NEW YORK (GenomeWeb) – MicroRNAs circulating in mothers' blood may predict outcomes in infants exposed prenatally to alcohol, according to a new study from researchers in the US and Ukraine.
Even with recommendations that advise against drinking during pregnancy, fetal alcohol syndrome is hard to prevent, the researchers noted, as, in the US, more than half of pregnancies are unplanned.
"It's a huge problem," co-senior author Rajesh Miranda from the Texas A&M College of Medicine said in a statement, "but we might not realize the full scope because infants born with normal-looking physical features may be missed, making many cases difficult to diagnose early."
As he and his colleagues reported this week in PLOS One, they found 11 maternal plasma microRNAs that could distinguish infants affected by alcohol exposure from infants that were not affected by exposure or not exposed. They further built a predictor that used microRNA profiles and other factors to classify exposed and affected groups.
In particular, Miranda and his colleagues recruited 68 pregnant women from two sites in Western Ukraine into their study. At enrollment, they asked the women whether they'd consumed alcohol in the month around conception — typically some 17 to 19 weeks prior — as well as if they'd had any alcohol in the previous two weeks of pregnancy, then collected blood samples.
They followed up with the women during their third trimester, and once the babies were born they were assessed for physical and other features of fetal alcohol syndrome.
Using microRNA real-time PCR arrays that examined 752 unique miRNAs, the researchers generated microRNA profiles for the women. In a two-way ANOVA that examined the effect of pregnancy stages and exposure group, Miranda and his colleagues uncovered 11 miRNAs that were significantly elevated in maternal plasma from the heavily prenatally exposed and affected group. They linked a further 21 miRNAs to the group, though they didn't reach as high significance.
Miranda and his colleagues then sought to gauge whether miRNAs could classify members of the exposed but unaffected group as being more like members of the exposed and affected or unexposed groups. They built a random forest model based on a set of 37 miRNAs that varied among group members as well as demographic and clinical data. While demographic factors like smoking and socioeconomic status contributed to the accuracy of the model's predictions, the researchers noted that miRNAs also had a key role.
The model could classify the exposed and affected group and the unexposed group with an overall misclassification rate of 13.3 percent. The model had a lower error rate, 8.7 percent, for the unexposed group and a higher rate, 18.2 percent, for the exposed and affected group.
When they turned this model on the exposed but unaffected group, they classified 17 percent of that group being more like the unexposed group throughout pregnancy and 83 percent to be more like the exposed and affected group during pregnancy.
While most remained stably classified as being like the exposed and affected group throughout pregnancy, a few moved toward unexposed-like classification toward the end of pregnancy, suggesting that there could be unknown resilience or protective factors.
Miranda and his colleagues noted that this signature could be uncovered during the second trimester of pregnancy, providing time for early maternal-fetal interventions.
"Early diagnosis is important because it permits early intervention to minimize the harm due to prenatal alcohol exposure," Wladimir Wertelecki, the research team leader in Ukraine, added. "Good nutrition, better perinatal health care, lowering stress levels, and infant care interventions can all improve the outcome of alcohol-affected pregnancies."
The researchers further noted that the circulating maternal miRNAs they identified could, down the line, be targets for interventions.