We have recently demonstrated that a novel n-3 long chain polyunsaturated fatty acid (PUFA) (β-oxa 21:3n-3) was a more potent and more selective anti-inflammatory agent than n-3 PUFA. To gain further insights into this technology, we synthesized other novel PUFA consisting of β-oxa, β-thia, and γ-thia compounds. All three types displayed anti-inflammatory activity. Each of the unsaturated β-oxa fatty acids showed similar inhibition of PHA-PMA-induced T cell proliferation with a parallel inhibition of TNF-β production. However, β-oxa 25:6n-3 and β-oxa 21:4n-3 displayed lower inhibitory action on IFN-γ production. Surprisingly, β-oxa 23:4n-6 and β-oxa 21:3n-6 had marginal effect on IL-2 production. Thus, structural variation can generate selectivity for different immunological parameters. The β-thia compounds 23:4n-6, 21:3n-6, and 21:3n-3 were highly effective in inhibiting all immunological responses. Of the two γ-thia PUFA tested, γ-thia 24:4n-6 was a strong inhibitor of all responses apart from IL-2, but γ-thia 22:3n-6 had very little inhibitory effect. Two of the most active compounds, β-thia 23:4n-6 and β-thia 21:3n-6, were studied in more detail and shown to have an IC50 of 1–2 μM under optimal conditions. Thus, these PUFA retain the immunosuppressive properties of the n-3 PUFAs, 20:5n-3 and 22:6n-3, but not the neutrophil-stimulating properties. Their action on T lymphocytes is independent of cyclooxygenase or lipoxygenase activity, and they act at a postreceptor-binding level by inhibiting the activation of protein kinase C and ERK1/ERK2 kinases.
The ability of n-3 polyunsaturated fatty acids (PUFA)3 to stimulate oxygen radical production from neutrophils ( 1, 2) and macrophages ( 3) even to a greater extent than arachidonic acid (20:4n-6) may in part explain some of the setbacks experienced with the application of n-3 PUFA as anti-inflammatory agents. We have proposed that attempts to treat inflammatory diseases through the use of n-3 PUFA are limited by the proinflammatory property of stimulating oxygen radical production ( 4).
Structural modifications of PUFA have been found to alter their biological activities so that PUFA with more selective biological properties can be generated. Thus, the addition of a hydroperoxy group gave rise to PUFA that failed to stimulate oxygen radical production ( 4). Yet, these molecules retained their ability to inhibit lymphocyte proliferation and cytokine production. Unfortunately, hydroperoxy fatty acids are highly unstable and are thus limited as pharmaceuticals. We therefore modified the PUFA structure to try to achieve the more selective properties of hydroperoxy PUFA as well as trying to improve the stability of the molecules. The published findings on fatty acid structure vs biological properties ( 4) were used as a basis for the synthesis of new types of PUFA and other long chain fatty acids. Fatty acids with either an oxygen atom in the β position (β-oxa) or sulfur atom in the β (β-thia) or γ (γ-thia) position were synthesized ( 5, 6).
Recently, we have demonstrated that one of these, β-oxa 21:3n-3, had similar biological properties as the 15-hydroperoxyeicosatetraenoic acid (15-HPETE), which markedly depressed T cell function and the inflammatory response, but was a poor stimulator of the respiratory burst in neutrophils ( 7). The availability of a range of novel PUFA with different structural modifications has enabled us to better define the relationship between specific structural elements of these novel fats and their biological properties.
Materials and Methods
β-oxa, β-thia, and γ-thia PUFA and their methyl esters were synthesized, as described previously ( 5, 6, 7, 8). All fatty acid stocks (20 mM in chloroform or ethanol) were stored at −20°C. Fatty acid purity was determined using 1H and 13C nuclear magnetic resonance spectroscopy, mass spectrometry, infrared spectroscopy, and microanalysis, as described previously ( 5), as well as HPLC and TLC. The maintenance of purity and absence of auto-oxidation products were determined at regular intervals using mass spectrometry and TLC.
Isolation of human leukocytes
Human leukocytes were isolated from the peripheral blood of healthy volunteers, as described previously ( 9). The neutrophil preparation was >99% viable, as judged by their ability to exclude trypan blue. T lymphocytes were purified, as described previously ( 10). The T lymphocyte preparation consisted of >98% CD3+ cells, and viability was >99%.
Presentation of fatty acids to cells
On the day of use, fatty acids were prepared in the presence of dl-α-dipalmitoyl phosphatidylcholine (DPC; Sigma-Aldrich, St. Louis, MO) as the carrier using a weight ratio of 4:1 of DPC to fatty acid, as described ( 11). Control cells received an equivalent amount of DPC. Fifty microliters containing 2 × 105 T lymphocytes were incubated with 50 μl of fatty acid (0.5–30 μM) in 96-well U-bottom plates (Linbro; Flow Laboratories) for the indicated periods of time. Cells were then stimulated with a variety of agonists for 48 or 72 h.
