The role of inflammation and SPARC (secreted protein acidic and rich in cysteine) in colorectal tumourigenesis

Chronic inflammation of the colon is associated with an increased risk of the development of colorectal cancer (CRC). Therefore, pathways that contribute to inflammation modifying tumourigenesis in the colon are of diagnostic and therapeutic interest. This thesis analyses the contribution of inflammation to tumourigenesis in a murine model of CRC in which sporadic tumourigenesis was produced by a colonotropic mutagen, azoxymethane (AOM). Colitis was induced by ingestion of dextran sulphate sodium (DSS) and mice treated with both agents (AOM/DSS) also developed colorectal tumours.

Both sporadic (AOM) and colitis-associated tumours (AOM/DSS) were found to consist of carcinomas-in-situ and in this study these are compared for the first time in detail. The number and size of tumours were studied macroscopically, and the extent of dysplasia and tumour morphology characterised by histology. The number of proliferating cells within tumours was determined by detection of Ki-67 and the number of apoptotic cells by fragmented DNA using TUNEL. The composition of the immune cell populations infiltrating the tumours was characterised by indirect immunofluorescent antibody staining for CD4, CD8a, CD11c, F4/80, and Ly6G. Finally, a comparison of gene expression levels was performed to examine the involvement of various pathways at a global level.

The presence of chronic colitis increased the number of tumours although these were smaller than those developing in the absence of inflammation. Chronic colitis associated tumours contained fewer apoptotic stromal cells and proliferating epithelial cells but increased numbers of CD4⁺ T-cells, macrophages, and dendritic cells, based on CD4⁺, F4/80⁺, and CD11c⁺ staining, than AOM-tumours. There were also fewer CD8⁺ and Ly6G⁺ cells, suggesting a decrease in CD8⁺ T-cells and neutrophils compared to AOM-only tumours. The majority of the immune infiltrate in both the sporadic and colitis associated tumours consisted of CD4⁺ cells and F4/80+ cells, suggesting a relative preponderance of CD4⁺ T-cells and macrophages in both models. Through a global gene expression microarray involving about 18 000 genes, 143 genes were found to be differentially expressed between sporadic and colitis-associated tumours. Of these, the expression levels of Cd164, Csf1r, Cd44, and Mmp10 were tested by a qRT-PCR and Cd164 and Csf1r were verified as significantly increased in AOM/DSS colitis-associated tumours. The products of Cd164 and Csf1r are linked to the phosphatidylinositol 3-kinase (PI3K)-Akt-pathway, suggesting a possible role in colitis-associated tumourigenesis.

The role of secreted protein acidic and rich in cysteine (SPARC) in colitis-associated colorectal tumourigenesis was examined as earlier work by my group had found an association of SPARC with inflammation and CRC tumourigenesis. This is the first study to examine the influence of SPARC in murine models of colitis-associated colorectal tumourigenesis. For these studies, tumours were induced as before with AOM/DSS treatment of SPARC-knockout (KO) mice and compared with those in wild-type (WT) mice.

More tumours developed in the SPARC-KO mice but these were found to be of comparable size and contained similar numbers of proliferating and apoptotic cells compared to those developing in WT mice. However, the presence of SPARC was associated with fewer CD11c⁺ and F4/80⁺ cells in the tumours, suggesting that they contained lower numbers of dendritic cells and macrophages. In contrast, there were increased numbers of T lymphocytes, based on CD4⁺ and CD8a⁺ cells, in the presence of SPARC while no difference in Ly6G⁺ cell number was detected. A global gene expression microarray compared the expression levels of about 18 000 genes between WT and SPARC-KO tumours and 6 genes, in addition to SPARC, were differentially expressed. The expression levels of 4 of these genes, Ido, Ifit2, Areg, and Erdr1, were tested via multiplex qRT-PCR but no significant changes were found. While 2 genes remain to be tested, the lack of differences in gene expression between the SPARC KO and WT tumours suggests that SPARC reduces tumour initiation and alters the intra-tumoural inflammatory population by other mechanisms.

In conclusion, chronic inflammation of the colon increases the number of tumours developing, perhaps through damage to the colon and possibly involving the Pi3K-Akt pathway. Notably, inflammation-associated tumours were smaller which did not appear to be explained by the extent of apoptosis. While they contained fewer neutrophils and CD8⁺ T-cells, there was an increase in the numbers of other inflammatory populations suggesting the possibility of suppression of tumour growth by cytokines produced by immune cells. The presence or absence of the matricellular protein SPARC also impacted on the number of tumours and the composition of the tumour infiltrating cell populations. Interestingly, this occurred without substantive changes in gene expression. The chronic colitis and SPARC-KO models have provided novel opportunities for investigating mechanisms that promote tumour development in the colon.