Vitamin E is an essential nutrient consumed through foods including vegetable oils, nuts, dark green leafy vegetables and egg yolk( 1 ) and also through supplemented foods. Vitamin E is believed to be particularly important in the nervous system, as deficiency in humans, although rare, leads to neuromuscular abnormalities with ataxia and myopathy, and can be reversed by α-tocopherol supplementation or consumption( Reference Sokol, Guggenheim and Iannaccone 2 – Reference Traber, Sokol and Ringel 4 ). Vitamin E activity can be derived from four tocopherols (α, β, γ and δ) and their corresponding four tocotrienols (α, β, γ and δ). However, α-tocopherol is the most biologically active form of vitamin E( 1 ). α-Tocopherol is widely accepted to play an important role in the protection of membrane-bound unsaturated fatty acids from oxidation through its potent lipid-soluble, chain-breaking antioxidant activity. In addition, it is thought to support diverse biological processes including regulation of enzymatic activity( Reference Mahoney and Azzi 5 ), gene expression( Reference Azzi, Gysin and Kempna 6 ) and cell signalling( Reference Azzi, Gysin and Kempna 6 , Reference Azzi 7 ).
The three chiral C atoms in the structure of α-tocopherol result in significant structural complexity. In nature, α-tocopherol only exists as the RRR stereoisomer (2R, 4'R, 8'R-α-tocopherol), commonly referred to as natural vitamin E. In contrast, synthetic α-tocopherol (all-rac-α-tocopherol) is a racaemic mixture of the eight possible stereoisomers. On the basis of these differences in stereochemistry, RRR-α-tocopherol has greater biological value than all-rac-α-tocopherol. Extensive research has indicated that RRR-α-tocopherol has 1·36–2 times the biological value of all-rac-α-tocopherol( Reference Brigelius-Flohe and Traber 8 , Reference Traber 9 ). Discrimination among stereoisomers is known to occur in the liver through the action of α-tocopherol transfer protein (α-TTP)( Reference Ferslew, Acuff and Daigneault 10 ), which preferentially binds 2R stereoisomers, resulting in degradation of the majority of 2S stereoisomers in all-rac-α-tocopherol( Reference Traber, Burton and Ingold 11 ).
Rodent and pig peripheral tissue α-tocopherol stereoisomer composition appears to reflect hepatic α-tocopherol stereoisomer discrimination. The brain and other peripheral tissues of rats fed all-rac-α-tocopherol contain dramatically higher levels of 2R than of 2S stereoisomers( Reference Ueda, Ichikawa and Igarashi 12 , Reference Nitta-Kiyose, Hayashi and Tadahiko 13 ). Consistent with enrichment of 2R stereoisomers via hepatic discrimination, mice fed equal quantities of d3-RRR-α-tocopherol and d6-all-rac-α-tocopherol had a brain d3:d6 ratio of approximately 1·8( Reference Leonard, Terasawa and Farese 14 ). Weiser et al.( Reference Weiser, Riss and Kormann 15 ) directly tested individual 2R stereoisomers in brain samples from α-tocopherol-deprived rats repleted with all-rac-α-tocopherol and found no discrimination among 2R stereoisomers. Piglets fed sow milk containing a 2:1 d3:d6 ratio had a 2:1 ratio in plasma and 2·6 and 2·4 ratios in the brain after 14 and 21 d of nursing, respectively( Reference Lauridsen, Engel and Jensen 16 ). Although others have reported total α-tocopherol concentrations in the adult and fetal human brain( Reference Johnson, Vishwanathan and Johnson 17 – Reference Metcalfe, Bowen and Muller 19 ), no information is available on the α-tocopherol stereoisomer composition in the infant human brain.
As human infants are generally born with low levels of plasma α-tocopherol, their vitamin E status is heavily dependent on the α-tocopherol content and composition of human milk or infant formula. Infant formula is widely supplemented with all-rac-α-tocopheryl-acetate in order to meet the vitamin E requirements of infants. Unlike breast-fed infants, formula-fed infants do not benefit from discrimination of α-tocopherol stereoisomers in the maternal liver, possibly increasing the likelihood of synthetic stereoisomers reaching the infant brain. Expression of hepatic α-TTP in neonatal mice is low at birth and increases to adult levels over the first 28 d of life( Reference Kim, Arai and Arita 20 ). However, it is not known whether the same developmental phenomenon occurs in human infants. Research in knockout mice has revealed that α-tocopherol is essential for Purkinje cell integrity in the cerebellum through the action of α-TTP( Reference Ulatowski, Parker and Warrier 21 ). As α-TTP is known to discriminate between the 2R stereoisomers of α-tocopherol, we considered it important to determine the α-tocopherol stereoisomer distribution in infant brain.
