Lipidomic profiling of amniotic fluid and its application in fetal lung maturity prediction

Background The pulmonary surfactant especially lipids in amniotic fluid can reflect the development stage of fetal lung maturity (FLM). However, the conventional lecithin/sphingomyelin (L/S) ratio method by thin layer chromatography (TLC) is insufficient and inconvenient for FLM prediction in clinical practice. Methods The amniotic fluid samples were collected from the pregnant women in labor or undergoing amniocentesis and analyzed for its lipid contents with the liquid chromatography coupled with high‐resolution mass spectrometry (LC‐HRMS) method and the lamellar body count (LBC) method. To reveal the lipidomic profiling of different FLM stages, three groups of amniotic fluid samples including 8 from premature group (gestational week (GW) < 37), 10 from mature group (GW < 37), and 10 from mature group (GW > 38) were compared with the control group (n = 6) of 18 GWs separately. Results In the FLM prediction study, the sensitivity of the LC‐HRMS method and LBC method was 91% and 73%, respectively; the specificity was 100% and 95%, respectively. The most significant metabolic pathway was linoleic acid metabolism between the premature group and the control group. Both glycerophospholipid metabolism and glycosylphosphatidylinositol‐anchor biosynthesis were enriched in the mature groups. In search of potential FLM prediction markers in amniotic fluid, 8 phosphatidylcholines, 1 sphingomyelin, and 1 phosphatidylethanolamine were significantly increased in the mature groups compared with the premature group. Conclusion An efficient LC‐HRMS method for L/S ratio in predicting FLM was established. The linoleic acid metabolism may play an important role in the fetal lung development.


| INTRODUC TI ON
Respiratory distress syndrome (RDS) is a major cause of morbidity and mortality in preterm infants. The deficiency of pulmonary surfactant resulting in lung prematurity is the most common cause of RDS in newborns. 1 The incidence of newborn RDS increases with decreasing gestational age as the lungs are the final fetal organs to mature. 2 Prediction for fetal lung maturity (FLM) plays an important role in the prevention of RDS in the preterm newborns. The conventional tests for FLM are based on the assessment of the amount of surfactant in the amniotic fluid, which results from the exchange of lipids between the developing lungs and the amniotic fluid. 3  during the late pregnancy. 3 However, this method is time-consuming, imprecise, and susceptible to blood and meconium interference. 4 PG is the last lipid to increase in fetal lung surfactants, and there are two methods currently available to detect PG: the quantitative TLC and the qualitative agglutination. The main advantage of PG method is that it is not affected by blood and meconium contamination. However, its relatively high false-positive rate presents a hurdle for wide application in clinical practice. 5 The principle of the S/A method is based on the fact that the fluorescent polarization is high when the dye binds to albumin and low when the dye binds to the lung surfactants. This method was once widely used in clinical prediction of FLM. Unfortunately, the instrument and reagents of the S/A assay were discontinued since 2011 and the interference caused by blood and meconium cannot be ignored. 4,5 Lamellar bodies are secreted into the alveolar space from type II pneumocytes and are further transported into the amniotic cavity. Although the similar size between lamellar bodies and platelets makes the automated hematological cell counters suitable for LBC with its platelet channel, the lack of appropriate quality control reagents and universal threshold value limits its application. 4,6 Lipidomics is a newly emerged research focus which studies cellular lipids on a large scale using advanced analytical technological tools, such as mass spectrometry. Lipidomics aims to study the structures and the functions of the complete set of lipids in a specific cell or organism as well as their interactions with other cellular components to elucidate the pathways and networks in biological systems. 7 As the existing laboratory methods of FLM are mostly based on the lipid components in amniotic fluid, we proposed that the lipidomics did not only encourage the establishment of an efficient L/S ratio method by liquid chromatography coupled with high-resolution mass spectrometry (LC-HRMS), but could also reveal the metabolic profiling difference in amniotic fluid between the fetal lung mature group and the premature group.

