α-Tocopherol Stereoisomer Profiles in Matched Human Maternal and Umbilical Cord Plasma

ABSTRACT Background α-Tocopherol (αT) is essential for fetal development. One study has shown that the human placenta preferentially transfers the natural stereoisomer, RRR-αT. But prenatal supplements generally contain synthetic αT (S-αT). Objectives We aimed to determine if umbilical cord plasma is enriched for RRR-αT in racially diverse neonates from both uncomplicated and complicated pregnancies and if cord RRR-αT enrichment is impacted by maternal αT stereoisomer profile. Methods We measured αT and αT stereoisomers in plasma from a randomly selected subset of 66 predominantly black and Hispanic maternal-fetal pairs from the Camden Study involving control (n = 28) and complicated pregnancies (n = 38). We collected maternal plasma at study entry (week 16 gestation; w16) and week 28 gestation (w28) and cord plasma at birth. Results RRR-αT was the predominant stereoisomer in all maternal and cord plasma samples, but S-αT stereoisomers were found in most samples and comprised a high percentage of αT in some maternal-neonate pairs. Cord plasma had a higher percentage RRR-αT (P < 0.05) and lower percentage S-αT (P < 0.0001) than w28 plasma. Pregnancy status did not impact maternal or cord plasma concentrations of αT, RRR-αT, or S-αT; except plasma from complicated pregnancies was higher in S-αT at w28 than at w16 (P < 0.05). Maternal w28 αT did not correlate with cord αT. However, both maternal w28 αT and S-αT positively correlated with both cord S-αT (r = 0.340, P = 0.0049; r = 0.538, P < 0.00001) and percentage S-αT (r = 0.399, P = 0.001; r = 0.786, P < 0.00001) but negatively correlated with cord percentage RRR-αT (r = −0.399, P = 0.0009; r = −0.786, P < 0.00001). Conclusions The proportion of RRR-αT was higher in cord compared with maternal plasma in both uncomplicated and complicated pregnancies. Our data suggest that maternal S-αT raises cord S-αT and decreases the proportion of RRR-αT in the neonatal circulation. Because the bioactivities of RRR-αT and S-αT differ, this warrants future research to determine the importance of our observations to neonatal αT status.


Introduction
α-Tocopherol (αT) is the only tocopherol structural isomer (α-, β-, δ-, and γ -tocopherol) that can meet vitamin E requirements (1). Supplementation studies in humans with vitamin E deficiency reveal that αT is critical to the function of the nervous system (2)(3)(4). Maternal αT insufficiency in early pregnancy leads to fetal resorption in rodents (5) and is associated with a higher risk of miscarriage in women (6).
The fetus acquires αT from maternal circulation via transplacental transport during gestation (7). It has been clearly established that cord plasma αT concentrations are significantly lower than, and generally do not correlate with, concentrations in matched maternal plasma (8)(9)(10)(11)(12)(13)(14)(15). It remains unclear whether this reflects poor placental αT transport or other fetal or placental metabolic effects. Maternal supplements are routinely supplemented with αT to improve maternal status to ensure that the developing fetus has adequate access to αT for normal development.
αT used in maternal supplements can be isolated from plant oils like soy oil, or it can be chemically synthesized. The stereoisomeric composition of αT derived from these processes is different. αT isolated from plant oils exists as a single stereoisomer, RRR-αT (2R, 4 R, 8 R-αT), commonly referred to as natural vitamin E. In contrast, synthetic αT (all rac-αT; S-αT) is a racemic mixture of the 8 possible αT stereoiso- mers equally divided amongst 2R (RRR, RRS, RSR, RSS) and 2S (SSS, SSR, SRS, SRR) stereoisomers (1). S-αT has between 1.36-and 2-fold less vitamin E activity than an equal amount of RRR-αT (16). Acuff et al. (17) found that when pregnant mothers were administered an equimolar mixture of RRR-αT and S-αT, the ratio of RRR-αT to S-αT in the fetal circulation was nearly double that of maternal circulation indicating that the fetal-placental unit selectively concentrated RRR-αT. This study was, however, small (n = 15 maternal-infant dyads) and individual αT stereoisomers were not analyzed. Because pregnant women are generally given supplements containing S-αT, 2 important questions were raised. First, is the preferential concentration of RRR-αT in umbilical cord plasma reported by Acuff et al. (17) also found in a larger, more diverse population with variable αT intake, potentially high maternal S-αT intake, and pregnancy complications? Secondly, in such a population, how much S-αT is found in umbilical cord plasma and is the concentration driven by the concentration of S-αT in maternal plasma? Our objective was to provide insight into these questions by analyzing individual αT stereoisomer concentrations in plasma from a subset of maternal-infant dyads from the Camden Study (18), a large, racially diverse population that included normal and high-risk pregnancies.

