Omega-6 and Omega-3 Fatty Acid-Derived Oxylipins from the Lipoxygenase Pathway in Maternal and Umbilical Cord Plasma at Delivery and Their Relationship with Infant Growth

Omega-3 and omega-6 fatty acids are important for neonatal development and health. One mechanism by which omega-3 and omega-6 fatty acids exert their effects is through their metabolism into oxylipins and specialized pro-resolving mediators. However, the influence of oxylipins on fetal growth is not well understood. Therefore, the objective of this study was to identify oxylipins present in maternal and umbilical cord plasma and investigate their relationship with infant growth. Liquid chromatography–tandem mass spectrometry was used to quantify oxylipin levels in plasma collected at the time of delivery. Spearman’s correlations highlighted significant correlations between metabolite levels and infant growth. They were then adjusted for maternal obesity (normal body mass index (BMI: ≤30 kg/m2) vs. obese BMI (>30 kg/m2) and smoking status (never vs. current/former smoker) using linear regression modeling. A p-value < 0.05 was considered statistically significant. Our study demonstrated a diverse panel of oxylipins from the lipoxygenase pathway present at the time of delivery. In addition, both omega-3 and omega-6 oxylipins demonstrated potential influences on the birth length and weight percentiles. The oxylipins present during pregnancy may influence fetal growth and development, suggesting potential metabolites to be used as biomarkers for infant outcomes.


Introduction
Pregnancy is a period of rapid fetal growth and cell differentiation in the womb, where diverse insults lead to acute diseases with untoward long-term consequences. The fetus is The median maternal daily dietary intake of total n-6 and n-3 fatty acids was 14.46 g/day [7.432] and 1.72 g/day [1.01], respectively. Total n-3 fatty acid intake includes fish oil supplementation, ALA, EPA, and DHA. However, ALA is minimally converted into EPA and DHA. Twenty percent of mothers reported fish oil supplementation during pregnancy.

Infant Demographics
Infants were, on average, born at 38.2 weeks (3.29) of gestation (Table 3). They weighed 3.2 kg at birth and measured at 34 cm for head circumference and 49 cm for birth length. Infants were in the~50th percentile for birth weight, length, and head circumference. There was no significant difference between the number of male vs. female infants included in this study.

Association between Metabolite Levels and Baseline Characteristics
Cord plasma n-6 metabolite levels differed significantly when delivery modes were compared. Umbilical cord samples from CS vs. vaginal deliveries had lower median levels of 9-HODE (7.26 nM vs. 10.52 nM; p = 0.010), 9-HOTrE (0.35 nM vs. 0.52 nM; p = 0.003), and 13-KODE (1.05 nM vs. 1.52 nM, p = 0.039). Further, umbilical cord plasma 5,15-DiHETE was higher in CS compared to vaginal deliveries (median: 0.24 nM vs. 0.14 nM; p = 0.036). There were no significant differences in plasma metabolite levels and infant sex. To unveil potential relationships between eicosanoid levels and outcomes, LTB4, 9-HEPE, and 8,15-DiHETE were classified as detectable and nondetectable. There was a significant difference in AA intake by subjects with detectable vs. nondetectable maternal LTB4 levels. A maternal intake of AA around 0.29 g/day was associated with detectable LTB4 in maternal plasma. Further, there was a relationship between maternal LTB4 levels and birth weight percentile. The mean birth weight percentile was 59.71 % (n = 103) in the nondetectable LTB4 group compared with 37.72% (n = 13) in the detectable LTB4 cohort. There appeared to be a relationship between BMI and the ability to detect LTB4 levels. Higher BMI status was associated with few samples with detectable LTB4 levels.

Correlations for Parent Nutrients and n-6 Metabolites in Maternal and Umbilical Cord Plasma
The relationships between omega-6 parent FAs and metabolites were explored ( Table 4). As maternal AA levels increased, cord 15-HETE and lipoxin A4 were negatively correlated. Appendix A illustrates the median concentrations of the parent FAs analyzed in this study. Table 4. Correlations between maternal plasma levels of omega-6 parent nutrients (µg/mL) and omega-6 maternal and umbilical cord metabolite levels.

