Investigating the interactions between tumour cells and the non-tumour cellular community in colorectal cancer

Susanti, Susanti (2021) Investigating the interactions between tumour cells and the non-tumour cellular community in colorectal cancer. PhD thesis, University of Nottingham.

[img] PDF (Final Submission (Revision after Viva)) (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (7MB)


Colorectal cancer (CRC) is one of the leading causes of cancer-related mortality. It is a heterogeneous disease and several different classifications for CRC have been developed over the years. The most recent classification has identified 4 tumour groups which has been called “Consensus Molecular Subtypes (CMS)”. As well as molecular differences in the epithelium, the CMS classification also identifies distinguishing features in the stroma. Thus, tumours belonging to CMS1 have hypermutation in the cancer epithelium and are characterized by a prominent lymphocytic infiltrate. In contrast, tumours belonging to CMS4 have activation of the TGF-b signalling pathway and a dense fibrous stroma. Since CRC arises as a consequence of gene mutation and these mutations only occur in the epithelial cells, it follows that signalling from the epithelial cells may influence the population profile of the cells in the stroma (i.e. the cancer community). This study hypothesized that (i) the cellular community in CRC is non-random and is dictated by signals from the epithelium and (ii) that, in turn, the cellular community can influence tumour kinetics.

To test the relationship between mutations in the cancer cell and the cancer community, it was necessary to build cancer mutation profiles and cancer community profiles of CRC. We aimed initially to classify tumours as either having microsatellite instability (MSI) or being microsatellite stable (MSS). Using PCR (Polymerase Chain Reaction) followed by HRMA (High Resolution Melting Analysis), a new panel of microsatellite markers, consisting of two established markers (BAT25, BAT26) and three novel markers (BCAT25, MYB and EWSR1) was developed and shown to have near perfect concordance with the IHC (immunohistochemistry) for mismatch repair (MMR) proteins and PCR followed by CE (capillary electrophoresis).

The HRMA was also used for mutation detection of other important driver genes including KRAS, BRAF (exon 11) and PIK3CA. The expression pattern of p53 protein was examined using IHC as a surrogate to detect the mutation of TP53 gene. The DNA content of tumour cells (the “ploidy” status) was assessed by flow cytometry.

A cohort of 99 cases of CRC was retrieved from the archives of the Nottingham University Hospitals NHS trust. This was selected to contain 44 cases which had been shown to be Mismatch Repair deficient (dMMR) and 45 cases which had been shown to be Mismatch Repair proficient (pMMR). These were confirmed as showing MSI and MSS respectively by the panel of markers described in this project. The frequencies of the mutations in the MSI and MSS (microsatellite stable) CRC were in concordance with published literatures. CRCs with MSI (MSI-CRCs) were enriched for BRAF mutation (p<0.001) and were all diploid or near-diploid tumours. Conversely, MSS-CRCs were enriched for TP53 mutation (p<0.001) compared to MSI-CRCs and 60% were aneuploid. The presence of diploidy in 40% of MSS-CRCs confirmed that the so-called MACS (microsatellite and chromosomal stable) CRC represented a significant proportion of MSS-CRCs but there was no association between TP53 mutation and aneuploidy.

Within tumours, there was a negative association between KRAS exon 2 and BRAF exon 15 mutations (p<0.05), Fisher’s exact test). Mutations in KRAS exon 2 were positively associated with mutations in PIK3CA exon 9 (P=0.046, Fisher’s exact test), whilst, in contrast, there was a significant negative association of mutations in BRAF codon 600 with mutation in PIK3CA exon 9 (p=0.026, Fisher’s exact test) when tested in larger cohort of CRC (from Nottingham and Edinburgh). As both KRAS and BRAF are part of the MAPK (mitogen activated protein kinase) signalling, these findings suggest that co-selection is gene specific but not pathway specific.