Measurement of the neutrophil respiratory burst
Purified T lymphocytes were stimulated with either 2 μg/ml PHA (Murex Diagnostics) and 10 ng/ml PMA (Sigma-Aldrich) for 48 h, 1 ng/ml PMA and 0.1 μM A23187 (Sigma-Aldrich) for 72 h, or anti-CD3 mAb (1/500 dilution; CLB) and anti-CD28 mAb (25 ng/ml; Beckman-Coulter) for 48 h, as described previously ( 14). Six hours before harvesting, 1 μCi of [methyl-3H]thymidine (Amersham Life Sciences) was added to the cultures. The supernatant was removed for cytokine estimation, the cells were harvested, and the incorporated radioactivity was measured using a Wallac liquid scintillation beta counter (Wallac 1409).
TNF-β, IFN-γ, and IL-2 levels were determined by ELISA, as described previously ( 14). Briefly, immobilized goat anti-mouse IgG (Cappel) was used to capture an anti-TNF-β, anti-IFN-γ, or anti-IL-2 mAb (Boehringer Mannheim). After addition of the supernatants, the wells were incubated with polyclonal rabbit anti-TNF-β, anti-IFN-γ (Boehringer Mannheim), or anti-IL-2 (Pierce) Ab. Detection was achieved using a HRP-conjugated goat anti-rabbit IgG Ab (BioSource International), using hydrogen peroxide as the substrate and 2,2′-azino-di(3-ethylbenzthiazoline sulfate) (Boehringer Mannheim) as the chromogen.
Determination of protein kinase C (PKC) translocation
PKC translocation was determined, as described previously ( 15). Briefly, T lymphocytes (10 ml of 1 × 106/ml) were pretreated for 30 min with 20 μM β-thia 23:0, β-oxa 21:3n-3, β-thia 23:4n-6, or β-thia 21:3n-6, or an equivalent amount of vehicle and stimulated with PHA (2 μg/ml) and PMA (10 ng/ml) for 5 min. The cells were harvested and sonicated (3 × 10 s) (Ystrom systems, setting 3), and particulate fractions were extracted by sonication with 2% Triton X-100. PKC activity was assayed, as described ( 15), and the activity was expressed as Ca2+/phosphatidylserine-dependent histone phosphorylation/minute.
ERK activity assay
ERK activity was assayed, as described previously ( 16). Briefly, T lymphocytes were pretreated with β-thia 23:0, β-oxa 21:3n-3, β-thia 23:4n-6, or β-thia 21:3n-6 fatty acid for 30 min and then stimulated with PHA and PMA for an additional 30 min. The cells were sonicated and centrifuged, and the supernatant was adsorbed onto phenyl-Sepharose CL4B (Pharmacia). ERK was batch eluted, and ERK activity was assayed by measuring the amount of 32P incorporated into myelin basic protein ( 16). The kinase activity in fractions prepared in this manner has previously been demonstrated to be indistinguishable from that obtained using immunoprecipitated ERK ( 16), and was almost undetectable in samples prepared from cells that had been pretreated with the MEK inhibitor, PD98059 ( 16).
Assessment of β-oxa-21:3n-3 incorporation into cellular lipids
Incorporation of the β-oxa 21:3n-3 into membrane phospholipids was performed, as described previously ( 17). Briefly, T lymphocytes (4 × 107) were incubated with or without 20 μM β-oxa-21:3n-3 in 4 ml of HBSS at 37°C for 30 min. The cells were centrifuged at 600 × g for 5 min at room temperature, and the medium was discarded. The cell pellet was resuspended in 2 ml of water, and lipids were extracted by the addition of 7.5 ml of chloroform/methanol/acetic acid (1/2/0.02, v/v/v) ( 18). The mixture was left at 4°C overnight and then partitioned by the addition of 2.5 ml of chloroform and 2.5 ml of water ( 19).
The amount of β-oxa-21:3n-3 associated with various lipid classes was determined, as follows. Concentrated lipid extracts were applied to silica gel 60 TLC plates and developed in the first dimension in chloroform/methanol/ammonia (65/25/5, v/v/v) and then in the second dimension in chloroform/acetone/methanol/acetic acid/water (40/30/10/10/5, v/v/v/v/v) to separate phospholipids. Concentrated lipid extracts were applied to silica gel 60 TLC plates and developed in hexane/diethyl ether/acetic acid (80/20/1, v/v/v) to resolve neutral lipids. Identification of the lipids was based on a comparison of their TLC mobility with that of authentic standards. The lipid zones were located under UV light after spraying the plates with dichlorofluorescein and were eluted from the silica gel. The oxa fatty acid is UV active, so on a TLC plate sprayed with the dichlorofluorescein and viewed under UV light, it quenches the fluorescence of the dichlorofluorescein to show up as a dark spot on a green background. The samples were transesterified, and the liberated fatty acid methyl esters were quantitated by GLC, as described above ( 17). The positional distribution of β-oxa-21:3n-3 in purified phosphatidylcholine, phosphatidylethanolamine, and phosphatidylinositol was assessed by degrading the compounds with phospholipase A2 and measuring the amount of the β-oxa-PUFA associated with the reaction products (unesterified fatty acids and lysophospholipids) after isolation by TLC ( 20).
