Abstract
The tumor microenvironment (TME) contains stromal cells in a complex interaction with cancer cells. This relationship has become better understood with the use of fluorescent proteins for in vivo imaging, originally developed by our laboratories. Spectrally-distinct fluorescent proteins can used for color-coded imaging of the complex interaction of the tumor microenvironment in the living state using cancer cells expressing a fluorescent protein of one color and host mice expressing another-color fluorescent protein. Cancer cells engineered in vitro to express a fluorescent protein were orthotopically implanted into transgenic mice expressing a fluorescent protein of a different color. Confocal microscopy was then used for color-coded imaging of the TME. Color-coded imaging of the TME has enabled us to discover that stromal cells are necessary for metastasis. Patient-derived orthotopic xenograft (PDOX) tumors were labeled by first passaging them orthotopically through transgenic nude mice expressing either green, red, or cyan fluorescent protein in order to label the stromal cells of the tumor. The colored stromal cells become stably associated with the PDOX tumors through multiple passages in transgenic colored mice or non-colored mice. The fluorescent protein-expressing stromal cells included cancer-associated fibroblasts and tumor-associated macrophages. The cancer cells in PDOX models can also be labeled with a telomerase-dependent adenovirus containing the gene for green fluorescent protein. Using this model, specific cancer-cell or stromal-cell targeting by potential therapeutics can be visualized. Color-coded imaging enabled the visualization of apparent fusion of cancer and stromal cells. Color-coded imaging is a powerful tool visualizing the interaction of cancer and stromal cells during cancer progression and treatment.
- Tumor microenvironment
- red fluorescent protein
- green fluorescent protein
- blue fluorescent protein
- niche
- cancer-associated fibroblasts
- tumor-associated macrophages
- collagen
- liver metastasis
- color-coded imaging
- review
In vivo imaging with fluorescent proteins was pioneered by our laboratories and has been particularly useful for studying tumor growth, invasion, cancer-cell trafficking, metastasis, angiogenesis, and other aspects of tumor progression (1-5). Multicolored proteins have allowed the color-coding of cancer cells growing in vivo such as between highly- and poorly-metastatic cells, cancer stem and non-stem cells, and gene transfer between cancer cells (3, 4, 6-13).
The present review focuses on color-coded imaging of cancer and stromal cells in the tumor microenvironment (TME).
Transgenic nude mice expressing fluorescent proteins for color-code imaging of the TME. The green fluorescent protein (GFP)-expressing athymic nude mouse was obtained by crossing nontransgenic nude mice with the transgenic C57/B6 GFP mouse in which the β-actin promoter drives GFP. In the adult GFP nude mice, most organs brightly expressed GFP, including the heart, lungs, spleen, pancreas, esophagus, stomach, duodenum, and the entire digestive system from the tongue to the anus; the male and female reproductive systems; brain and spinal cord; and the circulatory system, including the heart and major arteries and veins. The bared skeleton also highly expressed GFP (14).
The red fluorescent protein (RFP)-expressing nude mouse was obtained by crossing non-transgenic nude mice with the transgenic C57/B6 mouse in which the β-actin promoter drives DsRed2 including the heart, lungs, spleen, pancreas, esophagus, stomach, duodenum, bone marrow, spleen cells, the male and female reproductive systems; brain and spinal cord; and the circulatory system, including the heart, and major arteries and veins, and skeleton, all highly expressed RFP (15).
The cyan fluorescent protein (CFP)-expressing nude mouse was developed by crossing non-transgenic nude mice with the transgenic CK/ECFP mouse in which the β-actin promoter drives expression of CFP. In the CFP nude mice, the pancreas and reproductive organs had the strongest fluorescence (16).
Imaging the TME. With the use of multiple-colored fluorescent proteins, we developed imaging of the TME by color-coding cancer and stromal cells (3, 7, 15, 17-19). Color-coded fluorescence imaging of the TME is achieved, for example, by using RFP-expressing cancer cells growing in GFP-expressing transgenic mice. The host vasculature, expressing GFP, can be readily distinguished as interacting with the RFP-expressing cancer cells. GFP-expressing dendritic cells were shown to be contacting RFP-expressing cancer cells with their dendrites. GFP-expressing macrophages engulfed RFP-expressing cancer cells and GFP-expressing lymphocytes were visibly attaching to an RFP-expressing tumor, which eventually regressed (7).
