Vitamin E,⁷ RRR--tocopherol (-T)⁸ plays an essential role in protecting cell membrane polyunsaturated fatty acids (PUFA) against oxidation (Tappel 1962). Besides its antioxidant function, -T plays a role in stabilizing cell membranes and altering their permeability properties (Suzuki et al. 1993). Kagan and co-workers (1990) have shown that -T becomes associated with the membrane through a physiochemical interaction of the -T side chain with PUFA residues in membrane phospholipids. They have further demonstrated that -T helps protect membranes against the damaging effects of phospholipase A₂ and its products (i.e., free fatty acids and lysophospholipids). These actions of -T could be particularly important in immune cells in which arachidonic acid and phospholipase A₂ are integral parts of eicosanoid biosynthesis (Chilton et al. 1996).

Increasing consumption of PUFA increases the dietary requirement for vitamin E (Muggli 1994). A number of researchers have reported that (n-3) PUFA reduce vitamin E levels in the blood and tissues more than (n-6) PUFA (Fritsche et al. 1992, Javouhey-Donzel et al. 1993, Meydani, S. et al. 1987, 1988, Mouri et al. 1984). In large measure, these studies used a single source of fish oil to provide (n-3) PUFA. Adverse effects of fish oil on tissue vitamin E status have been attributed to (n-3) PUFA within fish oil rather than other factors, although few studies have simultaneously compared this effect in different sources of fish oil with similar levels of total (n-3) fatty acids but dissimilar in other lipid soluble components. Content of eicosapentaenoic acid [EPA, 20:5(n-3)] and docosahexaenoic acid [DHA, 22:6(n-3)], the primary (n-3) fatty acids in fish oil, varies with fish oil source. For example, menhaden fish oil and cod liver oil have approximately equivalent EPA levels (~10%), whereas sardine fish oil has ~20% of its total fatty acids occurring as EPA (Alexander et al. 1995). In addition, fish liver oils such as cod liver oil have higher vitamin A levels than do fish oils that are whole-body products. Increased dietary vitamin A has been shown to antagonize vitamin E status in poultry (Abawi and Sullivan 1989, Tengerdy and Brown 1977). In fact, inclusion of cod liver oil into vitamin E-free diets has historically been a common practice of researchers wishing to precipitate symptoms of vitamin E deficiency.

Our laboratory has been investigating the interaction between (n-3) PUFA and -T on the immune system. Relatively little is known about the -T content of isolated immune cells or factors that influence it, although immune cells contain at least one order of magnitude more -T than erythrocytes or platelets (Hatam and Kayden 1979). Furthermore, cellular -T concentration differs as much as 10-fold among immune cell populations (Alexander et al. 1995, Pacht et al. 1986). Recently, we measured the effect of three different (n-3) PUFA-rich fat sources on the -T content of rat immune cells. Fish oils stripped of endogenous -T could not be obtained; thus, the level of -T in each lipid source was equalized with free RRR--T to 150 mg/kg oil, the concentration of naturally occurring -T in our cod liver oil. Addition of AIN-76 vitamin mix to the semipurified diets provided total tocopherol of 60 mg/kg, or two times the AIN recommended level. We demonstrated that when dietary vitamin E levels were equalized in both amount and form in this fashion, immune cells from fish oil-fed rats maintained their vitamin E levels in the face of reduced circulating vitamin E (Alexander et al. 1995).

The purpose of the current study was to determine if similar high fat, fish oil diets containing a single source and form of vitamin E (i.e., RRR--T from the lipid source), supplied at a level consistent with the NRC requirement for growing rats (i.e., 30 mg/kg of diet), would compromise immune cell vitamin E status. Furthermore, we were interested in determining whether 5 wk was sufficient to observe a fish oil effect on immune cell vitamin E. Tissue samples were collected at both 5 and 10 wk after initiation of dietary treatments. These time points seemed appropriate in light of a previous report by Bieri (1972), who showed that, in general, diet-induced changes in tissue vitamin E status in rats stabilized after 4 wk although some tissues required as long as 8 wk to reach a stable level. In this study, we report the paradoxical finding that splenocyte -T concentration at both 5 and 10 wk and thymocyte -T concentration at 10 wk are higher in rats fed menhaden fish oil or sardine oil compared with those fed lard, a fat low in PUFA. The fact that these diets contained the same amount and chemical form of vitamin E and that there were no significant differences in circulating vitamin E among these three groups of rats at 10 wk post-treatment makes our observations all the more interesting.