In order to profile the cancer community, a novel approach called fm-IHC (faux multiplex immunohistochemistry) was employed. This method allows both the cellular populations to be identified as well as describing the interactions of the members of the community with each other and with the tumour epithelium. This method relies on IHC to identify cell populations and precise registration of multiple images to then allow virtual segmentation and spatial mapping of the cell populations.

In this study, although few significant features were found indicating active stage of TME of MSI tumour such as higher density of proliferating (Ki67+) CD3+ T-lymphocytes and low-apoptotic (BCL-2+) CD20+ B-lymphocytes within the stroma, the BRAF mutation are associated with stronger immune features in the TME. BRAF mutant tumour had higher density of CD3+ T-lymphocytes and CD20+ B-lymphocytes within the stroma and WTS (whole tissue section) and more proliferating (Ki67+) CD3+ lymphocytes and CD163+ M2-macrophages and BCL2+ (low apoptotic) cellular community, which includes CD3+ and CD8+ T-lymphocytes, CD20+ B-lymphocytes, CD163+ M2- macrophages and CD56+ NK cells in the stroma. BRAF mutant also showed higher correlation between CD8+ cytotoxic T-cell with CD68+ macrophages and CD163+ M2-macrophages in the epithelial.

We also found significance of TP53 mutation in inducing immune activation within the TME of CRC, indicated by higher intratumoural (epithelial) density of CD3+ T-lymphocytes, CD68+ macrophages, and CD56+ NK cells and higher density of proliferating (Ki67+) NK cells within the stroma of TP53 mutant tumour. Furthermore, MACS subgroup within the MSS tumour showed higher density of CD56+ NK cells and proliferating (Ki67+) macrophages in the stroma.

To investigate how the cancer community can influence cancer cells, a 3D (three-dimensional) spheroid co-culture model was developed. This model allows multiple cellular populations to be cultured together and, through specific labelling, each population can be tracked during the experiment. We initially sought to investigate the effect on cancer cells induced by the most common cell in the cancer community i.e., the cancer associated fibroblast (CAF). Two different CRC cell lines, HCT116 (classified as CMS4) and DLD1 (classified as CMS1) were co-cultured with CAFs. Proliferation of HCT116 was increased in the presence of CAFs whilst DLD-1 showed no response to the presence of CAFs. Since macrophages were identified as a common population in the cancer community, peripheral blood monocytes (precursors of tissue macrophages) and macrophages were added as a third population to the co-culture. However, any effect of addition of monocytes or macrophages to the growth of either HCT116 or DLD-1 was not observed.

The 3D co-culture model also allowed us to interrogate the specific interactions between genes expressed in the cancer cells and the cancer community. CD24 is a glycosylated cell-surface protein that acts as an oncogene in CRC. HCT116 has no expression of CD24 and thus provides a good model for testing the activity of CD24. Cells which had been transfected with a CD24 expression vector (HCT116CD24) were compared with cells transfected with empty vector (HCT116EV). Co-culture of HCT116CD24 with CAFs resulted in greater cell proliferation of the cancer cells than co-culture of HCT116EVC with CAFs. This apparent effect of CD24-CAF interaction on cell proliferation is a very interesting finding as previous studies in our laboratory showed no effect of CD24 in CRC proliferation using the 2D (two-dimensional) culture system.

Altogether, this study has established models which allow evaluation of gene-specific effects on the population profile of the cancer community and genotype-specific responses of cancer cells to the cancer community. We have made several interesting observations which need further investigation in order to fully understand the interactions between cancer cell and cancer community.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Ilyas, Mohammad
Martinez-Pomares, Luisa
Keywords: Tumour microenvironment; Driver gene mutations; Co-culture model; Cancer cells; Colorectal cancer
Subjects: W Medicine and related subjects (NLM Classification) > WI Digestive system
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 64240
Depositing User: Susanti, Susanti
Date Deposited: 04 Aug 2021 04:40
Last Modified: 04 Aug 2021 04:40

Actions (Archive Staff Only)

Edit View Edit View