Preclinical studies in oncology are frequently based on the use of mouse models. On average, the protein-coding regions of mouse and human genomes are 85% identical and the mouse genome can be easily manipulated.

But as the saying goes, close only counts in horseshoes.

The genomic identity between H. sapiens and M. musculus cannot compensate for significant species differences that, unfortunately, influence the outcome of experiments in not-so-beneficial ways. One good example is the failure of 9-AC in Phase II clinical trials despite very impressive preclinical data: the compound failed due to its small therapeutic window in humans as compared to mice.¹

To overcome this limitation, multiple efforts have been undertaken to establish and characterize large collections of patient-derived tumor xenograft (PDX) models—mice implanted with cancerous tissue from human patients—for cancer research.

Although this model system dates back to the late 1980s, it recently came into focus due to its ability to predict clinical outcomes and its utility in biomarker development. PDX models mainly retain the histological and genetic characteristics of the donor tumor and remain stable across passages, which other human tumor mouse models lack due to the selection process they undergo during their establishment, in vitro, on plastic. Since they can preserve cell-autonomous heterogeneity, human tumor xenografts in immunodeficient mice have also led to valuable insights into the biology of human cancers.

PDX and the tumor microenvironment
With that said, PDX models have their own set of speed bumps to contend with. For one, the high cost of developing such models has led to a “productivity gap.” PDX models have also been associated with high failure rates of new drugs particularly in early clinical development.^1-3 This lack of translational capability may originate from the fact that in PDX the human non-malignant cells, otherwise referred to as the tumor microenvironment (TME), get replaced within a few days by their murine counterparts.⁴ Fortunately, we’re gaining a much better understanding of the pivotal role TME plays in cancer biology. This is leading us to new drug targets that either modulate the TME in order to support more effective tumor cell killing or evoke an anti-tumor immune response.^5,6 Parts of the murine TME are being replaced with immune cells and stromal components of human origin to produce what is essentially a humanized version of a PDX mouse model.

This marriage between PDX and humanized mice has enormous relevance, especially in the development and testing of immune-oncology drugs. After struggling for years to make immunotherapies stick, researchers sparked a renaissance in cancer treatment with so-called checkpoint inhibitors that target the T-cell co-inhibitory receptors CTLA-4 and PD-1. Checkpoint inhibitors, which block the molecules that switch off immune cells, thus increasing tumor immunogenicity, are largely why Science magazine named cancer immunotherapy its breakthrough of the year in 2013.⁷

The pipeline is now filled with compounds targeting PD-1 and CTLA-4 in solid and blood tumors, and humanized mice engrafted with PDX tumors represent one of the most novel ways to measure the immune responses induced by such compounds against human tumors growing in a human tumor environment.

The reasons are clear. The co-culture of patient immune cells and patient derived tumors in an in vivo setting can theoretically trigger the study of novel therapies targeting tumor-immune interactions and in parallel allow for novel insights into tumor biology. No other model offers this much flexibility. But because the TME is made up of multiple components, it is also technically challenging to reproduce in a living animal. Triple immunodeficient mouse strains like NOD/Shi-scid/IL-2Rγnull (NOG) or NOD. Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) have made the process easier, as human non-malignant cells engraft to a high extent in these animals.

Given the basic role of the immune system in tumor biology, the need for immunocompromised mice might be the most obvious drawback of PDX models. Tumor cells are broadly thought to be antigenic due to their multiple mutations in coding exons resulting in a phenotypic make-up reasonably different from the normal cells they evolved.⁸ However, most tumors eventually progress and evade the immune surveillance by inducing immune tolerance. The pro-inflammatory microenvironment in the tumor tissue established by CD8+ T cells, tumor-associated macrophages (TAM), natural killer (NK) cells, and others leads to an immune suppressive signaling cascade enabling the tumor to evade immune surveillance.

The ability to influence the immune system is often related to prognosis: the cytotoxic to regulatory T-cell ratio is a strong prognostic marker in many solid tumors.⁹ The fact that macrophage and CD4+ T-cell recruitment following intensive chemotherapy in breast cancer patients is associated with significantly reduced recurrence-free survival follows the same line. In the light of these encouraging findings the need for predictive and robust preclinical models suitable for immune-oncology studies is getting more and more indispensable.

