点击显示 收起
1 From the Department of Biochemistry, University of Oxford, Oxford, United Kingdom; Unilever Research Colworth Laboratory, Sharnbrook, United Kingdom; The Hugh Sinclair Unit of Human Nutrition, Department of Food Science and Technology, University of Reading, Reading, United Kingdom; and The Institute of Human Nutrition, University of Southampton, Southampton, United Kingdom.
2 Supported by a grant to EAN and PCC from the Ministry of Agriculture, Fisheries and Food; by Unilever Research Colworth Laboratory; and by Nutricia Research under the Agro-Food LINK Programme (AFQ51). K562 cells were a gift from the Sir William Dunn School of Pathology, University of Oxford. 3 Address reprint requests to PC Calder, The Institute of Human Nutrition, University of Southampton, Bassett Crescent East, Southampton SO16 7PX, United Kingdom. E-mail: pcc{at}soton.ac.uk.
ABSTRACT
Background: Animal studies showed that dietary flaxseed oil [rich in the n-3 polyunsaturated fatty acid -linolenic acid (ALA)], evening primrose oil [rich in the n-6 polyunsaturated fatty acid -linolenic acid (GLA)], and fish oil [rich in the long-chain n-3 polyunsaturated fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)] can decrease natural killer (NK) cell activity. There have been no studies of the effect on NK cell activity of adding these oils to the diet of humans.
Objective: Our objective was to determine the effect of dietary supplementation with oil blends rich in ALA, GLA, arachidonic acid (AA), DHA, or EPA plus DHA (fish oil) on the NK cell activity of human peripheral blood mononuclear cells.
Design: A randomized, placebo-controlled, double-blind, parallel study was conducted. Healthy subjects aged 55–75 y consumed 9 capsules/d for 12 wk; the capsules contained placebo oil (an 80:20 mix of palm and sunflower seed oils) or blends of placebo oil and oils rich in ALA, GLA, AA, DHA, or EPA plus DHA. Subjects in these groups consumed 2 g ALA, 770 mg GLA, 680 mg AA, 720 mg DHA, or 1 g EPA plus DHA (720 mg EPA + 280 mg DHA) daily, respectively. Total fat intake from the capsules was 4 g/d.
Results: The fatty acid composition of plasma phospholipids changed significantly in the GLA, AA, DHA, and fish oil groups. NK cell activity was not significantly affected by the placebo, ALA, GLA, AA, or DHA treatment. Fish oil caused a significant reduction (mean decline: 48%) in NK cell activity that was fully reversed by 4 wk after supplementation had ceased.
Conclusion: A moderate amount of EPA but not of other n-6 or n-3 polyunsaturated fatty acids can decrease NK cell activity in healthy subjects.
Key Words: Fish oil • immunity • natural killer cell • n-6 fatty acid • n-3 fatty acid • polyunsaturated fatty acids • -linolenic acid • -linolenic acid • eicosapentaenoic acid • docosahexaenoic acid • arachidonic acid
INTRODUCTION
Natural killer (NK) cells are a class of lymphocytes found mainly in the blood and spleen. They are derived from the bone barrow but are not T or B lymphocytes and do not undergo thymic maturation. They are involved in the natural immune response to virus-infected and tumor cells and play a role in graft rejection (1). NK cell activity declines with age in both humans and laboratory animals (2, 3).
A reduction in fat intake (to <30% or 22% of total energy) is associated with a significant increase in the NK cell activity of human peripheral blood mononuclear cells (PBMCs; a mixture of lymphocytes including NK cells and monocytes) (4, 5), suggesting that high fat consumption suppresses NK cell activity. This is supported by some studies in laboratory rodents (6) but not by others (7). Animal feeding studies show that the type of fatty acid in the diet affects NK cell activity; n-3 polyunsaturated fatty acids (PUFAs) appear to have particularly potent effects. Increasing the amount of -linolenic acid (ALA; 18:3n-3) in a rat diet lowers spleen NK cell activity compared with that measured after feeding diets rich in linoleic acid (8, 9). Feeding rats or mice diets containing large amounts of fish oil [FO; which is rich in eicosapentaenoic acid (EPA; 20:5n-3) and docosahexaenoic acid (DHA; 22:6n-3)] results in suppressed spleen NK cell activity compared with that measured after feeding low-fat diets or high-fat diets rich in saturated fat or n-6 PUFAs (10–15). Feeding rats diets rich in -linolenic acid (GLA; 18:3n-6) also lowers spleen NK cell activity compared with that measured after feeding some other diets (13, 14). These studies are supported by cell culture studies in which human NK cell activity was suppressed by culture of the cells with GLA (16, 17), EPA (17–19), and DHA (17–19).
