Lipidomic signatures in patients with early-onset and late-onset Preeclampsia

Abstract

Background: Preeclampsia is a pregnancy-specific clinical syndrome and can be subdivided into early-onset preeclampsia (EOPE) and late-onset preeclampsia (LOPE) according to the gestational age of delivery. Patients with preeclampsia have aberrant lipid metabolism. This study aims to compare serum lipid profiles of normal pregnant women with EOPE or LOPE and screening potential biomarkers to diagnose EOPE or LOPE.

Methods: Twenty normal pregnant controls (NC), 19 EOPE, and 19 LOPE were recruited in this study. Untargeted lipidomics based on ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) was used to compare their serum lipid profiles.

Results: The lipid metabolism profiles significantly differ among the NC, EOPE, and LOPE. Compared to the NC, there were 256 and 275 distinct lipids in the EOPE and LOPE, respectively. Furthermore, there were 42 different lipids between the LOPE and EOPE, of which eight were significantly associated with fetal birth weight and maternal urine protein. The five lipids that both differed in the EOPE and LOPE were DGTS (16:3/16:3), LPC (20:3), LPC (22:6), LPE (22:6), PC (18:5e/4:0), and a combination of them were a potential biomarker for predicting EOPE or LOPE. The receiver operating characteristic analysis revealed that the diagnostic power of the combination for distinguishing the EOPE from the NC and for distinguishing the LOPE from the NC can reach 1.000 and 0.992, respectively. The association between the lipid modules and clinical characteristics of EOPE and LOPE was investigated by the weighted gene co-expression network analysis (WGCNA). The results demonstrated that the main different metabolism pathway between the EOPE and LOPE was enriched in glycerophospholipid metabolism.

Conclusions: Lipid metabolism disorders may be a potential mechanism of the pathogenesis of preeclampsia. Lipid metabolites have the potential to serve as biomarkers in patients with EOPE or LOPE. Furthermore, lipid metabolites correlate with clinical severity indicators for patients with EOPE and LOPE, including fetal birth weight and maternal urine protein levels.

Keywords: Biomarkers; Early-onset preeclampsia; Late-onset preeclampsia; Lipidomics; Preeclampsia.

Fig. 1

LC-MS/MS lipidomics analysis of NC, EOPE, and LOPE. a ~ c. The OPLS-DA model of the NC, EOPE and LOPE groups. The red dots represent NC, the blue dots represent EOPE, and the purple dots represent LOPE. d ~ f. Scatter plot of OPLS-DA model and validation model of permutation test between two groups. a, d. NC vs. EOPE. b, e. NC vs. LOPE c, f. EOPE vs. LOPE. OPLS-DA, orthogonal partial least square discriminant analysis; NC, normal control group; EOPE, early-onset preeclampsia; LOPE, late-onset preeclampsia

Fig. 2

Serum lipid species differed among the NC, EOPE, and LOPE groups. The volcano plots (a), pie charts (b), and (c) show differential lipids, differential lipid species composition, and corresponding subclasses between the EOPE and the NC groups, respectively. The volcano plots (d), pie charts (e), and (f) show differential lipids, differential lipid species composition, and corresponding subclasses between the LOPE and the NC groups, respectively. The volcano plots (g), pie charts (h), and (i) show differential lipids, differential lipid species composition, and corresponding subclasses between the LOPE and the EOPE groups, respectively

Fig. 3

The matchstick plots reveal the top 20 lipids significantly up-regulated and down-regulated among the NC, EOPE, and LOPE groups (a) The top 20 significantly changed lipids between the NC group and the EOPE group. (b) The top 20 significantly changed lipids between the NC group and the LOPE group. (c) The top 20 significantly changed lipids between the EOPE group and the LOPE group

Fig. 4

Clinical characteristics of the EOPE and the LOPE group correlate to distinct lipid network modules. (a) Module clustering trees were used to visualize the distribution of lipids of the EOPE group and the LOPE group in each module. (b) The correlation coefficient and p-value of lipid modules with the clinical characteristics of the EOPE and the LOPE groups. (c) The number of lipid metabolites species in the yellow module. (d) Metabolic pathways analysis of hub lipids

Fig. 5

Screening the potential lipid biomarkers for EOPE and LOPE diagnosis. (a) Venn plot depicts significant lipid numbers among the NC, EOPE, and LOPE groups. (b) Five different lipids distinguish the EOPE group from the NC group. (c) Five different lipids distinguish the LOPE group and the NC group. (d) Five different lipids distinguish the LOPE group and the EOPE group. e-i. The relative concentration of DGTS (16:3/16:3), LPC (20:3), LPC (22:6), LPE (22:6), and PC (18:5e/4:0) in the NC, EOPE, and LOPE groups, respectively

Fig. 6

Correlations between specific differential maternal serum lipids and clinical parameters of patients in the EOPE and the LOPE groups. (a) Scatterplots depict the correlation between lipids and fetal birth weight. (b) Scatterplots show the correlation between lipids and maternal urine protein