A new aggressive xenograft model of human colon cancer using cancer-associated fibroblasts

Abstract

Background: Colorectal cancer is the second leading cause of cancer death. Almost half of the patients present recurrence within 5 years after the treatment of the primary tumor, the majority, with metastasis. On the other hand, in the search for new animal models that simulate metastatic cancer, it has been suggested that fibroblasts immersed in the peritumoral stroma (cancer-associated fibroblasts (CAFs)), play a relevant role in the development of cancer. The objective of this study was to identify an adequate animal model to study metastatic colon cancer and the application of new treatments.

Methods: Human CAFs and normal fibroblasts (NF) for transplant and culture were obtained from surgical fresh samples of patients with adenocarcinoma of sigmoid colon. Stromal cell purity was evaluated by morphology and immunostaining with vimentin (VIM) as a fibroblast marker and anti-proColXIα1 as a specific human CAF marker. Phenotypic characterization of cultured stromal cells was performed by co-staining with mesenchymal and epithelial cell markers. For identification in mice, human CAFs were labeled with the PKH26 red fluorescence dye. Cell line HT-29 was used as tumor cells. Transplant in the head of the pancreas of 34 SCID mice was performed in four different groups, as follows: I. 150,000 CAFS (n = 12), IIa. 1.5 million HT29 cells (n = 7), IIb. 150,000 NF+1.5 million HT29 cells (n = 5), III. 150,000 CAFS+1.5 million HT29 cells (n = 10). After euthanasia performed one month later, histological analysis was made using hematoxylin-eosin and anti-proColXIα1. A histopathological score system based on three features (tumor volume, desmoplasia and number of metastasized organs) was established to compare the tumor severity.

Results: The CAFs and NF cultured were proColXIα1+/VIM+, proColXIα1/alphaSMA+ and proColXIα1+/CK19+ in different proportions without differences among them, but the CAFs growth curve was significantly larger than that of the NF (p < 0.05). No tumor developed in those animals that only received CAFs. When comparing group II (a + b) vs. group III, both groups showed 100% hepatic metastases. Median hepatic nodules, tumor burden, lung metastases and severity score were bigger in group III vs group II (a + b), although without being significant, except in the case of the median tumor volume, that was significantly higher in group III (154.8 (76.9-563.2) mm³) vs group II (46.7 (3.7-239.6) mm³), p = 0.04. A correlation was observed between the size of the tumor developed in the pancreas and the metastatic tumor burden in the liver and with the severity score.

Conclusion: Our experiments demonstrate that cultured CAFs have a higher growth than NF and that when human CAFs are associated to human tumor cells, larger tumors with liver and lung metastases are generated than if only colon cancer cells with/without NF are transplanted. This emphasizes the importance of the tumor stroma, and especially the CAFs, in the development of cancer.

Keywords: Animal model; CAFs; Fibroblasts culture; ProColXIα1; Xenograft.

Figure 1. Confocal microscopy of cultured CAFs.

Double fluorecence stain illustrates the presence of: (A) Cell proCOL11A1+/VIM+, (B) ProCOL11A1+/CK19+ and (C) ProCOL11A1+/alphaSMA+. Red, proCOL11A1; green, VIM, alphaSMA and CK19; blue, nuclei. Scale bar: (A and B) 20 µm, (C) 100 µm (X630).

Figure 2. Growth curves of fibroblasts.

Blue, normal fibroblasts; green, CAFs. Cultures obtained from three patients with adenocarcinoma of colon. Mean ± 2 SEM of three patients with duplicate determinations.

Figure 3. HT-29 in the pancreas model.

(A) Microphotograph of maximal tumor size in pancreatic tissue (maximal tumor size, score 1) 1.25×, H–E. Red circle, maximal tumor size in pancreatic tissue. (B) Histological image of poorly differentiated pancreatic adenocarcinoma, 10×, H–E. Malignant epithelial cells isolated or arranged in small or large clusters, with mild-moderate desmoplastic stromal reaction (score 2). (C) The image shows metastasis of a small cluster of malignant epithelial cells (score 1 of no. of metastasized organs) in: C, liver tissue; D, lung tissue (H–E, 4×). Red circle (C and D) nest of tumoral cells in liver and lung, respectively. Scale bar: (A) 500 µm; (B and D) 50 µm; (C) 200 µm.

Figure 4. HT29 + NF.

(A) Mouse necropsy. Great pancreatic tumor distention gallbladder (Courvoisier–Terrier signe), gastric infiltration. (B) Microphotograph of maximal tumor size in pancreatic tissue (maximal tumor size, score 2) 1.25×, H–E. (C) The image shows metastasis of small or medium clusters of malignant epithelial cells (score 1 of no. of metastasized organs) in: C, liver tissue; D, lung tissue (H–E, 4×). Scale bar: (B) 500 µm; (C) 200 µm; (D) 50 µm.

Figure 5. HT29 + CAFs.

(A) Mouse necropsy. Great pancreatic tumor with liver metastases, distention gallbladder (Courvoisier–Terrier signe), gastric infiltration. (B) Microphotograph of maximal tumor size in pancreatic tissue (maximal tumor size, score 3) 1.25×, H–E. Red circle, maximal tumor size in pancreatic tissue. (C) Histological image of poorly differentiated pancreatic adenocarcinoma, 10×, H–E. Malignant epithelial cells isolated or arranged in small or large clusters, with necrosis and moderate desmoplastic stromal reaction (score 2). Scale bar: (B) 500 µm; (C) 50 µm.

Figure 6. Relationship between pancreatic tumor and hepatic burden (A) or severity score (B).

Regression ecuations. (A) Pancreatic tumor size-hepatic burden and (B) pancreatic tumor size-severity score.