Despite these cell culture and animal feeding studies, few investigations have been conducted on the effect of dietary fatty acids on human NK cell activity. If effects such as those identified in cell culture and animal feeding experiments occur in humans, there could be implications for the safety and clinical applications of the fatty acids concerned. Thus, it seems important to ascertain the effects of fatty acid supplementation on human NK cell activity. We therefore studied the effect of moderate supplementation with encapsulated oil blends rich in ALA, GLA, arachidonic acid (AA; 20:4n-6), DHA, or EPA plus DHA on NK cell activity in healthy, free-living subjects aged >55 y.
SUBJECTS AND METHODS
Materials
Phosphate-buffered saline (PBS) tablets were obtained from Unipath Ltd (Basingstoke, United Kingdom). Histopaque, bovine serum albumin (fatty acid free), HEPES-buffered RPMI medium, glutamine, antibiotics (penicillin and streptomycin), and propidium iodide were obtained from Sigma Chemical Co Ltd (Poole, United Kingdom). Fluorescent-labeled monoclonal antibodies (anti-CD45/CD14 and anti-CD3/CD16/CD56) were purchased from Coulter Corp (Hialeah, FL). FACS-lysing solution was purchased from Becton Dickinson (Mountain View, CA). Cytotoxicity kits were purchased from Proteins International Inc (Rochester, MI). Solvents were purchased from Fisher Scientific (Loughborough, United Kingdom).
Subjects and study design
Ethical permission for all procedures involving human volunteers was obtained from the Central Oxford Research Ethics Committee (COREC no. 96.182). All volunteers completed a health and lifestyle questionnaire before entering the study and a doctor's consent for inclusion in the study was obtained. Volunteers were excluded if they were taking any prescribed medication; had diagnosed hypercholesterolemia, hypertriglyceridemia, coronary artery disease, or diabetes; had a chronic inflammatory disease; took aspirin regularly; were vegetarian; or consumed FO, evening primrose oil, or vitamin capsules. The characteristics of the 46 subjects who completed the study are given in Table 1; the subjects' mean age and body mass index did not differ significantly among the groups. All subjects were white and lived in their own homes; none were disabled or immobile in any way. Twenty-seven subjects were employed full-time and 19 were retired. All female subjects were postmenopausal.
View this table:
TABLE 1. Characteristics of the treatment groups1
Forty-eight subjects were randomly allocated to receive in a double-blind fashion 6 types of encapsulated oil blends (n = 8 per treatment group): placebo, ALA, GLA, AA, DHA, and FO. The types and amounts of oils used in the blends are shown in Table 2. The fatty acid composition of the oil blends (as determined by gas chromatography of total lipid extracts) is shown in Table 3. The capsules were gelatin-coated, nontransparent, and green. Each capsule contained 445 mg of the oil blend and subjects consumed 9 capsules/d (ie, 4 g encapsulated oil/d) for 12 wk. Therefore, subjects in the ALA group consumed an extra 2.0 g ALA/d; subjects in the GLA, AA, and DHA groups consumed an extra 700 mg (approximately) GLA, AA, or DHA/d, respectively; and subjects in the FO group consumed an extra 1 g EPA plus DHA/d (720 mg EPA + 280 mg DHA). All capsules contained 300 µg -tocopherol equivalents plus 180 µg ascorbyl-palmitate/g oil. Thus, all subjects consumed an extra 1.2 mg -tocopherol/d. Two female subjects dropped out of the study, one from the GLA group because of illness and one from the FO group because of an inability to comply as a result of stomach upset.
