Regular readers of this column will know that there are two main categories of nucleic acids—DNA and RNA. They’ll also know that while for living organisms DNA acts as the genetic data repository, RNA has a messenger role (mRNAs, transcribed from DNA to direct protein synthesis). Most will also recall that there are other classes of RNA molecules, particularly tRNAs (used to tag and identify amino acids for protein synthesis) and rRNAs (structural components of the ribosome, the cellular “machinery” for protein synthesis). In addition to these, there’s increasing interest in the molecular diagnostics community in a less widely known but no less common RNA form, the microRNA or miRNA. Where do miRNAs come from, how do we analyze them, and what do they have the potential to tell us about the health of patients?
Meet the miRNA
miRNAs are, as the name implies, tiny. They’re generally about 22 nucleotides in length and are most commonly coded for in “non-coding” regions of the DNA (non-coding for protein, that is). These regions can include introns within coding genes, or they can exist buried within other longer noncoding RNA molecules. In either case, the miRNA encoding region is marked by having an adjacent, self-complementary sequence such that after it is transcribed from DNA to a single-stranded RNA precursor it can fold back on itself to create a double-stranded hairpin configuration. Well-conserved cellular machinery, including the critical cytosolic enzyme Dicer, then act to cut out the miRNA from this precursor and release it in its active, single-stranded form.
In their active form, miRNAs act by binding to fully or partially complementary sequences within mature mRNAs. This in turn acts as a negative regulator on the mRNA being translated to functional protein, by one of several mechanisms: it can interfere with and slow down translation of the mRNA to protein, or it can destabilize the mRNA and trigger its degradation. Expression of an miRNA thus acts as a layer of post-transcriptional negative control on gene expression, and it turns out to be a very common one.
miRNAs are found and well conserved (that is, have genetically highly stable sequences over evolution) in both plants and animals, and even in some viruses. The ubiquity of this regulatory mechanism in humans can be best appreciated by considering that more than 5,500 unique miRNAs have been identified as being expressed, many in a tissue- and developmental stage-specific context.1 As miRNAs do not require target mRNA complementarity over their entire 22 bp length to function (a run of as little as six to eight bp can allow for interaction and mRNA regulation, meaning a single miRNA may interact with multiple targets), it becomes apparent that with only approximately 25,000 total human coding genes, many can be subject to miRNA control. In fact, current estimates suggest that nearly 60 percent of all human coding genes are regulated to some extent via miRNA.
As regulatory molecules, then, it stands to reason that miRNAs are expressed in controlled patterns based on cell type, developmental stage, and various external signals. Our interest from the MDx perspective, then, comes about from the idea that if miRNAs show particular reproducible expression patterns in healthy cells, disease states such as cancer or infection may lead to detectable perturbations in pattern of miRNAs expressed. With next-generation sequencing (NGS) technologies ideally suited to the identification of large numbers of short nucleic acid molecules simultaneously, we have technologies in hand which can rapidly produce miRNA expression profiles from input samples, giving the identity (and relative quantity) of the miRNAs found. While we may not know the target(s) of each miRNA, we can observe them as biomarkers whereby variations in the usual profile can indicate disease states, or even in some cases provide detailed insight to the disease state when the profile perturbation is itself statistically characteristic of a particular condition, or indicative of a particular suggested treatment strategy. In some cases, just the levels of a few specific miRNAs may be highly informative as biomarkers.
From a technological standpoint, miRNAs are generally observed either in bulk (NGS methods) or by targeted qPCR when only a small known handful are likely to be of interest. In either approach, specialized methods have to be applied to adapt these technologies to work with such short targets. Usually this includes the selective removal of longer nucleic acids from the input material, and the ligation of longer nucleic acid “labels” or “handles” onto the miRNAs. These handles then serve as the basis from which to perform sequencing or qPCR, with the miRNA body being selectively detected and quantified.
With that background on what miRNAs are and what lab methods are used to detect and quantify them, let’s briefly consider a few examples of how they are now coming into clinical utility.
