Mechanisms of Opioid-Mediated Inhibition of Human T Cell Receptor Signaling¹ Free

Opioids are the most potent analgesics and widely used for the treatment of severe pain such as cancer pain. However, a great number of studies have convincingly demonstrated that opioids, in particular morphine and derivatives, are immunosuppressive. This may cause unwanted side effects during pain treatment, especially in persons with an impaired immune system, e.g., patients suffering from cancer or AIDS. It is also reasonable to propose that the opioid-induced immunosuppression contributes to an increased prevalence of infections like pneumonia and HIV among opioid users and addicts (for reviews, see Refs. 1 and 2). Studies in mouse models have suggested that one of the mechanisms by which opioids cause immunosuppression is the inhibition of IL-2 transcription in activated T lymphocytes (3, 4, 5, 6). Thus, it was demonstrated that heroin and methadone inhibit the IL-2 production of activated murine splenocytes (3). Furthermore, it was shown that opioid-mediated inhibition of c-fos mRNA (4) and opioid-induced modulation of the cAMP response element-binding transcription factors CREB/CREM/ICER (5, 6) are involved in the inhibition of IL-2. However, detailed molecular mechanisms, which link opioids to TCR signaling and lead to its inhibition, have not been clearly defined. Furthermore, little is known about opioid-induced immunosuppression in human T cells.

It is known that the TCR-mediated activation of T cells induces a defined series of signaling events which culminate in the transcription of the gene for the T cell growth factor IL-2 (7, 8, 9, 10). In the nonactivated state, i.e., in resting T cells, the cascade is tonically repressed by constitutive phosphorylation of the negative regulatory site of Lck at Tyr⁵⁰⁵. This phosphorylation is mediated by the kinase C-terminal Src kinase (Csk),³ which is complexed in resting T cells with the transmembrane adapter protein Cbp/PAG (phosphoprotein associated with glycosphingolipid-enriched microdomain/Csk-binding protein). Upon engagement of the TCR, Cbp/PAG is dephosphorylated. This leads to a release of Csk from Cbp/PAG, resulting in a loss of the inhibitory effect of Csk on Lck. These events lead to activation of Lck and initiation of the TCR-induced signaling cascade (11, 12, 13). One of the earliest events in this cascade is the Lck-mediated phosphorylation and thereby activation of the protein tyrosine kinase Zap70. Activated Zap70 in turn phosphorylates the transmembrane adapter protein linker for activation of T cells (LAT) on at least four tyrosine residues, leading to the assembly of the calcium-initiation complex (14). These events lead to the induction of calcium flux and also to the activation of the MAPK cascades. In this way, the signal is transduced and finally results in the activation of the transcription factors AP-1, NF-κB, and NFAT. Together, they transactivate the IL-2 gene, which is almost entirely regulated at the level of transcription. After being released, IL-2 induces a variety of well-defined subsequent immune responses.

The modulation of the IL-2 production of activated T cells by opioids is pharmacologically of great importance, since it contributes to the immunosuppressive side effects of drugs like morphine. In addition, there may be a physiological regulation of activated T cells by endogenous opioid peptides like β-endorphin. However, the question of if and how β-endorphin regulates the TCR-mediated signaling cascade and whether this is involved in the regulation of physiological T cell responses are less well documented. Earlier studies using the opioid receptor antagonist naloxone have suggested that β-endorphin exerts an inhibitory effect on T cell functions like IL-2 production (15) and mitogen-induced splenocytes proliferation (16). In line with this observation, it was demonstrated more recently that mice lacking β-endorphin showed enhanced splenocyte proliferation and enhanced levels of IL-2 mRNA compared with wild-type mice (17).