Maternal Parent Nutrient
Metabolite n r s p-Value  To understand the presence of n-6 oxylipins at delivery, we explored the correlations between maternal and cord plasma. We observed significant correlations in the levels of n-6 PUFA metabolites in maternal and cord plasma. AA and DGLA's metabolites were higher in cord than maternal plasma, whereas LA precursors tended to be lower in cord vs. maternal specimens (Table 5).  Table 6 illustrates the correlations between maternal parent omega-3 FA nutrients with omega-3 metabolites in maternal and umbilical cord blood. We found that maternal EPA in plasma had a significant, negative correlation with maternal 17-HDHA. The same relationship was found between maternal EPA levels and cord 17-HDHA. Maternal parent nutrient concentrations are displayed in Appendix B. We identified EPA and DHA metabolites and found that these lipids tended to be higher in maternal than cord plasma. Table 7 shows maternal and cord metabolite levels and displays the significant, positive correlations between maternal and infant metabolites. ALA metabolite 9-HOTrE was not significantly correlated between maternal and cord plasma.

Metabolite Association with Maternal Diet
The influence of maternal dietary intake was analyzed to understand the impact of diet on metabolite levels in maternal and cord plasma. Total n-6 PUFA intake was positively correlated with cord 15-HETE levels. In contrast, the maternal intake ratio of n-6 to n-3 was negatively correlated with maternal 15-HETE (AA metabolite) and 17-HDHA (DHA metabolite). When analyzing the influence of DHA intake on metabolite levels, we found that DHA intake was significantly positively correlated with maternal 9-HEPE, 7-HDHA, and 17-HDHA plasma levels. Total omega-3 fatty acid intake, however, was only significantly positively correlated with cord 17-HDHA. Significantly lower median n-6 metabolite concentrations were observed in individuals who used fish oil supplements, including cord 15-HETrE (median: 1.05 nM vs. 1.50 nM, p = 0.032) and maternal 5,15-DiHETE (median: 0.08 nM vs. 0.28 nM, p = 0.027), both n-6 metabolites. Table 8 shows significant correlations between dietary intakes and metabolite plasma levels. The influence of n-6 FA metabolites on infant growth was analyzed using Spearman correlations (Table 9). Significant relationships were then adjusted for maternal smoking status and pre-pregnancy BMI. Cord 9-HODE, 13-HODE, and 13-KODE predicted birth length percentile. With every 1% increase in cord 9-HODE, 13-HODE, and 13-KODE, infant birth length percentile increased by 0.13, 0.20, and 0.025, respectively. However, the infant birth weight percentile decreased by 0.041 for every 1% increase in maternal 5,15-DiHETE. Figure 1 illustrates the relationship between the significant metabolites in the LOX pathway.

n-3 FA Metabolites and Infant Growth
We assessed potential relationships between infant growth and n-3 FA metabolites using Spearman correlations (Table 10). We found that the n-3 metabolites 5-HEPE in cord and maternal specimens, respectively, were associated with infant growth after adjustment for maternal smoking and pre-pregnancy BMI status. For every 1% increase in median cord 5-HEPE (EPA metabolite), the birth length percentile increased by about 0.12. In contrast, the expected birth weight percentile decreased by 0.062 for every 1% increase in maternal 7-HDHA. Figure 2 highlights the significant metabolites in the LOX pathway. No other significant relationships between n-3 metabolites and infant growth were identified.