MATERIAL AND METHODS

After a 3-d adjustment period to ensure that all rats were in good health, each rat was placed into one of five dietary treatment groups that were assigned in a stratified manner such that each diet was represented at each position and each level of the housing battery. The four lipid sources, tocopherol-stripped lard (LRD), menhaden fish oil (MFO), sardine fish oil (SFO) and cod liver oil (CLO) were mixed with tocopherol-stripped corn oil to provide sufficient dietary linoleic acid [18; 2(n-6)], which was equalized among dietary treatment groups. The MFO and SRD were a kind gift of Zapata Haynie Protein (Reedville, VA). Fatty acid composition of these four lipid blends is presented in Table 1.

Table 1. Fatty acid composition of fat blends used in experimental diets [View Table]

Endogenous -T in the lipid blends was measured in duplicate as described by Slover and Thompson (1983) using cold saponification followed by extraction and analyzed for -T by HPLC as described later. In this study, -T was the only form of vitamin E detected in the fish oils. Oil blends were then equalized with RRR--tocopherol (a gift from Eastman Kodak Chemical, Rochester, NY) to a final concentration of 150 mg/kg of oil, thereby supplying 30 mg/kg -T in a 20% fat diet. This approach was chosen because fish oils cannot be obtained stripped of endogenous vitamin E. To do so generally reduces their stability and quality. A fifth treatment group (a negative control) was fed a diet containing 20% tocopherol-stripped lard without any added vitamin E (LRD). We were unable to detect any vitamin E in the stripped lard (<0.1 mg/kg). The diets were isocaloric and formulated according to AIN guidelines (AIN 1980) with minor modifications. The composition of these diets was as follows (g/100 g): casein, 20; DL-methionine, 0.3; cornstarch, 20; dextrose, 29.3; -cellulose, 5; mineral mix (AIN-76), 4; vitamin mix (AIN-76A, vitamin E-omitted), 1.2; choline bitartrate, 0.2; and fat, 20. All dietary ingredients were purchased from U. S. Biochemical (Cleveland, OH), unless otherwise noted.

Rats had free access to fresh diet every day with any remaining diet being discarded and free access to distilled water. Auto-oxidation of the diets was prevented by the addition of 1.2 µmol/L tert-butyl hydroquinone (Eastman Kodak Chemical) to the oils as described by Fritsche and Johnston (1988). Oils were mixed into the dry components of the diet in small batches and stored at 4°C. Oxidative stability of these diets was assessed on three separate occasions during the study, by leaving aliquots of each diet in feed bowls for 48 h at room temperature. Following ether extraction, the total lipids were saponified and analyzed for -T by HPLC as described later. Furthermore, peroxide values were also determined on each extract following a standardized method (# Cd8-53, AOCS 1986). The peroxide value never exceeded 15 meQ/kg and no loss of -T in the diets was detected throughout the study.

All immune cell samples were enumerated electronically by a Coulter Counter, Model ZBI (Coulter Electronics, Hialeah, FL). A small aliquot of each cell preparation was subjected to centrifugation in a Cytospin 3 250 × g for 5 min (Shandon, Pittsburgh, PA) in order to deposit cells onto a microscope slide and stained in a Wescor 7100 slide stainer (Wescor, Logan, UT). Differential counts of immune cell preparations were done using a light microscope. Cell samples were pelleted and stored in 1 mL of 10 mmol/L EDTA at 80°C for vitamin E analysis.

