1b) Study on the genetic susceptibility to colorectal cancer (the CCC)

We are currently running the CCC (Colorectal Cancer Consortium) in collaboration with other European Centers from Barcelona (Spain), Prague (Czech Republic), and Heidelberg (Germany). Click here to obtain details on the CCC.

In spite the fact the Italian series of samples contributes for small numbers to the total project, the nice collaboration with the other foreigner P.Is enables us -following general agreement and reasoned plans- to design interesting research projects and develop new ideas.

The main research project (ongoing) agreed by all the member of the consortium is summarized below.

ABSTRACT.

Recent evidence indicate that small non-coding RNA molecules, called micro-RNAs (miRNA), can bind to the 3’UTRs of mRNAs and interfere with their translation, thereby regulating genes of cell growth, differentiation, apoptosis, and tumorigenesis at post-transcriptional level. Genetic polymorphisms can reside on miRNA binding sites. Thus, it is conceivable that the miRNA-regulation of genes may be affected by polymorphisms on the 3’ UTRs. It was clearly showed that colorectal cancer (CRC) has an altered pattern of expression of miRNAs.

Since gene de-regulation is one of the key mechanisms by which cells can progress to cancer, we hypothesize that common polymorphisms within miRNA target binding sites could play a role in the individual risk of cancer.

The principal goal of our research project is to identify new polymorphic genetic loci associated and affecting the individual risk of CRC. In addition, there are also two potential consequences deriving from this work: 1) In case-control studies carried out at whole-genome level the problem of multiple testing in wide-genome scans forces the investigators to correct for multiple testing (such as the Bonferroni’s correction), making very hard to reach the statistical significance, and reducing the overall power of the studies (with the possibility to discharge truly positive results). We will explore whether the approach of investigating specific and selected SNPs at whole-genome level based on reasoned a priori hypotheses allows to identify risk factors with a reduced number of statistical tests, making the correction for multiple testing (such as the Bonferroni’s correction) less stringent; 2) We will attempt to provide a proof-of-principle that routine and cost-effective case-control association studies could ease the screening of large number of putatively interesting SNPs predicted by algorithms, reducing the number of SNPs to be tested with time consuming molecular biology methods.

The project, will be composed of the following steps: 1) Identification of miRNAs involved in the carcinogenic process of CRC. This information will be obtained thanks in great part to the seminal work of Cummins, Vogelstein et al. (“The colorectal microRNAome”, PNAS, 2006) where a large number of miRNAs were identified to be deregulated in CRC. Also new miRNAs were described. 2) Genome-wide identification of the putative targets of all the miRNAs deregulated in CRC from step (1). This will be done in two ways: (a) using already established algorithms of target predictions (PicTar, Diana-MicroT, miRBase, miRanda, MicroInspector, TargetScan S), and (b) by a new algorithm currently under development in our laboratory. 3) The putative miRNA target regions within the 3’ UTR of all the genes identified in step (2) will be scanned in silico for the presence of known common polymorphisms (with a minimal frequency of the rare allele of 0.15%, for statistical power reasons). 4) The biological effect of each polymorphism within the putative miRNA targets from step (3) will be predicted in silico, by measuring the difference between the predicted strength of binding of the common and the rare alleles. This will be done using miRanda, that is specialized to calculate the Gibbs binding free energy (ΔG, expressed in KJ/mol). Thus, the effect of a genetic polymorphism will be measured as the difference of the free energies between the two alleles and computed as “variation of ΔG” (i.e. ΔΔG). This will be an objective parameter allowing to rank SNPs in the order from the most to the less effective. This approach was already used by our group to a limited number of genes and published on DNA cell Biol (Landi et al., 2007). 5) Top-of-the-list SNPs (depending on the total number of identified SNPs we could chose the top quartile of the distribution, or the top 10%) will be evaluated for their association with the risk of CRC. A first case-control association study will be run on a population from Czech Republic (600 cases and 600 controls). 6) Prioritized SNPs in predicted miRNA target regions found associated with the risk of CRC in step (5) will be replicated in an independent population, from Barcelona (Spain) (377 cases and 326 controls). 7) SNPs confirmed to be associated with the risk also in step (6) will be considered relevant for the biology of the gene and could be really implicated in the carcinogenesis process. SNPs positive in step (5) and (6) will be investigated to verify their actual biological activity. We will prepare chimeric constructs where the reporter gene luciferase has the 3’ UTR of the gene positive in (6). Both the common and the variant forms of each 3’ UTR will be assayed. The comparison of the luciferase activity between cells transfected with the two alternative 3’UTR will reveal whether a given polymorphism can affect the post-transcriptional regulation of the gene.

As proof-of-principle, in a previous work done on only a limited set of 104 candidate genes, we found that two polymorphic targets were associated with the risk of CRC (one held after Bonferroni’s correction).

This last work was published on Carcinogenesis (2008).