| Subjects
Thirty-three amniotic fluid samples without blood or meconium contamination were collected from the pregnant women who were in labor or undergoing amniocentesis in Beijing Obstetrics and Gynecology Hospital from January 2018 to December 2018. They were used in the comparison between the LC-HRMS-based L/S ratio method and the in-house LBC method.
Further, another 34 amniotic fluid samples were collected from pregnant women in labor of different gestational weeks (GWs) and were tested in the targeted lipidomics study. The FLM was judged by the Apgar scores of the newborns. The term "premature" referred to the newborns with penalty points for unsatisfactory or disturbed respiratory function, while the "mature" referred to the newborns with full Apgar scores. Of the 34 amniotic fluid samples, 6 samples were from 18 GWs, 8 premature samples from <37 GWs, 10 mature samples from <37 GWs, and 10 mature samples from >38 GWs.

| Amniotic fluid preparation
The amniotic fluid sample processing steps before the LC-HRMS were After ultrasonic treatment for 1 hour, 100 μL water was added and mixed thoroughly. Then, 300 μL supernatant was transferred to an empty vial and dried with nitrogen gas followed by reconstitution in 200 μL of isopropanol/acetonitrile (1:1, v/v) and LC-HRMS analysis.

| Data processing
The targeted lipid library applied in the metabolomic study was

| LBC method
The 3 mL of amniotic fluid from each participant was inverted for 1 minute before centrifugation (276 × g) for 5 minutes. Then, the supernatant was detected for LBC on the platelet channel of SYSMEX XN-2000/3000 automatic blood cell analyzer (Kobe, Japan) following the manufacturer's standard operation procedure for platelet counting.

| LC-HRMS method validation
In the validation study for the L/S ratio determination, 20 μL of all the amniotic fluid (AF) samples was mixed evenly and used as the quality control in the following steps. Briefly, the pure methanol solution and subsequent quality controls were used to balance the chromatographic column. In the actual experiments, every 5 AF sample injections were immediately followed by a quality control sample. The average coefficient of variations (CVs) of PC and SM were calculated for the precision evaluation. The cluster of the quality control samples in the principal component analysis (PCA) score scatter plot was used to present the overall stability and repeatability of this lipidomic analysis.

| Comparison of L/S ratio method and the LBC method
The LC-HRMS L/S ratio results and LBC results of 33 amniotic fluid samples (11 premature and 22 mature samples by Apgar scores) were compared. The clinical specificity and sensitivity were calculated for these two methods.

| Lipidomic profiling of amniotic fluid samples
As the 37 GWs is the cutoff used to define a preterm labor, 11 and to explore the changings of lipids in amniotic fluid with the development of fetal lung, three groups of amniotic fluid samples (8 from preterm newborns (<37 GWs) with premature lungs, 10 from preterm newborns (<37 GWs) with mature lungs, and 10 from term newborns (>38 GWs) with mature lungs) were compared with the control group (n = 6) of 18 GWs separately. Their corresponding significant lipid metabolic pathways were analyzed.

| Lipid biomarkers for predicting FLM in amniotic fluid
The FLM mature amniotic fluid samples were grouped together and compared with the premature group to identify any potential lipid biomarkers for predicting FLM. In this study, P value of t test < 0.05, VIP > 1.1, and fold change > 2 were used to search for significantly different lipid metabolites.

| Statistical analysis
For the pathway enrichment analysis, the P value was calculated by Holm-Bonferroni method and the whole process was performed on the MetaboAnalyst 3.0 (http://www.metab oanal yst.ca). For the study of potential lipid biomarkers for predicting FLM, the p value was calculated by t test method and the whole process was performed on the SIMCA 14.1 software (Umetrics AB, Umea, Sweden).