Discussion
The main findings of our study highlight the relationship between LOX pathway metabolites in maternal and cord plasma and the impact these metabolites have on infant growth metrics. Metabolites 5-HEPE, 9-HODE, 13-HODE, and 13-KODE positively predicted infant birth anthropometrics, while 7-HDHA and 5,15-DiHETE negatively predicted infant growth. Our study is the first to explore the relationship between n-6 and n-

Discussion
The main findings of our study highlight the relationship between LOX pathway metabolites in maternal and cord plasma and the impact these metabolites have on infant growth metrics. Metabolites 5-HEPE, 9-HODE, 13-HODE, and 13-KODE positively predicted infant birth anthropometrics, while 7-HDHA and 5,15-DiHETE negatively predicted infant growth. Our study is the first to explore the relationship between n-6 and n-3 LOX pathway metabolite exposure in utero and fetal growth outcomes at the time of delivery.

n-6 PUFA
AA metabolites play essential roles in the continued maintenance of pregnancy and delivery. Interestingly, we identified a trend toward higher circulating AA metabolite levels in cord plasma compared to maternal cord plasma. In addition, cord 15-HETE and lipoxin A4 were negatively associated with AA in maternal plasma. Previous studies reported decreased parent nutrient AA in maternal plasma accompanied by increases in cord plasma throughout pregnancy. Best et al. identified a similar trend in AA metabolites; this group reported lower 8-, 9-, 11-, 12-, and 15-HETE levels in maternal blood spots at 34 vs. 14 weeks of gestation. In contrast, maternal 5-HETE levels increase significantly throughout pregnancy [24] and have been shown to induce uterine contractility [25]. HETEs have essential functions: they function as PPAR ligands due to their binding capacity to nuclear receptor transcription factors. Additionally, they participate in protein kinase signaling activation, angiogenesis, and neuronal apoptosis [21]. Further, the ability of 12-and 15-HETE to activate the capsaicin-sensitive receptor correlates with pain signaling during inflammation [21]. We found similar increases in AA metabolites in cord vs. maternal plasma at the time of delivery. Arachidonic acid concentrations, the precursor for HETE metabolites, are much higher in fetal blood and tissues compared to maternal blood. However, there is little evidence that placental chain elongation and desaturation account for the observed increases in AA supplied to the fetus [26].
The impact of AA metabolites on pregnancy outcomes remains to be precisely defined. Both beneficial and adverse effects have been reported. Most studies investigated potential relationships between metabolite levels in maternal plasma collected before delivery and maternal/infant outcomes. Goveia-Figueira et al. found that 5-and 15-HETE levels were associated with premature labor [27]. Similarly, Ramsden et al. showed that above-median 5-HETE and 15-HETE concentrations were associated with a higher risk of preterm delivery when assessed at gestational week 14 [28]. Maternal plasma samples from women with pre-eclampsia had significantly higher 5-, 8-, 12-, 15-HETE, and LTB4 levels compared with women having normal pregnancies and nonpregnant women. Moreover, 5-HETE, 15-HETE, and LTB4 levels were significantly higher in plasma from women with severe vs. mild pre-eclampsia.
Few studies have addressed the importance of LOX metabolites in maternal and cord plasma on infant growth. Welch et al. used Bayesian modeling to demonstrate an association between the adjusted mean concentration of 12-HETE and size for gestational age. Among gestational age cases who were small, the adjusted mean concentration of maternal 12-HETE was 56.2% higher than that observed in mothers whose infants were classified as appropriate for gestational age [29]. We found no predictive value for maternal or cord 12-HETE on infant growth metrics. Still, we observed that maternal plasma levels of the AA metabolite 5,15-DiHETE were negatively associated with infant birth weight percentile at birth.
The impact of cord plasma AA metabolites on infant outcomes is not well understood. A previous study found that in pre-eclamptic pregnancies, umbilical cord constriction is dependent on 15-HETE levels [30]. In our research, cord AA metabolites were not significantly associated with infant growth, but we observed a trend toward lower LA metabolite levels in cord vs. maternal plasma. LA is an essential n-6 fatty acid metabolized into HODEs through the LOX pathway. HODEs, as with HETEs, are generally present in healthy pregnancies [21]; they function as PPAR ligands and have atherogenic properties. HODE activity is modulated by Th1 and Th2 lymphocytes with cytokine and chemokine participation, either pro-or anti-inflammatory, depending on the type of lymphocytes present [31]. Best et al. demonstrated that maternal 13-HODE significantly decreases throughout normal pregnancy, suggesting that less metabolite is transferred to the developing fetus [24]. On the other hand, pregnancies with maternal 9-HODE levels below the median concentration tended to have a higher risk of spontaneous preterm birth [28]. Our results show that cord 13-HODE, 9-HODE, and 13-KODE are positively associated with birth length percentile. Together, our experimental results lead us to propose that AA and LA metabolites likely regulate events necessary for normal delivery and potentially influence fetal growth and blood vessel development through their role as PPAR ligands. In addition, elevated HETEs appear to be associated with pregnancy pathologies such as pre-eclampsia, impaired infant growth, and preterm delivery.