View this table:
TABLE 2. Types and amounts of oils used to make the blends used in the study1
View this table:
TABLE 3. Fatty acid composition of the oil blends used in the study1
Blood was sampled 4 wk before and immediately before the start of supplementation, every 4 wk during supplementation, and after a 4-wk washout period. Plasma phospholipid fatty acid composition, plasma concentrations of thiobarbituric acid–reactive substances (TBARS), and blood leukocyte numbers and subsets were not measured in the first baseline sample. NK cell activities were measured in both baseline samples; there were no significant differences between the 2 baseline samples and the average of the 2 measurements is reported here. Throughout this article, week 0 represents the baseline measurements; weeks 4, 8, and 12 the supplementation period; and week 16 the end of the washout period. All treatment groups completed the study in parallel. The study ran from June 1997 (early summer) to December 1997 (early winter).
The capsules were provided to the subjects in blister packs (9 capsules per pack; 7 blister packs per box) along with clear instructions on how they should be administered (3 capsules 3 times daily immediately before breakfast, lunch, and dinner). During the supplementation period subjects received fresh blister packs of capsules every 4 wk. Compliance was assessed by questionnaire and biochemically by determining the plasma phospholipid fatty acid composition.
Subjects completed a 7-d food diary. Habitual nutrient intakes were determined by using FOODBASE (V1.3; Institute of Brain Chemistry, London).
Preparation of PBMCs
Blood samples were collected into heparin-treated tubes and diluted 1:1 with PBS. The diluted blood was layered onto histopaque (density: 1.077 g/L; ratio of diluted blood to histopaque: 4:3) and centrifuged for 15 min at 800 x g and 20°C. The cells were washed once with PBS, resuspended in 2.5 mL PBS, and layered onto 5 mL histopaque. They were then centrifuged once more to achieve a lower degree of erythrocyte contamination, washed with PBS, and finally resuspended in the appropriate medium.
Measurement of NK cell activity
The activity of NK cells, in preparations of PBMCs, against the K562 target cell line was determined by measuring the release of cellular lactate dehydrogenase (LDH) on target cell lysis. PBMCs were prepared as described above and resuspended in HEPES-buffered RPMI medium supplemented with 2 mmol glutamine/L and antibiotics. K562 cells were washed 3 times and resuspended in HEPES-buffered RPMI medium supplemented with 10 g bovine serum albumin/L and antibiotics. PBMCs (1 x 106, 5 x 105, or 2.5 x 105/100 µL) and K562 cells (1 x 104/50 µL) were added to each well of a 96-well microtiter plate to give ratios of PBMCs to K562 cells of 100:1, 50:1, or 25:1 in a final volume of 150 µL. Maximal target cell lysis was assessed by incubating K562 cells with the lysing reagent provided in the cytotoxicity kit. Spontaneous release of LDH by PBMCs was assessed by incubating the PBMCs in the absence of target cells; spontaneous release of LDH by K562 cells was assessed by incubating the K562 cells in the absence of effector cells. Plates were centrifuged for 3 min at 100 x g and room temperature to bring effector and target cells into contact and were then incubated at 37°C in air:carbon dioxide (19:1) for 4 h. After incubation, 50 µL ice-cold RPMI medium was added to each well and the plates were centrifuged for 5 min at 400 x g and room temperature. A portion (100 µL) of the medium was removed from each well and transferred to the wells of a flat-bottomed 96-well plate. The activity of LDH in the medium was assayed according to the instructions provided by the manufacturer of the kit. Percentage cytolysis was calculated as follows:
RESULTS
Habitual nutrient intakes
There were no significant differences between men and women with respect to habitual intakes of total energy; of fat, carbohydrate, and protein as a percentage of dietary energy; of total fat and individual fatty acids (g/d); or of cholesterol, -tocopherol, ß-carotene, and vitamin C (data not shown). Men consumed significantly more alcohol than did women (14.1 ± 2.4 compared with 7.0 ± 1.7 g/d).
There were no significant differences among treatment groups with respect to habitual intakes of total energy (11.6 ± 0.6 MJ/d for the 46 subjects who completed the study); of fat, carbohydrate, and protein as a percentage of energy (32.8 ± 0.8%, 45.0 ± 0.8%, and 19.5 ± 0.9%, respectively); or of cholesterol (284.5 ± 13.5 mg/d), -tocopherol (7.6 ± 0.5 mg/d), ß-carotene (2.7 ± 0.5 mg/d), and vitamin C (140 ± 18 mg/d). Habitual intakes of individual fatty acids did not differ significantly among the treatment groups; data for all subjects combined are shown in Table 4. Likewise, habitual intakes of total saturated fatty acids, total monounsaturated fatty acids, total PUFAs, total n-6 PUFAs, and total n-3 PUFAs did not differ significantly among treatment groups (data not shown); the ratio of n-6 to n-3 PUFAs in the habitual diet also did not differ significantly among groups (9.0 ± 0.7).