Imaging the critical role of stromal cells in metastasis. With the use of color-coded imaging of the TME, we have demonstrated the essential role of tumor-associated host cells in tumor progression and metastasis (20, 21). Color-coded imaging of the TME has enabled important discoveries of its cellular components (22, 23). Using color-coded imaging of the TME, Egeblad et al. observed that regulatory T-lymphocytes migrated to blood vessels. These authors observed that stromal cells had higher motility in the TME at the tumor periphery than within the tumor mass (22).
GFP-expressing spleen cells were found in the liver metastases that resulted from intra-splenic injection of colon cancer cells expressing RFP in the cytoplasm and GFP in the nucleus, in transgenic nude mice ubiquitously expressing GFP. When cancer cells were injected into the portal vein, no liver metastasis resulted. When GFP-expressing spleen cells and the RFP-GFP cancer cells were co-injected in the portal vein, liver metastasis resulted that contained GFP-expressing spleen cells. These results demonstrate the stromal (spleen) cells were indeed necessary for metastasis (20).
Confocal microscopy imaging demonstrated the close interaction of XPA-I-GFP-RFP pancreatic cancer cells and RFP-expressing pancreatic stellate cells in a liver metastasis, in an orthotopic model, suggesting the stellate cells were necessary for metastasis (Figure 1) (24).
Patient-derived orthotopic xenograft (PDOX) mouse models of cancer. Orthotopic implantation of intact tumor tissue leads to metastasis that mimics that seen in patients. PDOX models were pioneered by our laboratories in the early 1990s and recapitulate tumor behavior, including metastasis, in the patient in contrast to subcutaneous patient-derived xenograft models, which behave as benign tumors (25, 26).
PDOX models, which were implanted using intact tumor tissue with the technique of surgical orthotopic implantation (SOI) (25, 27), were established from patients with colonic (28-30), pancreatic (21, 31-41), breast (42), ovarian (43), lung (44), and stomach cancer (45), and mesothelioma (46) in the early 1990s, resulting in primary and metastatic tumor growth very similar to that in the patient.
Recently, PDOX models of sarcoma (47-64), cervical cancer (65-67) as well as melanoma (68-74) have also been developed.
PDOX tumors acquired fluorescent protein-expressing stroma while growing in nude mice. A pancreatic cancer PDOX acquired RFP, GFP, and CFP stroma while growing in transgenic nude mice expressing the corresponding fluorescent protein, as described above (Figure 2). The RFP-, GFP-, and CFP-expressing stromal cells acquired by the tumors included cancer-associated fibroblasts (CAFs), tumor-associated macrophages (TAMs), and blood vessels (Figure 3). The fluorescent protein-expressing tumor-acquired stromal cells persisted throughout at least three passages in other mice (21). These studies demonstrated the persistence of the mouse TME in patient tumors through multiple passages (21).
PDOX metastasis contains stromal cells from the primary tumor. A pancreatic cancer PDOX with GFP-expressing stroma metastasized to the lung, whereby the metastases maintained the GFP-expressing stroma from the primary tumor and apparently acquired stromal cells from the metastatic sites as well, resulting in very bright fluorescence of the metastasis (Figure 4) (16).
Noninvasive imaging of a pancreatic cancer imageable PDOX labeled with GFP and RFP stroma. After passage to RFP and GFP mice, a pancreatic cancer imageable PDOX was passaged to non-transgenic non-colored nude mice, enabling non-invasive imaging at days 21, 30, and 74 (Figure 5). The non-invasive imaging demonstrated extensive orthotopic growth of the pancreatic cancer PDOX (Figure 5) (34).
We subsequently showed that an undifferentiated pleomorphic sarcoma stably acquired imageable RFP-labeled stroma after only a single passage in a transgenic RFP nude mouse that was non-invasively imageable in non-colored mice (52).