Methods
Decedents and brain samples
Voluntarily donated, frozen brain samples were obtained from the National Institute of Child Health and Human Development Brain and Tissue Bank for Developmental Disorders at the University of Maryland. Samples from thirty-three decedents were collected between 1995 and 2005 and between 2007 and 2008 from the remaining three decedents. Samples were voluntarily donated by the parents and were identified using a unique numerical identifier, which obscured the identity of the decedent. Identity-obscured autopsy reports were reviewed to confirm that decedents were otherwise healthy and did not suffer from brain abnormalities or other systemic pathologies. A total of seventy-nine samples were obtained from the frontal cortex (FC) (n 28), hippocampus (HPC) (n 25) and visual cortex (VC) (n 26), and it is noteworthy that these regions of the brain are associated with memory (HPC), executive function (FC) and vision (VC). Decedent details are provided in Table 1. Samples were available from all three regions of the brain from fourteen of the thirty-six decedents. Only two regions were analysed for fifteen decedents, and a single region for the remaining seven decedents. Tissues were stored frozen (−70°C) until analysis for α-tocopherol stereoisomers. Term and preterm decedents were combined as values for total α-tocopherol and α-tocopherol stereoisomer profile for the preterm infants fell within the term infant range.
Descriptive characteristics of decedents (Numbers and percentages; mean values and standard deviations; thirty-six infants)
FC, frontal cortex; HPC, hippocampus; VC, visual cortex.
* Term, ≥37 weeks of gestation at birth, and preterm, <37 weeks of gestation birth.
α-Tocopherol analyses
The concentrations of tocopherols including γ-tocopherol were determined by HPLC after saponification and extraction into heptane. The stereochemical composition of α-tocopherol was determined after methylation of α-tocopherol stereoisomers into their corresponding methyl ethers and subsequent separation by chiral HPLC( Reference Jensen, Norgaard and Lauridsen 22 ). This method separates the eight stereoisomers of α-tocopherol into five peaks: peak 1 contains the four 2S stereoisomers (SSS/SSR/SRS/SRR), peak 2 contains RSS-α-tocopherol, peak 3 contains RRS-α-tocopherol, peak 4 contains RRR-α-tocopherol and peak 5 contains RSR-α-tocopherol. Results are reported as the ratio of the observed stereoisomer among five peaks and the concentration determined by calculation of each stereoisomer concentration from total α-tocopherol. Recovery of total α-tocopherol was 96 % with a CV (%) of 2·7, whereas recovery of the individual stereoisomers varied from 95·3 to 100·8 % with CV (%) varying from 1·2 to 7·9. Data were expressed per wet weight of tissue.
Statistical analysis
To detect significant main effects and interactions between main factors, data were analysed with one-way (brain region; stereoisomer) ANOVA using GraphPad Prism version 5.04 for Windows (GraphPad Software, www.graphpad.com). When appropriate, differences between treatment means were evaluated by Tukey’s multiple comparison test in GraphPad Prism. Means that differed significantly following a multiple comparison test were denoted using unique letter subscripts (P<0·05). To determine significant correlations between nutrient concentrations and age or other nutrients, data were analysed using Pearson’s correlation procedures of GraphPad Prism. All data are expressed as treatment means with their standard errors.