Effects of opioids on T lymphocytes require the expression of specific opioid receptors (for reviews, see Refs. 18 and 19). Drugs of the morphine type, and thus most of the clinically used opioids, bind preferentially to μ opioid receptors. The expression of functional μ opioid receptors on human T lymphocytes is induced by TNF and IL-4 (20, 21, 22). In addition, expression of μ opioid receptors is also induced indirectly by drugs and stimuli, which increase the expression of TNF and IL-4, like cannabinoids (23). As demonstrated recently by our group, also anti-CD3/anti-CD28-mediated activation of human T cells induces the expression of functional μ opioid receptors (24). Studies in mice have shown that morphine is only effective on mitogen-activated T cells, which may be explained by an induction of μ opioid receptors after T cell activation (4, 5, 25). Thus, there is reason to believe that the expression of μ opioid receptors on human T lymphocytes allows the modulation of T cell responses by ligands for μ opioid receptors, such as morphine. The endogenous opioid β-endorphin activates both μ and δ opioid receptors. It was reported that low levels of δ opioid receptors are expressed in Jurkat T lymphocytes and that their expression is induced in response to IL-4 (26), and anti-CD3/anti-CD28 (our unpublished observations). In murine T cells, it was shown that anti-CD3 increases the expression of δ opioid receptors (27). These observations suggest that also δ opioid receptor agonists can modulate T cell functions. In this report, we identified mechanisms by which opioids suppress the TCR-triggered signaling cascade in activated human T lymphocytes.

PBMC were isolated from heparinized blood collected from healthy volunteers as previously described (28). Primary human T cells and the human T cell line Jurkat were cultivated in RPMI 1640 medium (Lonza Verviers) supplemented with 10% FCS and antibiotics (100 U/ml penicillin and 100 mg/ml streptomycin; Lonza Verviers). For our experiments, no opioid receptor-overexpressing cells were used, but μ and δ opioid receptors were induced in physiological amounts by IL-4 according to an established procedure, which has been shown to produce functional receptors (21, 22, 26). Briefly, the cells received IL-4 (used at 5 ng/ml; R&D Systems) and were incubated for 3 days. Then the medium was replaced by fresh medium without IL-4 and the cells were cultured for an additional 24 h. An aliquot of the cells was tested to assure the induction of μ and δ opioid receptor transcripts. For these cells, the term “opioid receptor-expressing” cells is used throughout this publication. For activation of T cells, 100 μl each of anti-CD3 and anti-CD28 was used per 10⁶ cells. The sources for anti-CD3 (OKT3) and anti-CD28 (248.23.2) were the hybridoma supernatants produced in our laboratories (28). As opioids, morphine (Synopharm), β-endorphin (Sigma-Aldrich), and D-penicillamine2-D-penicillamine5-enkephalin (Sigma-Aldrich) were used at 1 μM. Antagonists were CTAP (μ opioid receptor specific) and naltrindole (δ opioid receptor specific), both used at 250 nM and purchased from Tocris. The protein kinase A (PKA) inhibitor (R)-adenosine cyclic 3′,5′-hydrogenphosphorothioate (cAMPS-Rp; Tocris) was used at 100 μM. The PKA activator 8-Br-cAMP (Tocris) was used at 5 μM. As phosphodiesterase inhibitors, IBMX (500 nM; Sigma-Aldrich), rolipram (10 μM; Tocris), BRL50481 (10 μM; Tocris), and dipyridamole (100 μM; Tocris) were used. PMA (Sigma-Aldrich) was used at 100 nM.

Experiments were performed similarly to a described protocol (29). Briefly, opioid receptor-expressing Jurkat cells (10⁶ per sample) were incubated for 4 days on anti-CD3/anti-CD28-coated dishes with opioids and opioid antagonists. Then cells were restimulated with anti-CD3/anti-CD28 for 15 min to induce IL-2 mRNA and lysed after 4 h for quantitative IL-2 RT-PCR. The restimulation and final 4-h incubation of the cells was performed in the presence of opioids to exclude “withdrawal”-like effects.

Oligonucleotides containing binding sequences (core-binding sequences are underlined) for the transcription factors AP-1 (5′-AAACATATGATTCACCAGGCA-3′), NF-κB (5′-AAAGTTGAGGGGACTTTCCCAGGCCT-3′), and NFAT (5′-GCCCAAAGAGGAAAATTTGTTTCATA-3′) were cloned 5′ to the herpes simplex thymidine kinase promoter of the reporter gene plasmid pBLCAT2 (30). These constructs were transfected into opioid receptor-expressing Jurkat cells by electroporation (210 V, 950 μF) using a Gene Pulser II (Bio-Rad). The day after transfection, cells were transferred to anti-CD3/anti-CD28- or anti-mouse Ig-coated dishes, with or without morphine. After 72 h of transient expression, the reporter gene was assayed via a chloroamphenol acetyl transferase-ELISA purchased from Roche.