View this table:
TABLE 4. Habitual fatty acid intakes of all subjects and of subjects in the different treatment groups during supplementation1
Fatty acid intakes during treatment
Intakes of individual fatty acids during the treatment period were calculated by adding habitual intakes to intakes from the supplements (Table 4). Intakes of myristic (14:0), palmitic (16:0), palmitoleic (16:1n-7), stearic (18:0), oleic (18:1n-9), and linoleic (18:2n-6) acids were higher than habitual intakes in at least some treatment groups during supplementation, but not significantly so. However, intakes of these fatty acids were not significantly different among the treatment groups. In contrast, there were significant differences in intakes of ALA, GLA, AA, EPA, and DHA among the different treatment groups.
Fatty acid composition of plasma phospholipids
The fatty acid composition of plasma phospholipids was not significantly affected by the placebo or ALA treatment. ALA and GLA did not appear in plasma phospholipids, even in those subjects in the ALA and GLA groups. Two-factor ANOVA did not detect any significant effects of time or treatment on the proportions of palmitic acid (30% by wt of total fatty acids) or stearic acid (13% by wt of total fatty acids). There were effects of time and treatment on the proportions of oleic acid (P = 0.015 for effect of time and P < 0.001 for effect of treatment) and linoleic acid (P < 0.001 for effects of both time and treatment) in plasma phospholipids, although there were no significant time x treatment interactions. In contrast, there were significant time-dependent effects of treatment on the proportions of dihomo--linolenic acid (DGLA; 20:3n-6), AA, EPA, and DHA in plasma phospholipids (two-factor ANOVA effects of both time and treatment, P < 0.001 in all cases; time x treatment interaction, P < 0.004 for DGLA and P < 0.001 for AA, EPA, and DHA). These effects were investigated further by one-factor ANOVA.
Effects of GLA supplementation
The proportion of DGLA in plasma phospholipids was significantly higher after 12 wk of GLA supplementation than at baseline (Table 5), and at this time point GLA was significantly higher in subjects in the GLA group than in each of the other groups, including the placebo group. The proportion of DGLA was significantly higher at weeks 4 and 8 in the GLA group than in the ALA, AA, DHA, and FO groups. GLA supplementation also increased the proportion of AA in plasma phospholipids, which was significantly higher after 12 wk supplementation than at baseline.
View this table:
TABLE 5. Effect of supplementation on the fatty acid composition of plasma phospholipids1
Effects of AA supplementation
The proportion of AA in plasma phospholipids was significantly higher after 4, 8, and 12 wk of AA supplementation than at baseline and after the washout period (Table 5). After 4 wk of washout the proportion of AA in plasma phospholipids was not significantly different from that observed at baseline. The proportion of AA was significantly higher at weeks 4, 8, and 12 in the AA group than in each of the other groups, including the placebo group. There was also a nonsignificant trend toward a lower proportion of EPA in the plasma phospholipids of subjects in the AA group after 12 wk of supplementation than at baseline.
Effects of DHA supplementation
The proportion of DHA in plasma phospholipids tended to increase after 4 wk of DHA supplementation (NS) and was significantly higher after 8 and 12 wk of DHA supplementation than at baseline and after washout (Table 5). After 4 wk of washout the proportion of DHA in plasma phospholipids was not significantly different from that at baseline. The proportion of DHA was significantly higher at weeks 8 and 12 in the DHA group than in each of the other groups, including the placebo group. The proportion of AA in plasma phospholipids after 8 wk was significantly lower in subjects in the DHA group than in each of the other groups.
Effects of FO supplementation
There was a significant increase in the proportion of EPA in the plasma phospholipids of subjects taking the FO supplement, such that the proportion of EPA was significantly higher after 4, 8, or 12 wk of supplementation than at baseline and was significantly higher in the FO group than in each of other groups, including the placebo group, after 4, 8, and 12 wk of supplementation (Table 5). This increase in the proportion of EPA in plasma phospholipids was near maximal after 4 wk of FO supplementation. The maximal increase in the proportion of EPA was 2-fold above baseline. There were trends toward an increase in the proportion of DHA and a decrease in the proportion of AA in plasma phospholipids but these were not significant.
