Workshop Report 32

Data analysis and biomarkers

Frank Slack, Department of Pathology, Director, Institute for RNA Medicine, Beth Israel Deaconess Medical Center – BIDMC Cancer Center Harvard Medical School, USA

MiRNAs have been found in all multicellular eukaryotes investigated so far, where they are important regulators of gene expression, usually acting post-transcriptionally. MiRNAs may be disease-related loci, e.g. oncogenic and tumour suppressor loci. As such, they may regulate the expression of important disease genes, e.g. oncogenes and tumour suppressors (Chen et al, 2008; Chin and Slack, 2008; Lawrie et al, 2008; Mitchell et al, 2008). Accordingly, miRNAs may be useful diagnostic and prognostic markers for diseases, and they are emerging as therapeutic and targeted therapeutics in different diseases, including cancer (Lu et al, 2005; Calin and Croce, 2006; Kasinski and Slack, 2011). For instance, RNA profiling points to miR-34a as important and prognostic in triple negative breast cancer (Kato et al, 2009; Adams et al, 2016).

A suitable starting point in investigating the applicability of miRNAs as therapeutic targets or agents is to measure their relative or absolute levels in normal and diseased tissues. MiRNA profiles may be obtained from (fresh, fixed or frozen) tissues, organs, cells and subcellular fractions as well as from all relevant body fluids. MiRNAs are stable in the blood and other body fluids (Weber et al, 2010; Mitchell et al, 2008). Technologies for detecting miRNAs include Northern blotting, in situ hybridization, (low density) qRT-PCR, microarray technologies, and miRNA sequencing (Johnson et al, 2007; Chiang, 2014). In a comparative assessment, each technology appeared to have its own strengths and limitations. Hence, even though miRNAs rankings were concordant between different technologies, their outcomes could not be compared quantitatively (Baker, 2010). One of the drawbacks of miRNA-sequencing is that many miRNAs have isomeric forms so that it is hard to directly compare with detection systems that are not always fully precise. Further challenges to the detection and diagnostic use of miRNAs include the small sizes of the molecules, delivery and specificity issues, the determination of ‘normal’ miRNA levels, and the fact that very many miRNAs are only expressed at very low levels (Cheng et al, 2015). Even though the available technologies allow detecting such low levels of miRNAs, their biological implications remain to be determined.

Generally, whereas miRNA expression profiling provides first information on biomarkers of interest, miRNA levels alone do not allow determining their functionalities (Pritchard et al, 2012). As a rule, such determinations require mouse knock-out or knock-in models or ‘therapeutic’ interventions, using, e.g. anti‑miRs that inhibit miRNA function and or expression, miRNA mimics that increase miRNA expression, viral pre-miR and miRNA sponges or miRNA luciferase sensors.


What are the implications that investigations on the functionality of miRNAs generally inhibit the miRNA? – Generally, when the miRNA is inhibited one sees that the corresponding target factors are increased in expression. This can be complemented with additional experiments where a miRNA mimic is used to increase the levels of a given miRNA. This should show the opposite effect on the expression of target factors.

What is the specificity of miRNA that can be detected in the blood? – The miRNA profiles of a given tumour and concordantly collected blood samples may not be the identical, because the tumour may release only some miRNAs. However, further comparative investigations are necessary to determine the biological implications of such findings.

How stable are miRNA in tissue samples or body fluids? – Generally, they are very stable and may be detected and recorded even after decades provided that samples were stored appropriately.