If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Nutrition, School of Public Health and School of Medicine, University of North Carolina, Chapel Hill, North Carolina 27599Nutrition Research Institute, University of North Carolina at Chapel Hill, Kannapolis, North Carolina 28081
* This work was supported, in whole or in part, by National Institutes of Health Grants AG09525, DK55865, and DK56350 (to S. H. Z.) and Grant DK034987 to the University of North Carolina Histology Core. The on-line version of this article (available at http://www.jbc.org) contains supplemental Tables S1–S8.
The development of fetal brain is influenced by nutrients such as docosahexaenoic acid (DHA, 22:6) and choline. Phosphatidylethanolamine-N-methyltransferase (PEMT) catalyzes the biosynthesis of phosphatidylcholine from phosphatidylethanolamine enriched in DHA and many humans have functional genetic polymorphisms in the PEMT gene. Previously, it was reported that Pemt−/− mice have altered hippocampal development. The present study explores whether abnormal phosphatidylcholine biosynthesis causes altered incorporation of DHA into membranes, thereby influencing brain development, and determines whether supplemental dietary DHA can reverse some of these changes. Pregnant C57BL/6 wild type (WT) and Pemt−/− mice were fed a control diet, or a diet supplemented with 3 g/kg of DHA, from gestational day 11 to 17. Brains from embryonic day 17 fetuses derived from Pemt−/− dams fed the control diet had 25–50% less phospholipid-DHA as compared with WT (p < 0.05). Also, they had 60% more neural progenitor cell proliferation (p < 0.05), 60% more neuronal apoptosis (p < 0.01), and 30% less calretinin expression (p < 0.05; a marker of neuronal differentiation) in the hippocampus compared with WT. The DHA-supplemented diet increased fetal brain Pemt−/− phospholipid-DHA to WT levels, and abrogated the neural progenitor cell proliferation and apoptosis differences. Although this diet did not change proliferation in the WT group, it halved the rate of apoptosis (p < 0.05). In both genotypes, the DHA-supplemented diet increased calretinin expression 2-fold (p < 0.05). These results suggest that the changes in hippocampal development in the Pemt−/− mouse could be mediated by altered DHA incorporation into membrane phospholipids, and that maternal dietary DHA can influence fetal brain development.
Choline is included in infant formulas and in the maternal diet because, during a critical span of time during fetal brain development, choline availability influences neural progenitor cell proliferation, apoptosis, and differentiation (
). However, there is a metabolic pathway that links choline with DHA, and there could be a common mechanism for the effects of these two nutrients on brain development.
DHA metabolism and choline metabolism are linked by the enzyme phosphatidylethanolamine-N-methyltransferase (PEMT), which catalyzes de novo biosynthesis of phosphatidylcholine (PtdCho) by methylation of PtdEtn (
). Because PtdCho is almost exclusively found in the outer leaflet of the cell membrane and PtdEtn and PtdSer are only found in the inner leaflet, any change in the activity of PEMT can modify not only PtdCho-DHA and PtdEtn-DHA concentrations, but may also lead to an imbalance in DHA distribution in the leaflets of the membrane. In fact, PtdCho-DHA concentrations in liver cell membranes and plasma were diminished in Pemt−/− mice compared with wild type (
), and because there was increased S-adenosylmethionine concentration as well as increased DNA and histone methylation in the hippocampus of Pemt−/− fetuses, it was suggested that the effects of deleting Pemt on fetal brain development were mediated by changes in gene methylation (
). Although this remains a likely mechanism mediating many of the observed changes, we now consider whether changes in DHA concentrations and/or DHA distribution in membrane leaflets might also mediate or enhance the observed effects of deleting Pemt. We hypothesized that dietary supplementation of DHA to Pemt−/− dams would restore normal PtdCho-DHA concentrations in cell membranes, and that this would reverse at least some of the abnormal cell proliferation, apoptosis, and differentiation observed in the knock-out fetal mouse hippocampus.
