Address correspondence and reprint requests to Dr Hee-Yong Kim, Section of Mass Spectrometry, Laboratory of Membrane Biochemistry and Biophysics, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, 12 420 Parklawn Drive, Rockville, Bethesda, MD 20892-8115, USA. E-mail: email@example.com
It has been suggested that hippocampus-related cognitive processes are especially sensitive to ethanol. To provide an insight into the biochemical mechanisms underlying the hippocampus-related functional deficits associated with prenatal ethanol exposure, we investigated the effects of chronic ethanol exposure on the phospholipid profile in developing rat hippocampi. High-performance liquid chromatography/electrospray ionization–mass spectrometry analysis revealed that ethanol lowered the levels of total phosphatidylserine (PS) by 15–20% at all ages examined, primarily owing to the reduction in 1-stearoyl-2-docosahexaenoyl-PS (18:0,22:6n-3-PS) species. Ethanol exposure also led to a decrease in phosphatidylcholine (PC) and an increase in phosphatidylethanolamine (PE), but the total phospholipid content was not significantly changed. At the fatty acid level, ethanol exposure significantly decreased the 22:6n-3 content at postnatal days 0 and 21, with a slight increase in 22:5n-6, without changing the total fatty acid content significantly. In conclusion, ethanol depleted PS, especially 22:6-containing species, and PC from hippocampal membranes with concomitant increase in PE. Alteration of the phospholipid profile in the hippocampus resulting from exposure to ethanol during prenatal and developmental stages may have significant implications with respect to the cognitive dysfunction observed in fetal alcohol syndrome.
To provide an insight into the biochemical basis for hippocampus-related functional deficits caused by prenatal ethanol exposure, we evaluated changes in the phospholipid molecular species and fatty acid composition caused by ethanol in developing rat hippocampus.
Materials and methods
Deuterium-labeled phospholipid standards used for quantitation were purchased from or custom synthesized by Avanti Polar Lipids (Alabaster, AL, USA) and calibrated by phosphorus assay and gas chromatography (GC). BF3/methanol used for fatty acid derivatization was obtained from Alltech (Deerfield, IL, USA). Bicinchoninic acid protein assay reagent was obtained from Pierce (Rockford, IL, USA). Fiske and SubbaRow reducer used for phosphorus assay was purchased from Sigma (St Louis, MO, USA). All solvents were HPLC grade and purchased from EM Scientific (Gibbstown, NJ, USA) or Burdick and Jackson (Muskegon, MI, USA).
Animals and diets
Pregnant Sprague–Dawley rats (250–300 g bodyweight) were fed with ethanol mixed in an AIN-93G (Reeves et al. 1993) purified liquid rodent diet from day 11 of gestation, gradually increasing the ethanol content in the diet over 3 days to 35 energy%. On the first, second and third day, the ethanol group was fed with 1/3 ethanol diet plus 2/3 control diet, 2/3 ethanol diet plus 1/3 control diet, and full ethanol diet, respectively. The control group was pair fed with an isocaloric control diet with maltose dextrin instead of ethanol. Dams were fed the same diet throughout the pregnancy and lactation period. The custom diets were obtained from Dyets (Bethlehem, PA, USA). The composition of the diets is listed in Table 1. Each dam was housed in a chamber with a 12-h light–dark cycle and controlled temperature (23°C) in the National Institute on Alcohol Abuse and Alcoholism (NIAAA) animal facility. All animal procedures were approved by the Animal Care and Use Committee of NIAAA, National Institutes of Health.
Table 1. Composition of experimental diets
Amount (g/L of diet)
Tissue preparation and lipid extraction
Rat pups were killed by decapitation on postnatal day 0 (P0) and P21. At P21, dams were also killed. Brains were quickly removed and hippocampi were collected, homogenized in ice-cold HEPES buffer (pH 7.4), and aliquoted for protein assay and lipid analysis. Total lipids were extracted according to the method of Bligh and Dyer (1959) for the determination of fatty acid composition and lipid phosphorus content. For phospholipid molecular species analysis by reversed-phase high-performance liquid chromatography/electrospray ionization–mass spectrometry (HPLC/ESI–MS), tissue lipids were extracted in the presence of deuterium-labeled phospholipid standards.