n-3 PUFA
EPA and DHA are n-3 PUFAs that compete with n-6 fatty acids for LOX pathway enzymes to produce a class of less potent anti-inflammatory specialized pro-resolving mediators (SPMs) [32]. EPA is converted to HEPEs and further metabolized into the SPM Resolvin E-series. HEPEs induce neutrophil chemotaxis, inhibit platelet aggregation, and play a role in mouse adipogenesis [33]. The resolvin E-series inhibit the further inflammation and release of cytokines and are involved in the resolution of inflammatory responses. A previous study reported higher EPA, SPMs, and metabolites in infant cord blood after maternal dietary n-3 supplementation [24]. However, these studies did not evaluate the relationship between metabolite levels and infant outcomes. In our research, EPA 5-HEPE levels in maternal and cord plasma were significantly correlated, but maternal plasma levels of EPA were not significantly correlated with EPA metabolite levels. Further, 5-HEPE positively predicted infant length percentile at birth. It remains to be established whether these correlations reflect the active participation of 5-HEPE in fetal outcomes. Mechanistic analyses have shown that 5-HEPE enhances glucose-dependent insulin secretion in mouse insulinoma cells and human intestinal carcinoma cells [33,34]. It is tempting to hypothesize that 5-HEPE is a fetal growth factor that regulates infant growth through its effects on insulin secretion.
Many DHA-derived resolvins are synthesized through metabolic steps catalyzed by 15-LOX and 5-LOX. 17-HDHA is a 15-LOX intermediate and a pathway marker for the D-series resolvins. Mozurkewich et al. demonstrated that umbilical cord 17-HDHA was significantly increased compared with maternal levels at 12-20 and 34-36 weeks of gestation [23]. In these investigations, maternal DHA levels were negatively correlated with maternal and cord 17-HDHA levels. Best et al. observed a significant decrease in 7-HDHA levels at 34 vs. 14 weeks of pregnancy [24]. Our study supported this trend as 7-HDHA at delivery was lower in cord plasma compared to maternal plasma. Interestingly, we found that maternal 7-HDHA levels were negatively associated with infant birth weight percentile. Best et al. found an increased risk of spontaneous preterm birth with higher levels of 7-HDHA in mothers who received supplementation and had higher n-3 baseline levels. This association was not seen in women who received supplementation and had moderate to low n-3 status [24]. A synthetic 7-HDHA has been identified as a PPAR ligand, potentially highlighting a mechanistic pathway for 7-HDHA to influence infant growth [35].

Study Characteristics
The study's eligibility for participation included mothers ≥19 years of age, delivering a live infant at Nebraska Medicine. All participating maternal-infant pairs were free of renal, metabolic, and hepatic diseases known to impair normal nutrient metabolism. Written informed consent was obtained from all subjects after admission to the labor and delivery unit. The UNMC's Institutional Review Board approved the study.

Maternal Dietary Intake
All participants completed the Harvard Food Frequency Questionnaire (FFQ) that assesses dietary intake over the previous year. An advantage of FFQ compared to other nutrient intake assessments (e.g., 24 h recalls) is that this survey evaluates intake over time and can be used during pregnancy [36,37]. De-identified surveys were sent to the Harvard T. H. Chan School of Public Health for quantification based on previously established food nutrient content. Notably, the survey assessed supplement use and reported nutrient intake with and without supplementation. This tool has been validated in numerous adult populations, including pregnant women [36].

