A Preclinical Murine Model for the Detection of Circulating Human Tumor Cells

Results and Discussion

We attempted to develop an in vivo murine model for studying human CTCs by taking advantage of previously described rodent models, including the orthotopic injection of highly metastatic human cancer cells into immunodeficient mice and the renal subcapsular injection of syngeneic cancer cells (12, 15). We first performed renal capsule implantation of GFP-labeled H226Br cells, a cell line that originated from a brain metastasis of primary tumor from which H226B cells were derived (19). At 1.5 months after cell implantation, H226Br cells exhibited very aggressive growth in the renal subcapsule of nude mice (Figure 1A). All primary tumors disfigured the abdomen, obliterating most of the renal parenchyma, and a large portion of the intestine adhered to the primary tumors. To confirm whether these tumors were indeed derived from the implanted human cells in the kidney, we analyzed the expression levels of human actin and GFP in the kidney tissues and tumors from the mice. The expression levels of human actin and GFP, indicators of the implanted human-derived cells, were remarkably higher in the cancer cell-injected kidney (Figure 1B) and tumor (Figure 1C) tissues than in normal kidney.

Close examination of blood samples revealed green fluorescence in the blood of tumor-bearing mice (Figure 1D, left). RT-PCR and real-time RT-PCR analyses of the cells collected from the blood further revealed the expression of GFP and human actin, confirming the presence of disseminated human CTCs in the blood of the tumor-bearing mice (Figure 1D, middle and right). Additionally, the expression of GFP and an epithelial cell surface marker, EpCAM (20), in the lungs of tumor-bearing mice further revealed lung metastases of the kidney-implanted cells (Figure 1E). These results suggest that NSCLC cells implanted under the renal capsule can appear as CTCs in the blood and induce metastatic tumor development in mice.

We next determined whether this animal model can be applied to the evaluation of the metastatic potential of cancer cells with anticancer drug resistance. To this end, we generated a subline of HCT116 cells with acquired resistance to 5-FU (HCT116/R) by the continuous exposure to increasing concentrations of 5-FU for more than six months. We confirmed that HCT116/R cells were significantly less sensitive to 5-FU treatment compared with their parental line (IC₅₀ of 5-FU: HCT116=7 μM; HCT116/R=50 μM). After labeling with GFP, we implanted these cells under the renal capsule of nude mice. We monitored the establishment of primary tumors in the mice 1.5 months after implantation using an IVIS system. As shown by image analysis of two representative animals from each group, we detected fluorescence only in the mice bearing HCT116/R tumors (Figure 2A), suggesting the HCT116/R subline grew more rapidly than the parental cells in vivo in the renal capsule. Similarly, the primary renal tumors established by the H226Br cells, most of the renal parenchyma was demolished by the primary tumors, which exhibited aggressive local invasion (Figure 2B, left). We also observed polygonal, odd nuclei in the HCT116/R cells and extensive blood vessels in the primary tumors (Figure 2B, right).

We next monitored the presence of CTCs by conducting a time-course study. Among five mice bearing HCT116/R cells, only one mouse had CTCs at three weeks after implantation, but the number of mice possessing CTCs gradually increased at four weeks (2 mice/total 5 mice) and five weeks (5 mice/total 5 mice) (Figure 2C). The increases in the CTCs were well-correlated with growth of primary tumors in the renal capsule (Figure 2D). We then assessed metastatic tumor formation in these mice using an IVIS system. At 1.5 months after implantation, metastatic tumor formation in the distal small intestine was observed only in mice with HCT116/R implantation (Figure 2E). At two months after implantation, metastatic tumor development in the liver, right kidney, intestine, and colon were observed (Figure 2F, top). Gross evaluation of the mice revealed obvious tumor nodules in the intestine and colon (Figure 2F, bottom). These findings suggested a positive correlation between the presence of CTCs and the appearance of metastatic tumor. We were not able to detect fluorescent signals in the lung. However, EpCAM-positive cell masses were found in the lung sections obtained from mice with HCT116/R cell implantation (Figure 2G, white arrow), implying the formation of lung metastases in HCT116/R-implanted mice. This might be due to a relatively weak signal in the lung caused by a small proportion of metastatic tumors. We observed that two over six and four over six of mice had tumor nodules in the colon and lung, respectively (Table II). Taken together, these results suggest that an increase in the number of CTCs is an indicative biomarker for metastatic tumors derived from colon cancer cells with acquired resistance to 5-FU.