Effect on plasma TBARS concentrations and on leukocyte numbers and subsets
There was no significant effect of any of the treatments on plasma TBARS concentrations (Table 6). There was also no significant effect of any of the treatments on the total number of leukocytes, lymphocytes, T lymphocytes, or NK cells/L blood (Table 7). Likewise, there was no significant effect of any of the treatments on the proportion of leukocytes as total lymphocytes or on the proportion of lymphocytes as T lymphocytes or NK cells.
View this table:
TABLE 6. Plasma concentrations of thiobarbituric acid–reactive substances (TBARS)1
View this table:
TABLE 7. Circulating total lymphocyte and lymphocyte subpopulations1
Effects on NK cell activity
NK cell activity measured at each PBMC-to-K562 target cell ratio did not differ significantly among the treatment groups at baseline; data for the PBMC:K562 of 100:1 are shown in Table 8. There were no significant effects of time or treatment nor was there a significant time x treatment interaction for NK cell activity measured at a PBMC:K562 of 25:1 or 50:1. However at the PBMC:K562 of 100:1 there were significant effects of time and treatment (P < 0.001 and P = 0.001, respectively) and a significant time x treatment interaction (P = 0.024). One-factor ANOVA indicated that at this ratio NK cell activity was not significantly affected by treatment with placebo, ALA, GLA, AA, or DHA (Table 8). However, treatment with FO resulted in a significant decrease in NK cell activity, such that activity was significantly lower after 12 wk of supplementation than after 0 and 4 wk (Table 8 and Figures 1 and 2). After 4 wk of washout (ie, week 16), NK cell activity in the FO group had returned to baseline values; activity at 16 wk was significantly higher than at the end of supplementation (12 wk; Table 8 and Figure 1).
View this table:
TABLE 8. Natural killer cell activity of peripheral blood mononuclear cells (PBMCs)1
FIGURE 1. . Time-dependent changes in natural killer (NK) cell activity in the placebo and fish oil (FO) groups. Data are expressed as mean (±SEM) NK cell activity (determined at a ratio of peripheral blood mononuclear cells to K562 target cells of 100:1) as a percentage of activity at baseline (n = 7 or 8). *Significantly different from placebo, P = 0.045 at week 8 and P = 0.028 at week 12. §Significantly different from baseline, week 4, and the end of the washout period, P = 0.031.
FIGURE 2. . Mean natural killer (NK) cell activity in the placebo and fish oil (FO) groups before supplementation (week 0) and at the end of supplementation (week 12). NK cell activity was determined at a ratio of peripheral blood mononuclear cells to K562 target cells (PBMC:K562) of 25:1, 50:1, and 100:1 (n = 7 or 8); error bars were omitted for clarity. *Significantly different from baseline (week 0), P = 0.034.
NK cell activity determined at a PBMC:K562 of 100:1 declined in all subjects in the FO group. The mean decline was 20% after 8 wk and 48% (range: 10–75%) after 12 wk of FO supplementation (Figure 1). The absolute change in NK cell activity determined at a PBMC:K562 of 100:1 after 12 wk of treatment with FO was significantly different from the absolute change in the placebo group (Table 8). Although NK cell activity determined at a PBMC:K562 of 50:1 and 25:1 was not significantly affected by treatment, there was a nonsignificant trend toward a decrease after 12 wk of FO supplementation (mean decrease of 25% at both ratios) (Figure 2).
DISCUSSION
Daily supplementation of the diet of healthy elderly subjects with 700 mg GLA, AA, or DHA or 1 g EPA plus DHA significantly altered plasma phospholipid fatty acid composition, with changes apparent after 4 wk of supplementation. Despite the marked changes in fatty acid composition caused by the GLA, AA, and DHA treatments, these treatments did not significantly alter the NK cell activity of PBMCs. In contrast, supplementation with FO for 12 wk significantly decreased NK cell activity measured at a PBMC-to-K562 target cell ratio of 100:1 and tended to decrease NK cell activity measured at the ratios of 50:1 and 25:1.