Protein and lipid phosphorus assay
Protein content was measured spectrophotometrically using bicinchoninic acid as the reaction reagent as described earlier (Smith et al. 1985). Phosphorus assay was performed according to the reported procedure (Bartlett 1959; Böttcher 1961) with slight modifications. This method was based upon the colorimetric measurement of a phosphomolybdenum complex formed during reduction of phosphomolybdic acid in the presence of sulfite. Briefly, organic solvents in samples were evaporated under nitrogen, and 100 μl 18 m H2SO4 were added. The mixture was heated at 300°C for 5 min. After cooling, 200 μl 30% H2O2 was added and the sample mixture was vortexed. The mixture was further heated at 300°C for 40 min. After cooling, 2 mL water, 100 μl 5% (w/v) ammonium molybdate and 100 μl of Fiske and SubbaRow reducer in water were added. The mixture was vortexed and heated in a boiling water bath for 15 min and the absorbance was read at 820 nm after cooling.
Phospholipid molecular species were separated and analysed using reversed-phase HPLC/ESI–MS with a C18 column (Prodigy, 150 × 2.0 mm, 5 μm; Phenomenex, Torrance, CA, USA) as described previously (Kim et al. 1994). The separation was accomplished using a linear solvent gradient (water:0.5% ammonium hydroxide in methanol:hexane), changing from 12:88:0 to 0:88:12 in 17 min after holding the initial solvent composition for 3 min at a flow rate of 0.4 mL/min (Ma and Kim 1995). An Agilent 1100 LC/MSD instrument (Palo Alto, CA, USA) was used to detect the separated phospholipid molecular species. For electrospray ionization, the drying gas temperature was 350°C; the drying gas flow rate and nebulizing gas pressure were 11 L/min and 45 p.s.i. respectively. The capillary and fragmentor voltages were set at 4500 and 300 V, respectively. Identification of individual phospholipid molecular species was based on the monoglyceride, diglyceride and protonated molecular ion peaks (Kim et al. 1994). As internal standards representing each phospholipid class we used 1-d35-stearoyl-2-docosahexaenoyl-glycerophosphoserine (d3518:0,22:6-PS), 1-d35-stearoyl-2-arachidonoyl-glycerophosphoethanolamine (d3518:0,20:4-PE) and 1-d35-stearoyl-2-linoleoyl-glycerophosphocholine (d3518:0,18:2-PC). Quantitation of phospholipid species was based on the area ratio calculated against the added deuterium-labeled internal standards using diglyceride ions for PS and PE, monoglyceride ions for plasmenylethanolamine (PLS) and protonated molecular ions for PC.
Fatty acid analysis
Fatty acids were determined by GC analysis after transmethylation (Morrison and Smith 1961), with the addition of tricosanoic acid (23:0) as internal standard. Briefly, aliquots of lipid extracts were transmethylated by the reaction with BF3/methanol (14%, w/v) at 100°C for 2 h under a nitrogen atmosphere. Fatty acid methyl esters were extracted with hexane and then analyzed using a Hewlett Packard 5890 gas chromatograph equipped with a flame ionization detector (Palo Alto, CA, USA) and a fused-silica DB-FFAP capillary column (30 m × 0.25 mm internal diameter, 0.25 µm; J & W Scientific, Folsom, CA, USA) as described earlier (Kim and Salem 1990). The oven temperature was programmed from 130 to 180°C at 4°C/min, from 180 to 215°C at 1°C/min and then raised to 245°C at 30°C/min, with a final hold for 15 min. The injector and detector temperatures were set at 250°C. Hydrogen was used as carrier gas with a linear velocity of 54 cm/s. Individual fatty acids were identified by comparing the retention times with known fatty acid standards GLC-411 (Nu-Chek Prep, Elysian, MN, USA). The content of each individual fatty acid was expressed as a percentage of the weight of total fatty acid.
All experiments were performed at least in triplicate. Statistical differences were assessed by using non-paired Student's t-test unless stated otherwise; p < 0.05 was considered to be significant.
