Differences between CAFs and their paired NCF from adjacent colonic mucosa reveal functional heterogeneity of CAFs, providing prognostic information

Abstract

Little is known about the difference in gene expression between carcinoma-associated fibroblasts (CAFs) and paired normal colonic fibroblasts (NCFs) in colorectal cancer. Paired CAFs and NCFs were isolated from eight primary human colorectal carcinoma specimens. In culture conditions, soluble factors secreted by CAFs in the conditioned media increased clonogenicity and migration of epithelial cancer cells lines to a greater extent than did NCF. In vivo, CAFs were more competent as tumour growth enhancers than paired NCFs when co-inoculated with colorectal cell lines. Gene expression analysis of microarrays of CAF and paired NCF populations enabled us to identify 108 deregulated genes (38 upregulated and 70 downregulated genes). Most of those genes are fibroblast-specific. This has been validated in silico in dataset GSE39396 and by qPCR in selected genes. GSEA analysis revealed a differential transcriptomic profile of CAFs, mainly involving the Wnt signallingsignalling pathway, focal adhesion and cell cycle. Both deregulated genes and biological processes involved depicted a considerable degree of overlap with deregulated genes reported in breast, lung, oesophagus and prostate CAFs. These observations suggest that similar transcriptomic programs may be active in the transition from normal fibroblast in adjacent tissues to CAFs, independently of their anatomic demarcation. Additionally NCF already depicted an activated pattern associated with inflammation. The deregulated genes signature score seemed to correlate with CAF tumour promoter abilities in vitro, suggesting a high degree of heterogeneity between CAFs, and it has also prognostic value in two independent datasets. Further characterization of the roles these biomarkers play in cancer will reveal how CAFs provide cancer cells with a suitable microenvironment and may help in the development of new therapeutic targets for cancer treatment.

Keywords: Biomarker; Carcinoma-associated fibroblasts; Colorectal; Microenvironment; Stroma.

Figure 1

Characterization of fibroblasts isolated from human colon cancer tissue. A) Cultured CAFs and their normal NCF counterparts visualized under the microscope, showing their characteristic cell morphology. B) Relative expression of specific mesenchymal markers such as vimentin and α‐SMA in the 8 paired NCF/CAF cell lines (samples included in the array analysis) and in four more pairs NCF/CAF. None of them showed a specific pattern between the two types of fibroblasts, indicating that although they were good mesenchymal biomarkers, they did not discard between NCFs and CAFs. Low levels of the epithelial‐specific marker cytokeratin 18 confirmed the purity of the fibroblast cultures. C) Protein levels of additional specific fibroblastic markers were tested by western blot, confirming the purity of the fibroblast cultures (samples are ordered in pairs). D) Immunohistochemical staining of α‐SMA in a specimen from a colorectal cancer patient, illustrating staining of myofibroblasts in both compartments, normal adjacent mucosa and adenocarcinoma. All the fibroblasts surrounding the non‐tumoral crypts are positive for α‐SMA.

Figure 2

Functional assays of CAFs vs. NCFs. A) In a wound‐healing assay, DLD1 cells significantly increased their migratory potential when cultured with CAF conditioned medium (CM), compared with their normal paired CM (left and right panels). B) The clonogenic capacity of four cell lines (DLD1, SW620, SW480 and SW116) was greater in the presence of 6 CMs from CAFs compared with 6 CMs from NCFs (two replicates per CM). C) Co‐injection of CAFs with DLD1 cancer cells into athymic mice significantly enhanced tumour growth compared with NCFs (P < 0.0001). D) Hematoxylin‐eosin and immunohistochemical staining. 21 days after coinjection, human specific vimentin were detected in almost all slides of tumours generated with DLD1 cells + CAFs. No staining was detected in tumours DLD1 cells + NCF. In addition, DLD1 cells in tumours generated with CAFs displayed high levels of Ki67 staining at the periphery of the tumours, whereas the staining was very mild in case of coinjection with NCF.