Changes in brain phospholipids were examined in two ways, as changes in the distribution of fatty acid species present in individual phospholipids, and as changes in the concentrations of phospholipids. We observed significant differences in the fatty acid composition of the phospholipids in fetal brain from the Pemt−/− mice compared with the WT mice (Table 1). Pemt−/− mice had lower proportional fractions of DHA in all four phospholipids measured in fetal brain than did WT mice on the control diet (p < 0.05; Tables 1 and supplemental S1, S3, S5, and S7). Maternal supplementation with DHA during E11–E17 increased the proportional fraction of PtdCho-DHA and PtdIns-DHA in the fetal brains of the WT mice by ∼150% (p < 0.001; Table 1), and increased the proportional fraction of PtdEtn-DHA and PtdSer-DHA in fetal brains by ∼10%. DHA supplementation of the Pemt−/− dams increased the proportional fraction of DHA in several membrane phospholipids in the fetal brain (PtdSer-DHA to 135% of the fraction present in control diet; PtdEtn-DHA to 158%; PtdCho-DHA to 192%; and PtdIns-DHA to 220%; all values p < 0.01 different from those on control diet Table 1) so that these fractions were similar to, or higher than, those measured in WT on the control diet. In addition, the fatty acid species docosapentaenoic acid (22:5) was detected in PtdEtn and PtdSer fractions of the Pemt−/− but not in WT fetal brains (Table 1). Brain PtdEtn-18:1 was higher in the fetuses from the Pemt−/− mice on the control diet compared with WT (p < 0.01; Table 1), but returned to WT levels with DHA supplementation. In PtdEtn, PtdSer, and PtdIns of fetal brains, the arachidonic acid (20:4) species was higher in Pemt−/− compared with WT on the control diet (p < 0.01; Table 1), but maternal DHA supplementation decreased the proportional amount in Pemt−/− fetal brains (p < 0.05).
TABLE 1Proportion of fatty acid species present in phospholipids of fetal mouse brain
Although we observed the above changes in the fatty species present in PtdCho, the concentrations of PtdCho in fetal mouse brain were unchanged by genotype or diets (Table 2). PtdEtn concentrations were lower in the Pemt−/− mice than in WT (p < 0.05), and the DHA diet restored normal concentrations. PtdSer concentrations were lower in WT on the control diet than in the Pemt−/− groups or in the DHA-supplemented WT fetuses (Table 2, p < 0.05). PtdIns concentrations were lower on the AIN76 diet than on the DHA-supplemented diet (p < 0.05). Most of the DHA in phospholipids was in the PtdEtn and PtdSer phospholipids in fetal brain (Tables 2 and supplemental S2, S4, S6, and S8). The concentration of DHA bound to PtdCho, PtdSer, and PtdIns increased in fetal brains of both genotypes (p < 0.05) when DHA was added to the maternal diet (Table 2). For PtdCho-DHA, there was an interaction between genotype and diet (p = 0.0007). This was also true for PtdSer-DHA (p = 0.02), whereas PtdIns had interactions for arachidonic acid (20:4, p = 0.0006) and DHA (p = 0.0014). PtdEtn had genotype x diet interactions for oleic acid (18:1, p = 0.04), DHA (p < 0.0001), and total nanamole/g (p = 0.0002).
TABLE 2Concentrations of phospholipids and their DHA species in fetal mouse brain
On the AIN-76A diet, Pemt−/− fetal brains had 60% more neural progenitor cell proliferation (assessed by histone-3 phosphorylation) in the ventricular zone of the developing hippocampus and cortex (p < 0.05; Fig. 1) as compared with WT. Feeding the DHA supplemented diet to dams did not change neural progenitor cell proliferation in WT fetal brains, but decreased this proliferation in Pemt−/− fetal brains in both the hippocampal (p < 0.05) and cortical (p < 0.05) ventricular zones, such that values were no longer higher than those in fetal brains from WT dams on the control diet. There was an interaction between genotype and diet for neural progenitor cell proliferation in the hippocampus (p = 0.012) and cortex (p = 0.044) of fetal brains. On the control diet, Pemt−/− fetal brains had 30% less calretinin protein levels (a marker for differentiation of GABAergic neurons) in the primordial dentate gyrus as compared with fetuses from WT dams (p < 0.05; Fig. 2), however, DHA supplementation increased calretinin expression 2-fold in both genotypes (p < 0.05, Fig. 2).