Results
Total α-tocopherol levels
Mean total α-tocopherol concentrations were not different across the three brain regions studied (Fig. 1). All samples analysed contained detectable concentrations of α-tocopherol, and inter-decedent variation was large with values ranging from <5 to approximately 17 μg α-tocopherol/g of tissue. Total α-tocopherol concentration positively correlated with age in each of the three brain regions tested (FC, r 2 0·31, P<0·01; HPC, r 2 0·36, P<0·01; FC, r 2 0·62, P<0·001). Intra-decedent variation in total α-tocopherol concentrations by brain region is shown for the fourteen decedents who provided samples for each region (online Supplementary Table S1). γ-Tocopherol was undetectable in the majority of samples (FC, twenty; HPC, fourteen; VC, twenty-one). Mean γ-tocopherol for samples with detectable levels was 0·30 (sem 0·06) μg γ-tocopherol/g tissue (n 8), 0·36 (sem 0·06) μg γ-tocopherol/g tissue (n 11) and 0·33 (sem 0·04) μg γ-tocopherol/g tissue (n 5) in FC, HPC and VC, respectively. Average ratios of α-tocopherol:γ-tocopherol were 39, 33 and 25 in FC, HPC and VC, respectively. Neither total α-tocopherol nor RRR-α-tocopherol (Table 1) was correlated with postmortem interval.
Total α-tocopherol concentrations in the frontal cortex (FC), hippocampus (HPC) and visual cortex (VC). Data are from thirty-six decedents with twenty-eight FC samples, twenty-five HPC samples and twenty-six VC samples. 
, 25th and 75th percentiles; 
, mean; 
, median; 
, minimum and maximum values.
α-Tocopherol stereoisomers
Mean concentrations of RRR-α-tocopherol were significantly higher than those of each of the synthetic α-tocopherol stereoisomers in each of the brain regions tested (Fig. 2(a–c)). Although RRR-α-tocopherol concentration positively correlated with age in each brain region studied (FC, r 2 0·22, P<0·05; HPC, r 2 0·20, P<0·05; VC, r 2 0·61, P<0·0001), individual synthetic stereoisomers did not. On average, across the brain regions, RRR-α-tocopherol accounted for 66 % (P<0·0001, compared with other stereoisomers) of total α-tocopherol, whereas none of the synthetic stereoisomers exceeded 14 %. RRR accounted for as much as 100 % of total α-tocopherol in two decedents. The minimum percentage of total α-tocopherol found as RRR was 58 %. More than half of the decedents had at least 30 % of total α-tocopherol as synthetic stereoisomers. Nonetheless, RRR-α-tocopherol was the most common stereoisomer in all samples and individuals on a per cent basis. In order to provide insight into whether the stereoisomer profile was consistent for different brain regions within individual decedents, the stereoisomer profile is presented by brain region for the fourteen decedents who provided samples for all three regions (online Supplementary Table S1).
α-Tocopherol stereoisomer profile in the frontal cortex (FC) (A), hippocampus (HPC) (B) and visual cortex (VC) (C) of the human infant brain. Data are from thirty-six decedents with twenty-eight FC samples, twenty-five HPC samples and twenty-six VC samples. The median values are not shown as they were very similar to the mean values. Σ2S stereoisomers include SRR, SSR, SRS and SSS. a,b,cStereoisomers with unlike superscript letters were significantly different: a
P<0·0001; b,c
P<0·05. *Different from other values for brain region (P<0·0001). , 25th and 75th percentile values, , mean; , maximum and minimum values.
α-Tocopherol stereoisomer ratios
The average ratio of the RRR stereoisomer:the sum of synthetic 2R stereoisomers was >2 (Fig. 3) with a range from a minimum of approximately 1 to a maximum of 6·1, 4·7 and 4·9 for FC, HPC and VC, respectively; two samples, one from HPC and one from VC, from different decedents, did not contribute ratios as the samples contained 100 % RRR.
The ratio of RRR - α-tocopherol:sum of RRS-α-tocopherol, RSR-α-tocopherol and RSS-α-tocopherol stereoisomers in infant decedent samples from the frontal cortex (FC), hippocampus (HPC) and visual cortex (VC). Scatter plot of all values by brain region; horizontal bar indicates mean value.