Statistical analysis of data
Experiments were conducted in triplicate using cells from at least three different donors. Statistical significance was evaluated using a two-tailed unpaired Student’s t test or Bonferroni multiple comparisons test. A value of p < 0.05 was considered significant.
Effect of β-oxa, β-thia, and γ-thia fatty acids on T lymphocyte function
The structures of the β-oxa, β-thia, and γ-thia fatty acids are shown in Fig. 1, A–C, respectively. The β-oxa and β-thia fatty acids have an oxygen or sulfur atom in the β position, while the γ-thia PUFA have a sulfur atom in the γ position. The β-substituted PUFA cannot undergo β-oxidation. These fatty acids differ in their chain lengths, number of double bonds, and position of the double bonds. These PUFA were examined for their ability to inhibit human T lymphocyte proliferation and the production of TNF-β, IFN-γ, and IL-2.
As previously reported by us ( 7), β-oxa 21:3n-3 was highly inhibitory for mitogen-induced T lymphocyte proliferation (Fig. 2,A). We now demonstrate that the n-6 form of this fatty acid was similarly effective in inhibiting this response. However, the addition of another double bond, 21:4n-3, significantly increased the amount of inhibition (Fig. 2,A). The others, β-oxa 25:6n-3 and 23:4n-6, were also highly active in depressing T lymphocyte proliferation (Fig. 2 A).
When comparing β-thia PUFA with β-oxa PUFA, the former were more active. Thus, all three, β-thia 21:3n-3, β-thia 21:3n-6, and β-thia 23:4n-6, completely prevented the T lymphocyte responses to PHA-PMA stimulation. When the two γ-thia compounds were examined, γ-thia 24:4n-6 almost completely prevented T cell lymphoproliferative response (Fig. 2 A). In contrast, γ-thia 22:3n-6 caused no significant inhibition of this response. Thus, this illustrates that slight modification in fatty acid structure can have a profound influence on their inhibitory properties on T lymphocyte proliferation.
The ability of the β-oxa, β-thia, and γ-thia PUFA to inhibit TNF-β production paralleled their ability to inhibit lymphocyte proliferation (Fig. 2,B). All β-oxa and β-thia fatty acids apart from β-thia 23:0 were very active in inhibiting mitogen-induced TNF-β production. The results showed that the β-thia PUFA were more active than their corresponding β-oxa analogues (Fig. 2,A) in inhibiting T lymphocyte responses. Some differences between the fatty acids were noted in relation to inhibition of IFN-γ production (Fig. 2,C). The γ-thia PUFA had minimal or no inhibitory activity toward IFN-γ production. Interestingly, the β-oxa fatty acids that strongly inhibited lymphocyte proliferation and TNF-β production (β-oxa 21:4n-3 and β-oxa 25:6n-6) showed much less inhibition of mitogen-induced IFN-γ production and vice versa for β-oxa 21:3n-3 and β-oxa 21:3n-6 (Fig. 2 C). β-oxa 23:4n-6 markedly inhibited both T cell proliferation and IFN-γ production.
IL-2 production was also inhibited (Fig. 2 D). However, the β-oxa fatty acid, 23:4n-6, did not inhibit IL-2 production. These results thus show that modification of fatty acid structure can generate molecules with selective activities, with some fatty acids inhibiting cytokines differentially.
The finding that inhibition of cytokine production did not necessarily parallel inhibition of cell proliferation suggests that the reduced cytokine production was not just a result in difference in cell numbers in the presence and absence of the fatty acid. In fact, our studies show that under the experimental conditions we mostly see lymphoblasts and only a 27% increase in cell number due to the stimulation.
Both β-thia 23:4n-6 and β-thia 21:3n-6 were significantly more active at inhibiting T lymphocyte proliferation than previously reported for β-oxa 21:3n-3 at 20 μM ( 7). β-thia 23:4n-6 inhibited T lymphocyte proliferation in a concentration-dependent manner (Fig. 3,A) with an IC50 of 5.4 μM. β-thia 21:3n-6 displayed a similar inhibition profile to β-thia 23:4n-6 (Fig. 3 A), and had an IC50 of 6.8 μM.
We next examined the ability of β-thia 21:3n-6 and β-thia 23:4n-6 to inhibit TNF-β, IFN-γ, and IL-2 production. β-thia 23:4n-6 inhibited TNF-β with an IC50 of 6.0 μM, while β-thia 21:3n-6 had an IC50 of 8.8 μM (Fig. 3,B). IFN-γ production was also inhibited in a concentration-dependent manner by both β-thia 23:4n-6 and β-thia 21:3n-6 (Fig. 3 C). β-thia 23:4n-6 had an IC50 of 8.8 μM, while β-thia 21:3n-3 had an IC50 of 11.0 μM. IL-2 production was similarly inhibited by both fatty acids with an IC50 of 10.6 μM for β-oxa 23:4n-6 and β-oxa 21:3n-6, respectively.