Pemt−/− fetal hippocampi, from dams fed the control diet, had 1.6–2.5-fold increased neural cell apoptosis (assessed by caspase 3 activation or by TUNEL assay, respectively) compared with WT fed the control diet (p < 0.01; FIGURE 3, FIGURE 4). Maternal dietary DHA supplementation decreased the high rates of apoptosis in the fetal hippocampus of the Pemt−/− mice so that they were similar to levels observed in WT on control diet (p < 0.01 for caspase results, p < 0.05 for TUNEL assay results; FIGURE 3, FIGURE 4).
) that there are more proliferating neural progenitor cells in the fetal brain of Pemt−/− mice compared with WT mice (Fig. 1), and for the first time show that numbers of apoptotic cells in fetal hippocampus are higher in Pemt−/− mice (Fig. 3). Apoptosis has a critical role in determining the size of neuronal subpopulations, and therefore affects morphogenesis in developing brain regions (
) that calretinin expression, a marker of neuronal differentiation, was diminished in Pemt−/− fetal brains compared with WT (Fig. 2). We suggest that these changes in fetal brain of Pemt−/− mice are, in part, related to differences in DHA-containing phospholipids in membranes (TABLE 1, TABLE 2) because maternal supplementation with DHA corrected the abnormally low amounts of DHA-containing phospholipids in Pemt−/− brain (Table 1) and prevented the change in neural progenitor cell proliferation (Fig. 1), as well as the increase in apoptosis seen in Pemt−/− fetal brains (FIGURE 3, FIGURE 4). Moreover, DHA supplementation increased calretinin expression by 2-fold in both knock-out and WT brains (Fig. 2).
), we suggested that the changes in fetal brain development seen in the Pemt−/− mouse were the result of changes in DNA methylation, which altered the gene expression required for neurogenesis and differentiation. However, there is a second possible mechanism for changes in fetal brain development seen in the Pemt−/− mouse. We had previously observed that these mice have substantially lower concentrations of DHA in plasma and liver PtdCho (
), and we now report that this also is the case in fetal brain (TABLE 1, TABLE 2). In addition, these mice have lower PtdEtn-DHA and PtdSer-DHA concentrations in brain (TABLE 1, TABLE 2). We hypothesized that diminished phospholipid-DHA, with resulting changes in membrane function and signaling, is an alternative mechanism for the changes in progenitor cell proliferation, differentiation, and apoptosis seen in the Pemt−/− mouse. If this is correct, then manipulations that correct phospholipid-DHA concentrations in the fetal brain would prevent the observed changes. Indeed, maternal supplementation with DHA abrogated the increase in neural progenitor cell proliferation seen in Pemt−/− fetal brains compared with WT (Fig. 1). This could be due to changes in membrane phospholipid-DHA concentrations, or to DHA interfering with a proliferation-signaling molecule (
). Our supplemented diet contained 0.3% DHA, which was just enough to normalize the fetal brain phospholipid DHA in the knock-out mice, but was far less than the 1–5% DHA given in the above studies on cell proliferation. The dose used in published human clinical studies is 3 g/day in the diet (
). As cells differentiate, they decrease cell proliferation, and we found that DHA increased neuronal differentiation (as measured by expression of the protein calretinin in GABAergic neurons) 2-fold (Fig. 2). Other investigators reported that DHA is important for neuronal differentiation (
); we found that the PtdSer that accumulated in the DHA-supplemented fetal brains had more 18:0 and DHA fatty acid species (supplemental Table S6), and suggest that this contributed to the decrease in apoptosis seen in DHA-treated Pemt−/− and WT fetal brains compared with those from dams fed the control diet (FIGURE 3, FIGURE 4). Neuroprotectin, which is formed in glial cells from DHA and induces expression of neuroprotective and anti-apoptotic genes (
), also could mediate decreased apoptosis after DHA treatment, but is unlikely to play a role here because we found no expression of the mature glial cell marker glial fibrillary acidic protein in our fetal brains (data not shown).
