Metabolomic Biomarkers of Preeclampsia: An NMR Protocol Development and Pilot Study

Abstract

Preeclampsia is a serious complication of pregnancy, which presents risk to both the expectant mother and to the fetus. It affects 3-8% of pregnancies, resolves only upon delivery, and is a major cause of preterm birth. The disease originates in the complex development of the placenta, and the immunological interplay between mother and fetus.

The purpose of this Masters thesis was to implement a high resolution nuclear magnetic resonance protocol for examining urine and serum, in order to study the metabolic signatures of preeclampsia on the biofluids. Following this, a pilot study was carried out on samples from pregnant women with preeclampsia, pregnant women with normal pregnancies and nonpregnant women (controls). Possible discriminatory metabolites were considered. Metabolomic studies on preeclampsia could give insight into the disease pathogenesis, and may provide early predictors of preeclampsia in the future.

For the method development, spot urine and venous blood samples were collected from women pregnant at gestational age 11+0 – 13+6. The samples were analyzed by 1D and 2D NMR spectroscopy and the resulting spectra studied for analytical reproducibility, effects of storage, and different collection and preparation methods. An assessment of the dieretic quantification method was done. For the pilot study, blood and spot urine samples from women with preeclampsia, and age- and gestational age-matched pregnant and nonpregnant controls were collected. The samples were analyzed by NMR and the spectra subjected to unsupervised and supervised multivariate analysis.

The implementation of a protocol for NMR analysis of urine and serum samples gave well resolved and highly reproducible spectra. The pilot study showed differing metabolic profiles between women with preeclampsia, women with normal pregnancies and nonpregnant controls. In the spectra of urine samples, the glycine levels were responsible for a major part of discrimination. In the spectra of serum samples, the general lipid levels and the lipoprotein distribution contributed to the discrimination. Levels of very low density lipoproteins were higher in serum samples from women with preeclampsia than those with normal pregnancies. Statistical models predicted samples from women with preeclampsia with 85–95% accuracy.

NMR produces excellent quality spectra for use in metabolomics. Minimal pretreatment and fast analysis make the method useful for analyzing large quantities of samples. The multivariate methods were able to effectively discriminate between samples from healthy and preeclamtic women in both urine and serum spectra. The specific metabolites responsible for the discrimination were related to disease-specific processes. The NMR metabolomic method clearly has potential for diagnostic use, and possibly future predictive use in preeclampsia.