A: Green fluorescent protein-red fluorescent protein (GFP-RFP)-expressing dual-color XPA-I human pancreatic cancer cancer cells in vitro. Cancer cells were initially transduced with RFP and the neomycin-resistance gene. The cells were subsequently transduced with histone H2B-GFP and the hygromycin- resistance gene. Double transformants were selected with G418 and hygromycin, and stable clones were established. Images were captured under FV1000 confocal microscopy (28). B: High-magnification imaging of XPA-I pancreatic cancer cells and stellate cells in the liver of cyan fluorescent protein (CFP)-expressing nude mice. A, B: XPA-I-GFP-RFP pancreatic cancer cells in the host liver 3 hours after splenic injection. White arrows indicate XPA-1 pancreatic cancer cells. C, D: Twenty-eight days after co-injection of XPA-I-GFP-RFP pancreatic cancer cells and RFP stellate cells, liver metastases were observed in the host CFP nude mice. White arrows indicate XPA-I-GFP-RFP pancreatic cancer cells. Yellow arrows indicate RFP stellate cells. White arrow in D shows dividing cancer cells. Imaging was performed with a FV1000 confocal microscope (Olympus) (24).
A: Flow diagram of the experimental protocol for imaging the tumor microenvironment. B: Pancreatic cancer patient-derived orthotopic xenograft (PDOX) transplanted to a red fluorescent protein (RFP) transgenic nude mouse. Yellow arrow indicates host RFP nude-mouse pancreas. Blue arrow indicates tumor with infiltrating RFP stroma (Bar=10 mm). Image obtained with Olympus OV100. C: Human pancreatic PDOX excised from RFP nude mouse with RFP stroma. The image is of a cross-section of the tumor. Blue arrow indicates RFP stroma (Bar=10 mm). D: Visualization of RFP tumor-associated macrophages (TAMs) in the human pancreatic cancer patient tumor (F2). High-magnification image. Yellow arrows indicate RFP TAMs (Bar=50 μm). E: Pancreatic cancer PDOX growing in transgenic GFP nude mouse for 30 days. Red arrow indicates host GFP nude mouse pancreas. Blue arrow indicates human pancreatic tumor with RFP stroma (Bar=10 mm). F: Pancreatic cancer PDOX growing in GFP-host model for 56 days. Red arrow indicates host GFP nude mouse pancreas. Blue arrow indicates human pancreatic tumor with RFP and GFP stroma (Bar=10 mm). G: Excised PDOX tumor with RFP and GFP stroma. The image is of a cross-section of the tumor. Yellow arrow indicates RFP stroma. Green arrow indicates GFP stroma (Bar=10 mm). H: PDOX with RFP and GFP stromal cells. Green arrows indicate GFP stromal cells from GFP mouse. Red arrows indicate RFP stromal cells from an RFP mouse. (Bar=50 μm) I: Pancreatic cancer PDOX with RFP stromal cells and GFP TAMs. (Bar=100 μm). J: High magnification image of PDOX shown in (I). RFP stromal cells and GFP-TAMs are readily observed (Bar=30 μm). All images obtained with Olympus FV1000 system (21).
A: Pancreatic cancer patient-derived orthotopic xenograft (PDOX) growing in a cyan fluorescent protein (CFP)-expressing host. White arrow indicates host CFP nude-mouse pancreas. Blue arrow indicates tumor (Bar=10 mm). B: Pancreatic cancer PDOX (blue arrow) with red fluorescent protein (RFP), green fluorescent protein(GFP)-expressing, and CFP-expressing stromal cells. Red arrow indicates CFP pancreas (Bar=10 mm). C: RFP, GFP, and CFP stromal cells were observed. Red arrow indicates RFP stromal cells. Green arrows indicate GFP stromal cells. White arrows indicate CFP stromal cells (Bar=100 μm). D: RFP tumor-associated macrophages (TAMs) (red arrow) and GFP blood vessel (green arrow) were observed in the tumor (Bar=100 μm). E: GFP blood vessels (green arrows) can be seen in the tumor. (Bar=100 μm). F: RFP cancer-associated fibroblasts (CAFs) (yellow arrow) and GFP TAMs (green arrows) present in the tumor. White arrow indicates CFP CAFs. (Bar=30 μm). Images A and B were obtained with an Olympus MVX10 microscope, C-F were obtained with an Olympus FV1000 confocal microscope (21).
