An Answer Out of Time

Liquid biopsy – a buzzword; a hot new technique; a minimally invasive way of taking a closer look at a patient’s disease. It’s exciting, certainly – but at the moment, it’s also flawed. Current liquid biopsy methods are plagued by low sensitivity and the potential for misinterpretation. When attempting to analyze circulating nucleic acids, and RNA in particular, we need a better way to ensure comprehensive, accurate analysis. Could an ancient enzyme – the progenitor of all reverse transcriptases – be the solution?

We recently published a paper in Molecular Cell that described the first determination of the molecular structure of a thermostable group II intron reverse transcriptase (TGIRT) enzyme (1), which not only reveals how these ancient enzymes work, but also suggests ways we might be able to engineer them to further improve comprehensive and targeted RNA sequencing (RNA-seq) applications. Our work also indicates that TGIRTs may be the ancestors of retroviral and all other eukaryotic reverse transcriptases (RTs), which means that retroviral RTs are actually highly degenerate enzymes that have lost entire regions and features of their active sites – features that contribute to the higher fidelity and processivity of TGIRTs.

An unanticipated solution

The use of TGIRT enzymes for liquid biopsy arose from our basic research on bacterial genetic elements called mobile group II introns. We were interested in those elements because of their ability to carry out novel biochemical reactions, such as RNA-guided integration of the intron RNA directly into DNA target sites for gene targeting applications, and because of their evolutionary relationship to retroviruses and other genetic elements called LINE-1 retrotransposons in humans. We found that mobile group II introns encode a novel form of reverse transcriptase (an enzyme that copies RNA into DNA) that differs from the widely studied retroviral RTs. The latter enzymes are commonly used in biotechnological and medical applications, including RNA-seq and real-time quantitative polymerase chain reactions (RT-qPCR) – key methods employed for liquid biopsy. Several years ago, we developed methods that overcame technical difficulties, particularly the insolubility of group II intron reverse transcriptases when freed from tightly bound endogenous RNAs, enabling us to produce large quantities of highly active TGIRTs from bacterial thermophiles. We found that TGIRTs have a surprising combination of properties that are lacking in retroviral RTs and are advantageous for liquid biopsy.

A giant leap for liquid biopsy?

Liquid biopsy requires the ability to characterize very small amounts of DNA or RNA that are released from cells in tumors and other diseased tissues into blood and other bodily fluids. Ideally, we do this via RNA-seq, which can provide a dynamic snapshot of exactly what’s happening in a diseased tissue, such as a tumor, and can be used to noninvasively monitor disease progression and how the tumor cells are responding to treatment. We found that TGIRTs enhance RNA-seq in several ways: first, they have a novel template switching activity that enables the direct attachment of RNA-seq adapters to small amounts of target RNAs or DNAs without an inefficient RNA ligation step. Second, they have higher fidelity than retroviral RTs, increasing their ability to detect mutations in RNA sequences. And third, they have extraordinarily high processivity and strand displacement activities, enabling them to copy structured RNA species that are not readily captured by currently used enzymes. These include transfer RNAs (tRNAs) and other structured small non-coding RNAs (sncRNAs), as well as mRNAs and long ncRNAs with stable secondary structures that block conventional RTs. Further, because of their higher processivity, TGIRTs also capture larger numbers of alternative splice junctions, particularly near the 5ʹ ends of RNAs, than do conventional RTs.

These attributes have allowed us to construct comprehensive RNA-seq libraries from the very small amounts of RNA present in human plasma (2) and in extracellular membrane vesicles called exosomes (3) – an exciting recent collaboration with Nobel laureate Randy Schekman’s laboratory at the University of California, Berkeley. We have also used TGIRTs to develop a streamlined method for single-stranded DNA sequencing of plasma DNA, which allows us to analyze nucleosome positioning and DNA methylation sites to identify tissues of origin in liquid biopsy applications (4).