Quantitation of -T in sample extracts was carried out by HPLC (Beckman System Gold HPLC with a 126A Programmable Pump and a 506 Autosampler, San Ramon, CA) equipped with a C-18 reverse-phase column (15 cm × 4.6 mm; 3 µm; Supelco, Bellefonte, PA). The mobile phase consisted of 98% methanol and 2% water (flow rate, 1.5 mL/min). Plasma and immune cell tocopherols were monitored at 292 nm using a Beckman 166 Programmable UV Detector. Oil and liver tocopherol determinations were conducted with fluorometric detection (excitation, 232 nm and emission, 328 nm) using a Perkin Elmer LS-3B Fluorescence Spectrometer (Norwalk, CT) to avoid interfering peaks. Sample -T concentrations were determined from peak area responses of known amounts of -T using a computer spreadsheet program (Microsoft Excel, v. 3.0, Redmond, WA). Values were corrected for losses during processing by following the recovery of the internal standard, which generally exceeded 90%. The lower level of detection for -T was 0.027 nmol per 100 µL of plasma. This procedure could also detect tocopherol quinone, a major oxidation product of vitamin E.

RESULTS

Table 2. Body, liver and immune organ weights for female rats fed diets containing 20% lard or different sources of fish oil for 5 or 10 wk [View Table]

After 10 wk of consuming experimental diets, there were no significant differences in body weight among treatment groups. However, relative liver weights continued to be greater for all fish oil-fed rats. Dietary fat source did not influence thymus weight after either 5 or 10 wk of dietary treatment.

Table 3. Plasma and liver -tocopherol (-T ) concentration in female rats fed diets containing 20% lard or different sources of fish oil for 5 or 10 wk [View Table]

The concentration of liver total lipids was significantly elevated in rats fed SRD for 5 wk compared with those fed all other diets except MFO, which was intermediate in value. By 10 wk, both SRD- and MFO-fed rats had significantly elevated liver lipids compared with rats fed the other fat sources (Table 3). Although the concentration of -T in the liver was significantly greater in rats fed LRD than in those fed LRD , other patterns of the effect of diet on liver tocopherol were not so clear. At 5 wk, LRD- and MFO-fed rats did not have significantly different liver -T concentrations, expressed on a lipid basis, although values for LRD-fed rats were significantly greater than those of SRD- and CLO-fed rats. Liver -T concentration from SRD- and CLO-fed rats did not different from each other. By wk 10, LRD-, MFO- and SRD-fed rats had significantly higher liver -T than did rats fed LRD , with CLO-fed rats having an intermediate value. At no time was tocopherol quinone, an oxidation product of -T, detected in any liver or plasma samples.

Table 4. Immune cell -tocopherol concentrations in female rats fed diets containing 20% lard or different sources of fish oil for 5 or 10 wk [View Table]

Unlike splenocytes, length of time consuming diet (i.e., 5 vs. 10 wk) did not significantly influence thymocyte -T content. Thymocyte -T content was generally lower on a per cell basis than that of isolated splenocytes. After 5 wk of consuming experimental diets, only thymocytes from MFO-fed rats had -T content significantly greater than those from rats fed the other fat sources. After 10 wk, thymocytes from both the SRD-fed and MFO-fed rats had significantly greater -T content than did other diet groups. At no time was the -T content of thymocytes from CLO-fed rats significantly different than that of rats fed LRD or LRD . Differential counts of immune cell preparations done using an aliquot of each cell preparation showed no differences in immune populations resulting from dietary treatment or time (data not shown).

DISCUSSION

In contrast to the previous study in which we found that feeding fish oil decreased plasma and hepatic -T concentrations at both 5 and 10 wk compared with LRD, in this study, MFO-fed rats did not have significantly lower hepatic or plasma -T concentration at either time point. In fact, after 10 weeks of consuming the fish oil diets, only CLO-fed rats had significantly lower plasma -T concentration than did rats fed LRD. Interestingly, including menhaden fish oil or sardine fish oil with only -T in the free alcohol form at the recommended level was also associated with significantly higher splenocyte -T at both time points and significantly greater thymocyte -T at 10 wk compared with LRD. These unique observations occurred at compromised levels of plasma and liver -T levels that were <50% of those reported in our previous study (Alexander et al. 1995). Examination of the plasma and hepatic -T concentrations in rats fed the vitamin E-free (LRD) diet supports the conclusion that the vitamin E status of these rats was compromised.