Effect of ethanol on the content of PS and total phospholipids
HPLC/EIS–MS analysis of hippocampal phospholipids indicated that the 22:6n-3-containing species were the most abundant polyunsaturates in PS and PE, from both control and ethanol-exposed rats. Figure 1 shows representative chromatograms of deuterium-labeled internal standards and typical phospholipid molecular species in rat brain hippocampi at P0. Diglyceride ions for 18:0,18:1-, 18:0,22:4n-6-, 18:0,22:5n-6- and 18:0,22:6n-3-PS and PE species are shown in the chromatograms as an example. The 18:0,22:6n-3 species was the most abundant species in PS, followed by 18:0,22:5n-6-PS. Mass spectra obtained from the chromatographic peaks indicated that diglyceride ions were the base peaks for PS and PE, whereas PC species were detected predominantly as protonated molecular ions (Fig. 2). Monoglyceride ions were also present for PS and PE, providing information regarding the identity of the fatty acyl composition. During development, the absolute levels of total phospholipids and PS were gradually increased both in the control and ethanol groups, although the most significant increase occurred during the first 3 weeks of development (Fig. 3). At all ages tested, the total PS levels in the ethanol groups (36.44, 49.61 and 54.83 pmol per µg protein for the P0, P21 pups and dams, respectively) were significantly lower than those of the control groups (43.70, 59.24 and 64.20 pmol per µg protein, respectively). The total phospholipid levels, however, were not significantly altered after ethanol exposure.
Effect of ethanol on individual phospholipid distribution
Ethanol exposure altered the profile of individual phospholipids, including PS, PE, PLS and PC, in rat brain hippocampi (Fig. 4). In P0 pups, the proportion of PS was decreased significantly from 11.28% in the control group to 9.15% in the ethanol group. Likewise, ethanol exposure significantly reduced the PC content from 38.96% in the control group to 35.40% in the ethanol group. The decreases in PS and PC were compensated by increases in PE and PLS content in the ethanol groups (33.92 and 21.53% respectively) in comparison to those of control groups (31.81 and 17.95% respectively). Hippocampal phospholipids in the P21 pups and dams showed similar changes in that the proportions of PS and PC in the ethanol groups were significantly lower, and that of PE significantly higher, than those of control groups. However, there was no significant effect of ethanol on the hippocampal PLS level in P21 rats and dams.
Effect of ethanol on PS molecular species composition
To determine whether the observed decrease in PS is associated with changes in specific molecular species, the effect of ethanol exposure on the distribution of individual PS molecular species was examined in rat brain hippocampi during the developmental periods (at P0 and P21) and in dams (Fig. 5). At all ages tested, 18:0,22:6n-3 was the major molecular species in PS. The levels of 18:0,22:5n-6 and 18:0,20:4n-6 were significant at P0 and these levels gradually declined with age. The 16:0,22:4n-6-PS species was detected only in P0 hippocampus, and its level was significantly reduced by ethanol exposure. In dam's hippocampi, 18:0,18:1 became the second dominant species in PS. In comparison to the control group, ethanol exposure significantly lowered the 18:0,22:4n-6- and 18:0,22:6n-3-PS levels in rat brain hippocampi at all ages tested. The most prominent reduction was observed for 18:0,22:6n-3-PS, indicating that the observed decrease in total PS was primarily due to the loss of this species. There was a tendency to decreased levels of 18:0,20:4n-6- and 18:0,18:1-PS, but 18:0,22:5n-6-PS appeared to increase in all ethanol-exposed hippocampi.
Effect of ethanol on PE and PC molecular species composition
Changes in the molecular species of PE and PC induced by ethanol treatment were also examined. At all ages tested, 16:0,22:6, 18:0,20:4 and 18:0,22:6 were the major species observed in PE (Fig. 6). In contrast to the the situation with PS, the observed increase in PE after exposure to ethanol was not specifically linked to a few particular molecular species. Instead, the increase was distributed across most PE molecular species although the increase in 18:0,22:6 became more prominent in adult hippocampi.
At P0, 16:0,18:1 was the major PC species followed by 16:0,16:0, 16:0,16:1 and 18:0,18:1 (Fig. 7). At day 21 and in adulthood, 18:0,18:1 and 16:0,18:1 were found to be the most abundant PC species, and the content of polyunsaturates such as 16:0,20:4-, 18:0,20:4-, 16:0,22:5- and 16:0,22:6-PC was increased. As observed with PE, levels of PC molecular species were generally decreased by ethanol, resulting in the reduction in total PC shown in Fig. 4. Although levels of some individual PC species showed a tendency to decrease after ethanol exposure at P21, the difference was statistically significant only for 18:0,22:6-PC.
