In 1969, one of the most important discoveries in the history of cancer research was made: athymic nude nu/nu mice could support the growth of human cancer cells (1). Until this discovery, there was no systematic experimental animal system to study human cancer. The nude mouse was subsequently demonstrated to enable growth of all types of human cancer cell lines as well as patient tumors (2). The initial nude-mouse cancer models involved subcutaneous implantation, in which metastasis rarely occurred (2). Sordat et al. (3) established the first orthotopic nude mouse model by injection of colon cancer cells into the colon of nude mice, which demonstrated cancer-cell invasion, but no metastasis.

Our laboratory began to develop orthotopic nude mouse models in the late 1980’s (4). In 1990, a breakthrough occurred when Dr. Sun Lee, the pioneer of experimental microsurgery (5), visited the laboratory at AntiCancer Inc. to demonstrate microsurgery on mice. This technique involved implanting tumor fragments, either from clinical sources or grown subcutaneously in nude mice, into the corresponding organs of other nude mice. This method became known as surgical orthotopic implantation (SOI) and led to clinical transplantation of tumors to the liver, kidney, lung, colon, breast and other organs.

Very soon, surprising results were obtained where a bladder cancer cell line was shown to be highly metastatic when transplanted by SOI in the bladder, compared to when cell suspensions were injected into the bladder (6,7). Similar results of high rates of metastasis were observed with SOI nude mouse models of the lung (8), breast (9), ovary (10), pancreas (11), colon (12), and prostate (13).

Another breakthrough occurred in the AntiCancer Laboratory when it was discovered that cancer cells expressing the jelly fish green fluorescent protein (GFP) could be imaged in vivo, even non-invasively (14,15).

At this time in 2001, one of us (NY) arrived at AntiCancer to develop imageable SOI metastatic models using GFP. The initial results and those of subsequent scientists from Kanazawa University visiting AntiCancer, are summarized below. These findings led to the development of new disruptive and revolutionary technology for the study of metastatic cancer.

Mixtures of GFP- and red fluorescent protein (RFP)-expressing HT1080 fibrosarcoma cells were injected into the tail vein (experimental metastasis) or foot pad (spontaneous metastasis) of nude or SCID mice. Colonies that developed in the lung that were pure GFP or pure RFP, were considered clonal. Colonies of mixed GFP and RFP fluorescence were considered not clonal. Rare spontaneous lung metastasis that arose after foot-pad injection were mostly single color, thereby clonal. The frequent lung colonies arising after tail-vein injection of cell mixtures were predominantly of mixed color, not clonal. These results suggested that rare lung metastases, which are the usual clinical pattern, are clonal (Figure 1A-C) (16,17).

Similar to the experiments described above, mixtures of GFP and RFP cancer cells were injected into the tail vein or foot pad to develop experimental or spontaneous metastasis, respectively, but in this case to image metastatic progression in real time. Imaging was performed by visualizing the colonies growing in the lung via a skin flap over the transparent chest wall. Pure and mixed colonies growing in the lung could be seen growing in real time, indicating clonal and non-clonal metastasis, respectively (Figure 2A and B, Figure 3A-F) (17-19).

After common-carotid-artery injection in nude mice, cancer cells expressing GFP in the nucleus and RFP in the cytoplasm were visualized through the transparent skull with a skin flap in the scalp, in micro vessels in the brain trafficking in single-file. Both the nuclei and cytoplasm appeared deformed, and the cancers cells became elongated in order to traffic in the very narrow micro vessels of the brain. The dual-color cancer cells were so bright, that a mitotic cell could be imaged non-invasively in the ear of the mouse with the nuclei and cytoplasm of the 2 cells clearly distinguishable (Figure 4A-C) (20).

With cancer cells labeled with GFP in the nucleus and RFP in the cytoplasm, it was possible, after injection of the cells into the heart, to visualize the dynamics of nuclear and cytoplasmic deformation, as well as the velocity of cancers cells trafficking in various-sized vessels in the inside surface of an abdominal skin flap. Before cancer-cell injection, the epigastrica cranialis vein in the skin flap was closed with a 6-0 suture.