Studies in humans report that DHA needed for fetal brain growth is derived from transplacental transfer of these fatty acids from the parent (
). Thus, feeding the pregnant dam more DHA should increase PtdCho-DHA concentrations in fetal brain. Indeed, PtdCho-DHA concentrations in fetal brain were increased by feeding DHA to pregnant dams of both genotypes (TABLE 1, TABLE 2). Increased availability of DHA increases PtdCho synthesis in fetal lung explants by a direct stimulatory effect of DHA on the rate-limiting enzyme in the pathway, CTP:choline cytidylyltransferase (
), and if the same effect occurs in fetal brain, it may contribute to the changes in PtdCho-DHA that we report. Other studies reported that increasing maternal dietary DHA increased DHA content of PtdEtn and PtdSer of whole brain glial cells in rat pups (
). We found this to be true, along with similar effects on PtdIns-DHA, in fetal brains from the Pemt−/− pups whose dams were fed a diet supplemented with DHA (TABLE 1, TABLE 2). PtdCho-DHA made in maternal liver is secreted as lipoproteins by liver and eventually delivered to the fetal brain, and is a major source of the DHA incorporated into the phospholipids of the fetal brain. In Pemt−/− dams this PtdCho-DHA is not formed in the liver and is not delivered to the fetus; thus, we find that docosapentaenoic acid (22:5) accumulates in PtdEtn and PtdSer of Pemt−/− but not in the WT fetal brain (Table 1); docosapentaenoic acid is incorporated in brain phospholipids when there is insufficient DHA available (
In the fetal brain from Pemt−/− mice, we observed a decrease in PtdCho-DHA concentrations without a concomitant increase in PtdEtn-DHA concentrations (Table 2). In liver this would be surprising, as PtdEtn-DHA is three times methylated by PEMT to form PtdCho (
). However, in brain, PEMT activity is low, and much of the PtdCho-DHA in the fetal brain is derived from uptake of phospholipids originally synthesized in the maternal liver. Once in the brain, the DHA in PtdCho is fungible, and can distribute among many phospholipids. PtdCho can be converted to PtdSer by phosphatidylserine synthase 1 (
). This could account for the observed increase in PtdSer pools in the Pemt−/− fetal brains (Table 2) despite no accumulation of PtdEtn, nor appreciable change in PtdCho concentration, in the absence of PEMT.
These observations suggest a potential common mechanism whereby the metabolic pathways for two nutrients interact to modify brain development. The activity of PEMT produces a membrane phospholipid that contains both choline and DHA. Varying maternal dietary choline intake in rats and mice during pregnancy causes significant and irreversible changes in hippocampal progenitor cell proliferation, neuronal apoptosis, and neuronal differentiation (
K. da Costa, L. M. Sanders, L. M. Fischer, and S. H. Zeisel, unpublished data.
as would be predicted from our current studies in mice. We do not know whether pregnant women with PEMT SNPs are more likely to have babies with abnormal brain development, or whether DHA supplementation would alter that development. It is interesting that women in the lowest quartile of dietary choline intake have 4-fold the risk of having a baby with a neural tube defect (
In conclusion, our study demonstrates that maternal supplementation with DHA can modulate some aspects of abnormal hippocampal development in the Pemt−/− mouse fetus. Our study suggests a novel mechanism whereby choline and DHA metabolism are intertwined, and this mechanism could underlie some of the observed effects of choline on brain development. We are reporting an interesting association between loss of PEMT activity and diminished brain phospholipid-DHA concentrations and changes in neurogenesis and apoptosis. These observations could be important in humans because a large portion of the population have functional SNPs in the PEMT gene.
We thank Alyssa Gulledge, Meng Hong, and Joseph Galanko for assistance in this study.