Study subjects
The current data were generated from a subset of 66 mother-infant pairs enrolled between January 1998 and April 2005 as part of the Camden Study (18), which was designed to study the effects of maternal nutrition in pregnant women from Camden, New Jersey, USA. The characteristics of 1231 pairs, which include the 66 pairs studied here, were previously provided in detail by Scholl et al. (19). The Institutional Review Board of the University of Medicine and Dentistry of New Jersey, School of Osteopathic Medicine (which later became Rowan University School of Osteopathic Medicine in 2013) approved the study procedures. Descriptive data for the 66 pairs studied here are presented in Table 1 combined and by pregnancy status: control or complicated. The 66 pairs were selected because each had an available plasma sample for maternal weeks 16 (w16) and 28 (w28) of gestation and an umbilical cord plasma sample. Demographic information and maternal weight and height were obtained at w16 and w28 of gestation as previously described (19). Most of the study population was either black or Hispanic, with the combination accounting for ∼85% of the study population. Most of the mothers were single, experiencing their first pregnancy, were high school educated, did not smoke, and were not working. Pregnancy complications included preterm delivery (delivered <37 wk of gestation), small for gestational age (birth weight <10th percentile), low birth weight (birth weight <2500 g), and pre-eclampsia. Large for gestational age was defined by a birth weight >90th percentile of Zhang's standard that adjusts for maternal parity, ethnicity, and fetal sex (19), but was not categorized as a pregnancy complication. Information on current pregnancy outcome, complications, and infant abnormalities was abstracted from the prenatal record, the delivery record, delivery logbooks, and the infant's chart. Most of the mothers did not consume a maternal supplement at study w16, but the majority reported taking a maternal supplement at w28 of gestation. Supplemental αT intake was not different between mothers with control pregnancies and those with complicated pregnancies and changed from 4-5 mg/d at w16 to ∼14 mg/d during w28 of gestation. Infant sex and birth anthropometrics are presented in Table 2 for the 66 pairs studied here. More than half the infants were female, and they were distributed evenly into the control and complicated pregnancies ( Table 2). Gestation duration was based on the gravid woman's last normal menstrual period confirmed or modified by ultrasound scan. Gestational age averaged 38.5 wk and was similar for control and complicated pregnancies. However, birth weight was significantly lower for complicated pregnancies compared with control pregnancies. In contrast, birth length and head circumference were similar for control and complicated pregnancies. Samples of the mother's blood were obtained at w16 of the study and at w28 gestation, aliquoted, and stored at −80 • C until analysis.

Analysis of αT and its stereoisomers
αT and its stereoisomers were assessed as we described previously (20). In brief, samples were extracted with hexane following saponification as described (21). A portion of the hexane was dried and reconstituted to measure αT by HPLC with electrochemical chemical detection (ECD) as described (20). αT was quantified at the dominant oxidation potential relative to an external αT standard (Sigma) that was validated against certified reference material (National Institute of Standards and Technology SRM 968f). To assess αT stereoisomers, the remaining portion of the aforementioned hexane extract was used to measure the percentage distribution of αT stereoisomers as we described (20). In brief, the hexane extract was dried under nitrogen gas, resolubilized, and the reconstituted sample was methylated under basic conditions prior to extracting with hexane. Samples were then separated and detected by HPLC with fluorescence detection using a chiral separation column and excitation/emission settings of (290 nm /330 nm ). Under these conditions, each specific 2R stereoisomer of αT (RRR-, RRS-, RSR-, and RSS-αT) was determined along with a single peak for total 2S stereoisomers. Peak area of each stereoisomer was calculated to determine percentage distribution, and their molar concentration determined based on the concentration of total α-T obtained by HPLC-ECD. The detection limit was ∼0.02 μmol/L for S stereoisomers and 0.05 μmol/L for R stereoisomers.