Effect of ethanol on fatty acid composition of phospholipids
The effect of ethanol exposure on the fatty acid composition in rat brain hippocampi is summarized in Table 2. Both in the control and ethanol groups, 16:0 (22.15–33.81%) and 18:0 (19.94–24.29%) were the major saturated fatty acids (SFAs). There were no significant differences in total SFA between the control and ethanol groups at all ages tested. The major monounsaturated fatty acid (MUFA) species was 18:1n-9, and its content increased from 13% at P0 to 18% in adulthood. No significant differences were observed between the control and ethanol groups for MUFAs. The major PUFA observed in P0 hippocampus was 20:4n-6 (9.26 ± 0.80%) followed by 22:6n-3 (7.53 ± 0.45%) and 22:5n-6 (3.04 ± 0.16%). Levels of both 20:4n-6 and 22:6n-3 increased significantly with age whereas 22:5n-6 decreased. At all ages tested, ethanol did not alter the 20:4n-6 content. On the contrary, the percentage of 22:6n-3 was significantly reduced by ethanol at P0 and P21, although the extent of the reduction was lessened with age. In dams' hippocampi, the difference in the 22:6n-3 content was not statistically significant. Despite low levels, a concomitant increase in 22:5n-6 was observed in ethanol-treated animals. As a consequence, the ratio of 22:6n-3/22:5n-6 decreased significantly in the ethanol groups. The total PUFA content was not altered significantly by ethanol exposure at any age tested.
Table 2. Effect of ethanol exposure on total fatty acid composition in rat hippocampi
Values are mean ± SD (n = 3) percentage of total identified fatty acids by weight. ND, not detected. *p < 0.05, †p < 0.01 versus control (Student's t-test).
The absolute total fatty acid levels were not significantly different between the control (825, 1249 and 1132 pmol per μg protein for P0, P21 pups and dams, respectively) and ethanol groups (814, 1287 and 1113 pmol per μg protein, respectively). These results are consistent with the levels of total phospholipids, which also showed no significant differences between the control (387, 528 and 567 pmol per μg protein for P0, P21 pups and dams, respectively) and ethanol (398, 537 and 576 pmol per μg protein, respectively) groups.
We have reported previously that 22:6n-3 promotes PS biosynthesis (Garcia et al. 1998; Kim and Hamilton 2000), and that this process is selectively inhibited by chronic ethanol exposure in C-6 glioma cells (Kim and Hamilton 2000). To test whether hippocampus is affected by ethanol in a similar manner, we evaluated the changes in phospholipid profile caused by ethanol exposure in rat hippocampus during the developmental period when 22:6n-3 accumulates predominantly. Detailed analysis of phospholipid molecular species by HPLC/ESI–MS revealed that ethanol depletes PS, especially 22:6-containing species, from hippocampal membranes.
The reduction of PS by ethanol exposure was observed concomitantly with an increase in PE and a decrease in PC, but no change in total phospholipids, suggesting a possible effect of ethanol on phospholipid remodeling processes in hippocampus. It is known that in mammalian tissues PS is synthesized from pre-existing PC and PE mainly by a l-serine base-exchange reaction catalysed by PSS1 and PSS2 respectively (Dennis and Kennedy 1972; Vance 1990, 1998; Kuge and Nishijima 1997). On the other hand, PS can be converted to PE by mitochondrial PS decarboxylation (Borkenhagen et al. 1961; Dowhan 1997; Voelker 1997; Vance 1998). It is conceivable that the reduction in PS caused by ethanol treatment results from inhibition of PS synthesis by PSS1 and/or PSS2 pathways. The concomitant increase in PE observed in this study suggests that inhibition of PSS2 may have a particular role in reducing PS in ethanol-exposed hippocampi. Indeed, in dam's hippocampus 18:0,22:6-PE was increased by ethanol to a greater extent than other PE species, although most PE species were generally increased at other ages. Alternatively, the PSD pathway may be stimulated by exposure to ethanol, contributing to the observed decrease in PS.
