Skip to main content



Colorectal cancer (CRC) is one of the most common cancers in both men and women in the EU. Although some studies in the past 10 years have identified a large proportion of the genetic factors that predispose to CRC development, we are still not at a point where these markers are predictive enough to be applied in clinical practice. Nevertheless, these discoveries have also led to the identification of important biological pathways that are essential to CRC development, and raise the possibility of using other “intermediate” biomarkers (molecular phenotypes) that allow for better prediction of CRC risk. The main aim of this project was to assess a number of these intermediate phenotypes, and investigate their use as potential biomarkers of CRC risk.

The project was divided into 4 specific objectives:
1. Establishment of a bank of specimens from normal colorectal mucosa.
Most genetic studies are routinely performed on peripheral blood samples. Nevertheless, for other types of molecular biomarkers (RNA-based, for instance), the use of the tissue-of-origin is essential, since they can vary or be present only in certain tissue types. Thus, we collected normal colonic biopsies from patients undergoing colonoscopy with the help of collaborators in the Department of Endoscopy the John Radcliffe Hospital. All patients received informed consent, meaning they were explained the purpose and aims of the study, and agreed to donate a sample of blood and/or normal bowel biopsies. A total of 408 patients were finally recruited in the study between October 2012 and March 2014. Of these, 107 patients showed no sign of polyps or CRC in the bowel, and had no personal or family disease of CRC. These were regarded as the control group. Then, 101 presented with Adenomatous Polyps (the most common form of benign bowel polyps), 38 with serrated polyps, and 30 with CRC. I believe ours is possibly one of the largest cohorts of normal bowel tissue that has so far been described in the literature.
2. Assay development.
We aimed to assess three types of markers: genetic risk SNPs, Telomere Length TL) and Pathway Expression. The risk SNPs are single-nucleotide changes in the normal DNA sequence that are very common in European populations and confer only an slight increase in CRC risk. Telomeres are the DNA sequences that constitute the ends of chromosomes, and are related to their appropriate replication, stability and cell ageing. Pathways are networks of proteins interacting with one another to provide response to cellular processes, such as proliferation, response to external stimuli or cell death. By assessing these three types of markers we aimed to find a biomarker that could summarize all processes leading into the malignant neoplasia transformation, and that could be helpful in determining which people in the population are at risk of developing CRC and should therefore be followed-up carefully.
3. Assessment of intermediate marker levels in cases and controls.
In order to assess the abovementioned hypothesis, one of the most widely accepted techniques consists in comparing levels of the given biomarker in a group of patients with the disease, and a group of disease-free people, or controls. For this purpose, we extracted DNA and RNA from both the bowel tissue and the blood from each patient, and performed the assays described in objective 2, in order to find out any potential differences in these markers between healthy and diseased individuals. Once the individual assays were analysed, we also created a combined model including all 3 types of markers to check for any potential overlap or relatedness amongst the biomarkers. By these means, we obtained consistent and significant associations of both colonic TL as well as a polygenic risk score constructed on SNP genotypes.

4. Translation of markers from bowel to blood.
Whereas findings in bowel tissue are of utmost importance, and bowel biopsies can nowadays be easily and steadily collected from endoscopic procedures (also on a population basis, such as from Bowel Cancer Screening National Programs), a test based on peripheral blood would be simpler and quicker. Therefore, significant markers in the bowel assays were also trialled in blood from the same patients. In our case, I found blood TL to be correlated to bowel TL measurements. This means it would be possible to indirectly assess bowel TL through its blood counterpart, which would be a less invasive means of risk assessment.

By assessing 3 different types of intermediate phenotypes, we aimed to detect molecular
biomarkers that would act as surrogate markers of genetic and/or other types of variation, thus allowing much better prediction of CRC risk. We would hope that these phenotypes would represent pathway endpoints and as such would constitute a summatory measure of all previous events leading to neoplastic transformation that would provide with a clinically useful means of assessing CRC risk.
We have discovered that both bowel telomere length and polygenic risk scores may be used as good measures of case/control status in polyp/CRC development. Moreover, blood telomere length seems to be related to that of the bowel, which would allow for the assessment of the latter via the former in a much simpler way. We have also seen a strong effect of age as a population modulator, which could be an indicator that strategies need to be targeted or refined depending on patient age.
Although the positive predictive value of these biomarkers is relatively small when used to assess individual CRC risk, they could be easily modelled into binning the population in different risk groups. For instance, individuals at the top of the distribution of the polygenic risk score would be seven times more likely to develop CRC than those in the bottom, and 3 times more likely than the median. This fact could be then potentially implemented into National Bowel Cancer Screening Programs in an algorithm that would ideally give preference for screening to those individuals at higher risk of bowel neoplasia.
Given the importance of CRC burden in the EU, we am happy that our study contributes to general understanding of CRC development, and gives relevant data to be used in the clinical field in terms of both CRC prevention and early detection and prognosis. It also fulfils one of the FP7 framework aims, which was to enhance and promote personalized medicine.