In summary, we have successfully demonstrated that i) the renal capsule model is applicable to detecting CTCs and primary and metastatic tumor formation derived from different human cancer cell types and that ii) resistance to anticancer drugs may result in increased metastatic potential in cancer cells, contributing to an increase in CTCs and metastatic tumor formation. Although this model requires for the use of immunodeficient mice and has thus limited relevance to host-tumor interaction during metastasis, this model might be useful for various preclinical studies, including detection and capture of CTCs and their genetic, epigenetic, and biochemical characterization. The expression profiling of CTCs would further provide with useful information about biomarkers involved in the highly selective metastatic process and mechanisms of anticancer drug resistance. Knowledge of the distinct pattern of metastatic tumor formation, depending on the implanted cells, may also aid different aspects of experimental research, such as the investigation of the target organs of specific cancer types in metastatic disease.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Establishment of a murine model using H226Br cells. Green fluorescence protein (GFP)-labeled H226Br cells were implanted under the capsule of the left kidney in BALB/c nude mice. At 1.5 months after inoculation, the formation of primary tumors in the kidney was determined (A). The expression levels of human actin (hActin) and GFP in the kidney tissues that had been injected with cancer cells (B) and tumors surgically removed from the kidney tissues (C) of H226Br-implanted nude mice. The mRNA expression levels of these genes were evaluated by real-time polymerase chain reaction (PCR) (B, C) and reverse transcription-polymerase chain reaction (RT-PCR) (C). The presence of circulating tumor cells (CTCs) and metastatic tumor nodules in mice with H226Br cell implantation as determined by detection under a fluorescence microscope (D, left). RT-PCR (middle) and real-time PCR (right) analyses of the expression levels of GFP, hActin, and mouse glycerahdeyde 3-phosphate dehydrogenase (GAPDH in the blood of H226Br-implanted nude mice. Real-time PCR analysis of the expression levels of GFP and mGAPDH in the lung of H226Br-implanted nude mice (E, left). Secondary lung metastasis in H226Br-implanted mice (E, right). The expression of epithelial cell adhesion molecule (EpCAM) and GFP was observed under a fluorescence microscope after immunofluorescence staining.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Generation of primary tumors, secondary metastasis, and the presence of circulating tumor cells (CTCs) in mice with HCT116/R cells implanted in the renal subcapsule. Parental HCT116 and 5-fluorouracil (5-FU)-resistant HCT116/R cells were implanted under the capsule of the left kidney in BALB/c nude mice. A: After 1.5 months, the formation of primary tumors was evaluated using an IVIS system. B: Formation of primary tumors in the kidney and local invasion were photographed (left). Photomicrograph of the hematoxylin&eosin (H&E)-stained primary tumors (right). C: The time-dependent increase in the number of CTCs in the blood of mice implanted with HCT116/R cells. The GFP-positive cells in blood were counted. D: Primary tumor formation at each time point (3 to 5 weeks). At 1.5 (E) or 2 (F) months after inoculation, metastatic tumor formation was detected using an IVIS system. G: Metastatic tumor formation in the lung of HCT116/R-implanted nude mice. Epithelial cell adhesion molecule (EpCAM)-positive cell masses (indicated by a white arrow) were observed under a fluorescence microscope. H&E staining further confirmed secondary tumor formation in the lung.