| Validation of the L/S ratio method by LC-HRMS
The L/S ratio assay by LC-HRMS was a semi-quantitative method calibrated with the internal standards that are exogenous compounds used to assess and adjust the stability of the analytical methods. The

| Method comparison of the L/S ratio with the LC-HRMS and the LBC
Of the 33 samples, 11 samples were premature and 22 samples were mature determined by Apgar scores. Ten out of the 11 premature samples and all the 22 mature samples were accurately predicted for FLM by the LC-HRMS L/S ratio method using the cutoff value of 10.0 that was previously reported in a mass spectrometry FLM study. 12 In comparison, with the cutoff value of <50 × 10 9 /L, 13 the LBC method was able to make correct diagnosis for 8 of 11 premature samples and for 21 of 22 mature samples (Table 1). As a result, the sensitivity of LC-HRMS L/S ratio method and LBC method was 91% and 73%, respectively; the specificity of LC-HRMS L/S ratio method and LBC method was 100% and 95%, respectively ( Table 2).

| The lipid metabolic profiling of amniotic fluid from different FLM and GWs
In the pathway enrichment analysis, the impact scores were a series of normalized results obtained from the pathway topology analysis, which indicate the location and importance of the significant me-  Table 3).

| Lipid biomarkers for predicting FLM in amniotic fluid
Four major lipid classes including PC, PE, SM, and Cer were detected and compared in the mature groups (GW < 37 or GW > 38) and the premature group (GW < 37). Of the top ten significant lipids identified, there were 8 PCs, 1 SM, and 1 PE of which the fold changes were ranged from 2.26 to 6.74. (Table 4).

| D ISCUSS I ON
As the fetal lung contributes to the formation of the amniotic fluid, its composition may well reflect the stage of fetal lung development. 14 The lung surfactants consist of approximately 90% lipids in which lecithin and PG are the most abundant (76%-86% and 6%-13%, respectively). The sphingomyelin accounts for approximately 2% of total lipids and keeps constant during the late pregnancy. 4 The lecithin is synthesized from 28 GWs with a rapid rise around 36 GWs and continues to increase until delivery. 4 The L/S ratio by TLC was the first biochemical test of assessing FLM development in F I G U R E 1 The principal component analysis (PCA) score plot of the 6 quality control (QC) and amniotic fluid samples. QC samples were applied to exhibit the stability of the LC-HRMS system. The cluster of the QC samples in the PCA score scatter plot showed a satisfactory stability and repeatability of this lipidomic analysis approach. The colors display the subjects from different groups. The green dots represent QC samples, and the blue dots represent amniotic fluid samples the early 1970s and was considered as the gold standard in the past few decades. 4 Nowadays, this testing method was much less commonly used in clinical laboratories due to the technical barriers of TLC that is often inaccurate, time-consuming, and labor intensive. 5 Alternatively, there had been some studies about the measurement of the L/S ratio by fast atomic bombardment mass spectrometry. 15,16 Later, Kwak and his colleagues established new methods to measure the L/S ratio on liquid chromatography-tandem mass spectrometry   (Table 2), although larger patient cohort is warranted to further validate the cutoff values with the LC-HRMS method.

F I G U R E 2
The pathway analysis of lipid metabolism among amniotic fluid of different FLM development status, presenting metabolic pathways arranged according to the scores from enrichment analysis (y-axis) and from topology analysis (x-axis). The darkness of the circles stands for the statistical significance in the corresponding pathway, and the radius represents the centrality measures. The pathway analysis details were elaborated in Table 4   infectivity. 22 Besides, PC(16:0/18:1) has been reported to be significantly increased in thyroid papillary and breast cancers compared with healthy controls. 23,24 However, as the studies of the above lipid candidates in FLM prediction are rather scarce, further studies with larger samples size are required to verify their clinical values in such area.
To decrease the incidence of RDS of preterm infants, lung promotion drugs such as corticosteroids are usually be given to mother before the planned delivery. 25 Nevertheless, there may be long-term adverse consequences of antenatal exposure to corticosteroids.
Studies have shown that exposure to excess corticosteroids before birth may be linked to impaired fetal neurodevelopment, cardiovascular disease, and type 2 diabetes. 25 Therefore, it is beneficial to timely and accurately predict FLM and avoid excessive pulmonary promotion drugs for mothers before delivery.
In this study, an efficient and accurate LC-HRMS method was es-