SAN FRANCISCO (GenomeWeb) – Researchers from Stanford University have found a preliminary RNA signature that may predict preterm birth and that can be detected noninvasively in blood.
For a pilot study published today in Science, the researchers sequenced cell-free RNA from the blood of a cohort of pregnant women, including some who delivered prematurely and others who delivered at full term, and identified seven differentially expressed transcripts. When they evaluated expression of those transcripts in a third group, they found that they were able to predict which women would deliver prematurely up to two months before birth.
The next step, according to co-lead author Mira Moufarrej, a graduate student in Stephen Quake's laboratory at Stanford, is to validate the signature in a larger cohort. "These are preliminary results in a small cohort," she said, "but it's encouraging that we see that changes in cell-free RNA are predictive."
If further research validates the findings, the goal is to develop a noninvasive test that could both predict both preterm birth and establish gestational age. Such a test would likely be based on PCR, since only a handful of RNA transcripts would need to be evaluated and because it would keep costs low, Moufarrej said.
The Stanford team has also filed a patent application on the method.
The study builds on previous work in which the researchers had analyzed blood samples from a cohort of 31 pregnant women throughout the course of their pregnancies and foundthat levels of certain cell-free RNA transcripts changed over time. Moufarrej said that work had identified transcripts associated with fetal tissue, immune-related genes, and the placenta. Monitoring quantitative changes of those transcripts over time enables researchers to establish the gestational age, she said. For instance, the level of transcripts associated with fetal liver tissue increases throughout pregnancy as the fetus develops and then drops after birth. Similarly, transcripts associated with placental genes also increase throughout the course of pregnancy and then decline after birth.
Tracking those changes over time enabled the team to develop an algorithm that could determine gestational age. They found that the algorithm could estimate it within 14 days with an accuracy of 45 percent in trimesters two and three. By contrast, ultrasound has an accuracy of around 48 percent in the first trimester.
In a second portion of the study, the researchers sought to see whether cell-free RNA could also predict preterm birth. Women were recruited from two cohorts of pregnant women known to be at risk of preterm birth due to premature contractions. The researchers performed RNA sequencing on seven women who ultimately delivered at full term and eight women who did delivered prematurely and identified 38 genes with differential expression. Next, they created a quantitative reverse transcription PCR panel to measure the 38 cell-free RNA transcripts and confirmed that the differential expression could still be detected. Next, the team developed a classifier to identify women at risk for preterm birth and found that seven of those transcripts were the most predictive, accurately classifying six out of eight of the women who gave birth prematurely and misclassifying one of the 26 women who gave birth at full term.
Finally, the team validated the test on a set of 23 samples collected up to two months prior to birth. They found it predicted four out of the five preterm births and misclassified three of the 18 full-term births are premature births.
Louise Laurent, director of perinatal research in the department of reproductive medicine at the University of California, San Diego, who was not affiliated with the study but who has also worked on developing better predictors of preterm birth, said that she thought the method "does have potential," but noted that the sample sizes in the study were very small.
Laurent noted that the Stanford team would have to demonstrate its test in a larger cohort of women that is not enriched for those at high risk and show how it compares to a currently available mass spectrometry-based test that measures two serum proteins — IBH4 and SHBG. In the Science study, the Stanford researchers said that their cfRNA test was comparable in accuracy to Sera Prognostics' mass spec test, but Laurent said that it is too early to make such a comparison, given the small sample size and the fact that the samples tested by the Stanford team were all from women considered to be at higher risk for preterm birth. By contrast, Sera validated its test in more than 5,000 women. Laurent noted that she was a participating clinical site in Sera's validation trial, but said that she was not involved in data analysis for the study.
"If it panned out that [the new test] was accurate in identifying preterm birth well ahead of any symptoms, that would be very useful," she said. "And, if that could also be correlated with time to delivery, that would be extremely valuable."
She also noted that for any preterm birth diagnostic test to be widely adopted, it would have to show that it improved outcomes. There are currently prophylactic therapies available, Laurent said, so a better diagnostic could inform who to treat, but such clinical utility studies have not yet been conducted. In addition, she said, testing could help reassure women who have had a prior preterm birth. Also, molecular biomarkers of risk could aid in future drug discovery and development, both to be able to identify a sub-population more likely to respond and to potentially identify targets, she added.
Around 15 million babies are born prematurely globally each year, according to the World Health Organization, and preterm birth is the most common pregnancy complication. It can be difficult to predict in advance who is at risk, however. Women who've had previous preterm births or who experience premature contractions are considered to be at risk, but not all do give birth prematurely.
Laurent and other researchers are also working on identifying predictive biomarkers of preterm birth. Laurent's lab has been evaluating circulating cell-free microRNAs as potential biomarkers, for example. She said her group chose to focus on microRNAs since those appear to be enriched in cell-free RNA, potentially making them a more robust analyte. "There's definitely some similarities," between her work and the Stanford team's, she said.
Researchers at the Lunenfeld-Tanenbaum Research Institute in Toronto are also working to develop a preterm birth diagnostic test based on RNA sequencing. However, rather than a universal screening test, that team is focused on distinguishing between false preterm labor and actual preterm labor, since women often present with symptoms of labor but do not actually go into labor.
Moufarrej said that the Stanford group's next steps are to do a larger study. The team is interested in sequencing cell-free RNA across a larger cohort of women to see whether there may be additional transcripts that are predictive and whether the transcripts found in this study hold up.
If the findings are confirmed in a larger study, she said, the group would be interested in developing a clinical test. The team would also like to study whether circulating cell-free RNA could track fetal development. "You could imagine being able to track development of different fetal organs over time and understanding at a molecular levels what's happening," she said.