Effect of varying pretreatment time
When long chain fatty acids are added to cells, they become esterified in different lipid pools, including membrane phospholipids ( 17). The incorporation and composition of phospholipids may vary with time. Thus, we examined the effect of varying pretreatment time on the effects of β-thia 21:3n-6 and β-thia 23:4n-6 on T cell function. Both were found to inhibit lymphocyte proliferation to a greater extent when a pretreatment time of 24 h was used (Fig. 4,A). β-thia 23:4n-6 significantly inhibited lymphocyte proliferation at 2.5 μM with maximal inhibition at 7.5 μM. The IC50 for β-thia 23:4n-6 was 1.5 μM with the 24-h pretreatment, a 3.6-fold increase in potency compared with the 30-min pretreatment time. The potency of β-thia 21:3n-6 was also increased (Fig. 4 A). Maximal inhibition was observed at 7.5 μM, and was found to have an IC50 of 4 μM, a 1.7-fold increase in potency compared with the shorter pretreatment period.
The ability to inhibit cytokine production was also increased following extended pretreatment with both β-thia 23:4n-6 and β-thia 21:3n-6. Maximal inhibition of TNF-β production by β-thia 23:4n-6 occurred at 7.5 μM, with an IC50 of 0.8 μM for inhibition of TNF-β production, a 7.5-fold increase in activity compared with the shorter pretreatment period (Fig. 4,B). β-thia 23:4n-6 was significantly more active than β-thia 21:3n-6 at 0.5 μM, even though the maximum inhibitory action was attained at 7.5 μM. β-thia 21:3n-6 had an IC50 of 1.7 μM for inhibition of TNF-β production, a 5.1-fold increase in potency compared with the shorter pretreatment time (Fig. 4 B).
β-thia 23:4n-6 inhibited IFN-γ production in a linear fashion from 0.5 to 2.5 μM, reaching a maximum at 7.5 μM (Fig. 4,C) with an IC50 of 0.9 μM, a 9.7-fold increase in potency compared with the shorter pretreatment time. β-thia 21:3n-6 inhibited IFN-γ production with a similar profile to β-thia 23:4n-6 (Fig. 4 C). Significant inhibition was evident between 0.5 and 2.5 μM, with maximum inhibition at 7.5 μM and having an IC50 of 1.3 μM, an 8.4-fold increase in potency compared with the shorter pretreatment time.
IL-2 production was inhibited over a concentration range of 0.5–2.5 μM by β-thia 23:4n-6, reaching a maximum by 2.5 μM (Fig. 4,D) and having an IC50 of 1.25 μM, an 8.5-fold increase in activity compared with the shorter pretreatment time. β-thia 21:3n-6 inhibited IL-2 production in a linear fashion from 0.5 to 7.5 μM (Fig. 4 D), reaching a maximum at ∼7.5 μM. β-thia 23:4n-6 was significantly more active than β-thia 21:3n-6 at 2.5 μM. β-thia 21:3n-6 had an IC50 of 2.1 μM for IL-2 production, a 5-fold increase in potency over the shorter pretreatment time.
We also investigated the effects of shorter pretreatment times with the β-thia PUFA on T cell responses. T lymphocytes were incubated with β-thia 23:4n-6 or β-thia 21:3n-6 (20 μM) for 0, 5, 10, 20, or 30 min before the addition of PHA-PMA. The degree of inhibition of T lymphocyte proliferation was time dependent over the 30-min period (Fig. 5,A). Both β-thia 23:4n-6 and β-thia 21:3n-6 caused significant levels of inhibition within 10 min. The degree of inhibition increased as the pretreatment time increased (Fig. 5,A). Inhibition of cytokine production was also found to be similarly time dependent. Both β-thia 23:4n-6 and β-thia 21:3n-3 significantly inhibited TNF-β (Fig. 5,B), IFN-γ (Fig. 5,C), and IL-2 (Fig. 5 D) production after just 5 min, with the degree of inhibition increasing with pretreatment time.
The kinetics of pretreatment showed that there was a steady increase in inhibition of lymphocyte proliferation between a 30-min to 8-h pretreatment and a dramatic increase in inhibition between 18- and 24-h preincubation (data not shown). Given the rapid inhibition of lymphocyte proliferation, we examined the cellular incorporation of a representative engineered fatty acid, β-oxa 21:3n-3, after 30-min pretreatment. After 30 min, the cells had incorporated 77% of the fatty acid. Most of the fatty acid was found in phosphatidylglycerol, with lesser amounts present in phosphatidylcholine, cholesterol ester, and sphingolipids (Table I).
|Lipid .||Recovered β-oxa 21:3n-3 (% of total) .|
|Free fatty acid||0.1|
|Lipid .||Recovered β-oxa 21:3n-3 (% of total) .|
|Free fatty acid||0.1|
T lymphocytes (4 × 107) were incubated with β-oxa-21:3n-3 (20 μM) or diluent (control) in 4 ml of HBSS at 37°C for 30 min. The amount of β-oxa-21:3n-3 associated with phospholipids and neutral lipids of the cells was determined, as described in Materials and Methods. No β-oxa-21:3n-3 was found in phosphatidic acid, phosphatidylethanolamine, cardiolipin, monoacylglycerol, diacylglycerol, or triacylglycerol. The results are of a representative experiment and are expressed as a percentage of the total recovered β-oxa-21:3n-3.