The cytoplasm and nuclei of the cancer cells elongated to fit within the capillary they occupied. The cytoplasm could deform extensively, up to four times its normal length, and could fracture if the deformation was too extensive. The nuclei were more rigid and much less deformable than the cytoplasm and could extend to only 1.6 times their normal length, in narrow capillaries. In capillaries of approximately 8 μm in diameter, cancer cells could migrate 48.3 μm/h. The minimum capillary diameter that allowed cell migration was approximately 8 μm (Figure 5A-G) (21).

Cancer cells expressing GFP in the nucleus and RFP in the cytoplasm were injected through a vascular route in an abdominal skin flap and imaged in real time. The cancer cells migrated by various means in the vessels or adhered to the vessel walls and some adhered to the vessel outer surface surface after extravasation. During extravasation the cancer cells exited the vessels with the nuclei following a cytoplasmic projection that perforated the blood vessel to begin the exit process from the vessel. Both cytoplasm and nuclei underwent deformation during extravasation. Extravasation was rare for HT1080-GFP-RFP-cells but frequent for the mouse mammalian tumor cell line MMT-GFP-RFP. The trafficking cancer cells were observed to aggregate, and aggregates could collide and adhere to each other in large vessels. Cancer cells migrated at 24 μm/second in large vessels.

Post extravasation, Lewis lung cancer (LLC) -GFP-RFP cells remained closely associated with the outer vessel wall and migrated and surrounded the vessel to occupy as much of the vessel surface as possible by elongating their cytoplasm and nuclei (Figure 6A-K) (22).

HT1080 fibrosarcoma cells expressing GFP and RFP died rapidly after injection into the epigastric cranialis vein in an abdominal skin flap made in nude mice. However, if 24 h before cell injection, the mice were pre-treated with the cancer-chemotherapy drug cyclophosphamide, the HT-1080-GFP-RFP cells grew extensively within blood vessels of the cyclophosphamide pre-treated mice and extravasated at high frequency and formed large extravasated colonies. It was speculated that a host-based cancer-cell killing process was inhibited by cyclophosphamide (Figure 7A-L) (23).

HT-1080-GFP-RFP cells were seeded on the surface of a skin flap of nude mice, and the cells were imaged 24 h after seeding. The nucleoli of the dual color cells could be imaged very clearly on the surface of the skin flap. The nucleoli were not labeled with GFP or RFP and could be clearly visualized by contrast to the GFP expression in the nucleus. When the cancer cells were in mitosis, the chromatin was highly condensed, and the nucleoli could not be visualized. After mitosis, in early G1, 4 small nucleoli were visualized in each nucleus. During the subsequent phases of the cell cycle, the nucleoli increased in size, but decreased in number. The nucleoli are a potential marker of the cell-cycle phase of cancer cells, which is important since cytotoxic chemotherapy drugs depend on the S-phase of the cell cycle for efficacy (Figure 8A-F, A’-F’) (24).

Currently “immuno-oncology” is a very active area of clinical research for the treatment of cancer. Previously, we developed a relevant imageable mouse model to study the efficacy of immune-oncology agents. The model used HT1080-GFP-RFP human fibrosarcoma cells and transgenic mice expressing GFP in all immune-cell types. HT-1080-GFP-RFP cells were sprinkled on a skin flap of a GFP-transgenic mouse. After 24 h, GFP-expressing macrophages could be imaged contacting the HT1080-GFP-RFP cells and phagocytizing them. Macrophages could be imaged digesting the HT1080-GFP-RFP cells because RFP cytoplasmic fragments could be visualized within the GFP macrophages. T-cells could kill the dual-color cancer cells by fragmenting them. GFP expressing nuclei of HT1080-GFP-RFP cells which were confronted with T-cells were fragmented and appeared to be at an early stage of apoptosis. This model can be used to determine whether immuno-oncology agents, such as immune check-point inhibitors, could enhance caner-cell killing (Figure 9A-J) (25).