Discussion
Although others have previously reported the concentration of total α-tocopherol in human adult and fetal brain( Reference Johnson, Vishwanathan and Johnson 17 – Reference Metcalfe, Bowen and Muller 19 ), we believe we are the first to report the α-tocopherol stereoisomer profile in the human infant brain. Our data reveal that RRR-α-tocopherol was the predominant stereoisomer of α-tocopherol in each of the three regions of the infant brain studied. In fact, RRR-α-tocopherol was the predominant stereoisomer in each sample tested, despite high inter-decedent variation in α-tocopherol. We found that the stereoisomer patterns were relatively consistent across the three brain regions in decedents with samples available for each region. The degree of discrimination in favour of RRR-α-tocopherol is striking as most infant formulae were supplemented with all-rac-α-tocopherol. The extent of exposure of the decedents to all-rac-α-tocopherol is emphasised by the finding that the brain samples of only one decedent was devoid of synthetic stereoisomers. That both total α-tocopherol and RRR-α-tocopherol were positively correlated with age in each brain region might provide further evidence of the importance of α-tocopherol stereoisomer discrimination.
We believe the α-tocopherol stereoisomer profile may be of particular importance to infants. First, vitamin E is an essential nutrient that plays an important role in neural function and development. Second, vitamin E requirements of formula-fed infants are routinely met using the less biologically active, synthetic source of vitamin E – all-rac-α-tocopherol. Third, the state of development of the hepatic stereoisomer selection mechanism in infants has not been established.
Total α-tocopherol concentrations reported here are lower than, but in line with, those previously reported for autopsied adult human brain( Reference Johnson, Vishwanathan and Johnson 17 – Reference Metcalfe, Bowen and Muller 19 ) and very similar to those for rat( Reference Weiser, Riss and Kormann 15 ) and piglet( Reference Lauridsen, Engel and Jensen 16 ) brain. Although we cannot rule out the potential that tissue degradation before sample preservation could introduce artifact, no relationship was found between total α-tocopherol concentration and time until tissue freezing. We believe that the slow turnover rate of α-tocopherol in the brain( Reference Lauridsen, Engel and Jensen 16 , Reference Vatassery, Angerhofer and Knox 23 , Reference Vatassery, Angerhofer and Paterson 24 ) may have reduced the potential for preservation artifact.
We found that all individuals tested had very low levels of 2S stereoisomers and corresponding high levels of the 2R stereoisomers of α-tocopherol. This observation is largely consistent with previous animal research, revealing that rodent and piglet brains preferentially contain the 2R stereoisomers of all-rac-α-tocopherol( Reference Nitta-Kiyose, Hayashi and Tadahiko 13 , Reference Weiser, Riss and Kormann 15 , Reference Lauridsen, Engel and Jensen 16 ). Discrimination against 2S stereoisomers is believed to occur in the liver. Hepatic α-TTP-mediated discrimination in the infant and mother likely explains the low concentrations of γ-tocopherol we found in the infant brain samples. Despite high levels of γ-tocopherol in the American diet, breast milk generally has a 5:1 ratio of α-tocopherol:γ-tocopherol. In addition, some infant formulae contain relatively high levels of γ-tocopherol, indicating that the infant liver also discriminates against γ-tocopherol in favour of α-tocopherol( Reference Stone, LeClair and Ponder 25 ).
Previous research in animals does not provide definitive insight into whether there is discrimination between the 2R stereoisomers in the brain. The single study published to date that directly quantified 2R α-tocopherol stereoisomers in the brain revealed no selection among 2R stereoisomers in α-tocopherol-deficient rats repleted with all-rac-α-tocopherol( Reference Weiser, Riss and Kormann 15 ). Our data, however, are consistent with preferential discrimination for RRR-α-tocopherol in the human infant brain. Extensive research in human adults and corroborative evidence in infants indicate that plasma from individuals consuming all-rac-α-tocopherol has roughly equal concentrations of the 2R stereoisomers( Reference Brigelius-Flohe and Traber 8 , Reference Traber 9 , Reference Kiyose, Muramatsu and Kameyama 26 ). Equal concentrations of 2R stereoisomers result in a ratio of RRR-α-tocopherol:the sum of RRS-α-tocopherol, RSR-α-tocopherol and RSS-α-tocopherol of approximately 0·33( Reference Brigelius-Flohe and Traber 8 , Reference Kiyose, Muramatsu and Kameyama 26 , Reference Traber, Elsnerb and Brigelius-Floheèb 27 ). Therefore, if there was no stereoisomer discrimination in the infant brain, the minimum RRR:(RRS+RSR+RSS) would be approximately 0·33( Reference Weiser, Riss and Kormann 15 ). It is highly likely that some of the decedents studied received predominantly all-rac-α-tocopherol; however, the minimum ratio we observed was 0·7 with median brain values of 1·8 or higher in each brain region studied. These observations are consistent with significant selection among the 2R stereoisomers in favour of RRR-α-tocopherol.
