Evolution of the Umbilical Cord Blood Proteome Across Gestational Development

Abstract

Neonatal health is dependent on early risk stratification, diagnosis, and timely management of potentially devastating conditions, particularly in the setting of prematurity. Many of these conditions are poorly predicted in real-time by clinical data and current diagnostics. Umbilical cord blood may represent a novel source of molecular signatures that provides a window into the state of the fetus at birth. In this study, we comprehensively characterized the cord blood proteome of infants born between 24 to 42 weeks using untargeted mass spectrometry and functional enrichment analysis. We determined that the cord blood proteome at birth varies significantly across gestational development. Proteins that function in structural development and growth (e.g., extracellular matrix organization, lipid particle remodeling, and blood vessel development) are more abundant earlier in gestation. In later gestations, proteins with increased abundance are in immune response and inflammatory pathways, including complements and calcium-binding proteins. Furthermore, these data contribute to the knowledge of the physiologic state of neonates across gestational age, which is crucial to understand as we strive to best support postnatal development in preterm infants, determine mechanisms of pathology causing adverse health outcomes, and develop cord blood biomarkers to help tailor our diagnosis and therapeutics for critical neonatal conditions.

Keywords: biomarker development; cord blood; immune development; neonatal immunology; prematurity; proteomics.

Figure 1.. Total protein concentration in each plasma sample.

Box and whisker plot of protein concentration (ug/uL) distribution across gestational age (GA) categories. The lower and upper ends of each box correspond to the 25th and 75^th percentiles for a given group [shaded area is the interquartile range (IQR)]. The black line in each box is the median. The whiskers represent the largest and smallest observed data points that are no further than +/−1.5 times the IQR, respectively. Points outside of the boundary of the whiskers are outliers. Kruskal-Wallis across GA categories p<0.0001.

Figure 2.. Representative plots of relative protein abundance by gestational age.

The scatter plots show the observed values by sex of newborn. The blue line and shaded blue show the fitted linear model and 95% confidence interval (CI) of the association respectively. The purple dashed line shows a loess smoothed line of the association, and the 95% CI is the shaded gray region (most appeared approximately linear). A) Plasminogen model included n=74 samples. B) Alpha-fetoprotein included n=143 unique samples in the model.

Figure 3.. Volcano plot of protein abundance association with gestational age.

Shown are −log₁₀ of the FDR adjusted P values and betas from linear regression models for a unit increase in continuous gestational age term on a standard deviation increase in protein abundance for a given protein adjusted for sex, preeclampsia, labor route of delivery. Colors show direction of linear associations (positive [blue] indicates increasing GA associated with increasing protein abundance and negative [red] indicates decreasing GA associated with decreasing protein abundance). Seventy proteins were found to be significantly associated with gestational age in adjusted models. Proteins with FDR adjusted p <0.0001 are labelled. The full list of proteins with FDR adjusted p<0.05 can be found in Supplementary Table 1.

Figure 4.. Functional enrichment analysis of proteins in neonatal cord blood that change over gestational age.

A) Heatmap of the protein Z-scores detected in neonatal cord blood samples arranged by gestational age from left to right. Proteins are grouped top to bottom by hierarchical clustering. Functional enrichment analysis and protein-protein interaction networks of proteins significantly B) decreased or C) increased over gestational age are shown below. Network nodes are shaded by abundance change over gestational age with edge width reflecting protein-protein interaction confidence. Significantly enriched pathways are highlighted in colored bar charts to the left; each protein that maps to the identified pathways is indicated by a color-matched box beneath the network node.