Importance of the terminal carboxyl group toward biological activity
To further elucidate the importance of fatty acid structure toward biological activity, the carboxyl group of the fatty acids was modified by generating methyl ester derivatives. The β-thia 23:4n-6 and β-thia 21:3n-6 methyl esters showed significantly reduced inhibitory activity toward lymphocyte proliferation (Fig. 6,A), although these still retained the ability to significantly inhibit lymphocyte proliferation (Fig. 6,A). The methyl ester derivatives exhibited reduced or no inhibitory effect on TNF-β (Fig. 6,B), IFN-γ (Fig. 6,C), and IL-2 (Fig. 6 D) production. Thus, the presence of a carboxyl group was critical for the inhibitory effects of both β-thia 23:4n-6 and β-thia 21:3n-6.
Effect of β-thia PUFA on anti-CD3/CD28-induced lymphoproliferation
To provide further evidence for the inhibitory effects of these novel fatty acids on immune responses, we examined their effects on responses when engaging the TCR and costimulatory signal receptor. mAbs directed against CD3 and CD28 mimic physiological activation and have been demonstrated to stimulate lymphocyte proliferation ( 21). Both β-thia 21:3n-6 and 23:4n-6 significantly inhibited lymphocyte proliferation induced through CD3/CD28 (percentage of inhibition, 98.5 ± 2.1 and 98.4 ± 2.6, respectively).
Effect of β-oxa and β-thia PUFA on CD3, CD4, and CD8 expression
Alterations in the expression of cell surface molecules involved in lymphocyte activation could result in reduced lymphocyte responses. To examine this possibility, T lymphocytes were treated with β-thia 23:4n-6 and β-thia 21:3n-6 (both 20 μM) alone for 30 min, and the expression of CD3, CD4, and CD8 was determined by FACScan analysis. β-thia 23:4n-6 did not significantly alter the expression of CD3 (96.4 ± 0.9, percentage of control cells ± SEM), CD4 (94.1 ± 1.0), or CD8 (90.3 ± 1.5). In contrast, β-thia 21:3n-6 caused a significant reduction in CD3 (78.4 ± 1.1), CD4 (77.1 ± 5.8), and CD8 (80.7 ± 1.4) expression (p < 0.05). Although β-thia 23:4n-6 reduced CD3, CD4, and CD8 expression, this is unlikely to explain the inhibitory activity, as β-thia 23:4n-6 was as active at inhibiting lymphocyte responses, yet failed to affect CD3, CD4, or CD8 expression. This suggests that processes postreceptor binding were likely to have been affected by the fatty acids.
Effect of β-oxa and β-thia PUFA on PMA-A23187-induced lymphoproliferation and cytokine production
To determine whether the inhibitory effects were mediated at a postreceptor level, we examined their effects on PMA- and A23187-mediated T lymphocyte responses. These agents bypass cell surface receptors and stimulate T cells by activating PKC and Ca2+/calmodulin-dependent effectors ( 22, 23). β-thia 23:4n-6 and β-thia 21:3n-6 significantly inhibited lymphocyte proliferation (stimulation index, 43.6 ± 11.9, n = 3), TNF-β, IFN-γ, and IL-2 production in response to PMA-A23187 (Fig. 7). Thus, modulation of cell surface receptors was not a mechanism involved in the inhibition of lymphocyte function by the β-thia PUFA.
Effect of β-thia PUFA on PHA-PMA-induced activation of PKC
The inhibition of PMA-A23187-induced T cell effector functions by β-thia 23:4n-6 and β-thia 21:3n-6 suggests that these fatty acids blocked PKC- and/or Ca2+-dependent signaling pathways. This action could be exerted at the level of PKC and/or increases in intracellular calcium mobilization or at events downstream of PKC and changes in intracellular Ca2+. PKC activation is associated with a rapid and transient translocation of PKC from the soluble to a particulate fraction. We examined whether β-thia 23:4n-6 and β-thia 21:3n-6 affected PKC activation. β-thia 23:0 reduced PHA-PMA-induced translocation of PKC weakly in T lymphocytes (Fig. 8,A). In contrast, β-thia 23:4n-6 and β-thia 21:3n-6 reduced particulate fraction-associated PKC by 87.1 ± 4.45 and 78.9 ± 1.10%, respectively, compared with the DPC control (Fig. 8 A). This finding suggests that the effects of the fatty acids may be exerted at the level of PKC.