Potential over-representation of breast-fed infants in the study cohort does not explain our observations. Consistent with observations in sow milk( Reference Lauridsen, Engel and Jensen 16 ), unpublished observations from our group (M Kuchan) suggest that breast milk from North American women contains a significant amount of synthetic 2R α-tocopherol stereoisomers. This observation is based on 300 samples from more than 200 women and likely reflects the use of synthetic α-tocopherol in maternal supplements. In addition, the results of this study indicate that both total and RRR-α-tocopherol increased with age during the 1st year, despite the rapid increase in brain volume known to occur over this period. We believe this diminishes the likelihood of possible confounding effects of maternal liver–placenta–infant liver tandem selection before birth. Enhanced discrimination between stereoisomers in the infant liver compared with that known to occur in the adult liver is inconsistent with published results in the human infant( Reference Stone, LeClair and Ponder 25 ). For these reasons, we believe that these data are best explained by a combination of discrimination for RRR-α-tocopherol in the infant brain and the source of α-tocopherol consumed in the diet.
The predominance of RRR-α-tocopherol in the infant brain reported here might be explained by the presence of α-TTP and tocopherol-associated protein (TAP) in the human brain. Extra-hepatic stereoisomer-specific discrimination has been previously demonstrated in the human fetus/fetal-placental unit( Reference Acuff, Dunworth and Webb 28 , Reference Schenker, Yang and Perez 29 ) and is believed to be at least partially explained by placental expression of α-TTP ( Reference Gordon, Campbell and Dutta-Roy 30 ). The human cerebellum expresses α-TTP ( Reference Copp, Wisniewski and Hentati 31 ), and multiple regions of the human brain express relatively high levels of TAP ( Reference Zimmer, Stocker and Sarbolouki 32 ). TAP has a high degree of binding-site sequence homology to α-TTP, and similar to α-TTP is known to possess differential affinity for tocopherol structural isomers( Reference Panagabko, Morley and Hernandez 33 ). These observations suggest that TAP may be capable of discrimination among α-tocopherol stereoisomers.
In conclusion, these first-of-kind data reveal a clear, preferential accumulation of the naturally occurring stereoisomer RRR-α-tocopherol in the human infant brain. This is a surprising observation given the extensive exposure of the decedents to all-rac-α-tocopherol. Our data are consistent with discrimination among the 2R stereoisomers of all-rac-α-tocopherol in favour of RRR-α-tocopherol. We believe this finding raises important questions, as biological discrimination often reflects biological importance. Emerging evidence suggests that RRR-α-tocopherol and all-rac-α-tocopherol may have differential cellular effects mediated through gene expression( Reference Han, Pang and Zingg 34 ). It is not known whether there is an advantage of increased RRR-α-tocopherol, or if accumulation of synthetic stereoisomers is unfavourable for brain development. However, as vitamin E is essential for the integrity of cerebellar Purkinje neurons( Reference Ulatowski, Parker and Warrier 21 , Reference Ulatowski and Manor 35 , Reference Hosomi, Goto and Kondo 36 ) through the action of α-TTP, a stereo-selective protein, we believe this is an important question for future research.
Acknowledgements
The authors express their deep appreciation to the National Institute of Child Health and Human Development Brain and Tissue Bank and parents for donating the extremely valuable samples tested in this study. Elsebeth Lyng Pedersen is greatly acknowledged for conducting the tocopherol analyses and Chron-si Lai for technical consultation.
This work was supported by Abbott Nutrition. E. J. J. was also supported by the United States Department of Agriculture 58-1950-1-014.
M. J. K.: conceived, designed, interpreted, drafted and critically reviewed the work. S. K. J.: analysis and critical review of the work. E. J. J.: acquisition of samples and critical review of the work. J. C. L.-B.: drafted the manuscript and critically reviewed the work. All the authors are accountable for the results.
None of the authors has any conflicts of interest to declare.
Supplementary Material
For supplementary material/s referred to in this article, please visit http://dx.doi.org/doi:10.1017/S0007114516001719