The habitual intake of ALA by the subjects in the current study was 0.7–1 g/d, which is consistent with other reports for the UK adult population (20, 21). The amount of ALA included in the supplement (2 g/d) increased total ALA intake by 2- to 2.8-fold to 3 g/d. ALA was absent from plasma phospholipids in most subjects at baseline and did not increase significantly in the subjects in the ALA group. The products of ALA elongation and desaturation (EPA and DHA) were also not significantly elevated in plasma phospholipids in the ALA group. These observations are consistent with those of a recent study in which an increase in ALA intake from 0.9 to 3.7 g/d for 28 d did not significantly alter the proportions of ALA, EPA, or DHA in plasma phospholipids (22). Thus, it appears that when ALA is included in the diet at moderate amounts it is not incorporated into plasma phospholipids in significant amounts. Furthermore, if it is elongated and desaturated, the products of this are not preferentially incorporated into plasma phospholipids.
Habitual intakes of GLA have not been reported and GLA is not mentioned in the UK Department of Health dietary reference values (21). It seems likely that GLA intake is normally negligible; habitual intake of GLA among subjects in the current study was 5 mg/d. Thus, the GLA supplement used in the current study (770 mg/d) represented a significant increase (>150-fold) in GLA intake. Nevertheless, GLA did not appear in plasma phospholipids. Amounts of the elongation product of GLA, DGLA, were increased during GLA supplementation, however, suggesting that some of the GLA is elongated before incorporation into phospholipids. These observations are consistent with those of Johnson et al (23), who found no increase in serum GLA concentrations after 3 wk of supplementation with 1.5 g GLA/d but an increase in serum DGLA. Wu et al (24) recently reported a small increase in the proportion of GLA in total plasma lipids (from 0.5% to 0.8% by wt of total fatty acids) in healthy subjects given 700 mg GLA/d for 8 wk. The reason for the difference between the results of Wu et al (24) and those of the current study may relate to the nature of the lipid fraction studied: Wu et al reported the fatty acid composition of total plasma lipids, whereas we report the composition of the plasma phospholipid fraction. It is possible that GLA is found in higher proportions in the triacylglycerol and cholesteryl ester fractions than in the phospholipid fraction, in which case it would be more readily detected in total lipids than in purified phospholipids. Observations of Johnson et al (23) support this explanation: in that study, the amount of GLA in neutral lipids (tri- and diacylglycerols and cholesteryl esters) was higher than the amount in the phospholipid fraction.
The habitual AA intake of subjects in the current study was in the range of 100–200 mg/d, which is consistent with earlier reports for other Western populations (25–27). The amount of AA provided in the supplement in the current study (680 mg/d) increased AA intake almost 5-fold. This resulted in significant enrichment of plasma phospholipids with AA, which was increased from 9% to 17% by wt of total fatty acids. This change is consistent with a recent study that reported that providing healthy young men with 1.5 g AA/d for 50 d increased the proportion of AA in plasma phospholipids from 10% to 18% by wt of total fatty acids (28).
Habitual intake of EPA plus DHA by subjects in the current study was 250–300 mg/d, which is in accordance with other reports for UK adults (20, 21). The amount of EPA plus DHA provided in the FO supplement (1 g/d) increased total EPA plus DHA intake by 4-fold. Intake of both EPA and DHA was increased in the FO group (by 8- and 1.7-fold, respectively) and there was a 2-fold increase in the proportion of EPA in plasma phospholipids and a nonsignificant trend toward increased DHA in this group. In the subjects in the DHA group, the intake of DHA increased 5.5-fold, whereas that of EPA did not change. Supplementation with this amount of DHA increased the proportion of DHA in plasma phospholipids without affecting that of EPA. Although it is often thought that FO supplementation decreases the proportion of AA in phospholipids, this was not observed in this study. This agrees with the observations of Allard et al (29), who found that supplementation of the diet with 5.3 g EPA plus DHA/d (ie, 5 times more EPA plus DHA than used in the current study) for 6 wk did not significantly decrease the proportion of AA in plasma phospholipids.