Statistics
All data were analyzed with GraphPad Prism version 5.04 for Windows (GraphPad Software; www.graphpad.com). Normality and homogeneity of variance were evaluated using the D' Agostino-Pearson test using a cutoff of −2 and 2 before further testing. Prior to comparison, data that were not normal were log transformed (Y + 1) if zero was a value, or log transformed (Y) if not. Comparisons of birth anthropometrics between control and complicated pregnancies and maternal w16 plasma, maternal w28 plasma, and cord plasma αT values were done using 1-factor ANOVA, and when appropriate, differences between means were eval- stereoisomers. a-c Labeled means within a row for μmol αT/L or % αT without a common lowercase superscript letter differ: a,c; b,c P < 0.0001; a,b P < 0.05 by 1-factor ANOVA and Tukey post hoc test. Data were normalized using log(y). αT, α-tocopherol.
uated using Tukey multiple comparison test. Means that differed significantly following Tukey multiple comparison test are denoted using unique letter subscripts (P < 0.05). Pearson correlation test was used to test for significant correlations. A false discovery rate analysis using a 1% rate was used to adjust for multiple comparisons within correlation matrices. Unless indicated otherwise, data are expressed as means ± SD of the mean.
Umbilical cord plasma had a higher mean percentage RRR-αT than did maternal w28 plasma, but not w16 maternal plasma (Table 3). Percentage RRR-αT ranged from a minimum of 24% and 34% in maternal and cord plasma, respectively, to a maximum of 100% in both. Umbilical cord plasma had a lower mean percentage S-αT than maternal plasma at either w16 or w28. The percentage S-αT ranged from 0% to as high as 76% and 66% in maternal and cord plasma, respectively. Individual data for plasma percentage S-αT are provided in Figure 1 and reveal that the majority of maternal and umbilical cord samples contained S-αT. Approximately one-quarter to one-third of plasma samples had >20% S-αT, an arbitrary cutoff value for illustrative purposes. Consistent with these observations, umbilical cord plasma had a higher mean RRR-αT/S-αT ratio (9.59 ± 8.86; range: 0.523-40.2) than w28 maternal plasma (6.63 ± 5.22; range: 0.368-26.0) (P < 0.001).
Control and complicated pregnancies were not different for maternal plasma αT, RRR-αT, or S-αT concentrations ( Table 4). Whereas ma-ternal values generally did not change from w16 to w28, mean S-αT concentration for w28 complicated pregnancies was higher than complicated pregnancy w16 S-αT (P < 0.05). There were no differences in αT, RRR-αT, or S-αT concentrations in cord plasma from control or complicated pregnancies (Table 4).
Maternal w28 plasma total αT concentration did not correlate with cord plasma total αT or RRR-αT concentration, but did correlate positively with both cord plasma S-αT concentration and percentage S-αT ( Table 5). Maternal plasma total αT concentration negatively correlated with cord plasma percentage RRR-αT. In contrast, the concentration of RRR-αT in maternal w28 plasma was not related to the concentration of αT, RRR-αT, or S-αT in cord plasma. Maternal w28 plasma S-αT concentration correlated positively with both cord plasma S-αT and percentage S-αT, and negatively with cord percentage RRR-αT. Maternal w28 plasma αT and S-αT correlations with individual S-αT stereoisomers in umbilical cord are shown in Supplemental Table 2.

Discussion
We measured the αT stereoisomer profile in maternal plasma and matched umbilical cord plasma in a cohort that was racially diverse and included both complicated and uncomplicated pregnancies. We found that RRR-αT was the most common αT stereoisomer in all maternal and umbilical cord plasma samples analyzed, and that the proportion of RRR-αT was higher in cord plasma than in w28 maternal plasma. This latter finding is consistent with that of Acuff et al. (17), who found in a small study involving deuterated αT that the fetal placental unit concentrated RRR-αT compared with maternal plasma. Because our study population was heavily weighted with minority and low socioeconomic/poorly educated mothers together with complicated pregnancies, our findings suggest that the enrichment of fetal circulation with RRR-αT is a general phenomenon in humans. This is likely explained by the presence of α-TPP at the placental interface between the maternal and fetal circulation (22). α-TPP preferentially binds RRR-αT compared with other stereoisomers of αT (16).
Our study population included pregnancies with and without complications. Complications included preterm delivery, small for gestational age, pre-eclampsia, and gestational diabetes. Pregnancy status had no effect on maternal and umbilical cord plasma αT measures except that maternal w28 plasma samples had a higher concentration of S-αT. We cannot explain this difference, but speculate it might reflect differential adherence to prenatal supplement use.