Lymphatic trafficking is a common route of metastatic spread of HT1080-GFP-RFP fibrosarcoma cells. In one experiment HT1080-GFP-RFP cells were injected into nude mice in the inguinal lymph node in a skin flap, along with FITC dextran to label lymphatic channels, enabling the imaging of labeled cancer cells in lymphatic channels. The cancer cells could be imaged later in the axillary lymph nodes, also labeled with FITC dextran. Spontaneous lymphatic trafficking of cancer cells was also observed, after HT-1080-GFP-RFP cells were injected into the footpad of mice from where they subsequently metastasized to the popliteal lymph node. When the metastatic popliteal lymph node was exposed, dual-color cancer cell trafficking in lymphatic channels connected to the metastatic lymph node could be imaged. When 25 g or 100 g weights were applied to a footpad tumor formed from HT-1080-RFP-GFP cells, cancer cells were imaged shedding in the lymphatic channel leading to the popliteal lymph node, with the greater weight causing much more cancer-cell shedding than the lesser weight (Figure 10A-N) (26).

In order to image the trafficking of cancer cells to the lung, the chest wall of the nude mouse had to be open for long periods. Thus, assisted breathing was necessary to keep the mice alive. Therefore, a novel retrograde wire-guided endotracheal intubation procedure for mice was developed. An intravenous catheter of 25 mm length was used as an endotracheal tube. A guide wire was inserted into the trachea through a small hole. The endotracheal catheter could then be accurately introduced into the trachea over the guide wire. The intubation tube was connected to an oxygen tube. At this point, the chest wall could be opened. Anesthesia was maintained with isoflurane applied via a vaporizer. A positive end expiratory pressure (PEEP) system was used to regulate lung inflation and deflation. After each imaging session, the chest wall was closed with 6-0 sutures. Imaging sessions were carried out for as long as 8 h and repeated up to 6 times per mouse. Cancer cells expressing GFP in the nuclei and RFP in the cytoplasm were injected into the tail vein, and HT1080-GFP-RFP or MMT-GFP-RFP cells could be imaged arriving to the lung in real time. Surviving cells in the lung could be imaged to form colonies in the lung with multiple imaging sessions over 10 days. Only rare cells among these arriving in the lung after tail vein injection could survive and form colonies in the lung (Figure 11A-Q) (27).

PC-3 human prostate cancer cells were injected into the heart ventricle of nude mice. Subsequently multiple metastasis were observed in the skull, femur, rib, and vertebral bones. Time-course imaging via skin flaps demonstrated the formation of metastatic colonies in the bones. Twenty min after cardiac injection of PC-3 GFP cells in RFP transgenic mice, the PC-3-GFP cells could be imaged at the single-cell level in the bone marrow of the skull.

After intra-tibial injection of PC-3 cells, RFP-expressing PC-3 cells were imaged growing in the bone marrow and spreading to an inguinal lymph node. This model can thus be used to image the lethal aspect of prostate cancer, metastasis to the bone (Figure 12A-C) (28).

Exquisite application of experimental microsurgery and fluorescent genetic reporters enabled the establishment of unique mouse models of cancer, that along with appropriate imaging technologies, could be used to visualize all the critical aspects of cancer progression, in particular the mechanism of metastatic spread at the single-cell or sub-cellular level that distinguished the cellular cytoplasm and nuclei in vivo.

Imaging of the cancer cells was enabled by their strong expression of GFP and RFP, including the capability of cancer cells to deform and traffic in narrow vessels, extravasate, and subsequently form colonies in distant organs, all of which are the most critical steps of metastases.

Real-time in vivo imaging demonstrated the interaction of cancer cells and immune cells which can be a model of the mode of action of immune checkpoint inhibitors (ICI).

The models described in this brief perspective represent a revolutionary and disruptive improvement over previous cancer models. The unique collaboration of young physician-scientists from Kanazawa University, AntiCancer Inc., and UCSD made this possible.

The Authors declare no competing interests.

SM, NY, and RMH wrote the paper. KY, KH, HK, SM, KI, TH, HT, and SD critically read and approved the final manuscript.

This paper is dedicated to the memory of A. R. Moossa, MD, Sun Lee, MD, Professor Gordon H. Sato, Professor Li Jiaxi, Masaki Kitajima, MD, Shigeo Yagi, PhD, Jack Geller, MD, Joseph R. Bertino, MD, J.A.R. Mead PhD, Eugene P. Frenkel, MD, Professor Sheldon Penman, Professor John R. Raper and Joseph Leighton, MD.