Isolation of RNA and quantitative real-time RT-PCR for β-actin, μ, and δ opioid receptor transcripts have been described earlier (20, 22, 26). IL-2-specific RT-PCR was performed with 5′-GAAGGCCACAGAACTGAAACATCT-3′ and 5′-CTGTTCAGAAATTCTACAATGGTTG-3′ primers, with a preincubation for 8 min at 95°C and 50 cycles with 5 s at 95°C, 5 s at 65°C, and 10 s at 72°C (specific T_m = 81, 75°C). Quantitative real-time RT-PCR was done in a total volume of 20 μl on a LightCycler instrument using a LightCycler-Fast Start DNA Master SYBR Green I kit (both from Roche) according to the manufacturer’s suggestions.

After washing in PBS, cells were incubated with murine anti-CD3 for 30 min at 4°C. Secondary goat anti-mouse-FITC Abs (Dianova) were incubated for 30 min at 4°C and the surface expressions of the TCR were measured using a FACSCalibur flow cytometer (BD Biosciences).

Western blots were performed as previously described (22, 23, 31). For activation, 2.5 × 10⁶ opioid receptor-expressing cells per sample were pelleted, resuspended in anti-CD3/anti-CD28, and incubated at 37°C for 2 min. The activation was stopped with 1 ml of ice-cold PBS, cells were pelleted again, and lysed. For protein detection, the following Abs were used: primary Abs were actin C-11 (Santa Cruz Biotechnology); phospho-p44/42 MAPK Tyr²⁰²/Tyr²⁰⁴_; phospho-LAT Tyr¹⁹¹, phospho-Zap70 Tyr³¹⁹, and phospho-Lck Tyr⁵⁰⁵ (all from Cell Signaling Technology/New England Biolabs). All Abs were from rabbits; the secondary Ab was anti-rabbit IgG (GE Healthcare). Anti-PAG (MEM255) hybridoma was made in the laboratory of Dr. V. Horejsi (Institute of Molecular Genetics, Prague, Czech Republic) and grown and supernatant was collected in our institute. The secondary Ab anti-mouse-HRP was obtained from Dianova.

Opioid receptor-expressing cells were incubated with morphine and β-endorphin for 0 h (controls) up to 24 h. Then cells were lysed with 50 mM HCl for 30 min on ice. The competitive cAMP-ELISA was performed according to a described procedure (32).

For the construction of silencing plasmids, the sequences targeting human Cbp/PAG (5′-AAGCGATACAGACTCTCAACA-3′), mouse Cbp/PAG (5′-AAGCCATACAGACTCTAAACA-3′), and a scrambled sequence (5′-AACGCACGAACTATCAACATA-3′) were cloned into the vector pCMS3-EGFP. All constructs were sequenced to ensure integrity. These constructs were transfected into Jurkat T cells by electroporation (described above). The cells were then cultured in RPMI 1640 supplemented with IL-4 for 4 days to induce opioid receptor expression (described above) and ensure Cbp/PAG protein down-regulation.

Cells (three times 10⁵ cells/ml) in RPMI 1640 medium (phenol red free; Invitrogen) containing 10% FCS were loaded with 5 μg/ml Indo-1-AM (MoBiTec) at 37°C for 45 min. After washing, cells were incubated in RPMI 1640 medium (phenol red free) supplemented with 10% FCS at 37°C for an additional 45 min. Cells were measured using an LSR I flow cytometer (BD Biosciences) equipped with a 325-nm laser. Cells were stimulated with anti-CD3 and the ratio of the two emission wavelengths (512/20 nm and 400/40 nm) was recorded over time. To confirm proper Indo-1 loading and the viability of cells, ionomycin (10 μg/ml; Sigma-Aldrich) was used. The kinetics of the data were analyzed using FlowJo software (Tree Star).

For statistical evaluations between two groups of samples, Student’s t tests were performed. Multiple comparisons were performed with ANOVA, followed by Tukey’s multiple comparison posttest. Asterisks indicate significantly different values (∗, p < 0.05; ∗∗, p < 0.01; and ∗∗∗, p < 0.001).

Activation of the TCR leads to the induction of IL-2. A number of reports demonstrated that chronic morphine treatment of activated T cells inhibits this induction in mice (3, 4, 6). Fig. 1,A shows the effects of chronic opioid treatment on IL-2 transcription in opioid receptor-expressing Jurkat cells (i.e., cells in which opioid receptors were induced; see Materials and Methods) after stimulation with a combination of anti-CD3 and anti-CD28. Similar to murine cells, morphine caused a significant inhibition of anti-CD3/anti-CD28-induced IL-2 mRNA production in the human cells. This effect was mediated by μ opioid receptors, since it was blocked by CTAP, a specific μ opioid receptor antagonist. Similarly, β-endorphin, which is a ligand for μ and δ opioid receptors, caused inhibition of anti-CD3/anti-CD28-induced IL-2 transcription. This effect was also mediated by μ opioid receptors, since it was blocked by CTAP, but not by naltrindole, a specific δ opioid receptor antagonist. To test possible effects on the anti-CD3/anti-CD28-induced IL-2 mRNA production mediated by δ opioid receptors more specifically, we additionally tested the specific δ opioid receptor agonist d-penicillamine2-d-penicillamine5-enkephalin (DPDPE). However, treatment of the cells with DPDPE had no significant effect on IL-2 transcription. Additionally, the effect of morphine on the activities of the transcription factors AP-1, NF-κB, and NFAT were investigated. As shown in Fig. 1 B, morphine significantly inhibited the anti-CD3/anti-CD28-induced activities of all three transcription factors in opioid receptor-expressing Jurkat cells.

One of the early events in T cell activation is mobilization of intracellular calcium. As shown in Fig. 2,A, incubation of opioid receptor-expressing Jurkat cells with morphine for up to 2 h gradually inhibited the TCR-induced calcium flux by >60% (Fig. 2,B). In contrast, morphine had no effect on the calcium flux in naive T cells, which do not express μ opioid receptors (Fig. 2 C).

To obtain further insight into the mechanism(s), by which opioids inhibit anti-CD3/anti-CD28-induced IL-2 transcription, we next investigated their effects on the signaling cascade downstream of the TCR. The activation states of MAPK, LAT, and Zap70 were determined in Western blot experiments using phospho-specific Abs (Fig. 3). Activation of opioid receptor-expressing primary human T cells with anti-CD3/anti-CD28 for 2 min induced a robust phosphorylation of MAPK, LAT, and Zap70. However, this phosphorylation was markedly reduced when the cells were incubated with morphine for 2–8 h before their activation (Fig. 3,A). Simultaneous incubation of the T cells with morphine and the μ opioid receptor-specific antagonist CTAP abolished the effect, indicating that it is mediated via μ opioid receptors (Fig. 3,B). Similar effects were observed when opioid receptor-expressing primary T cells were incubated with β-endorphin (Fig. 3,C). The inhibitory effect of β-endorphin on anti-CD3/anti-CD28-induced MAPK, LAT, and Zap70 phosphorylation was abolished after coincubation of the cells with β-endorphin and CTAP, but not after coincubation with β-endorphin and naltrindole, a specific antagonist for the δ opioid receptor (Fig. 3,D). This indicates that the inhibitory effect of β-endorphin is also mediated via μ opioid receptors. As shown in Fig. 3, E and F, the inhibitory effects of morphine and β-endorphin on the TCR-triggered activation of MAPK are clearly visible at time points later than 80 min of incubation of the cells with the drugs. The effects of the opioids were not observed in naive T cells, which did not express μ opioid receptors (Fig. 3 G).

