MicroRNA Biomarkers: Pitfalls and Potential

Progress has been made in establishing the role of miRNAs as biomarkers for disease, but challenges remain with clinical translation

Thirty years ago, the small molecules we now know as microRNAs (miRNAs) would have been regarded as “junk” genetic material. Today, miRNAs have been hailed as promising biomarkers for diseases including cancer, neurological disorders, heart conditions, and infections. Although progress has been made toward characterizing the role of miRNAs in different diseases, there remain challenges in bringing them into clinical use. Researchers continue to explore where miRNAs will make the most impact.

Circulating miRNAs

Since their discovery in 1993 and formal classification in the early 2000s, miRNAs have been the focus of intense study. miRNAs are small, noncoding RNAs—only around 22 nucleotides long—cleaved from larger precursor molecules. Humans express approximately 2,500 miRNAs, which play important roles in cell activity. Their main role is to regulate gene expression by binding to messenger RNAs and inhibiting translation [1].

When physician–scientist Muneesh Tewari started his lab at the Fred Hutchinson Cancer Research Center in 2005, it was known that certain miRNAs were dysregulated in cancer cells versus healthy cells. But Tewari wondered if there might be an even easier way to detect miRNAs associated with cancer; that is, what if one could find miRNAs circulating in the blood?

In 2008, Tewari’s lab was one of three independent groups that reported that miRNAs are present in human plasma in a remarkably stable form [2]. “What’s more, we found that there were specific miRNAs, detectable in plasma, that could distinguish patients with metastatic prostate cancer from healthy controls,” says Tewari, who is now at the University of Michigan (Figure 1). “It showed that miRNAs may represent a whole new type of blood-based biomarker.”

Over the next decade, scientists documented changes in miRNA expression in other cancers, as well as cardiovascular diseases, autoimmune diseases, neurological disorders, tuberculosis, and viral respiratory diseases, among others. Further studies revealed that miRNAs are not only present in blood; they can be detected in any body fluid, including saliva, urine, tears, breast milk, and cerebrospinal fluid [1].

This led to a flurry of research interest in miRNAs as potential noninvasive biomarkers in various diseases. But measuring and interpreting levels of circulating miRNAs turned out to present unique challenges. According to Tewari, the field had to go through some “growing pains.”

Early challenges

After initial promising results, it became clear that there were some major factors holding back miRNA biomarker research. Principal among these is the issue of specificity. Researchers soon found that differentiating related diseases based on circulating miRNA signatures could be challenging. For instance, studies found the same miRNA, miR-21, is upregulated in patients with colorectal, lung, breast, prostate, liver, esophageal, and endometrial cancers [3]. Other miRNAs, such as miR-146a, are upregulated in multiple infections and seem to be indicative of a generalized immune response [4].

Some researchers, like Ryan Farr, a molecular biologist at the Commonwealth Scientific and Industrial Research Organization (CSIRO) Australian Center for Disease Preparedness (Figure 2), are leveraging advanced bioinformatics tools to address this apparent lack of specificity. “Using machine learning enables us to identify and evaluate the pattern of miRNAs with the highest predictive power,” he says. This approach has been used to find miRNA signatures of different specific pathogens.

However, in its early days the field had an even more basic problem than specificity. According to Séverine Lamon, a molecular biologist at Deakin University (Figure 3), methodological differences and a lack of relevant quality controls led to inconsistent results and problems with reproducibility. “Because of its huge potential, miRNA research evolved very quickly, and people jumped into it without thinking about optimizing the methodology,” she says.

In fact, the experimental setup and processes can drastically influence the measurement of circulating miRNAs. The source of samples—whether they are extracted from blood, plasma, or another body fluid—is critical, as is the sampling method and preservation and processing of the samples [5].

Lamon says that after the initial research rush, there are now standardized protocols for sample collection, storage, and processing for miRNA. But she still thinks something is missing: mechanistic studies. “99% of miRNA studies are observational, meaning they look at differences in miRNA signatures between people with a disease and healthy people without specific interventions,” she says. “They are interesting but all they tell us is that there is a dysregulation somewhere.”

According to Lamon, there is still much to learn about the function and biology of circulating miRNAs, including the molecules and signals involved in their secretion, their effects on upstream and downstream biological pathways, and their precise role in different diseases processes.

Biomarker potential

Even as questions remain about basic miRNA biology, researchers are further probing their usefulness as biomarkers—and finding promising applications.