A: Upper image shows pancreatic cancer patient-derived orthotopic xenograft (PDOX) liver metastasis. Red arrows indicate liver metastasis with green fluorescent protein (GFP)-expressing stroma. Bar=10 mm. Lower panel shows high-magnification image of liver metastasis. Bar=1 mm. Red arrows indicate GFP stromal cells. B: Image of liver metastasis. Yellow arrows indicate tumor-associated macrophages (TAMs). White arrows indicate cancer-associated fibroblasts Image was obtained with an Olympus FV1000 confocal microscope (Bar=20 μm). C: Liver metastasis stained with hematoxylin and eosin. Blue arrows indicate pancreatic tubular adenocarcinoma. Yellow arrows indicate stromal cells. Red arrows indicate hepatocytes (Bar=100 μm) (33).
Color-coded imaging of TME enables effective fluorescence-guided surgery in a pancreatic cancer PDOX model. The telomerase-dependent GFP-containing adenovirus OBP-401 was used to label the cancer cells of a pancreatic cancer PDOX previously grown in an RFP transgenic mouse, where it acquired bright red stroma. Color-coded fluorescence-guided surgery (FGS) allowed complete resection of the pancreatic tumors, including stroma, preventing local recurrence, which bright-light surgery and single-color fluorescence-guided surgery were unable to do (41).
Color-coded imaging demonstrates recombination between cancer and stromal cells. Subcutaneous EL4 mouse lymphoma tumors were harvested and transplanted to the abdominal cavity of GFP-expressing nude mice. Metastases to the liver, perigastric lymph node, ascites, bone marrow, and primary tumor were imaged. In addition to EL4-RFP cells and GFP-host cells, genetically-recombinant yellow fluorescent cells were observed only in the ascites and bone marrow. These results indicate genetic exchange between the stromal and cancer cells (75).
Non-invasive imaging of a fluorescent tumor originally from a patient with pancreatic cancer growing orthotopically in nude mice. A: Flow diagram of the experimental protocol. B: Whole-body non-invasive imaging of a pancreatic cancer patient-derived orthotopic xenograft (PDOX) in non-transgenic nude mice. Mice were non-invasively imaged at day 21 (upper row), day 30 (middle row) and day 74 (lower row). PDOX tumors were initially grown in green fluorescent protein (GFP) and red fluorescent protein (RFP) -transgenic mice. Green arrows indicate tumor with RFP stromal cells. Red arrows indicate tumor with GFP stromal cells. Images were obtained with the Olympus OV100 Small Animal Imaging System. C: Resected PDOX with RFP and GFP stroma. Images were obtained with an FV1000 confocal microscope. Left panel, RFP-expressing and GFP-expressing cancer-associated fibroblasts (CAFs). Right panel, RFP-expressing blood vessels and GFP-expressing tumor-associated macrophages (TAMs). Yellow arrows indicate RFP-expressing CAFs. Blue arrows indicate GFP-expressing CAFs. White arrows indicate GFP-expressing TAMs. Red arrows indicate RFP-expressing blood vessels (34).
Color-coded imaging to visualize therapeutic targeting of stromal cells. An inhibitor of transforming growth factor-β (TGF-β) was shown to target stroma in an orthotopic mouse model of pancreatic cancer with a color-coded TME. The BxPC-3 human pancreatic adenocarcinoma cell line expressing GFP was used in an orthotopic model in transgenic nude mice ubiquitously expressing RFP. The area of RFP fluorescence from the stromal cells relative to the area of GFP fluorescence of the cancer cells was significantly reduced by a TGF-β inhibitor indicating targeting of the stroma by the inhibitor. Color-coded imaging in an orthotopic pancreatic-cancer cell-line mouse model thus readily enabled detection of the selective targeting of stromal cells by a TGF-β inhibitor (76).
Conclusion
High-resolution color-coded fluorescence imaging is a powerful technique for differentially labeling cancer and stromal cells in the tumor microenvironment (TME), allowing observations of the behavior of each cell type and their interaction within tumors, thereby better defining the specific role of cancer and stromal cells in tumor progression (77).
Footnotes
This article is freely accessible online.
- Received February 1, 2018.
- Revision received February 28, 2018.
- Accepted March 6, 2018.
- Copyright© 2018, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved