Bioluminescent imaging and organ-specific metastasis of human cancer cells

Cancer metastasis is the end result of a complex series of biological events that lead to the formation of clinically significant secondary tumors at distant sites. Clinical evidence demonstrates that such sites are not randomly located, and certain malignant tumors show a tendency to develop metastases in specific organs (e.g., brain, liver, and lungs).¹ However, an appropriate animal model to characterize the hematogenous nature (i.e., originating in blood) of transplantable human-cancer-cell lines is not available for metastatic cells, while profiling data of hematogenous metastasis has not attracted sufficient scientific attention to improve understanding.

Recent advances in bioluminescent-imaging (BLI)² technologies have facilitated quantitative analysis of cellular processes in vivo. BLI reporters have significant signal-to-noise ratios in mammalian tissues, and emitted-light signals can be quantified in intact animals using noninvasive assays. To obtain profiling data of the metastatic fate in human transplantable tumor-cell lines, we have been generating a luciferase-expressing human-cancer-cell library (including melanoma, colon, breast, and prostate cancer) since mid-2007. (Luciferase is an enzyme present in the cells of bioluminescent organisms that produces light by catalyzing the oxidation of luciferin and adenosine triphosphate, ATP, a nucleotide found in the mitochondria of all plant and animal cells.) We created these cells using a retroviral gene-transfer technique. In the presence of D-luciferin, as few as 50 luciferase-transduced cells can be detected in vitro against the background's linear-dose-dependent output of light. Although expression levels among cell lines are not always the same, selected cells provide sufficient numbers of photons in vivo for real-time luminescent imaging.

To date, cancer-cell injection into mice has been done easily through the tail vein. However, this route is not beneficial for systemic cell delivery because most of the sizable injected cells are trapped in the lung capillaries. To overcome this and systemically deliver cells via arteries, many scientists have blindly tried injections into the left ventricle (heart chamber). Even veteran technicians have blindly been injecting cells into a very narrow space in the mouse heart. However, high-resolution ultrasonography (US), developed specifically for small-animal imaging, now provides clear identification of areas of interest within the myocardial wall (i.e., of the heart muscle) and allows precise, site-directed cell injection. Therefore, using fine-US guidance, we have successfully realized accurate and reproducible cardiac-cell injection into mice.

We have generated more than 30 luciferase-expressing human-cancer-cell lines using retroviral transduction. We inoculated these cells into the left cardiac ventricle of nonobese diabetic/severe combined-immunodeficiency (NOD/SCID) mice under fine-US guidance. BLI was conducted for each cell line, and representative organs (e.g., brain, liver, lungs, lymph nodes, bones, and gastrointestinal tract) were then inspected ex vivo. We observed cancer-cell-type-dependent metastasis to specific organs even in mice (beyond the species): see Figure 1. For instance, human colon-cancer HT-29 cells accumulated significantly in the liver of mice, while human-melanoma cell lines showed frequent metastasis to brain, lungs, and lymph nodes in the mouse model. For breast-cancer MDA-MB-231 cells, metastasis was observed to the bone in addition to the brain, lungs, and lymph nodes. Notably, reflecting the clinical features of melanoma, breast, and lung cancer, some cell lines showed preferential metastasis to the brain of mice.

Figure 1. Representative metastatic images of human-cancer-cell lines in nonobese diabetic/severe combined-immunodeficiency mice. Luciferase-expressing cells were injected into the left cardiac ventricle under fine-ultrasonography guidance. In and ex vivo bioluminescence imaging was conducted for (left) HT-29 colon and (middle) MDA-MB-231 breast cancer, and (right) Mewo melanoma cells 30–40 days after tumor implantation. (right) Black dots in the brain represent metastasis of Mewo cells.

Characteristics common to both tumor cells and normal stem cells appear to exist, referred to as stemness. The hallmark traits of stem cells—self-renewal and differentiation capacity—are reflected by the high proliferative capacity and phenotypic plasticity of tumor cells.³ Since the initial concept of cancer stem cells in solid tumors was established using NOD/SCID mice, we have had to employ animals to apply luciferase-expressing cell behavior to the theory of cancer stem cells. Our recent BLI-based experimentation suggests that a subpopulation of cancer stem cells is essential for organ-selective cancer metastasis.⁴

Approximately 10–20% of all systemic malignancies will eventually metastasize to the brain.⁵ Despite this high frequency of brain tumors, an accepted approach for effective treatment is still lacking. Accumulating clinical data suggest that the interaction between chemokines (proteins) and their receptors is a critical component for regulation of tumor progression and metastasis in many cancer types,¹ and that the CXCR4/CXCL12 pathway is involved.¹ However, the pathophysiology in brain metastasis is not fully understood because of the difficulty of creating appropriate animal models. Therefore, BLI combined with high-frequency US imaging should allow various preclinical studies at the tumor/normal-brain interface.

Combining cell resources with an appropriate animal model, our goal for the immediate future, will promote a better and more profound understanding of human-cancer-cell biology. Advances in optical imaging should provide a new platform to accelerate development of therapeutic strategies for human cancer.

This study was supported by a Health and Labor Science Research Grant from the Ministry of Health, Labor, and Welfare (Research on Biological Resources) and the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan, and a grant from the ‘Strategic Research Platform’ for Private Universities: Matching Fund Subsidy from MEXT. We thank Masafumi Takahashi for helpful suggestions.