Nearly a year after the US Food and Drug Administration (FDA) withdrew its approval of a progesterone drug to treat preterm birth due to a lack of efficacy, March of Dimes scientists may have discovered why the drug only works for some women.
In a recently published paper in Science Advances, a group of March of Dimes-funded scientists from the Stanford University Prematurity Research Center (PRC) and the University of California San Francisco (UCSF) used Machine Learning (ML) to find that common mutations in uterine muscle genes are likely the reason many women don’t have a favorable response to progesterone to slow or prevent preterm birth.
The team also used computational drug repurposing to find and validate five existing drugs they showed could be used to treat preterm birth, based on drug interactions with 315 new genes they found to be associated with spontaneous preterm birth.
Taken together, the three-fold discovery—that there are hundreds of new genes associated with preterm birth, that a woman’s personal genome dictates her progesterone response, and that existing drugs can be applied to slow or prevent preterm birth—add a wealth of understanding to the mystery of early labor and a powerful boost of ammunition in the fight to treat it.
“We’ve been advocating for personalized medicine for years, and this work only strengthens the evidence behind that call,” said senior study author Dr. Jingjing Li, an Associate Professor at the University of California, San Francisco’s Department of Neurology and a close collaborator of the Stanford PRC, a March of Dimes-funded group that led the effort. “Although the medical community suspected that genetic diversity could be a reason why there were such inconsistent clinical outcomes with progesterone therapy, this work now suggests that mutations in uterine muscle genes in each personal genome is likely a factor.”
“Knowing this, it’s hard not to see the benefit of personalized medicine based on a patient’s personal genome baseline—a goal that’s well within reach and could make an enormous difference for women at risk of preterm birth.”
While exciting, the discovery regarding genetic determinants of response to progesterone therapy, along with the discovery of a potential role for pre-existing drugs to treat preterm birth by reducing labor contractions, must still be validated in clinical trials to ensure the findings are true in larger, more diverse patient populations.
“We are excited to validate our findings and move closer to the realm of personalized medicine and reliable therapeutics and interventions to treat preterm birth,” said Dr. Cheng Wang, a post-doctoral fellow in Dr. Li’s UCSF genomics laboratory who was the study’s first author.
Along with Drs. Cheng and Li, Stanford PRC investigators Drs. David Stevenson and Gary Shaw contributed substantially to the work, especially with study design and patient recruitment.
While the study consisted of two separate parts—the progesterone-response analysis and the drug analysis—it was made possible with the same newly uncovered genetic data from the team.
First, the group’s ML model identified 1,079 new genetic variants that are associated with spontaneous preterm birth. Mapping the genomic variants onto genes, the researchers identified 315 genes in preterm birth linked to uterine contractions and inflammatory response regulation.
With that foundation, the team enrolled into the study a group of pregnant women of European and African American descent with a history of preterm birth. All of these women received progesterone in an effort to slow or prevent preterm birth with their current pregnancy.
From these women, the researchers studied the responders (those whose current pregnancies did not result in a preterm birth) and non-responders (those who went on to have a preterm birth) to progesterone therapy, and examined whether genomic mutations in the identified genes displayed different patterns between the two groups.
“We observed that individuals not responding to progesterone therapy tended to have increased mutation levels, specifically in the identified smooth muscle genes,” Dr. Li said. “This finding suggests that the different [progesterone] responses were likely resultant from different mutational effects on these genes regulating smooth muscle contractility.”
“This is further evidence that we need a precision framework for future therapeutic development, where drug efficacy should be evaluated at a personal level rather than at the level of population averages.”
The second part of the study aimed to expand therapeutic options by identifying new small molecules that could be used for treating preterm birth. Specifically, the researchers asked whether existing and pre-approved drugs could be repurposed to treat preterm birth.
After examining about 4,000 compounds using computational drug repurposing, the team prioritized 10 drugs that show the strongest interaction with the identified new genes in preterm birth. Of those 10, they experimentally validated five that significantly reduced uterine muscle contractions. They plan to test those five in a clinical trial with pregnant women in the near future to ensure efficacy and hone in on dosing.
While Dr. Li admitted the experimental validation of the drugs was based on cell models and it’s possible the new drugs may not work on all women, due to each woman’s unique genetic makeup (the same way progesterone has not worked for all women), he said the team’s cell-based experiments were promising.
“Our work has shown that the top ranked drugs can indeed reprogram smooth muscle contractility in human uterine cells, so we are hopeful these drugs could be used as new treatment options for spontaneous preterm birth,” Dr. Li said.
In addition to clinical trials to validate the progesterone response findings and the drug repurposing work, Dr. Li is also working on identifying the “genetic architecture in preterm birth that results from placental dysfunction.”