We believe that the decreased -T status of rats in the present study plays an important part in the (n-3) PUFA-induced enrichment in immune cell -T we observed and may also explain why it has not been previously reported in immune cells. Consistent with this proposal are the findings of Verdon and Blumberg (1988), who reported that protein-mediated transfer of -T by hepatic liver microsomes and mitochondria was enhanced only when vitamin E status is compromised. The newly identified intracellular 15-kDa tocopherol binding protein and the plasma membrane tocopherol binding protein are thought to provide for vitamin E uptake, intracellular distribution and retention of -T in tissues (Dutta-Roy et al. 1993 and 1994, Traber et al. 1993). Differential expression of these tocopherol binding proteins, which could up-regulate -T levels in tissues, has not yet been demonstrated and it is not clear that immune cells possess -T binding protein; however, speculation about such a mechanism is intriguing. Many tissues such as kidney, erythrocytes, platelets and liver have been shown to up-regulate cytosolic antioxidant defenses in response to dietary (n-3) PUFA (Chandrasekar and Fernandez 1994, Christon et al 1995, Joulain et al 1994, Nanji et al 1995). Immune cells, which have high PUFA content as well as increased oxidative metabolism during an immune response, provide an excellent model for such investigations.

Interestingly, other investigators have reported that greater -T concentration in several tissues may be observed when the tissues are enriched with (n-3) PUFA. Chautan et al. (1990) reported that, as the ratio of (n-3) to (n-6) PUFA increased in the diet,-T concentration in the heart, but not the liver of rats, was greater. They found a strong positive correlation between heart -T concentration and docosahexaenoic acid concentration (r = 0.86, P = 0.0001). Additionally, Croset and co-workers (1990) reported that human platelet -T was elevated in elderly people consuming 100 mg of eicosapentaenoic acid daily. Berlin et al. (1992) reported that in adult men, fish oil supplementation significantly increased both - and -tocopherol levels in erythrocyte membranes. In contrast, others have reported that vitamin E levels in a variety of tissues of rodents fed (n-3) PUFA and (n-6) PUFA are similar to or lower than the levels in rodents fed low PUFA diets (Cho and Choi 1994, Leibovitz et al. 1990, Meydani, M. et al. 1991, Meydani, S. et al. 1987 and 1988).

There may be several reasons for these discrepancies, including differences in tissues examined, length of time on diets, and amount of PUFA and total fat being fed. However, we believe an important point for consideration to be the common practice of assuming that the biopotency estimates for various chemical forms of vitamin E remain the same regardless of differences in dietary fat source. The biopotency of various forms of vitamin E has been standardized for animal and human nutrition and is the basis for the international unit (IU) designation. By this convention, one IU of vitamin E is equal to the biological response obtained with 1 mg of dl--tocopheryl acetate (dl--TA; "dl" refers to a mixture of equal parts RRR- and SRR- stereoisomers). The naturally occurring free form of vitamin E (RRR--tocopherol) and the racemic mixture of the synthetic free form (dl--T) are generally accepted to have biopotencies equal to 1.46 and 1.1 IU/mg, respectively. However, studies with rats using simultaneous determination of fetal resorption, myopathy (i.e., plasma pyruvate kinase) and liver storage capacity have shown that dl--T was only half as active as dl--TA (Leth and Sondergaard, 1983, Weiser and Vecchi, 1985). These estimates were obtained when vitamin E sources were gavaged in an oil carrier. However, when the free and acetate forms of the vitamin were administered in conjunction with food, the free tocopherol had a biopotency similar to the acetate form of the vitamin (Burton et al. 1988). Studies in weaned piglets also support the availability of free tocopherol when it is delivered in combination with feed (Chung et al. 1992). Thus, it would seem from our data that in high fat diets, -T behaves as if it were being gavaged in an oil carrier, rather than being fed in association with the diet. It is clear from these data and our previous findings (Alexander et al. 1995) that both the amount and form of vitamin E in the diet will affect the impact of (n-3) PUFA on the vitamin E status of immune cells.