Building a Mighty Mouse
What’s involved in building these humanized PDX models? Mice supplemented with human immune cells should meet at least three criteria:
(a) All lineages and subsets of human hematopoietic cells should develop in the mouse host, and be maintained in proportions and localizations similar to those observed in healthy humans
(b) The cells should be functional, that is they should be capable of mounting innate and adaptive immune responses in vivo
(c) The cells should allow for testing of therapeutic interventions and faithfully predict the outcome in clinical settings¹⁰

Two basic approaches are used to generate humanized mice. One involves the engraftment of hematopoietic stem cells (HSC) from different sources, the other approach involves injecting mice with human peripheral blood mononuclear cells (PBMC) (see Figure 2).

In the latter, one can encounter challenges when engrafted human T-lymphocytes show xeno-reactivity against foreign major histocompatibility (MHC) class I and II molecules and other antigens of murine origin. As a result, inflammatory cells, namely T lymphocytes, infiltrate different organs and kill the animals after several weeks, a process known as xenograft-versus-host disease (xGVHD).¹¹ However, the co-transplantation of human hematopoietic stem cells together with cancer cells seems to enable the co-existence of MHC-mismatched cells without inducing rejection.¹²

Despite the risk of xGVHD, treatment experiments are mainly performed using PBMC-humanized mice. As long as the influence of xGVHD on tumor growth is taken into account, these experiments provide valuable insights into the possible benefit of new immunomodulatory compounds.¹³ On the other hand, HSC-humanized mice, which are rising in popularity, offer a broader time window of up to four months for treatment experiments. The lack of xGVHD also facilitates the read-out of these studies with respect to anti-tumor activity and side effects.¹⁴

Humanized PDX performance
Studies by our group and others proved the feasibility of this model for drug screening purposes in an immune-oncology setting. The studies, which focused on non-small cell lung cancer (NSCLC), found that growth behavior of the subcutaneously implanted PDX models of NSCLC were not affected at all by the engraftment of HSC in the murine host, meaning that take rates and doubling times were not influenced by presence of human immune cells. The histological structure of the tissue and cells of the investigated PDX models were similar when implanted subcutaneously in humanized or immunodeficient mice, and still closely resembled the patient donor material.¹⁵

In several studies, the experimental checkpoint inhibitors targeting CTLA-4 and PD-1 depicted anti-tumor activity in traditional xenografts as well as PDX of NSCLC. Combination therapy promoted T cell expansion in most of the examined models. And the use of humanized mice in a single mouse trial format (Figure 3),¹⁶ turned out to be a cost-effective in vivo screening approach, with one tumor-bearing animal per tumor model and treatment scheme. The approach also perfectly mirrored clinical diversity of tumor responses to a specific treatment.

At the end of the day, our investigations confirmed the general benefit of PDX in mice engrafted with human immune cells, and in combination with the well described advantages of the PDX-based platform, such as preserved tumor heterogeneity and clinical relevant molecular make-up, this assay format provided an additional step toward the development of new immunotherapies.

The investigation of immune cell subsets like monocytes or natural killer cells is another area of preclinical investigation being done in humanized PDX models, with encouraging results. The co-injection of human monocytes in parallel with the lung cancer drug bevacizumab in PDX-bearing animals improved the anti-tumor activity of the monoclonal antibody markedly. In contrast, the application of human monocytes alone did not influence the tumor growth of the NSCLC PDX investigated in this study.¹⁷ Monocytes and macrophages have been reported to induce antibody-dependent cellular cytotoxicity and antibody-dependent cellular phagocytosis of tumor cells in the presence of IgG anti-tumor monoclonal antibodies, like bevacizumab. By confirming these observations in a PDX-based NSCLC in vivo model, the study highlighted the suitability of PDX for immuno-oncology approaches by supplementation of the murine host with human immune cells.

Chimeric antigen receptor T-cell therapies, where immune cells are genetically modified to hunt down and destroy cancer cells, have also proved to be highly efficient in hematological malignancies as well as solid cancers.^18,19 CAR-T, meaning a T cell expressing an antigen-specific or antibody-based chimeric receptor with antibody specificity and T-cell effector or regulatory function, was first described in the 1980s by Eshhar et al.²⁰ Clinical^21,22 as well as preclinical studies^23,24 now show encouraging data in hematological as well as solid tumors. The availability of large PDX collections comprising a broad range of different specific tumor types enables the preclinical testing of CAR-T cells targeted against most of the tumor-specific antigens currently under investigation. 

