significantly reduce measurement variability. These considerations extend to consistency in sampling during normal cyclic patterns (circadian, menstrual, etc.), location and fraction of sampling (for example in blood: sublingual, aortic, jugular, etc. and whole blood, serum, plasma, or EV fraction), minimization of hemolysis of collected blood samples, and short-term and long-term storage considerations. Also, miRNAs in biofluids may vary in stability owing to interactions with RNA-binding proteins, EVs, and lipoproteins that may impart protection from high RNase activity in these matrices. In freshly isolated human serum, miRs-16, -21, and -142–3p demonstrated greater stability than miRs-122 and -1 after twenty-four hours at room temperature (Koberle et al. 2013). This difference was affected by RNase activity in the serum, whereas miRNAs in EVs were protected against this activity.
If isolating EVs, the techniques used can influence the content measured. First, there is confusion in the literature on how to distinguish different subclasses of EVs, and this subsequently complicates the identification of RNA cargo (Mathieu et al. 2019). Further, RNA can bind other types of carriers such as lipoprotein and ribonucleoproteins, which are difficult to separate from membrane-enclosed RNAs (O’Brien et al. 2020). Not surprisingly, vesicle purification can lower the RNA yield and integrity and, because of the lower RNA concentration, result in more interindividual variability and increased difficulty in measuring some miRNA biomarkers. To compound the issue, obtaining enough miRNA from vesicles is also a challenge, as studies have demonstrated an average of one or less than one miRNA per EV (Chevillet et al. 2014; Li et al. 2014). Therefore miRNA biomarkers released through active and passive mechanisms may have different stability levels in stored samples and recovery may be inherently low, requiring optimization and consistency in sample collection for specific miRNAs. There are a variety of methods that can be used to measure miRNAs in biofluids, and the selection depends on application, cost considerations, expertise, and available resources. These methods include lower-throughput techniques such as quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) or digital drop PCR (ddPCR), medium-throughput techniques such as bead hybridization (e.g., NanoString™ or FirePlex™ assays), and higher-throughput techniques such as microarray and next generation sequencing. A comparative study that used qRT-PCR across eleven research institutions to examine miRNA biomarker alterations due to toxicant exposure found that, when outliers were removed, miRNA measurements were very consistent (only contributing 2% of the measured variance). This supports the idea that standardized protocols can lead to consistent measurements regardless of different site locations (Thompson et al. 2016). Comparison across methods shows consistent correlation when using fold change calculations: concordance ranges from 0.6 to 0.95 (Giraldez et al. 2018; Git et al. 2010; Mestdagh et al. 2014; Yauk et al. 2010).
While miRNAs can be consistently measured methodologically, a primary post-analytical challenge is data normalization. The normalization of measured miRNA data would theoretically correct for natural inter-individual fluctuations in fluid volume and rhythmic or cyclic release, as well as for technical variations attributed to miRNA extraction methods, assay inhibition due to biological factors, and other variables. This would allow precise site-to-site, temporal, and methodological comparisons, alongside increasing confidence in interindividual evaluations. Endogenous “housekeeping” calibrators measured in biofluids such as miRNAs or other nucleic acids would be ideal; however, no universally invariant calibrator has been found (Saliminejad et al. 2019). Where endogenous normalizers may not be available or optimized, synthetic exogenous normalizers or miRNAs “spiked-in” during or after small RNA extraction are commonly used to correct technical variations introduced during sample processing and measurement. A recent study analyzed the intra- and inter-individual variability of circulating miRNAs among healthy and cancer patients (Vigneron et al. 2016). Intra-individual variability significantly improved with a geometric mean of three exogenous measurements by comparison with endogenous normalization. These same normalizers—and, interestingly, the endogenous normalizer miR-16-5p—also improved cross-platform correlation of qRT-PCR and microarray measurements. Much like pre-analytical variables, normalization procedures must be optimized for each individual system and application to increase one’s confidence in the measurements taken.