To more precisely identify the level at which μ opioid receptors and the TCR communicate, we examined Lck, the kinase responsible for initiating TCR signaling and the activation of Zap70. Lck is phosphorylated in unstimulated T cells at Tyr⁵⁰⁵, by which signaling downstream of the TCR is prevented. We observed that in opioid receptor-expressing primary T cells (Fig. 4) and Jurkat cells (data not shown) the dephosphorylation of Lck at Tyr⁵⁰⁵, which occurs after TCR engagement, is inhibited by morphine, as well as by β-endorphin. In contrast, the opioids did not inhibit the TCR-induced dephosphorylation of Lck at Tyr⁵⁰⁵ in naive cells, which do not express μ opioid receptors (Fig. 4). As a control, and to rule out the possibility that the effect of the opioids might be due to down-regulation or internalization of the TCR, the presence of TCR on the cell surface was determined by flow cytometry. It was shown that 98% of the opioid receptor-expressing Jurkat cells were positive for the TCR. After 2 h of morphine treatment, the number of TCR-positive cells did not change significantly, indicating that morphine does not influence the surface expression of the TCR (data not shown).

It is known that the phosphorylation of the inhibitory Tyr⁵⁰⁵ of Lck is mediated by Csk (33). Csk is recruited to the membrane by binding to Cbp/PAG (11), which, in turn, enhances its activity (34). The activity of Csk can be further enhanced by PKA-mediated phosphorylation of Csk (35). Thus, activation of PKA via elevated levels of cAMP exhibits an inhibitory effect upon TCR signaling by stabilizing the inhibited state of Lck (36). Therefore, we monitored the intracellular cAMP concentration in Jurkat cells after morphine and β-endorphin treatment (Fig. 5). In opioid receptor-expressing cells, both substances caused an initial decrease in the cAMP formation at the 15-min time point. However, longer incubation of the cells with the opioids caused a significant elevation of cAMP, which reached a maximum of ∼550% of the controls (Fig. 5,A). In contrast, the opioids did not affect the cAMP concentration in naive T cells, which do not express μ opioid receptors (Fig. 5,B). In the opioid receptor-expressing cells, the μ opioid receptor-specific antagonist CTAP blocked both the morphine- and β-endorphin-induced increase in cAMP levels, whereas the δ opioid receptor-specific antagonist naltrindole had no effect (Fig. 5,C). Furthermore, pertussis toxin inhibited the morphine- and β-endorphin-induced increase in cAMP levels, indicating that the effect is mediated by G_i/o proteins (Fig. 5,C). Next, we aimed to find a possible explanation for the terminal decrease in cAMP that occurs between the 4- and 24-h time points and that brings cAMP levels back to control levels. Speculating that phosphodiesterases might be involved, we found that coincubation of cells with morphine and the unspecific phosphodiesterase inhibitor IBMX inhibited this decrease. Thus, cAMP levels remained at maximal levels at the 8- and 16-h time points (data not shown) and the 24-h time point (Fig. 5,D). To more specifically determine the types of phosphodiesterases, which are involved in this cAMP decrease, we next tested rolipram (inhibitor of phosphodiesterase 4), BRL50481 (inhibitor of phosphodiesterase 7), and dipyridamole (inhibitor of phosphodiesterases 5, 6, 8, 10, and 11) (Fig. 5,D). Our results indicate that the phosphodiesterases 4 and 7 are involved in the terminal decrease in cAMP, since rolipram and BRL50481 significantly reduced the cAMP decrease. Applied together, both substances inhibited the cAMP decrease to a similar extent as IBMX did alone. To test whether the ability of opioids to inhibit TCR signaling is dependent on cAMP-activated PKA, we applied the PKA antagonist cAMPS-Rp. Indeed, preincubation of opioid receptor-expressing primary T cells (Fig. 6) and Jurkat cells (data not shown) with cAMPS-Rp for 1 h before opioid incubation completely abolished the inhibitory effects of morphine and β-endorphin on the anti-CD3/anti-CD28-induced phosphorylation of MAPK and Zap70 and dephosphorylation of Lck. To investigate whether Cbp/PAG/Csk is involved in the inhibition of TCR signaling by μ opioid receptor agonists, we down-regulated the expression of Cbp/PAG using siRNA (Fig. 7). As shown in Fig. 7,A, transient transfection of siRNA directed against human Cbp/PAG resulted in an ∼90% down-regulation of the protein compared with a control (scrambled) siRNA. Transfection of siRNA specific for the mouse protein, which has two nucleotide exchanges compared with the human gene, resulted in an inhibition of about one-third of Cbp/PAG expression compared with the control siRNA. Transfection of control siRNA had no detectable effect on Cbp/PAG expression compared with untransfected, naive cells (data not shown). In opioid receptor-expressing Jurkat cells transfected with siRNA directed against human Cbp/PAG, the inhibitory effect of morphine and β-endorphin and 8-Br-cAMP, which was used as a control that directly activates PKA, on TCR signaling was almost completely blocked as demonstrated by monitoring the phosphorylation of Zap70 (Fig. 7,B). Applying siRNA specific for the mouse protein resulted in a weak blockade of the inhibitory opioid effect. In contrast, applying control siRNA had no influence on the inhibitory effect of the opioids. Together, these data suggest that μ opioid receptor agonists inhibit TCR signaling via the PKA-Cbp/PAG/Csk-Lck pathway. If this was indeed the case, then the PMA/ionomycin-induced activation of MAPK, in which Lck-regulated TCR activation is bypassed, should not be influenced by opioids. Indeed, although the incubation of opioid receptor-expressing Jurkat cells with morphine and β-endorphin for 4 h produced elevated cAMP levels, which are required for the activation of PKA, the PMA/ionomycin-induced activation of MAPK was not influenced (Fig. 8).