Conclusion
The establishment of PDX models that recapitulate the complexity and genetic heterogeneity of human cancer is an invaluable research tool, enabling insights into tumor biology and strategies for developing new innovative approaches for anticancer treatment. Combining highly immunosuppressed mouse strains like NOG/NSG and second generation mouse models opens up new possibilities for innovative and predictive preclinical platforms reflecting clinical reality specifically in the context of immuno-oncology research.

At the end of the day, the quality of animal research will have a major impact on the translational value of all PDX-based experiments. Quality in this context begins at the animal facility, which should meet internationally accepted standards provided by AALAC (Association for Assessment and Accreditation of Laboratory Animal Care).

The read-out and interpretation of the data is also key. In many research groups, the design of animal studies, the experimental execution, and the evaluation of the data are under the purview of one, non-masked, person. Several studies elucidated that certain weaknesses, including this lack of masking, limit their translational value to human application.²⁵ Moreover, rigorous criteria have to be implemented to define anti-tumor activity. Repeated measures of tumor burden by caliper are common practice irrespective of the investigated tumor type or study context. These simplistic criteria often used in mouse models do not match the challenging criteria of clinical or pathological investigations. These obvious discrepancies unfortunately contribute to conflicting opinions about the reliability of mouse models in predicting clinical outcomes³ and necessitate the implementation of additional metrics, such as histopathological examinations and overall survival for immune-modulatory drugs, which primarily target immune cells rather than tumor cells.

On the other hand, among all innovative models and read-outs, the subcutaneously implanted patient derived tumor xenograft remains the core area of preclinical in vivo models because it is easy to follow and still discovers as many drugs as more sophisticated tumor models.

What will the future look like for preclinical drug testing of immunotherapies? PDX models are an important part of the preclinical tool box, and a good complement to genetically engineered mouse models or cell line derived xenografts. Although establishment, characterization and validation of PDX models has been in existence for decades, their value in oncology drug development and tumor biology research is really just starting to emerge. Continued improvements in the PDX platform should bring us closer to more clinically relevant models and help to secure the future of PDX research. 