TGIRTs allow us to analyze RNAs that are inaccessible to conventional methods and remain untapped as biomarkers for disease. We found that we could simultaneously analyze more than 40,000 different RNA species in as little as 0.5 mL of human plasma – including RNA fragments from large numbers of protein-coding genes and long and small ncRNAs. Surprisingly, we found that both human plasma and exosomes contain full-length sncRNAs – including mature tRNAs. Previous methods could detect only segments of these RNAs because of conventional reverse transcriptase’s inability to copy through highly structured and post-transcriptionally modified RNA species.

A number of the RNA species we encountered had not previously been detected in human plasma. In a collaborative study with Charles Thornton, a physician at the University of Rochester, we found that TGIRTs could reverse-transcribe through regions of RNA containing GC-rich repeat expansions, such as those in myotonic dystrophy and some forms of ALS. This enabled the development of quantitative assays for GC-rich repeat mass in mouse models and could potentially lead to improved diagnostic methods for these diseases (5).

Taking the first steps

Initial studies using TGIRTs for liquid biopsy are currently underway. At the moment, we are focusing on two types of disease: inflammatory breast cancer (with physicians at MD Anderson Cancer Center), and multiple myeloma (with scientists and physicians at City of Hope National Medical Center and the Ohio State University). In these studies, we are using TGIRTs to construct comprehensive RNA-seq libraries for formalin-fixed paraffin-embedded (FFPE) tumor slices, peripheral blood mononuclear cells (PBMCs), and cell-free RNAs in plasma or circulating membrane vesicles from cancer patients. The goal is to identify optimal biomarkers and combinations thereof that correlate with tumor progression and immune responses – and that, we hope, can be used to predict therapeutic response.

We believe that RNA-seq using TGIRTs (TGIRT-seq) is the best method currently available for comprehensive biomarker identification because it can simultaneously analyze a large number of different types of RNA, including RNAs invisible to other methods. In parallel, we are developing TGIRT-based, targeted RNA-seq and multiplex RT-qPCR methods that we hope will enable rapid and relatively inexpensive assay of the most instructive biomarkers for the diagnosis and routine monitoring of disease progression and treatment response. Kevin Ho, a very talented undergraduate researcher in our laboratory, has recently made progress toward developing a single-tube RT-PCR protocol in which the TGIRT initiates reverse transcription of targeted regions of RNAs at high temperature and the resulting cDNAs are then directly amplified by a PCR polymerase. We anticipate that these studies will be completed in the next year or two. We are also interested in extending the research to a variety of other diseases.

Additionally, because TGIRT-based liquid biopsies can be used to comprehensively analyze large numbers of different types of RNAs simultaneously, they could be used as a routine laboratory test in conjunction with annual medical examinations. Such a test would potentially permit the early detection of a large number of different diseases well before symptoms ever appeared.

What will our tests look like when they reach the clinic? Pathologists and laboratory medical professionals would need to draw blood and prepare plasma or PBMCs for RNA extraction – tasks already routine in clinical work. We have, however, noticed considerable variability in the preparation protocols for plasma and PBMCs in different laboratories, as well as in those used for RNA extraction. All of these methods would need to become standardized for any form of liquid biopsy. TGIRT-based RNA and DNA sequencing and data analysis are specialized, big data-driven tasks and could well be delivered by third parties with Clinical Laboratory Improvement Amendments certification and medical genetics expertise. When the information reaches primary care providers, pathologists would again become involved, because there’s always a need for interpretation – helping clinicians and patients understand how the genetic information obtained from liquid biopsies can be used to inform treatment and predict outcomes.

TGIRT-based liquid biopsies have a lot to offer – sensitivity, simplicity, accuracy and a wide variety of applications. It’s our hope that, one day, this new approach could change the face of diagnostic and prognostic liquid biopsy.

Conflict-of-interest disclosure:TGIRT enzymes and methods for their use are the subject of patents and patent applications that have been exclusively licensed by the University of Texas and East Tennessee State University to InGex, LLC. A.M. Lambowitz and the University of Texas are minority equity holders in InGex, and A.M. Lambowitz, the University of Texas, and some present and former members of the Lambowitz lab receive royalty payments from the sale of TGIRT enzymes and kits and from the sublicensing of the intellectual property by InGex to other companies. Y.Q. is a consultant to InGex.