Figure 3

Unsupervised heatmap of deregulated genes in CAF vs. NCF. A) Genes differentially expressed in 8 paired CAFs vs. NCFs. A total of 108 differentially expressed genes, 70 downregulated (yellow side bar) and 38 upregulated (green side bar), were identified by SAM analysis. Fibroblasts were clustered independently of their patient origin. B) Expression of some upregulated genes in CAFs was tested and confirmed at the protein level by western blot, except for the pair highlighted with the red dashed line in the densitometry analysis (CDH2, SLC7A2 and TGFB2). C) Quantitative RT‐PCR validation of 11 of the differentially expressed genes in 10 healthy colonic mucosas (blue bars), 10 adjacent normal mucosas (red bars) and 10 paired tumours (green bars). Expression values are expressed as z‐scores obtained from the relative expression values normalised for two housekeeping genes (PMM1 and ACTB).

Figure 4

CAF specificity of deregulated genes. A‐B) Unsupervised heatmaps of deregulated genes (overexpressed and underexpressed, respectively) in the GSE39396 dataset. The expression profile of 108 deregulated genes was checked in 4 CRC cell populations: FAPα+ (fibroblasts), EPCAM+ (epithelial cells), CD45+ (leukocytes) and CD31+ (endothelial cells). Gene expression levels clearly classified cells into the correct class. Of the complete list of genes, almost half are specific to the fibroblast cell type population (green box). A set of stromal genes heterogeneously expressed in all cell types but not in epithelia are depicted in black as a mixed group of stromal genes. A group of genes is mostly expressed in endothelial cells (blue box) but not in the other cell types. C) Expression levels of selected deregulated genes was validated by qRT‐PCR in paired (n = 12), as well as in a set of 13 colorectal cancer cells lines. Results confirmed the microarray data and demonstrated stromal specificity in the vast majority of selected genes. D) We expanded PCR validation to some genes with a >2‐fold change in expression but with a false discovery rate q‐value < 0.1.

Figure 5

Signature score. A) NCFs and CAFs used in the microarray were plotted according to their signature score (expression level of all 108 deregulated genes). Expression levels of those genes clearly separated the two fibroblast populations even though they came from the same patient. Fibroblasts used in subsequent functional assays are highlighted with a blue “L” (low score), and a red “H” (high score). B) Migration rates of CAFs were proportional to signature scores, the high score CAFs being the fastest to close the wound. C) CAFs with higher 108‐signature scores had a higher proliferation rate than those with lower expression levels. D) When conditioned media from the same fibroblasts were added to a colorectal cancer cell line (DLD1), significant differences were observed between CM from a high score CAF vs CM from low score CAF (Mann–Whitney test, p = 0.034). E) Prognostic value of the differentially expressed genes in a pooled cohort (GSE14333‐GSE17537) of stages I‐to‐III colorectal patients. Left and middle panels show Kaplan–Meier plots for disease‐free survival, by the two risk groups (left panel; using the mean value as a cut off) or stratifying into three risk groups, using a smooth function of risk of recurrence based on the signature score (middle panel). The smooth function correlates the gene signature score with the relative risk of recurrence. Red dashed lines: 95% confidence interval (CI). Grey dashed lines: thresholds for patient selection into groups depicted in the middle panel: low (L; blue, 40 patients), medium (M; gray, 187 patients), and high (H; red, 40 patients); p values and increases in HR per standard deviation increase in expression (+1 SD) are shown (right panel). F) The prognostic value was corroborated using 90 colorectal patients stage II (GSE33113). G) Quantitative RT‐PCR validation of some differentially expressed genes (four upregulated and four downregulated in CAFs vs NCFs) in 30 non‐recurrent stage II/III colorectal tumours and 30 recurrent stage II/III colorectal tumours maintaining stages proportionality between the two groups.