In this report, we addressed the question how opioids inhibit the TCR-mediated production of IL-2 in human T lymphocytes. We investigated the effects of morphine, as the pharmacologically most important opioid, and β-endorphin, as an endogenous opioid. The mechanisms are summarized in a schematic diagram presented in Fig. 9.

One of the first steps in these opioid-T cell interactions was an increase in cAMP, which was evoked by both opioids and mediated by μ opioid receptors. Thus far, detailed mechanisms for the regulation of cAMP by μ opioid receptor ligands are poorly understood. It has been suggested from studies in other cells that the activation of μ opioid receptors leads to an initial decease in cAMP levels, which is due to their coupling to the inhibitory G_i/o proteins. Under certain experimental conditions, e.g., removal of the ligand, a compensatory, sometimes overshooting increase in cAMP levels occurring thereafter was observed (37, 38). However, most studies were performed in μ opioid receptor-overexpressing cells of various tissues that normally do not express the receptor and, therefore, are of limited relevance to the physiological regulation of cAMP levels by opioids in human T cells. In activated murine T cells, it was found that morphine treatment for 18 h leads to a weak increase in cAMP (6). However, the determination of cAMP levels at a single, fixed time point only may be insufficient to establish a role for this molecule in the opioid-mediated modulation of TCR signaling, because of the biphasic, time-dependent modulation of the cAMP levels caused by opioids. For the first time, we demonstrated that simple incubation of human T cells with opioids, without any further manipulation like removal of the opioid, induced a short, initial decrease in cAMP levels, which was followed by a marked increase in cAMP. This increase was sensitive to pertussis toxin and thus mediated by G_i/o proteins. The elevated cAMP levels were most pronounced ∼2–4 h after opioid addition. Thereafter, they decreased again, reaching control levels at the 24-h time point. This final decrease in cAMP levels was due to phosphodiesterase activity, catalyzed by the phosphodiesterases 4 and 7.