A common and well known test for acute cardiac injury is the presence of circulating cardiac-specific troponin T. Damage to the cardiac muscle cells leads to membrane permeability and thus leakage of this marker into peripheral blood. A similar approach can be taken based on the observation that some miRNAs are almost exclusively expressed in cardiac muscle. MicroRNAs miR-499-5p, miR-1, miR-133a, miR-208b, and miR-499 appear to be the most promising in this application, with improved predictive value as compared to troponin T testing. Readers interested in more detail on this application are directed to Reference 2, a recent review.
Since miRNAs have a regulatory role, it stands to reason that their appearance and levels are likely to be perturbed in states of deregulation of normal cellular development, such as tumors. Indeed, studies have indicated that those miRNAs shown to repress known tumor suppressor genes are often amplified in cancers, while those shown to repress proto-oncogenes are often deleted. Different tumor types have been observed to have distinctive miRNA expression profiles. One study utilizing a panel of just 48 selected miRNAs demonstrated accurate classification of root cancer type even in metastatic sites in nearly 80 percent of cases.
In addition to classifying cancer based on type and tissue of origin, some attempts have been made to utilize miRNA profiling to classify subtypes of cancers (such as luminal versus basal breast cancer, or squamous versus non-squamous non-small cell lung cancer). miRNAs have also shown promise as biomarkers in the early detection of some cancers; in the case of ductal adenocarcinoma, one study has indicated that detectable overexpression of just two markers (miR-21, miR-205) is both diagnostic and detectable before overtly visible phenotypic changes detectable by traditional histopathology.
Additionally, work has been done on using miRNA profiling as a means to classify cancers based on likelihood of response to a given therapeutic approach; an example of this is in hepatocellular carcinoma, where suppressed levels of miR-26 has been suggested as a marker for good response to interferon-α treatment. Readers interested in more on this category of miRNA application (and a more detailed look at the history of the discovery of, biogenesis of, and modes of biological action of miRNAs) are directed to Reference 3 as a good starting point.
Infections represent another situation where abnormal cell function may become apparent through alterations in normal miRNA expression profiles. For instance, chronic hepatitis B as the underlying cause of liver failure has been shown to be detectable at 85 percent sensitivity and 70 percent specificity—less than ideal as a diagnostic, but a promising start—by analysis of just 10 miRNAs in peripheral blood.4 Analysis of just three miRNAs in peripheral blood (miR-361-5p, miR-889, and miR-576-3p) was demonstrated in another study as approaching 90 percent specific for the detection of pulmonary tuberculosis as distinct from other infections or uninfected controls.5
These few examples highlight the diversity of applications to which miRNA profiling, following sufficient validation as biomarkers in particular instances, may be usefully applied. Since this class of molecules can be detected and characterized with existing MDx lab equipment—in some of the examples covered here, even with basic real-time qPCR instrumentation—their adoption into routine diagnostic streams has few technical barriers. Validation and regulatory complexities common to all biomarker strategies remain, but as regulatory flows for approving multi-analyte testing become more common, miRNA-based approaches will likely take on many more diverse and useful roles as diagnostic tools for the MDx lab to offer.
- Londin E, Loher P, Telonis AG, et al. Analysis of 13 cell types reveals evidence for the expression of numerous novel primate- and tissue-specific microRNAs. PNAS. 2015;112(10):E1106-E1115.
- Schulte C, Zeller T. microRNA-based diagnostics and therapy in cardiovascular disease—Summing up the facts. Cardiovasc Diagn Ther. 2015;5(1):17–36.
- Lorio M, Croce C. MicroRNA dysregulation in cancer: diagnostics, monitoring and therapeutics. A comprehensive review. EMBO Mol Med. 2012;4(3):143–159.
- Jin B-X, Zhang Y-H, Jin W-J, et al. MicroRNA panels as disease biomarkers distinguishing hepatitis B virus infection caused hepatitis and liver cirrhosis. Nature Scientific Reports. 2015;5. Article number: 15026.
- Qi Y, Cui L, Ge Y, et al. Altered serum microRNAs as biomarkers for the early diagnosis of pulmonary tuberculosis infection. BMC Infectious Diseases. 2012 Dec 28;12:384. DOI: 10.1186/1471-2334-12-384.
John Brunstein, PhD, is a member of the MLO Editorial Advisory Board. He serves as President and Chief Science Officer for British Columbia-based PathoID, Inc., which provides consulting for development and validation of molecular assays.