Immune-deficient mice engrafted with primary human tumors or cell lines have long been the traditional preclinical models for evaluating candidate cancer drugs. This strategy, while useful for understanding some key aspects of tumor cell behavior, has had only limited success in predicting the efficacy of promising therapeutic candidates clinically. One reason for the limited success of these models may be the animals’ immune deficiency because the immune system likely influences both cancer disease development and response(s) to treatment; therefore, efficacy studies using immune-competent models or models with engrafted human immune cells may have better predictive value. Bioluminescent or fluorescent proteins, like firefly luciferase (ffLuc) and enhanced green fluorescent protein (EGFP) often are used to label tumor cells to monitor tumor growth, metastases, and treatment responses in live mice and in real time. Because these proteins are foreign to mammals, however, they can elicit immune responses that may cause inconsistent reporter expression, suppress tumor growth, limit metastasis, and lead to therapeutic response misinterpretations. In late 2014, an investigative team lead by Dr. Glen Merlino, National Cancer Institute, developed genetically-engineered mouse strains that are immune-tolerant to ffLuc and EGFP. These “Glowing Head” mice eliminate allogeneic immune responses in immune-competent mice engrafted with labeled, syngeneic tumors and may provide an important research tool for evaluating candidate cancer therapeutics (Day et al. 2014).
Reporter activity of ffLUC-EGFP labeled tumors in Glowing Head mice
In their 2014 study, the Merlino team labeled Lewis Lung Carcinoma cells (LLC, C57BL/6 background), Mtv-1 breast cancer cells (FVB/N background), and two mouse melanoma cell lines (one derived from a B6X129 hybrid and the other from FVB/N) with an ffLuc-EGFP reporter-expressing lentivirus ex vivo. In vivo bioluminescent imaging of all four tumor cell lines following engraftment and passaging in immunocompetent wild type (WT) syngeneic mice revealed variable ffLuc-EGFP expression between mice and/or passages. This inconsistency was observed regardless of the tumor type, genetic background, or transplantation site (subcutaneous or orthotopic (Mtv-1 cells)). These results suggest that ffLuc-EGFP immunogenicity accounts for the inconsistent expression.
To eliminate this problem, the Merlino group genetically-engineered mice to express a ffLUC-EGFP fusion protein in the anterior pituitary gland. They named their new strain “Glowing Head” (GH) mice due to the reporter’s expression site. The anterior pituitary was chosen as the tissue in which to express the bioluminescent/fluorescent fusion protein because tumors rarely metastasize to this tissue. Further, choosing founders that displayed low but consistent bioluminescent transgene expression minimized instances of interference between transgene-derived bioluminescence and that derived from labeled cells in both primary and metastatic tumors. Following engraftment of ffLUC-EGFP-labeled LLC tumors into GH, WT, and immune-deficient NOD-scid mice, no anti-GFP antibodies were detected in the sera of GH mice or NOD-scid mice, which can’t make antibodies. In contrast, sera from engrafted WT mice had high levels of anti-GFP antibodies. These results demonstrate that ffLUC-EGFP fusion protein is recognized as “self” in GH mice and can tolerize the host to those foreign proteins.
ffLUC-EGFP-expressing tumor cells show improved growth and metastases in GH Mice
ffLUC-EGFP-labeled metastatic Mtv-1 breast cancer cells engrafted orthotopically into mammary fat pads of syngeneic GH mice demonstrated a strong increase in bioluminescence signal over time. In contrast, signal diminished from labeled Mtv-1 cells engrafted into syngeneic WT mice. Following primary tumor resection, bioluminescence imaging of engrafted GH host mice showed that metastases were present and growing within a few days following surgery whereas bioluminescence imaging detected metastases in only a small percentage of the WT host mice, and then only after 2 months post-resection. Metastases developed in multiple sites in the engrafted and resected GH mice but only in the lungs of the WT mice. Wild-type-engrafted mice also survived for significantly longer than engrafted GH hosts. Syngeneic GH and WT mice engrafted with ffLUC-EGFP cells subcutaneously showed differences similar to those observed in the orthotopically-engrafted mice. Together, these data clearly demonstrate that expression of ffLUC-EGFP reporters compromises the growth and metastasis of labelled tumor cells in immunocompetent mice, limiting these models’ utility as a preclinical cancer model. Moreover, the data show that GH mice represent a more appropriate host for such studies.
Immune responses to reporter expression influence therapeutic outcomes
Not only can a tumor-expressed bioluminescent tag affect the tumor’s growth and metastases, it also can impact the interpretation of therapeutic testing data. The Merlino team next evaluated the therapeutic outcomes of labeled or unlabeled cells engrafted into syngeneic WT or GH mice. Treating unlabeled LLC tumor-engrafted, syngeneic WT mice with the chemotherapeutic paclitaxel did not significantly affect tumor growth compared to treatment with vehicle. Treating ffLUC-EGFP labeled LLC tumors engrafted into syngeneic GH mice with paclitaxel also did not affect tumor growth. The growth of ffLUC-EGFP-labeled LLC tumors, however, was significantly delayed in paclitaxel-treated syngeneic WT mice as compared to vehicle treated WT mice. The spleens of the paclitaxel-treated WT mice were substantially larger than those in the other groups, had enlarged and disrupted lymphatic follicles, and had an increased CD4/CD8 splenocyte ratio. These results illustrate a potential false-positive therapeutic response to paclitaxel resulting not from the drug’s efficacy, but to an immunogenic response in the mice to the ffLUC-EGFP reporter.
The Merlino team next labeled melanoma cells derived from a HGF/CDK4R24C transgenic mouse with ffLUC-EGFP and engrafted them into syngeneic GH and WT mice. Treatment with crizotinib, which targets the HGF receptor, did not affect tumor growth in the GH mice and slightly delayed tumor growth in WT mice at the higher dosage. Crizotinib treatment, however, dramatically reduced the number of metastases to the lungs in a dose dependent fashion in the GH- but not the WT-engrafted mice. Decreased inflammation and tumor invasiveness at the engraftment site was also observed in engrafted and treated GH mice, but not in engrafted and treated WT mice. These results illustrate a false negative therapeutic response in the non-tolerized WT mice.
Collectively, these data demonstrate that the immunocompetent Glowing Head mice are a superior host for syngeneic engraftment using ffLUC and/or EGFP labeled tumor cells. They eliminate the interference with tumor growth and metastasis and confounding therapeutic outcomes that occur when tumor cells expressing allogeneic markers are engrafted into immunocompetent WT mice. Glowing Head mice are available from The Jackson Laboratory on C57BL/6 (027662), BALB/c (027848), and FVB/N (027665) genetic backgrounds.