RNA profiles reveal signatures of future health and disease in pregnancy

Maternal morbidity and mortality are increasing, with pre-eclampsia significantly contributing to this burden. Early identification of at-risk pregnancies remains a challenge due to the difficulty in assessing underlying pathophysiology before clinical presentation. Human pregnancy research has ethical constraints, and the unique nature of human gestation limits understanding of pregnancy physiology and pathophysiology. Conventional animal models have limited value due to differences in placentation. Pre-eclampsia, characterized by maternal endothelial dysfunction and hypertension, complicates a significant number of pregnancies and increases the risk of long-term cardiovascular disease. This research explores the use of cell-free RNA (cfRNA) transcripts in maternal blood to characterize normal pregnancy progression and predict pre-eclampsia before diagnosis, independent of clinical variables like race, BMI, and maternal comorbidities.

The literature review section is implicit within the introduction. The authors cite several studies highlighting the challenges in pre-eclampsia prediction, the limitations of animal models, and the need for non-invasive methods to assess pregnancy progression and risk of adverse outcomes. Existing methods for predicting pre-eclampsia rely heavily on clinical factors and maternal characteristics, often leading to inaccuracies and biases. This study aims to overcome these limitations by utilizing cfRNA analysis.

This study utilized a large and diverse dataset of maternal transcriptomes from eight prospectively collected cohorts, comprising 2,539 plasma samples from 1,840 pregnancies across various ethnicities, nationalities, geographic locations, and socioeconomic contexts. The data covered a range of gestational ages. The researchers developed a machine learning model to predict gestational age using cfRNA transcript data. The data was split into training and test sets, stratified by gestational age. Gene counts were standardized across cohorts. A Lasso linear model was used, and model performance was assessed using mean absolute error. The researchers investigated whether including clinical variables (BMI, maternal age, and race) improved model performance using ANOVA. They also explored the association between cfRNA signatures and tissue of origin (fetal organs and maternal tissues) using Gene Ontology and the Molecular Signatures Database. Longitudinal data from one cohort were used to track changes in cfRNA profiles over time. To predict pre-eclampsia, a case-control study was conducted using samples from two independent cohorts. Spearman correlation tests identified cfRNA signatures that distinguished cases from controls. A logistic regression model was used in a leave-one-out cross-validation setup to predict pre-eclampsia risk. Pathway analysis was performed to understand the molecular mechanisms involved. The researchers compared gene expression in groups with chronic or gestational hypertension to the normotensive group to assess the specificity of the pre-eclampsia signature. They also examined the effect of excluding non-case samples with preterm deliveries on pre-eclampsia risk.

The study's key findings include: 1. A cfRNA signature accurately predicted gestational age with a mean absolute error of 1.4 days, comparable to second-trimester ultrasound. Inclusion of clinical variables did not significantly improve accuracy. 2. cfRNA profiles dynamically changed with gestational age, reflecting changes in both maternal and fetal tissues. Gene sets associated with fetal development and maternal tissue growth were identified. 3. A cfRNA signature predicted pre-eclampsia with high sensitivity (82.3%) and specificity, outperforming conventional methods. The model performed well across diverse populations, with clinical variables having negligible impact. 4. Pathway analysis of pre-eclampsia-associated genes revealed upregulation of pathways related to placental blood vessel development and artery remodeling, and downregulation of immune pathways. 5. The pre-eclampsia signature was specific to pre-eclampsia and independent of signals associated with chronic hypertension. 6. The test correctly identified 73% of individuals destined to have a medically indicated preterm birth over 3 months in advance.

The findings address the research question by demonstrating the potential of cfRNA profiles as non-invasive predictors of pregnancy complications. The high sensitivity and specificity of the pre-eclampsia prediction model, coupled with its independence from clinical variables, represent a significant advancement. The identification of specific cfRNA signatures associated with gestational age and pre-eclampsia provides valuable insights into the biological processes involved. This approach offers the potential for personalized assessments of pregnancy risk and could lead to improved interventions and outcomes. The study's strength lies in its large and diverse sample size, reducing bias and increasing generalizability. The focus on molecular mechanisms allows for the stratification of risk independently of potentially biased clinical factors.

This study demonstrates the utility of cfRNA profiles as accurate and unbiased predictors of gestational age and pre-eclampsia risk. The findings suggest a new approach for personalized pregnancy care and improved maternal and neonatal outcomes. Future research could focus on identifying the specific drivers of the identified pathophysiological pathways and developing targeted interventions.

While the study included a large and diverse population, the generalizability to all populations may still have some limitations. Further validation in independent cohorts is needed to confirm the findings. The study primarily focused on pre-eclampsia, and further research is necessary to determine the applicability of cfRNA profiling to other pregnancy complications. The biological mechanisms underlying the observed cfRNA signatures require further investigation.