References

1. Takimoto CH. Why drugs fail: of mice and men revisited. Clinical cancer research : an official journal of the American Association for Cancer Research. 2001 Feb;7(2):229-30.
2. Johnson JI, Decker S, Zaharevitz D, Rubinstein LV, Venditti JM, Schepartz S, et al. Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. British journal of cancer. 2001 May 18;84(10):1424-31.
3. Talmadge JE, Singh RK, Fidler IJ, Raz A. Murine models to evaluate novel and conventional therapeutic strategies for cancer. The American journal of pathology. 2007 Mar;170(3):793-804.
4. Hylander BL, Punt N, Tang H, Hillman J, Vaughan M, Bshara W, et al. Origin of the vasculature supporting growth of primary patient tumor xenografts. Journal of translational medicine. 2013;11:110.
5. Cavazzana-Calvo M. Nature. 2010;467:318-22.
6. Mullard A. Nature Rev Drug Discov. 2010;9:905-6.
7. Couzin-Frankel J. Breakthrough of the year 2013. Cancer immunotherapy. Science (New York, NY). 2013 Dec 20;342(6165):1432-3.
8. Heubeck B, Wendler O, Bumm K, Schafer R, Muller-Vogt U, Hausler M, et al. Tumor-associated antigenic pattern in squamous cell carcinomas of the head and neck--analysed by SEREX. European journal of cancer (Oxford, England : 1990). 2013 Mar;49(4):e1-7.
9. Liu K, Yang K, Wu B, Chen H, Chen X, Chen X, et al. Tumor-Infiltrating Immune Cells Are Associated With Prognosis of Gastric Cancer. Medicine. 2015 Sep;94(39):e1631.
10. Rongvaux A, Takizawa H, Strowig T, Willinger T, Eynon EE, Flavell RA, et al. Human hemato-lymphoid system mice: current use and future potential for medicine. Annual review of immunology. 2013;31:635-74.
11. King MA, Covassin L, Brehm MA, Racki W, Pearson T, Leif J, et al. Human peripheral blood leucocyte non-obese diabetic-severe combined immunodeficiency interleukin-2 receptor gamma chain gene mouse model of xenogeneic graft-versus-host-like disease and the role of host major histocompatibility complex. Clinical and experimental immunology. 2009 Jul;157(1):104-18.
12. Wege AK, Schmidt M, Ueberham E, Ponnath M, Ortmann O, Brockhoff G, et al. Co-transplantation of human hematopoietic stem cells and human breast cancer cells in NSG mice: a novel approach to generate tumor cell specific human antibodies. mAbs. 2014 Jul-Aug;6(4):968-77.
13. Sanmamed MF, Rodriguez I, Schalper KA, Onate C, Azpilikueta A, Rodriguez-Ruiz ME, et al. Nivolumab and Urelumab Enhance Antitumor Activity of Human T Lymphocytes Engrafted in Rag2-/-IL2Rgammanull Immunodeficient Mice. Cancer research. 2015 Sep 1;75(17):3466-78.
14. Wege AK, Ernst W, Eckl J, Frankenberger B, Vollmann-Zwerenz A, Mannel DN, et al. Humanized tumor mice--a new model to study and manipulate the immune response in advanced cancer therapy. International journal of cancer Journal international du cancer. 2011 Nov 1;129(9):2194-206.
15. Oswald E, Klingner K, Lenhard D, Niedermann G, Schüler JB. Abstract 5023: NSCLC PDX model for the evaluation of immuno-oncological treatment strategies. Cancer research. 2015 August 1, 2015;75(15 Supplement):5023.
16. Christina Gredy JBS, Nina Zanella, Heinz-Herbert Fiebig, Thomas Metz. Abstract 2890: Single mouse trials, a concept using patient-derived tumor xenografts for large scale in vivo screens. .  AACR Annual Meeting 2015, 04/2015; 2015; Philadelphia, PA, USA; 2015.
17. Cordula Tschuch KK, Anne Löhr, Yana Raeva, Teppo Haapaniemi, Eva Oswald, Julia B. Schüler. Abstract 590: Co-injection of human monocytes improves the in vivo antitumoral activity of bevacizumab in two NSCLC PDX models. Cancer research. 2016; 07/2016; (76(14 Supplement). ).
18. Kakarla S, Gottschalk S. CAR T cells for solid tumors: armed and ready to go? Cancer journal (Sudbury, Mass). 2014 Mar-Apr;20(2):151-5.
19. Maus MV, Grupp SA, Porter DL, June CH. Antibody-modified T cells: CARs take the front seat for hematologic malignancies. Blood. 2014 2014-04-24 00:00:00;123(17):2625-35.
20. Gross G, Waks T, Eshhar Z. Expression of immunoglobulin-T-cell receptor chimeric molecules as functional receptors with antibody-type specificity. Proceedings of the National Academy of Sciences of the United States of America. 1989 Dec;86(24):10024-8.
21. Yang Y, Jacoby E, Fry TJ. Challenges and opportunities of allogeneic donor-derived CAR T cells. Current opinion in hematology. 2015 Nov;22(6):509-15.
22. Zhang Y, Zhang W, Dai H, Wang Y, Shi F, Wang C, et al. An analytical biomarker for treatment of patients with recurrent B-ALL after remission induced by infusion of anti-CD19 chimeric antigen receptor T (CAR-T) cells. Science China Life sciences. 2016 Apr;59(4):379-85.
23. Cherkassky L, Morello A, Villena-Vargas J, Feng Y, Dimitrov DS, Jones DR, et al. Human CAR T cells with cell-intrinsic PD-1 checkpoint blockade resist tumor-mediated inhibition. The Journal of clinical investigation. 2016 Aug 1;126(8):3130-44.
24. Shiina S, Ohno M, Ohka F, Kuramitsu S, Yamamichi A, Kato A, et al. CAR T Cells Targeting Podoplanin Reduce Orthotopic Glioblastomas in Mouse Brains. Cancer immunology research. 2016 Mar;4(3):259-68.
25. de Jong M, Maina T. Of mice and humans: are they the same?--Implications in cancer translational research. J Nucl Med. 2010 Apr;51(4):501-4.

---

Julia Schueler, DVM, PhD, is head of in vivo operations at Oncotest, a Charles River company based in Freiburg, Germany.