Recently the microRNA Biomarkers Working Group of the Health and Environmental Sciences Institute’s (HESI) Committee on Emerging Systems Toxicology for the Assessment of Risk (eSTAR) examined the pre- and post-analytical variables that have impeded the development of miRNA biomarkers for toxicological and regulatory use (Chorley et al. 2021a). As a result of examining both the technical and the biological variables that influence the measurement of miRNA biomarkers in biofluids, the group made six recommendations for future research. These recommendations were: determining the pre-analytical stability of miRNAs; establishing standard operating procedures for collection and measurement; establishing a reporting framework; determining the mechanisms of action of extracellular release of miRNAs; determining the linkage of miRNA variants (also called isomeric miRNAs or “isomiRs”) to adverse outcomes; and identifying the cellular or tissue source of extracellular miRNAs. Establishing these unknowns for a miRNA biomarker application strengthens one’s confidence not only in the data produced but also in the mechanism by which the miRNA is released into the biofluid, thereby indicating clear linkages to the adverse outcome of interest and an increased specificity and sensitivity of the biomarker. Of the recommendations, the establishment of standard operating procedures was the highest priority. Procedures that follow standard guidelines for biofluid collection, storage, and handling (such as those outlined by the Early Detection Research Network (ERDN)), as well as the automation of procedures, where applicable, will ensure consistent, reproducible results (Farina et al. 2014).
Conclusions
The identification of miRNA biomarkers that can distinguish the biological effects of toxicant and other environmental chemical exposures is a much needed tool for toxicological and regulatory science. Studies have suggested that these biomarkers, which are stable and present in most biofluids, may indicate general toxicity in a tissue-specific manner. Further, the RNA content of EVs may reflect the type and the physiological or pathological state of the source cells. Extracellular vesicle-mediated functional transfer of miRNAs to recipient cells has been demonstrated both in vitro and in vivo for a variety of diseases and physiological states, but the mechanisms by which EVs are selectively packaged and “addressed” to different cell types are still largely unknown. In addition to specificity of tissue and MOA, in vitro and in vivo studies suggest early, dose-responsive alterations in expression that can indicate breaking points that may lead to adverse health outcomes. Recent development of miRNA biomarkers in biofluids for clinical indications provide promise that similar uses can be leveraged in toxicological sciences. Despite these qualities, there remain significant technical challenges to reducing variability in measurements and to increasing our confidence in the application of these biomarkers. A standardization of protocols and reporting efforts analogous to the Transcriptomics Reporting Framework (Gant et al. 2017), and large-scale database efforts to categorize EV contents relating to source cells will be important for biomarker identification and for linking biofluid measured miRNAs to specific tissues, cells, and biological effects (see exRNA Atlas: Murillo et al. 2019). These and other concerted efforts will ultimately lead to a consistent and widespread use of miRNAs in regulatory and toxicological sciences.
Acknowledgements
The authors wish to thank Drs. Sheau-Fung Thai, Chris Corton, Sid Hunter and Kimberly Slentz-Kesler for their critical reviews of this manuscript. This manuscript has been reviewed by the US Environmental Protection Agency and approved for publication. Approval does not signify that the contents reflect the views of the Agency, nor does the mentioning of trade names or commercial products constitute endorsement or recommendation of use.
References
1 Amorim, M.G., Valieris, R., Drummond, R.D., Pizzi, M.P., Freitas, V.M., Sinigaglia-Coimbra, R., Calin, G.A., Pasqualini, R., Arap, W., Silva, I.T., Dias-neto, E., and Nunes, D.N. (2017). A total transcriptome profiling method for plasma-derived extracellular vesicles: Applications for liquid biopsies. Sci. Rep. 7: 14395.
2 Arroyo, J.D., Chevillet, J.R., Kroh, E.M., Ruf, I.K., Pritchard, C.C., Gibson, D.F.,