Although the total phospholipid level did not change, it is also possible that the de novo synthesis of PC and PE by the Kennedy pathway (Kennedy and Weiss 1956) was affected by ethanol. It has been reported that PC synthesis from CDP-choline and diacylglycerol (DAG) is inhibited by ethanol in rat hepatocytes (Carrasco et al. 2002). As both PC and PE are synthesized from DAG, a decrease in PC synthesis by ethanol may increase the availability of DAG substrate for the CDP-ethanolamine pathway, increasing the levels of PE. Our observation that most molecular species were generally increased in PE or decreased in PC after ethanol exposure is consistent with this view. The reduction in PC substrate for PSS1 may in turn reduce PS synthesis, leading to a decrease in PS levels as observed in this study. Furthermore, inhibition of PE methylation by PE N-methyltransferase (Vance 1998) cannot be excluded as a possible explanation for the observed reduction in PC and the concomitant increase in PE. In addition, it is important to consider that ethanol may alter phospholipid retailoring processes by deacylation/reacylation of phospholipids. Because a phospholipid species containing 22:6n-3 serves as the best substrate for the PS synthesis (Kim et al. 2004), modification of the fatty acyl composition in the substrate pool by ethanol may consequently lead to alteration in the PS content. Detailed investigation of the effects of ethanol at the level of enzymes involved in each metabolic pathway is required to elucidate the mechanism underlying the ethanol-induced alteration in phospholipid composition.
The PS concentration as well as 22:6n-3 content in mammalian brain varies according to age, brain region, type of cell and subcellular components. It has been shown that the content of membrane 22:6n-3 influences the synthesis and accumulation of PS (Garcia et al. 1998; Hamilton et al. 2000; Murthy et al. 2002; Mozzi et al. 2003). Cells with high levels of 22:6n-3 appeared to contain high levels of PS, possibly because 22:6n-3-containing phospholipids are favorable substrates for the PS synthesis (Hamilton et al. 2000; Kim et al. 2004). In the present study, the increase of 22:6n-3 from P0 to P21 at the total fatty acid level (Table 2) was well reflected by the increase in PS level (Fig. 3). In this context, the reduction in PS level at P0 and P21 after ethanol exposure was also in agreement with slight but significant decreases in 22:6n-3 content at the total fatty acid level. However, the decrease in PS induced by ethanol was not necessarily associated with significant changes in 22:6n-3, as observed in dams' hippocampi, suggesting that mechanisms other than availability of favorable substrates contribute to the negative effect of ethanol on the PS level in neuronal membranes.
It is well established that membrane PS plays an important role in various signaling pathways supporting cellular function (McPherson et al. 1999; Bittova et al. 2001). Neuronal survival under adverse conditions has been shown to be promoted by PS accumulation in cell membranes (Kim et al. 2000; Akbar and Kim 2002). It is also well documented that disruption of PC homeostasis can induce growth arrest or cell death (Cui and Houweling 2002). In addition, a proper PE composition in membranes has been shown to be important for differentiation of epithelial cells (Kano-Sueoka et al. 2001). It is therefore conceivable that the particular reduction of PS, together with alterations in PC and PE, induced by ethanol may have significant functional consequences in hippocampus.
In conclusion, the present study demonstrated that ethanol exposure during prenatal and developmental stages significantly alters the phospholipid profile, particularly reducing the level of 22:6n-3-containing PS, in rat brain hippocampus. According to our data, long-term ethanol exposure in adulthood can also generate similar effects. Although the exact molecular mechanisms underlying the effect of membrane phospholipid profile on cellular function have yet to be established, it has been demonstrated that receptor function (Litman and Mitchell 1996) as well as activation of signaling proteins (McPherson et al. 1999; Kim et al. 2000; Bittova et al. 2001) can be influenced by the membrane phospholipid bilayer. Therefore, alterations in the hippocampal phospholipid profile resulting from ethanol exposure may have a significant impact on signaling processes mediated by receptors such as GABAA, glutamate, glycine and 5-HT3 receptors, which are known to be localized in the hippocampus (White et al. 2000). In this context, the observed changes in hippocampal phospholipids due to ethanol including the particular loss of 22:6n-3-containing PS together with the alteration in total PC and PE levels, may underlie the cognitive deficits associated with fetal alcohol syndrome.
The authors are grateful to Dr Frances Calderon and Raouf Kechrid for their assistance with the collection of rat brain hippocampi.