According to Farr, circulating miRNA biomarkers may prove valuable as a diagnostic tool for infectious diseases. His group and others have identified miRNA signatures associated with a range of pathogens, including Hendra virus, HIV, tuberculosis, malaria, and Ebola [4]. Farr says that one advantage of miRNAs is that changes can be observed early in the course of a disease. “In the context of infectious diseases, there are many benefits in diagnosing earlier,” he says. “We can detect changes in miRNAs within hours to days of an infection, before the onset of symptoms and before antibodies or the pathogen itself can be directly detected.”

The potential for early detection via a noninvasive blood test also makes miRNAs an attractive biomarker for various cancers. Srinivasulu Yerukala Sathipati, a research scientist at the Marshfield Clinic Research Institute (Figure 4), is working on machine learning-based cancer prediction models, with the goal of identifying miRNA biomarkers for early detection and survival estimation in different cancer types. He and his team developed an artificial intelligence-based diagnostic prediction method, called BSig, for identification of early biomarkers in patients with breast cancer. “Using serum samples from a gene expression omnibus database, we found that BSig distinguished patients with cancer and healthy individuals with an accuracy of 99%,” he says.

In a recent study, Sathipati and colleagues described the development of an evolutionary machine learning method called CancerSig to identify miRNA signatures that could serve as early biomarkers for multiple cancer types. Using this method, the researchers identified 15 cancer stage-specific miRNA signatures for 15 different cancer types [6]. “We are hoping that in the near future, we can screen and detect an early cancer before the onset of symptoms just by a blood test,” says Sathipati.

Future of miRNAs in medicine

The discovery of circulating miRNAs led to a rush to explore the potential biomarker applications, some of which are beginning to bear fruit. But Tewari and others say there are other, more targeted applications for these molecules. “Small, noncoding RNAs, in general, are promising therapeutic targets,” says Tewari. Researchers are exploring ways to inhibit or replace miRNAs in different diseases where they are either overexpressed or repressed. This approach may make it possible to target the activity of a specific gene without affecting the genetic material of the host cell. Already, drugs based on miRNAs are undergoing clinical trials [1].

While Lamon appreciates the biomarker and therapeutic applications of miRNAs, she is still interested in teasing out what they do in the body. “We know that miRNAs are basically fine tuners of gene regulation,” she says. “But there are still many mechanistic questions about how they function in health and disease.”

“There are a lot of molecular tools we can use,” she says. “We can overexpress or repress expression of miRNAs in a mouse model or in various types of cell culture and look at the downstream effects—not only the genes and proteins that it regulates but also the functional outcome that is associated with it.”

As the research on miRNA biomarkers continues to be refined and move toward the clinic, Farr says it’s important to target those contexts where they will make the biggest difference. “It’s easy to extol the virtues of miRNAs, but we are not looking to replace everything with a miRNA diagnostic,” he says. “We are looking at filling gaps where conventional diagnostics aren’t meeting our needs.”

“We have shown that they have promise. Now, we want to prove that an assay based on circulating miRNAs can work.”

References

1. T. Pozniak, D. Shcharbin, and M. Bryszewska, “Circulating microRNAs in medicine,” Int. J. Mol. Sci., vol. 23, no. 7, p. 3996, Apr. 2022, doi: 10.3390/ijms23073996.
2. P. S. Mitchell et al., “Circulating microRNAs as stable blood-based markers for cancer detection,” Proc. Nat. Acad. Sci. USA, vol. 105, no. 30, pp. 10513–10518, Jul. 2008, doi: 10.1073/pnas.0804549105.
3. K. Saliminejad, H. R. K. Khorshid, and S. H. Ghaffari, “Why have microRNA biomarkers not been translated from bench to clinic?” Future Oncol., vol. 15, no. 8, pp. 801–803, Mar. 2019, doi: 10.2217/fon-2018-0812.
4. L. Tribolet et al., “MicroRNA biomarkers for infectious diseases: From basic research to biosensing,” Frontiers Microbiol., vol. 11, p. 1197, Jun. 2020, doi: 10.3389/fmicb.2020.01197.
5. D. Hiam and S. Lamon, “Circulating microRNAs: Let’s not waste the potential,” Amer. J. Physiol.-Cell Physiol., vol. 319, no. 2, pp. C313–C315, Aug. 2020, doi: 10.1152/ajpcell.00175.2020.
6. S. Y. Sathipati et al., “Artificial intelligence-driven pan-cancer analysis reveals miRNA signatures for cancer stage prediction,” Hum. Genet. Genomics Adv., vol. 4, no. 3, Jul. 2023, Art. no. 100190, doi: 10.1016/j.xhgg.2023.100190.