Effect of β-thia PUFA on MAPKs
In human T lymphocytes, PKC is an upstream regulator of the MAPKs ERK and JNK ( 24, 25) and is also likely to regulate p38. Efficient activation of these MAPKs requires either PMA (ERK) or PMA + A23187 (JNK and p38) ( 26). As the β-thia PUFA were found to inhibit PHA-PMA- and PMA-A23187-induced lymphocyte function, it was important to determine whether activation of these MAPKs was also blocked by β-thia 23:4n-6 and β-thia 21:3n-3. Both β-thia PUFA inhibited total ERK activity by 40.5 ± 11.6 (n = 7) and 43.5 ± 5.8% (n = 10), respectively (Fig. 8 B). Interestingly, the reduction in ERK activity was not as great as for PKC. In contrast, pretreatment with β-thia 23:0, β-thia 23:4n-6, or β-thia 21:3n-6 did not cause a reduction in the activity of JNK or p38 compared with PMA-A23187 alone (data not shown).
Effect of indomethacin (IM), nordihydroguaiaretic acid (NDGA), and vitamin E (Vit E) on β-thia PUFA-mediated inhibition of lymphoproliferation
Metabolism of natural fatty acids, particularly 20:4n-6 via the cyclooxygenase (COX) and lipoxygenase (LOX) enzymes, is believed to be one mechanism by which natural fatty acids mediate their activities ( 27, 28). It was thus important to determine whether metabolism of the β-thia PUFA was required for their inhibitory activity in vitro. T lymphocytes were pretreated with IM (100 μM), a COX inhibitor; or NDGA (10 μM), a LOX inhibitor; or Vit E, an antioxidant, before being treated with β-thia 23:4n-6 or β-thia 21:3n-6 (both at 20 μM) and stimulated with PHA-PMA. The results show that neither IM, NDGA, nor Vit E significantly prevented β-thia 23:4n-6 or β-thia 21:3n-6 from inhibiting lymphocyte proliferation (Table II). Thus, the inhibitory effects of the β-thia PUFA on T lymphocyte function are independent of their metabolism via COX or LOX enzymes and lipid peroxidation.
|.||% Inhibition .||.|
|.||β-thia 21:3n-6 .||β-thia 23:4n-6 .|
|Diluent||94 ± 0.6||60 ± 4.7|
|IM||94.8 ± 0.7||46.4 ± 12.3|
|Diluent||92.6 ± 1||57 ± 6.2|
|NDGA||92.8 ± 1.1||48.4 ± 1.1|
|Diluent||85 ± 4.2||73 ± 2.8|
|Vit E||83 ± 1.9||70 ± 1.9|
|.||% Inhibition .||.|
|.||β-thia 21:3n-6 .||β-thia 23:4n-6 .|
|Diluent||94 ± 0.6||60 ± 4.7|
|IM||94.8 ± 0.7||46.4 ± 12.3|
|Diluent||92.6 ± 1||57 ± 6.2|
|NDGA||92.8 ± 1.1||48.4 ± 1.1|
|Diluent||85 ± 4.2||73 ± 2.8|
|Vit E||83 ± 1.9||70 ± 1.9|
T lymphocytes were treated with IM, NDGA, or Vit E for 30 min, followed by β-thia 21:3n-6 or β-thia 23:4n-6 for 30 min, and then stimulated with PHA/PMA for 48 h. Data are presented as the mean ± SEM of three experiments.
Ability of β-oxa 21:3n-3, β-thia 21:3n-6, and β-thia 23:4n-6 to stimulate superoxide production in human neutrophils
The natural PUFA 20:4n-6, 20:5n-3, and 22:6n-3 have all been shown to strongly stimulate the neutrophil respiratory burst ( 1, 2). Thus, the novel PUFA were examined for their ability to stimulate superoxide production at 0.5–20 μM. Consistent with previously published data ( 1, 2, 4), 22:6n-3 was a strong inducer of superoxide production in human neutrophils. The unsaturated β-oxa and β-thia fatty acids were significantly less active than 22:6n-3 in stimulating superoxide production (Fig. 9). Thus, the presence of an oxygen or sulfur atom can significantly influence the ability of these novel fatty acids to stimulate superoxide production.
The ability of the β-oxa and β-thia fatty acids to inhibit lymphocyte proliferation and cytokine production was, to some degree, dependent on their structures. Although the presence of double bonds was critical for activity, the number of double bonds did not correlate with biological activity. For example, β-thia 21:3n-6 was as effective as β-thia 23:4n-6 at inhibiting lymphocyte proliferation and cytokine production. No relationship between the series (n-3 vs n-6) of a fatty acid and its ability to inhibit T cell responses was found. The only exception to this was that β-oxa 21:3n-3 was more active at inhibiting IL-2 production than β-oxa 21:3n-6. Of the two γ-thia PUFA tested, γ-thia 24:4n-6 was a stronger inhibitor of lymphocyte proliferation, TNF-β, and IL-2 production. In comparison, γ-thia 22:3n-6 was not very active. The presence of an oxygen or sulfur atom was found to influence the ability of these PUFA to inhibit lymphocyte proliferation and cytokine production. For example, β-oxa 21:3n-3 and β-oxa 23:4n-6 were less active than their β-thia analogues at inhibiting lymphocyte responsiveness. The main focus of the activity of the oxa and thia PUFA has been their effects on the Th1 cytokines because many chronic inflammatory diseases such as rheumatoid arthritis have a prominent production of these. In addition, we have previously shown that β-oxa 21:3n-3 strongly inhibits the DTH response in mice ( 7). However, knowing the effects on Th2 cytokines would allow a better perspective to be placed on their usefulness as anti-inflammatory agents.