Furthermore, our data suggest that -T has a much lower than expected bioavailability in rats. It could be argued that the low availability of -T in this study was a result of its oxidative destruction in the diet or within the digestive tract prior to absorption. Although such an explanation would be consistent with the rationale (i.e., greater stability) typically cited for using esterified forms of vitamin E (e.g., -T acetate and -T succinate) in diets, our data do not support this argument. For example, we regularly measured the concentration of -T in our diets and no loss was detected throughout the study. Furthermore, if oxidative destruction of dietary -T was occurring either in the diets or in the digestive tract, we would anticipate a greater loss of -T in the high PUFA fish oil diets than in the lard diet. Our plasma and hepatic -T data are not consistent with this possibility. Thus, oxidative loss of vitamin E activity in the diets cannot explain our findings.

Our demonstration of cod liver oil's ability to antagonize vitamin E status in rats is not novel. However, it should be pointed out that cod liver oil's dramatic effect on rat vitamin E status is unlikely to be solely a consequence of the (n-3) PUFA content of this oil because menhaden fish oil has a very similar (n-3) PUFA profile, but quite different effects. Cod liver oil was included primarily as an historical reference group because much of the early research establishing the concept that PUFA antagonized vitamin E status used cod liver oil as a means of precipitating vitamin E deficiency symptoms (Moore et al. 1959). The high content of fat-soluble vitamins in liver vs. whole-body fish oils makes it inappropriate to draw conclusions about possible effects of (n-3) PUFA from these fat sources on vitamin E status. In other words, dietary fats should not be considered simply as a source of esterified fatty acids.

It is possible that the three different sources of fish oil differentially enriched immune cells with (n-3) PUFA and that this, in turn, might explain why immune cell -tocopherol concentration varied among the three fish oil treatments. Regrettably, sample size precluded evaluation of both -tocopherol concentration and fatty acid composition of the immune cells. Without these data, it is not possible to entirely discount the scenario described above. However, we believe that two lines of evidence suggest that this is unlikely. First, a relatively large body of literature exists describing the qualitative and quantitative effects of various sources of (n-3) PUFA on immune cell fatty acid composition. Although no single study has documented the effect of different fish oils on the fatty acid composition of immune cells, comparisons among studies are possible. Based on the findings of Chapkin et al. (1988), Brouard and Pascaud (1990), Broughton et al. (1991) and Hinds and Sanders (1993) we believe the following general conclusions can be reached: 1) Immune cells from rats fed all three of the fish oil sources used in our study (i.e., menhaden fish oil, sardine oil and cod liver oil) would have been enriched with (n-3) PUFA, particularly EPA and DHA. 2) The difference in tissue (n-3) PUFA enrichment caused by feeding rats the various fish oil sources would be small in comparison with the total enrichment associated with any one of these three sources of long-chain (n-3) PUFA. Furthermore, the liver phospholipid fatty acid data from rats in our study indicated that tissue (n-3) PUFA enrichment was similar for rats fed cod liver oil and those fed menhaden fish oil or sardine oil. Thus, we believe it is unlikely that the differential effect of the three fish oils on immune cell -T results from differential immune cell (n-3) fatty acid enrichment.

In summary, this study investigated the effect of three different fish oil sources of (n-3) PUFA on immune cell vitamin E status when the sole source of vitamin E in the diet was free -T, supplied at 30 mg/kg of diet. Our data demonstrate the complexity of the relationship between dietary fish oil and vitamin E status in rats. Factors such as the source of fish oil, the level of fat in the diet, the amount and form of vitamin E in the diet and the tissue examined must be considered in both experimental design and conclusions drawn from data. That some mechanism exists by which immune cell vitamin E concentration may be maintained at higher levels in rats fed some sources of fish oil compared with those fed a low PUFA fat is both paradoxical and novel. Further research is required to delineate the mechanism(s) through which (n-3) PUFA affect immune cell vitamin E metabolism. Finally, it should be pointed out that we are not suggesting that cellular -T concentration in these immune cells is sufficient to meet the potentially higher antioxidant needs associated with lipid peroxidation in PUFA-enriched cells. Additional studies directed at examining immune cell susceptibility to lipid peroxidation as a result of increased (n-3) content will be necessary to address that question.

ACKNOWLEDGMENTS