NEW YORK – Researchers from Stanford University are planning to commercialize a recently published method for measuring gene expression in blood through the use of cell-free RNA.
Veteran liquid biopsy researchers Max Diehn and Ash Alizadeh led the development of RARE-seq (random priming and affinity capture of cfRNA fragments for enrichment analysis by sequencing), a process that took the better part of a decade to complete.
"If you just take sort of off-the-shelf, RNA-seq kits and apply them to cell-free RNA, you can generate data, but you don't get very meaningful data," Diehn said. That's because they're not optimized for the "sensitivity that's required given the low concentrations of the transcripts you're usually interested in."
They made improvements along the entire workflow, he said, from the sample collection tubes to bioinformatic analysis, where a computational method for removing RNA signatures from platelets in blood proved key.
Using RARE-seq helped the team identify a set of nearly 5,000 so-called "rare abundance genes" (RAGs) which they built into a hybrid capture panel developed with Twist Bioscience. In proof-of-concept studies, the researchers suggested that their RAG panel is approximately 50 times more sensitive than a whole-transcriptome approach — even when using RARE-seq — for analyzing tumor RNA in blood.
In a small study of patients with non-small cell lung cancer who had developed resistance to EGFR tyrosine kinase inhibitors, they were able to detect gene expression profiles associated with histological transformation to small cell lung cancer. Other studies suggested the ability to inform tumor tissue of origin and even to provide information for noncancerous lung conditions.
Along with researchers from Massachusetts General Hospital, Memorial Sloan Kettering Cancer Center, Fred Hutchinson Cancer Center, and South Korea's Sungkyunkwan University, they published their results Wednesday in Nature.
"Accurately measuring cell-free RNA has been notoriously challenging," said Jacob Berchuck, a medical oncologist and liquid biopsy researcher at Emory University. Researchers have been driven by the "transformative" nature of liquid biopsies in obtaining tumor information, he said, but have often resorted to measuring "indirect surrogates" of gene expression. He helped develop a method for cell-free chromatin immunoprecipitation sequencing (ChIP-seq) and is on the scientific advisory board of Precede Biosciences, a firm commercializing that approach.
"What's exciting to me about this paper is the proof of concept of new technology that can accurately and directly measure tumor gene expression, noninvasively."
Diehn and Alizadeh are planning a spinout based on RARE-seq. They have already commercialized two other liquid biopsy technologies in this way. They cofounded CAPP Medical, a spinout offering cancer profiling by sequencing circulating tumor DNA, which was acquired by Roche in 2015 for an undisclosed amount.
They also developed PhasEd-seq, a ctDNA sequencing method with applications in minimal residual disease testing that targets so-called "phased variants" — two or more SNVs located on the same DNA fragment. PhasED-seq is being commercialized by Foresight Diagnostics, which closed a $73.8 million Series B financing round last summer.
"We wanted to go after mRNAs because they're the most therapeutically and clinically important part of the RNA world," Diehn said. "If you could do that, then you can leverage the knowledgebase that's been built around [them]: different gene expression pathways, different resistance mechanisms that are based on RNA signatures, identification of drug targets, which of course are generally protein coding genes, which are coded by the mRNAs."
One technical hurdle was dealing with RNA from platelet cells, a "major confounder" for obtaining clean data from cfRNA. "That has never been an issue for the cell-free DNA field, because platelets don't have DNA," Diehn said. Most plasma preparation protocols using a centrifuge don't spin hard enough to get rid of the platelets. "If you compare, say, different banked samples from different speed spin protocols, you get massive differences in the amount of platelet signatures, such that most of the RNA that you're isolating from the plasma can often be from platelets, not the cell-free compartment."
Spinning samples properly only partially solved the problem, though, because frozen samples can lyse platelets, spilling their RNA into the rest of the sample. The solution turned out to be at the bioinformatics stage. "We can, in effect, subtract the platelet gene expression from any sample and then reconstitute the expression matrix as if there had been no platelets in the sample," Diehn said. "And that then lets you really uncover the transcripts that are coming from all the solid organs."
Alizadeh said they don't have an itemized breakdown of pricing but estimated that the cost of consumables per samples was "on the order of high hundreds of dollars," and excluded fixed costs of equipment and labor. They used Illumina sequencing in their study.
One potential application is to make liquid biopsies in oncology more sensitive. In their paper, they ran RARE-seq for matched cell-free DNA samples. For every sample where the cancer was detectable by DNA, it was also detectable with RNA, Diehn, said; moreover, about a third of the cancers were only detectable with RNA. "For tumor-naïve testing, RNA may be more sensitive," he said, noting that this would require more study.
The histologic transformation in NSCLC is a case where "resistance is driven by changes in gene expression," said Aaron Hata, a researcher at Massachusetts General Hospital and a coauthor on the Nature paper. "Because RARE-seq has potential to read out changes in gene expression, this allows us to see resistance that is not mutations."
One next step is to see how well it works when one doesn't already know what to look for. Hata is using RARE-seq on expanded cohorts of EGFR patients and hopes to use it to develop new gene expression signatures relevant to cancer treatment that have not yet been discovered.
How RARE-seq will compare with other assays that have been developed to look at non-mutational signatures of cancer remains to be seen. "A major limitation in the field is that there haven't been head-to-head or comparative studies to understand how the various technologies provide similar or distinct information," Berchuck said. "There are now several platforms [such as] cell-free ChIP-seq, fragmentomics, and now cfRNA that likely will be able to achieve or provide similar insights into tumor gene regulation."
"And as a nascent tech, it remains to be seen how RARE-seq will augment the current capabilities of liquid biopsy," he said.