IO therapies and combinations.
2) Biomarker/target/ molecular targets based studies; e.g., Her2 expression, MSI high models.

Peripheral blood mononuclear cell (PBMC) based studies

These are fast and aggressive models where PBMCs are adoptively transferred into immunodeficient mice and PDX tumors are implanted within a week of PBMC engraftment. These models have developed

T cells and are not dependent on mouse thymus education, but they do have a drawback of succumbing to faster graft vs. host disease. (4) Beta2microglobulin deficient mice strain in these studies can reduce the rate of graft vs host disease.

Future directions (R&D initiatives):

In order to overcome limitations associated with haplotype HIS models, Champions Oncology is developing and evaluating autologous human immune system (HIS) PDX models.

Mobilized peripheral blood derived CD34 models – are from the same patients from whom PDX are developed, eliminating graft vs host disease progression and allowing evaluation of immune response to PDX from the same patient — a much more suitable translational preclinical model to assess therapeutic efficacy of IO therapies.

Autologus PBMC models – A matched PBMC and PDX platform allows assessing therapeutic efficacy against IO agents in a relevant immune background. We would anticipate reduced graft vs host disease due to autologous PBMC matching, enabling a longer duration on study. This platform also allows for manipulation of the immune cells present in PBMCs ex vivo for expansion of T cells and dendritic cells among others (5). CART and adoptive cell therapies can also be performed using this platform.

Autologus models derived from TILs – Successful isolation of tumor infiltrating lymphocytes (TILs) and rapid large scale expansion ex vivo will allow for development of this platform. TILs from the matched Patient tumor (PDX derived from tumor) have been selected in the patient to recognize the tumor cells and may be specific; which will allow for better understanding of resistance mechanisms to immune therapy and therapeutic responses to combinations consisting of cancer immune therapy and targeted therapies. (6)

Ex vivo expansion of CD34 cells

Ex vivo expansion of CD34 cells derived from different sources allows for/to:

1) Repeat haplotype HIS PDX experiments from individual donors due to expanded UCB derived CD34 cells

2) Promote expansion of autologous CD34 cells

Clinical Correlation Studies

By using PDX models with known clinical responses to IO agents, we can interrogate relevant humanized platforms to show the degree of correlation between patients and PDX outcomes.  This not only extends to tumor growth inhibition, but also immune cell activation and/or tumor infiltration.

Biomarker Analysis – Studies to under stand cancer genome alterations after IO therapy

As we continue to characterize the genomics of our PDX models (WES and RNASeq) we can assess markers of response/non-response to IO agents, in addition to identifying any changes in mRNA expression, mutation profiles, copy numbers and/or gene fusion events.


1. Ito, et al, J Immunol, 2013
2. Katano, et al, J Immunol, 2015
3. Walsch, et al, Annu.Rev.Pathol, 2017
4. Sanmamed, et al, Cancer Res, 2015
5. Spranger, et al, Journal of Translational Medicine, 2012
6. Jespersen, et al, Nature communications, 2017