Despite the predominance of RRR-αT in all plasma samples studied, we also generally found measurable and variable concentrations of S-αT in plasma. More than three-quarters of both maternal and cord samples contained measurable concentrations of S-αT. Most maternal and umbilical cord plasma samples had between 10% and 40% of αT as S-αT. One maternal-neonate pair had more than two-thirds S-αT, which might be explained by a specific nucleotide polymorphism previously reported for α-tocopherol transfer protein (αTTP) (23). Prenatal maternal supplements in the United States generally contain S-αT and generally provide a higher daily intake of αT than diet. Indeed, maternal plasma concentrations of αT increased concomitant with increased reports of supplement use. The increase in maternal plasma αT appeared to have been driven by increased concentrations of plasma S-αT. Therefore, we consider it likely that prenatal vitamin consumption explains the widespread presence of S-αT in the plasma samples studied. Consistent with this, maternal plasma αT and S-αT concentrations were positively correlated with S-αT in umbilical cord plasma. This suggests that despite the presence of αTTP in the placenta, a significant proportion of the fetal-placental units studied could not fully compensate for maternal plasma S-αT concentrations. This in turn led to lower concentrations of RRR-αT based on the negative correlation found between maternal plasma S-αT and umbilical cord RRR-αT. Taken together, this suggests that increasing maternal αT status through prenatal supplement use is likely to be associated with a decrease in umbilical cord plasma RRR-αT despite a selective placental bias for RRR-αT.
Our data are not entirely consistent with the interpretation by Acuff et al. (17) of their findings. They speculated that their results were numerically consistent with the selective transport of RRR-αT and 1 S-αT stereoisomer by the fetal-placental unit. Here, 35 of the 66 cord plasma samples studied had >1 S-αT stereoisomer. Our data are consistent with increased tissue proportions of S-αT stereoisomers in animals that received doses of S-αT that exceeded the dose of RRR-αT (24)(25)(26).
Consistent with previous reports, we found that mean maternal plasma αT concentration was markedly higher than that of cord plasma (8)(9)(10)(11)(12)(13)15). In addition, the maternal plasma αT concentrations reported here during pregnancy are similar to values reported by others during pregnancy, including the complete cohort by Scholl et al. (19) and elsewhere (8)(9)(11)(12)(13)(14)(15), but are higher than those reported by Kiely et al. (10). In addition, the αT concentrations reported here are similar to those reported for healthy nonpregnant women (27)(28)(29) indicating that the maternal concentrations were not unusually high or low. The few women at the lower end of the range in this study approached the median maternal plasma αT concentration that Shamim et al. (6) found positively related with elevated risk of miscarriage. The cord plasma αT concentrations reported here, including those from complicated pregnancies, are directionally higher than those reported from a large cohort-4.3 μmol αT/L (15)-and a smaller cohort reporting 3.6 μmol αT/L (13), but are similar to those reported for smaller cohorts with mean values ranging from 5.5 to 7.7 μmol/L (8)(9)(10)(11)(12). Thus, the αT concentrations reported here are in line with those of previous reports, despite the poor economic and education status of the current cohort.
A limitation of this study is that maternal plasma was not collected at delivery. A second limitation is that the sample size did not allow for incorporation of potential confounders such as maternal age, gestational length, neonate sex, and pregnancy status into our statistical analyses. Strengths of the study include the heavy representation of black and Hispanic maternal-neonate pairs, the inclusion of complicated pregnancies, and the direct quantitation of αT stereoisomers.
In conclusion, we found that RRR-αT was the most common αT stereoisomer in all maternal and umbilical cord plasma samples, and that umbilical cord plasma contained a higher proportion of RRR-αT than maternal plasma. Nonetheless, most samples had measurable concentrations of S-αT, which could account for a substantial percentage of total αT. These observations were not influenced by pregnancy status. Maternal plasma concentrations of αT and S-αT were positively correlated with umbilical cord plasma S-αT, but negatively correlated with cord RRR-αT. We believe these finding to be important for several reasons beyond the well-documented differences in vitamin E activity between RRR-αT and S-αT (5). αT has increasingly been found to influence gene expression (30)(31)(32)(33)(34)(35), and more recently, evidence is accumulating that S-αT stereoisomers and RRR-αT have differential impacts on gene expression in the brain (20,36,37). In addition, improved mater-nal αT status appears to positively influence fetal growth (19,(38)(39)(40)(41)(42), but recently cord plasma αT concentration was reported to be negatively related to neonate 5-min Apgar (physician ratings of the newborn infant's appearance, pulse, grimace, activity, respiration) (38). Because umbilical cord proportions of RRR-αT and S-αT were related to maternal plasma αT measures, we believe future research is warranted to explore the relative impact of RRR-αT and S-αT on fetal and neonatal health.