It is known that agents, which lead to increased cAMP levels, may inhibit the TCR-activated signaling cascade of T lymphocytes via the PKA-Cbp/PAG/Csk-Lck pathway (8). In line with this model, our experiments using a PKA inhibitor strongly suggest that the effect is mediated via this enzyme. The data using siRNA demonstrated the involvement of Cbp/PAG. Since interactions between Cbp/PAG and Csk are crucial for the function of the “master regulator” Lck, these data suggest that μ opioid receptor ligands inhibit TCR signaling by superactivating Csk and thereby inhibiting Lck. The stable phosphorylation of the inhibitory Tyr⁵⁰⁵ within the C terminus of Lck produced by the opioids is likely to lead to intramolecular binding between the Src homology 2 domain and the inhibitory phospho-tyrosine of Lck resulting in the formation of a closed conformation (discussed in Ref. 8). As a consequence, this may prevent interaction of Lck with its substrates, e.g., the TCR complex, resulting in a blockade of TCR-mediated signaling and IL-2 induction. In line with this model, we demonstrated that downstream activation of the TCR pathway with PMA/ionomycin was not inhibited by opioids. The opioid-inhibited signaling cascade then results in decreased activities of the transcription factors AP-1 and NF-κB, which contribute to the induction of IL-2. In addition, we demonstrated inhibition of the TCR-activated calcium flux by morphine. This is likely to result in the inhibition of the calcium/calmodulin-dependent transcription factor NFAT, which also contributes to the induction of the IL-2 gene (5). In our cellular model, the time course for the inhibition of TCR signaling by μ opioid receptor ligands is in good accordance with the kinetics of cAMP accumulation. Noteworthy, it is reasonable to assume that an application of morphine will cause inhibition of the function of activated T cells also in vivo in humans. Since the biological half-life of morphine is ∼2–4 h, considerable concentrations of morphine will be present in an organism for several hours, which may cause the inhibition of IL-2 production. One should recall that orally applied morphine has an even longer half-life of ∼8–12 h. Furthermore, also the metabolite morphine-6-glucuronide is still biologically active. Together, this might markedly increase the time window within which inhibitory effects on T cells might be observed in vivo after morphine treatment, compared with the cellular model shown in this study. To sum up, our data are likely to be directly relevant for pharmacological opioid applications. Similar to morphine, the endogenous opioid β-endorphin had comparable effects on anti-CD3/anti-CD28-induced IL-2 mRNA and the TCR-activated signaling cascade. Interestingly, these effects were mediated by μ opioid receptors only, although β-endorphin binds to μ as well as δ opioid receptors. In line with these findings, the δ opioid receptor-specific agonist DPDPE had no effect on the TCR-mediated induction of IL-2. This suggests that TCR-induced IL-2 production is not influenced by activation of δ opioid receptors. Nevertheless, it was reported by others that specific δ opioid receptor agonists caused robust activation of AP-1 and NFAT transcription factors and induction of IL-2 in δ opioid receptor-overexpressing Jurkat cells (39). However, such strong effects may be due to an unphysiologically high amount of receptors present in the overexpressing cells.

It is becoming increasingly clear that β-endorphin is expressed in T cells and up-regulated in response to several conditions and stimuli (40, 41). Most notably, up-regulation of β-endorphin in T cells was demonstrated in response to T cell activation (40, 42), which may cause an inhibitory feedback regulation within the processes of T cell activation. Furthermore, it was shown that β-endorphin itself induces up-regulation of its own expression in T cells (43). In addition, the peptide is up-regulated in various immune effector cells in response to inflammation (41, 44, 45). Taken together, it is very likely that under certain conditions and, at least locally, a remarkable amount of β-endorphin is present. Our data suggest that β-endorphin has a physiological regulatory function in the immune system by inhibiting T cell activation. They thus support earlier findings obtained in mice, which demonstrated enhanced splenocyte proliferation and enhanced levels of IL-2 mRNA after application of the opioid receptor antagonist naloxone (15, 16) and in knockout mice lacking β-endorphin (17). At the moment, the physiological relevance of the β-endorphin inhibition in TCR signaling is unclear. Since high amounts of the opioid are released from the pituitary during stress, it might contribute to stress-induced immunosuppression. It is interesting to speculate about effects of β-endorphin on activation of T cells in the brain. Being a classic neuromodulator, it can be assumed that remarkable amounts of the opioid are present in the brain. Therefore, activation of T cells infiltrating the brain, e.g., after inflammation may be reduced by β-endorphin, which would establish a neuroprotective function of the opioid within the neuroimmune cross-talk.

We thank Camilla Merten and Helga Tischmeyer for excellent technical assistance. Plates and reagents for the cAMP-ELISA were a gift from Hermann Ammer (Institute of Veterinary Medicine, University of Munich, Munich, Germany).

The authors have no financial conflict of interest.