Raf Kinase Inhibitor Protein (RKIP) was originally described as an inhibitor of the Ras-Raf-MEK-ERK pathway, exerting its action by the physical inhibition of the interaction of Raf with MEK. It has subsequently been shown to play important roles in a number of other signalling pathways, including the NFκB pathway and in the stability of the mitotic spindle. Not surprisingly given that it impacts on many important signalling pathways RKIP levels have been shown to be important in the progression of a number of different cancers. RKIP expression is lost or decreased in a number of common human cancers and decreased still further in tumour metastases. One of the tumours in which RKIP is downregulated is colorectal cancer (CRC). Importantly it has been shown that not only is RKIP depleted in tumour tissue when compared with normal tissue but that the level of RKIP within a tumour is inversely correlated with the likelihood of metastatic relapse and with patient prognosis. Although we already have a number of very good prognostic indicators in CRC, one group of patients for whom new prognostic indicators would be useful are patients with Dukes B CRC. These are patients with locally advanced but non-metastatic disease and at present there is no firm consensus on their correct post-operative management. Therefore we set out to examine whether RKIP is a useful prognosticator in this particular group using a tissue microarray (TMA) with samples from over 200 patients with Dukes B CRC. The analysis revealed a strong inverse correlation between RKIP levels and disease specific survival. Moreover, in a multivariate analysis RKIP emerged as an independent prognostic indicator along with lympho-vascular invasion and peritoneal invasion, two well-known and powerful prognosticators. This allowed for the generation of a simple prognostic index, using information from the different independent indicators, allowing for improved patient risk stratification. This led us to examine whether RKIP could also function as a predictive marker in CRC. To do this we again used a TMA, this time consisting of a much larger cohort of patients across the whole range of tumour stages. The results confirmed the prognostic utility of RKIP and indicated that patients whose tumours have low levels of RKIP may derive a greater benefit from chemotherapy than those patients whose tumours have high levels, although this result did not reach statistical significance. In the second part of the thesis I have examined the effect of RKIP in previously characterised mouse models of CRC. To do this I have used a germline RKIP knockout mouse and in the first instance crossed it to the APC580S mouse. In this mouse APC is lost conditionally within the intestine and liver. RKIP knockout did not have any effect on the rate of tumourigenesis or on the invasiveness of tumours in this model. However, in the setting of acute homozygous deletion of APC, RKIP knockout resulted in a decrease in apoptoses in the small intestine and an increase in aberrant mitotic activity in the liver. To follow this up I have examined the effect of RKIP knockout in a mouse model of superficially invasive CRC, specifically to see if RKIP knockout can promote invasive and metastatic behaviour. In this model the APC580S mouse is crossed to mice which conditionally express oncogenic KRas. Although RKIP knockout did not result in an increase in invasive tumours in this model there was a shift in tumour location from the small intestine to the colon. This shift appeared to be due, at least in part to an increase in chromosomal instability in the tumours. The final aim of the thesis was to develop a mouse model of CRC which more closely recapitulates the late stages of the human disease, specifically invasion and metastasis. To do this we have crossed the APC580S mouse with either a conditional p53 knockout or with a mouse that conditionally expresses a point mutation of p53 (p53R172H). In human tumours the majority of abnormalities of p53 are point mutations that result in the production of mutant protein that accumulates in tumour cells. There is evidence that this mutant protein may have oncogenic properties beyond the simple loss of normal p53 protein function. Therefore we have also used this model to study the differing effects of p53 loss and point mutation in CRC. We found that mice homozygous for p53 deletion (p53fl/fl) and those expressing a single copy of the mutant allele with loss of the second copy (p53R172H/fl) developed invasive tumours with nearly 100% penetrance and indeed metastasis was observed. Remarkably, although mice that were heterozygous for p53 deletion (p53fl/+) only rarely developed invasive tumours almost 100% of mice expressing a single copy of the mutant allele (p53R172H/+) developed invasive tumours. We went on to show that the increase in invasion seen in this model is related to an increase in Wnt signalling, which is associated with increased expression of pro-invasive Wnt targets such as fascin. We also showed a novel pro-invasive role for ARF in this process. This is also an excellent model of Dukes B CRC and therefore the ideal model to test the effect of RKIP deletion on invasion and metastasis. These studies led us to examine the differences in effect between knockout and mutant p53 in another tumour model. In this we used a novel model of the aggressive tumour pleomorphic rhabdomyosarcoma to demonstrate that mutant p53 can both promote both tumourigenesis and metastasis more potently than p53 knockout. These studies have demonstrated the value of RKIP in the clinically important Dukes B CRC population and shown its possible utility as a predictive marker in this group. Although we have not seen an effect of RKIP knockout in traditional mouse models of CRC we have developed a novel model which closely recapitulates Dukes B CRC and may be useful in elucidating the effect of RKIP knockout. We have also used this model to gain novel insights into the invasive process, in particular into the role played by mutant p53.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:523342 |
| Date | January 2010 |
| Creators | Doyle, Brendan |
| Publisher | University of Glasgow |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://theses.gla.ac.uk/2201/ |
Page generated in 0.0027 seconds