Exciting progress is being made in the development of mice that support tumor growth in the presence of a human immune system. A new publication in Oncogene (Morton et al., 2015) demonstrates the importance of human immune cells in maintaining the microenvironment of patient derived xenografts (PDX). The human immune cells infiltrate the tumor, stimulate lymphangiogenesis, and help to maintain the original genetic expression profile of the PDX. These observations highlight the complex interactions between tumors and the immune system and demonstrate the importance of these interactions in maintaining original tumor fidelity. Maintaining all aspects of human tumor growth in a mouse may enable identification of new treatment strategies that more accurately translate into clinical applications.
Although cancer cell lines grown in mice provide valuable data about tumor cell biology and therapeutic target validation, they are homogeneous cell populations that do not recapitulate all aspects of tumor biology. Further, cancer treatment strategies that have been developed using them have a poor record of therapeutic translation to the clinic. PDX tumors are a heterogeneous population of cells providing a more complete model of human tumor biology, but are typically propagated in highly immunodeficient mice, limiting analysis of human tumor immunobiology. There is a growing body of literature describing the complex role of immune cells and their influence on both tumors and their microenvironment. Therefore, achieving a comprehensive understanding of tumor biology necessary for the development of treatments, in particular immunotherapeutics, requires a model system that includes both human tumors and human immune cells.
Highly immunodeficient NSG mice (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ, 005557) are capable of supporting both the growth of PDX and the engraftment of human hematopoietic stem and progenitor cells (HSPC) that establish a highly functional human immune system. Morton et al. transplanted head and neck squamous cell PDX into HSPCengrafted NSG™ mice, a platform they termed XactMice. These researchers then examined whether the human immune cells provided a more appropriate tumor microenvironment and analyzed how these cells influenced the PDX’s gene expression profile.
In a prior publication, Morton and colleagues described a method to ex vivo expand human HSPCs derived from cord blood, G-CSF-mobilized peripheral blood, and bone marrow. The culture conditions included Il-3, Il-6, SCF, TPO and GM-CSF as growth factors along with Tat-MYC and Tat-BCL2 fusion proteins. Tat, transactivator of transcription, is a protein transduction domain of the HIV-1 transactivation protein. MYC is the myelocytomatosis oncogene transcription factor that is associated with cell cycle progression, histone acylation, regulation of stem cell self-renewal, and B cell proliferation. BCL2, or B cell leukemia/lymphoma 2, is thought to diminish apoptosis. These fusion proteins in combination with growth factors expand HSPC numbers while maintaining their capacity for multi-lineage repopulation and establishment of human immunity following engraftment in NSG™ mice.
The growth characteristics of two samples of human head and neck squamous cell carcinomas (CUHN004 and CUHN013, both initially established in nude mice) were compared following engraftment in either nude, NSG™, or humanized NSG™ (XactMice). Although no attempt was made to HLA match these tumors to the human HSPCs engrafted into the XactMice, the tumors grew at similar rates in XactMice and in the non-humanized NSG™.
Although PDXs at P0 contain passenger patient-derived immune cells, these cells generally are lost during successive passages of the tumors in mice. The CUHN004 and CUHN013 PDXs used in the study had been passaged fourteen (P14) and five (P5) times prior to this experiment, respectively. Following their engraftment in NSG™ or XactMice, neither tumor in NSG™ mice contained any human immune cells, whereas those engrafted in XactMice mice contained CD3+ T cells and CD19+ B cells. Therefore, donor HSPC-derived immune cells were infiltrating the tumors in the engrafted XactMice. Interestingly, the PDX-engrafted XactMice also contained cells in their marrow, blood, spleen and in the tumors themselves that were human CD45+ and CD151+. These markers are associated with bone marrow-derived mesenchymal and immune cells, fibroblasts, and vascular endothelium. Further studies using DNA polymorphisms between the PDX tumor and the donor HPSCs confirmed the CD45+ CD151+ cells were derived from the donor HSPC population. In addition, immunofluorescence and immunohistochemistry identified human HSPC derived cells within the stroma surrounding the human tumor cells.
As PDX tumors are passaged in mice, their associated human stroma cells are gradually lost and replaced by mouse-derived stroma. The change in stroma coupled with the loss of signaling factors no longer supplied from passenger patient-derived immune cells are thought to contribute significantly to changes in the tumors’ gene expression profiles as they are passaged. Based on their findings that human HSPC-derived cells infiltrated the tumor and surrounding stroma in the PDX-engrafted XactMice, the authors examined whether these cells influenced the tumors’ gene expression profiles. Next generation sequencing was used to compare the transcriptomes of the initial (P0) CUHN004 and CUHN013 tumors and those engrafted in nude, NSG™, and humanized XactMice to the normal human genome. The results revealed CUHN013 at P0 and the P5 samples isolated from XactMice showed greater transcriptome alignment to the human genome than those isolated from nude or NSG™ mice, suggesting the HSPC derived cells helped diminish genetic drift in the tumors. However, while the initial (P0) CUHN004 tumors displayed the greatest alignment to the human genome as expected, the P14 samples from the nude recipient showed greater alignment to the human genome than those from the NSG or XactMice. CUHN004 tumor from the XactMice did show a higherpercent alignment to the human genome than tumor from the NSG™ recipients. One possible reason for lower alignment from the XactMice tumor may be the greater overall high passage number of the sample prior to engraftment.
The authors also examined the gene families that differed between tumors grown in different hosts. Tumors from the XactMice expressed genes associated with the immune system, extracellular matrix, or epithelial-mesenchymal transition, but tumors from the non-humanized hosts did not. RNA sequencing data also identified increased expression of several genes associated with lymphangiogenesis. Indeed, subsequent immunohistochemical evaluation of the engrafted tumors for the lymphatic vascular marker LYVE-1 revealed that lymphatic vessel density was higher in XactMice-engrafted tumors compared to tumors harvested from the NSG™ hosts. Taken together, these data provide evidence that human bone marrow-derived cells infiltrate tumors and the surrounding stroma and strongly influence the gene expression profile of those tumors. Clearly, humanized, tumor-bearing mice are an important and relevant platform for studying human tumor biology, and moving forward, will be an important tool for preclinical efficacy testing.