Comprehensive Circulating DNA Profiling in Maternal Plasma Enabled the Identification of a Molecular Signature for the Detection of Fetal Growth Disorders at the First Pregnancy Trimester

Background: Early diagnosis, close follow-up and timely delivery constitute the main elements for the appropriate detection and management of pregnancies with fetal growth disorders (FGD). Besides the unprecedented technological development in screening of fetal growth abnormalities based on fetal biometry by ultrasound, inaccuracies still hinder FGD detection and assessment especially at early stage. Hence, there is an unmet need for reliable prenatal biomarkers of fetal growth, which can be assessed using low-invasive approaches. We hypothesize that the dysregulation of epigenetic mechanisms underly the etiology of FGD in infants impacting the shedding of genetic material into maternal blood. As a corollary, we propose that markers in circulating fetal DNA (cfDNA) in maternal blood can be used as early detection markers of FGDs. Methods: 56 pregnant women were recruited in this study. Maternal blood sample was prospectively collected from each subject at first trimester and plasma circulating DNA (cirDNA) was isolated. The number of Small, Large and Adequate for Gestational Age (SGA n=11, LGA n=18, and AGA n=27, respectively) were determined at birth according to weight and gestational age. The total amount of plasma cirDNA in each subject was quantified using qPCR. cirDNA fragmentation was assessed using specific qPCR assays corresponding to intact, nucleosome-size, and fragmented DNA, respectively. The mitochondrial/nuclear cirDNA was calculated using specific qPCR assays. cirDNA methylation profiles were studied in 10 genes associated with placental homeostasis. Furthermore, machine learning approaches were applied to build a molecular signature for the prediction of LGA and SGA occurrence using first trimester samples. Prediction accuracy was assessed by Receiving-Operating Curve (ROC) analysis and Positive and Negative Predictive values (PPV and NPV, respectively) were calculated. Results: We observed an increase in the total concentration of plasma cirDNA in SGA (mean cirDNA=9.40 ± 5.49 ng/mL) and LGA (mean cirDNA=12.36 ± 4.24 ng/mL) compared to AGA pregnancies (mean cirDNA=8.48 ± 3.16 ng/mL), although the differences were not statistically significant (p=0283; Kruskal Wallis test). The fraction of fragmented DNA was significantly increased (p=0.002) in SGA (mean DNA fragmentation index, DFI=2.38 ± 0.14), and SGA (mean DFI=2.47± 0.25) compared with AGA pregnancies (mean DFI=1.56 ± 0.15). Likewise, the ratio of mitochondrial/nuclear DNA was significantly increased (p=0.005) in LGA (mean FC=0.78 ± 0.31) and LGA (mean FC=1.20 ± 0.32 compared with AGA pregnancies (mean FC=0.23 ± 0.05). DNA methylation profiles also shown distinctive patterns. Out of the 10 selected loci, we detected 4 genes showing significant differential methylation differences (p<0.05) across the SGA, LGA and SGA samples at first trimester. Using a machine learning approach, we combined these molecular and epigenetic cirDNA markers in a single signature which discriminates between FGD and AGA pregnancies with high accuracy (AUC=0.93) In addition, the performance of the signature was further evaluated according to positive predictive and negative predictive values (PPV and NPV, respectively), achieving 93% PPV and 82% NPV. Conclusions: Our findings demonstrate that cirDNA epigenetic variables in maternal blood cirDNA can serve as FGD markers. Our novel marker panels will enable accurate prediction of FGDs using a low invasive approach that can be implemented as early as the first gestational trimester. The application of these panels in a clinical setting holds the potential to enable a disruptive path toward precision medicine in FGD.