Structural effects of cross-kingdom miRNA on regulation of RNA function in cancer

MicroRNAs (miRNAs) are a class of short (17-25 nucleotides), non-coding RNA that regulate post-transcriptional silencing of target genes. A single miRNA can target multiple mRNAs and impact the expression of many genes. In general, miRNAs are key regulators of numerous biological processes such as development, stress response, growth and other physiological processes. In 2012, it was found that plant miRNAs are able to inhibit mammalian gene expression in liver through food intake; demonstrating the first case of cross-kingdom regulation. Countinuing this research path, in 2015, it was found that oral administered mixtures of tumour suppressor miRNAs reduced tumour burden in a mouse model of colon cancer. These observations demonstrate that plant miRNA is absorbed by the mammalian digestive system; therefore, target mammalian gene activity. Moreover, they suggest the hypothesis that engineered edible plant miRNAs with tumour repressor properties could be a novel approach for cancer treatment. This would provide an effective, non-toxic, chemo-preventitive treatment plan for humans.

Recently, Shu and collaborators highlighted the potential computational approach to identify plant miRNA and mammalian mRNA interactions. A comperative analysis was applied focusing on inferring the likelihood of exogenous miRNA in human circulation. High-throughput sequencing technologies have made the computational approach potentially quicker and more cost-effective than standard laboratory methods. Using these technologies the knowledge gained about the miRNA universe rapidly increases. Several computational technologies have been developed to assess and predict miRNA:mRNA interactions such as miRanda, TargetScan, PITA and PicTar. Most algorithms use experimental knowledge to develop a specific scoring system such as miRNA:mRNA partial complementarity, seed region, target position and sequence convservation features. In 2013, Coronello and Benos developed a web tool for combinatorial miRNA target prediction called COMIR, which combines four scoring schemes (miRanda, PITA, TargetScan and mirSVF) to predict potential target genes. In 2016, the industrial partner MirNat developed MirCompare, a web application with the aim to identify plant and mammalian miRNA functional homologies as plant miRNA tend to have stringent selectivity for possible binding sites.

In this project, the aim is to advance these computational approaches working together with the industrial partner on MirCompare and design a pipeline that detects miRNA sites eligible for mRNA binding.