Maternal and Infant Outcomes
We used participants' electronic medical records (EMRs) to extract relevant information and metrics, including maternal and infant outcomes, birth weight and length, head circumference, and associated percentiles (Table 11). Maternal smoking status and body mass index (BMI) were chosen as covariates due to previously reported associations between these parameters and inflammation, nutrient status, and infant outcomes [38,39].

Blood Sample Collection
Maternal and umbilical cord blood samples were collected by trained nurse personnel upon maternal admission to the hospital and at delivery. The research team obtained the samples within 12 h of collection, processed the specimens, and stored plasma fractions at −80 • C until metabolite analysis.

Quantification Methods for n-3 and n-6 FA Oxylipin
The levels of lipoxygenase pathway metabolites derived from linoleic acid (LA), dihomo-γ-linolenic acid (DGLA), arachidonic acid (AA), alpha-linolenic acid (ALA), eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) were quantified in maternal and umbilical cord plasma using liquid chromatography-tandem mass spectrometric (LC-MS) approaches. Mediators assessed included those formed through lipoxygenase actions (including 5-, 12-, and 15-LOX) with or without additional modifications from other enzymes (e.g., COX metabolism)- Figure 3. Tables 12 and 13 list the parent n-3 and n-6 FA nutrients and their metabolites included in the lipid panel.     FAs were extracted as described by Yang et al. [40]. Waters Oasis-HLB cartridges (30 mg/30 µm) were treated with ethyl acetate (1 mL), methanol (2 × 1 mL), and 95:5 v/v water/methanol containing 0.1% acetic acid (1 mL). We mixed 100 µL of plasma with 7 µL of internal standard solution (stock concentration: 500 nM), 10 µL of butylated hydroxytoluene (BHT; stock concentration: 2 mg/mL), and 120 µL of H 2 O:methanol (MeOH) (95:5) containing 0.1% acetic acid. The resulting samples (240 µL) were then loaded onto pre-treated cartridges and washed twice with 750 µL of H 2 O/MeOH (95:5) containing 0.1% acetic acid. The aqueous plug was pulled from the cartridges using a high vacuum, and the cartridges were then dried further under a low vacuum for about 20 min. Waters Oasis-HLB cartridges were eluted into tubes with 250 µL of methanol followed by 1 mL of ethyl acetate into 2 mL tubes containing 6 µL of 30% glycerol in MeOH, a trap solution. The samples were dried under nitrogen and then dissolved in 70 µL of methanol containing 20 nM of 1-cyclohexyl-dodecanoic acid urea. The samples were then vortexed for 5 min, transferred to autosampler vials with low-volume inserts, and stored at −20 • C until further analysis.

Liquid Chromatography-Mass Spectrometry
Samples (3 µL extracts) were analyzed by liquid chromatography coupled to electrospray ionization on a triple quadrupole mass spectrometer (Waters XEVO TQ-XS); the autosampler was cooled to 4 • C. Chromatographic separation was achieved on an Ascentis Express column (2.1 × 150 mm, 2.7 µm particles; Sigma-Aldrich Supelco, Darmstadt, Germany) at a flow rate of 0.35 mL/min at 40 • C using a gradient of 0.1% acetic acid and acetonitrile-isopropanol (90:10 v/v), as described [41]. All standards were dissolved in MeOH to achieve a final concentration of 1 µM. A standard preparation was used to build a 12-point calibration curve (97 pM-200 nM). Each point included 50 nM of deuterated internal standards. Individual calibration curves were generated by plotting oxylipin standard concentration vs. calculated response ratio (i.e., the ratio of oxylipin standard peak area and the corresponding internal standard peak area).
Data were processed using Skyline [42] software and Microsoft Excel. Calibration curves for each oxylipin were built using standards and deuterated internal standards. Calibration curves were calculated by linear regression with 1/x 2 weighing. The calibration result was then corrected for dilution to determine the original oxylipin concentration in plasma. Results falling above or below the calibration curve were excluded from data analysis. With 60% or more of their samples above or below the standard curve, oxylipins were categorized as not detectable (ND). Appendix B includes the oxylipin library used and a detailed list of multiple reaction monitoring mode transitions. The chromatographs for the oxylipins are included in Appendix C.