The anti-T lymphocyte effects of these novel PUFA were highly dependent on the pretreatment time. This was most likely to be related to their incorporation into various lipid pools. This is supported by the finding that conversion of β-thia 23:4n-6 and β-thia 21:3n-6 into their methyl esters significantly reduced their ability to inhibit lymphocyte responses. Thus, the methyl ester of β-thia 23:4n-6 and β-thia 21:3n-6 did not inhibit IFN-γ and IL-2 production, while their ability to inhibit lymphocyte proliferation and TNF-β production was evident, but greatly reduced. The possibility that this remnant activity could have been due to the hydrolysis of esters back to the fatty acids under the assay conditions was considered unlikely because the level of IFN-γ and IL-2 production was not affected. These results are consistent with previously published data demonstrating that the carboxyl group is critical for the biological activity of fatty acids ( 1, 4). Conversion of the β-thia fatty acids into their methyl ester derivatives is likely to affect fatty acid activity in at least two ways. First, the binding of the fatty acids to lipid carrier proteins requires the presence of a carboxyl group. Elimination of this group will thus reduce the amount of fatty acid being able to enter the cell and the sites of action ( 29). A more likely explanation is that methyl esters are not readily converted into their acyl-CoA derivative due to the lack of the terminal carboxyl group. This means that the fatty acid is poorly incorporated into membrane phospholipids, and hence, unlikely to interfere with downstream signaling events.
The metabolism of fatty acids via the LOX and COX pathways can lead to the generation of highly active metabolites. These molecules may be responsible for the inhibitory effects of the engineered PUFA. However, we found that the inhibitory action of the fatty acids was not affected by pretreating cells with the LOX and COX inhibitors NDGA and IM, respectively. These data are consistent with previously published findings that the ability of 22:6n-3 to inhibit lymphocyte proliferation was not blocked by IM or NDGA ( 30, 31). At this concentration, NDGA is probably also a nonspecific antioxidant. This is consistent with the fact that the antioxidant, Vit E, did not block the inhibitory effects of β-thia 21:3n-6 or β-thia 23:4n-6 toward lymphocyte proliferation. This suggests that lipid peroxidation did not play a role in fatty acid-mediated inhibition of lymphocyte proliferation because the β-thia fatty acids are highly unsaturated and are thus susceptible to peroxidation. Indeed, lipid hydroperoxides in the presence of Fe2+ ions generate the OH ·, lipid alkoxyl, and peroxyl radicals, which disrupt nucleic acids, proteins, and membrane lipids ( 32). These free radicals have been shown to depress DNA, RNA, and protein synthesis in murine thymic lymphocytes ( 33). The ability of 22:6n-3 to inhibit lymphocyte proliferation has also been reported not to be abrogated by Vit E or catalase ( 34, 35). These results provide additional evidence that lipid peroxidation did not play a role in the inhibition of lymphocyte responsiveness by the PUFA.
The ability of β-thia 23:4n-6 and β-thia 21:3n-6 to inhibit T lymphocyte proliferation was not unique to the agonists, PHA-PMA. When T lymphocytes were activated via the TCR and costimulatory molecule, using anti-CD3 and anti-CD28 Abs, inhibition of lymphocyte proliferation was also observed. This suggests that the fatty acids may be mediating their effects at a target that is used by all agonists.
It is unlikely that the inhibitory effects of the engineered fatty acids are mediated through alterations in cell surface receptor expression on T lymphocytes. β-thia 23:4n-6 induced only a minor and nonsignificant decrease in the surface expression of CD3, CD4, and CD8. Although β-thia 21:3n-6 reduced CD3, CD4, and CD8 expression by ∼20%, this reduction is unlikely to explain its inhibitory activity, because β-thia 23:4n-6 was as active as β-thia 21:3n-6 in inhibiting T lymphocyte responses. The above arguments are further supported by the finding that the β-thia fatty acids inhibited lymphocyte proliferation in response to anti-CD3 and anti-CD28 Ab and to PMA-A23187, the latter acting directly on PKC and the calcium signaling pathways.