There was no significant effect of ALA, GLA, AA, or DHA on human NK cell activity. The lack of effect of AA is consistent with the observations of Kelley et al (30), who reported no significant effect on NK cell activity after 55 d of supplementation of the diet of young men with 1.5 g AA/d. Also, 6 g DHA/d as part of a low-fat diet for 55 d did not alter human NK cell activity, although NK cell activity decreased (by 20%) after 83 d (31). In the current study, we observed no significant effect on NK cell activity of 720 mg DHA/d for 12 wk. Another study found no significant effect of GLA, AA, or DHA in amounts of 4.4% by wt of total dietary fatty acids on rat spleen NK cell activity after 6 wk (32). The results of that animal study (32), the earlier human studies involving AA and DHA (30, 31), and the current study indicate that significantly increasing the amounts of ALA, GLA, AA, and DHA in the human diet will not compromise NK cell activity. Nevertheless, higher amounts of these fatty acids in the diet might impair NK cell activity, as shown by Kelley et al (31) for DHA.
The current study is the first to report decreased NK cell activity after FO supplementation of the human diet, although Yamashita et al (19) reported a significant decrease in NK cell activity 24 h after infusion of an EPA-rich triacylglycerol into human volunteers. The reduction we observed in NK cell activity after FO supplementation (providing 720 mg EPA plus 280 mg DHA/d) but not after supplementation with 720 mg DHA/d strongly suggests that EPA but not DHA was responsible for the effect. Alternatively, the effect of FO on NK cell activity may have been due to a component found in FO other than the long chain n-3 PUFAs. However, in support of an effect of EPA, rather than of DHA or a contaminant, are data showing that a diet containing EPA as 4.4% by wt of total fatty acids inhibited rat spleen NK cell activity, whereas a diet containing the same amount of DHA did not (32).
One concern with recommendations to increase intakes of certain long-chain PUFAs (either in the general population or in specific groups such as pregnant women and newborns) and with the widespread availability of products rich in some of these PUFAs is the potential for adverse immunologic effects resulting from excessive intakes. However, the current study indicates that significantly increased intakes of ALA, GLA, AA, and DHA may not adversely affect immune function. However, this study was of subjects aged >55 y and extrapolation of our results to other groups such as pregnant women and newborn infants may not be appropriate.
Although the inhibition of NK cell activity by EPA (in FO) was fully reversible, this effect suggests that increased EPA intake could have some adverse immunologic effects. NK cells play an important role in host defense against virus infection (33) and in immunosurveillance against tumor cells (34). Thus, it might be inappropriate for groups at risk of viral infection and some cancers to increase their intake of EPA. Other studies showed that increased intakes of EPA can decrease T lymphocyte functions such as proliferation and interleukin 2 production in certain populations (35, 36). It will be important to conduct studies to identify whether increased intakes of FO alter rates of infection in humans. The results of some animal studies indicated that a high intake of FO results in decreased resistance to infectious agents (37–41), although some other animal studies reported that FO does not alter (42), or even increases (43), resistance to some pathogens.
There may be some clinical situations in which an EPA-induced decrease in NK cell activity might be beneficial. Rejection of transplanted bone marrow and organs is in part mediated by NK cells (44) and so a transient decrease in NK cell activity might be useful before and soon after transplantation, although this might provide an opportunity for viral infection and tumor growth. FO prolongs the survival of cardiac transplants in rats (45–47) and renal transplants survive or function better in patients who receive FO (48–51). It is tempting to suggest that a part of this improvement is due to EPA-induced inhibition of NK cell activity, but none of these studies measured NK cell activity in the patients.
The plasma TBARS concentrations observed in this study compare well with the value of 4.0 ± 0.4 µmol/L reported by Meydani et al (52) for 10 women aged 51–71 y. In the current study, none of the treatments, including FO, significantly changed plasma TBARS concentrations. Therefore, it appears that oxidative stress resulting in lipid peroxidation was not induced by any of the treatments in the current study and that lipid peroxidation did not play a role in the alteration in NK cell activity observed after treatment with FO. However, some studies reported that the suppressive effects of FO on T lymphocyte functions can be prevented by adding suitable amounts of -tocopherol to the diet (53, 54). Therefore, a role of oxidant stress in eliciting the observed decrease in NK cell activity cannot be totally excluded and warrants further investigation.
ACKNOWLEDGMENTS
We thank Paul Quinlan (Unilever Research Colworth Laboratory), Gert Meijer (Unilever Vlaardingen), and Jacques Bindels and Monique Al [Nutricia (now Numico) Research] for their contributions to the study and Ria Banerjee for technical assistance.
REFERENCES