Fatty Acid Quantification in Maternal Plasma
Gas chromatography with flame ionization detection (GC-FID) was used to analyze the plasma fatty acid composition at OmegaQuant Analytics LLC (Sioux Falls, SD, USA). FAs were extracted as described previously; in brief, plasma was mixed with boron trifluoride-methanol solution, vortexed, and heated in a hot bath at 100 • C [43]. Hexane and high-performance liquid chromatography-grade water was added after the solution was cooled. The tubes were centrifuged and the hexane layer was transferred to a GC vial. GC was performed using a GC-2010 Gas Chromatograph (Shimadzu Corporation, Columbia, MD, USA). A standard mixture of fatty acids (GLC OQ-A, NuCheck Prep, Elysian, MN, USA) was compared to the fatty acids of interest.

Statistics
For mothers who had twins, only information from the first twin was considered in our analysis. Descriptive statistics included means, standard deviations, medians, interquartile range (IQR), minimums and maximums for continuous data, and counts and percentages for categorical data. The Mann-Whitney U test was used to compare continuous measures between dichotomous groups. The Kruskal-Wallis test was used to compare measures between more than two groups. Fisher's exact test was used to associate dichotomous categorical variables. Spearman correlation coefficients were used to assess relationships between continuous variables. The eicosanoids LTB4, 9-HEPE, and 8,15-DiHETE were detected in less than 40% of the samples. These metabolites were dichotomized to detectable vs. nondetectable to identify significant differences between the groups. McNemar's test was used to assess potential associations between matching maternal-cord dichotomized metabolite values (i.e., ND vs. Determinate).
Linear regression modeling was performed on metabolites that correlated with birth weight, length, and head circumference at the p < 0.05 level in the univariate analysis. Associations of the metabolites with these birth measures were adjusted for obesity (>30 or ≤30 BMI (kg/m 2 )) and maternal smoking status (current/former vs. never) in the models. Metabolites were log-transformed in the regression analyses to meet the statistical assumptions of the models. p < 0.05 was considered statistically significant.

Conclusions
This study highlights the presence and potential influence of LOX pathway metabolites of both n-6 and n-3 FAs during pregnancy. Further, we demonstrate that LOX pathway metabolites potentially influence infant growth outcomes at delivery. Our study highlights the presence and potential influence of LOX pathway metabolites of both n-6 and n-3 PUFA on infant growth outcomes at delivery. Omega-6 and n-3 FAs have not been described in detail in maternal and cord plasma at the time of delivery. Therefore, this study adds to the breadth of metabolites described during pregnancy and related to infant growth. A limitation of this study is not identifying the direct causation of these eicosanoids on infant growth. Further studies are needed to understand the change in eicosanoids across pregnancy and their role in fetal development.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the study's design, in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish results.

Appendix B
The mass spectrometer was operated in multiple reaction monitoring (MRM) mode, and electrospray ionization was performed in negative ion mode. A detailed list of MRM transitions is presented in Table A2. Source and desolvation temperatures were 150 • C and 500 • C, respectively. Desolvation gas was set to 1000 L/h and cone gas to 150 L/h. Collision gas was set to 0.15 mL/min. All gases were nitrogen except the collision gas, which was argon. The capillary voltage was 2 kV in negative ion mode. Samples were analyzed in random order. A quality control sample was analyzed every eight injections to monitor system stability and performance. We achieved 100% recovery of the internal standards. Table A2. Library of 52 targeted oxylipins with their corresponding retention time, product and metabolite m/z, and deuterated internal standard used for quantification.

Appendix C
The chromatographs from the metabolites analyzed in this study are provided below.

Appendix C
The chromatographs from the metabolites analyzed in this study are provided below.