PKC is essential for lymphocyte proliferation and IL-2 production in response to antigenic stimulation and anti-CD3-CD28 stimulation ( 36, 37). Thus, inhibition of PKC translocation and activity has been demonstrated to dramatically affect lymphocyte responses. Pretreatment of T lymphocytes with the β-thia PUFA, followed by stimulation with PHA-PMA, resulted in a significant decrease in the amount of PKC translocation from the cytosolic to the particulate fraction. Our results are consistent with previous work in our laboratory showing that the 20:4n-6 15-HPETE decreased the amount of membrane-bound PKC in HUVEC ( 38). The reduction in PKC translocation was not due to a nonspecific lipid effect because β-thia 23:0 did not suppress PKC translocation in response to PHA-PMA. The inability of β-thia 23:0 to affect PKC translocation is consistent with its lack of inhibition of lymphocyte responsiveness. The degree of inhibition of PKC translocation by β-thia 23:4n-6 and β-thia 21:3n-6 correlated with the level of the inhibitory activity of the fatty acid. It is possible that the reduced amount of PKC associated with the particulate fraction was due to an increase in the degradation rate of particulate fraction-associated PKC. However, Huang et al. ( 38) demonstrated that fatty acids do not accelerate the degradation of particulate fraction-associated PKC, but prevented its association with the particulate fraction.
PKC, consisting of a family of ∼11 isozymes, has been demonstrated to regulate the activity of several members of the MAPKs, including the ERK cascade ( 16). Consequently, it would be expected that a reduction in PKC activity should lead to some reduction in ERK activity. Activation of the ERK cascade is critical for T lymphocyte proliferation and cytokine production ( 16, 39). Constitutively active mutants of MEK1, the upstream activator of ERK1 and ERK2, in conjunction with A23187 leads to the production of IL-3, GM-CSF, IFN-γ, and IL-4 ( 39). Inhibition of MEK1, with the specific inhibitor, PD98059, has been shown to inhibit lymphocyte proliferation and expression of IL-2 mRNA, IL-2, IFN-γ ( 40, 41), and TNF-β (16) in response to antigenic or CD3-CD28 stimulation. These data suggest that MEK1, ERK1, and ERK2 are important for the production of both Th1- and Th2-type cytokines ( 39).
Both β-thia PUFA reduced ERK activity; however, the reduction in ERK activity was not as great as for the reduction in PKC translocation to the particulate fraction. This may be explained in two ways. First, the ERK module can be activated in a PKC-dependent as well as a PKC-independent manner. Thus, the possibility exists that activation of the ERK module via the PKC-independent pathway is unaffected by the fatty acids. Alternatively, PKC isozyme whose translocation is not affected by β-oxa 21:3n-3 may also contribute to the activation of the ERK module.
In contrast to the results obtained for PKC and ERK, the β-thia fatty acids did not suppress the PMA-A23187-induced activity of JNK or p38. These results exclude suppression of JNK or p38 activity as a mechanism of action of the β-thia fatty acids. Thus, of the three MAPK cascades that play important roles in T cell responses, only the ERK cascades were affected by the novel PUFA, thereby demonstrating selectivity of action.
Activated T lymphocytes with a Th1 cytokine profile have been implicated in the pathogenesis of several autoimmune diseases ( 42, 43, 44, 45, 46). Dietary supplementation with n-3 fatty acids has been used as a means to treat rheumatoid arthritis ( 42, 43), multiple sclerosis (44), insulin-dependent diabetes ( 45), psoriasis, and atopic dermatitis ( 46). However, such treatments have generally provided only modest improvements in disease severity ( 42, 43, 44, 45, 46). The less than expected amelioration following n-3 PUFA treatment may in part be explained by the ability of n-3 fatty acids to activate the neutrophil respiratory burst, which is likely to cause tissue damage. Based on our knowledge of the structure-function relationship of fatty acids toward biological activity ( 4, 47), novel chemically engineered fatty acids were made in an attempt to synthesize molecules that displayed differential activities, that is, fatty acids that did not activate neutrophils, but still depressed T lymphocyte function. This work illustrates that PUFA with desirable biological activities can be generated by modifying specific structural elements. By introducing an oxygen atom into the β-position, preferential anti-inflammatory properties were achieved, giving rise to molecules with very poor ability to stimulate neutrophil oxygen radical production compared with 20:4n-6, 20:5n-3, and 22:6n-3, but still possess the ability to inhibit T lymphocyte function. This skewing of the fatty acids toward anti-inflammatory activity was similar to that seen with metabolites of 20:4n-6, such as 15-HPETE, hydroxyeicosatetraenoic acid, and lipoxins ( 48, 49, 50, 51). These novel fatty acids may be useful as anti-inflammatory agents, demonstrating increased biological activity and selectivity compared with the natural long chain PUFA. This selectivity could be due in part to its ability to selectively target the PKC-ERK1/ERK2 module that is required for T lymphokine production.
We are grateful to Dr. Michael Pitt for synthesizing the fatty acids, and Drs. Harmeet Singh and Neil Trout for lipid analysis.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work received financial support from the National Health and Medical Research Council of Australia, The Heart Foundation of Australia, Channel 7, Women’s and Children’s Research Foundation, and the United Nations Development Program/World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases.
Abbreviations used in this paper: PUFA, polyunsaturated fatty acid; COX, cyclooxygenase; DPC, dl-α-dipalmitoyl phosphatidylcholine; 15-HPETE, 15-hydroperoxyeicosatetraenoic acid; IM, indomethacin; LOX, lipoxygenase; NDGA, nordihydroguaiaretic acid; PKC, protein kinase C; Vit E, vitamin E.