Tumor Necrosis Factor α Enhances Nicotinic Receptor Up-regulation via a p38^MAPK-dependent Pathway*

Neuronal nicotinic acetylcholine receptors (nAChRs)³ play an important role in modulating normal neurotransmission in the central nervous system. These receptors also have a direct impact upon behavioral and physiological pathologies ranging from addiction to their early and selective loss in Alzheimer disease (1). Addiction to nicotine correlates with the curious trait of up-regulation in response to chronic exposure to receptor ligands (2–4). The majority of high affinity nicotine receptors that undergo up-regulation in the mammalian brain are composed of at least nAChR α4 and β2 subunits as demonstrated by high affinity ligand binding and genetic studies (1, 5–7). Although multiple mechanisms contribute to up-regulation (1, 8, 9), the normal physiological reason for this response is poorly understood despite its being conserved in animals ranging from Caenorhabditis to mammals. Since most of these organisms have never been exposed to nicotine either acutely or at any time throughout their evolutionary history, up-regulation probably reflects a normal physiological response preserved by processes of natural selection. It is also an intrinsic property of nAChRs composed of α4 + β2 subunits (termed α4β2-nAChR; see Refs. 10 and 11 for additional nomenclature). Even when expressed in heterologous systems, such as human embryonic kidney cells, this receptor responds to nicotine with pharmacokinetics that are almost identical to those of the mammalian forebrain (1, 12).

Possible reasons for endogenous up-regulation of nAChR may involve the recent recognition of the participation by nicotinic receptors in regulating proinflammatory processes. In this context, α7-nAChR has received the greatest attention since its antagonism of the proinflammatory cytokine, TNFα, was documented over a decade ago (13). Subsequent investigations have supported the physiological relevance of this interaction in regulating numerous proinflammatory processes (14). However, a more complex interaction between nAChRs and inflammation is suggested by studies using both tissue culture (13, 15, 16) and animal models (14, 17). For example, proinflammatory cytokines can impact upon nAChR expression through promoting efficient receptor assembly and altering relative subunit composition of mature nAChRs when HEK293 (293) cells are co-transfected with cDNA encoding α4 and β2 and/or β4, respectively (16). In these earlier studies, we did note a small but persistent increase of [³H]epibatidine ([³H]Eb) binding in 293 cells transiently expressing α4β2-nAChR that were also treated with TNFα (16). This has been examined further with 293 cells stably expressing nAChRs. Our findings show that the proinflammatory cytokine, TNFα, dramatically enhances α4β2-nAChR up-regulation induced by nAChR ligands, including nicotine, cytisine, and carbachol. The pathway to enhancement of up-regulation is both actinomycin D- and cycloheximidesensitive, and it is inhibited by SB 202190, a highly specific inhibitor of p38^MAPK. These findings offer a novel insight into how proinflammatory cytokines can impact upon mechanisms of up-regulation and strongly suggest that reciprocal regulatory interactions between the nAChR and inflammatory systems are likely.

EXPERIMENTAL PROCEDURES

Cell Lines and Culture Conditions—The 293 cell lines stably co-transfected with nAChR subunits α4 + β2, α4 + β4, α3 + β2, or α3 + β4, respectively, were generously provided by Drs. Ken Kellar and Yingxian Xiao (Department of Pharmacology, Georgetown University). These cells were maintained as described (12, 18, 19). Additional analysis of these cells by real time PCR showed expression of RNA encoding the TNFR subtype 1 (cycle thresholds of 22) and the less abundant TNFR type 2 (cycle threshold of 35) compared with β-actin (cycle threshold of 17; not shown). Application of TNFα (human recombinant TNFα; BioSource) to these cells (25 ng/ml for 2–4 h) revealed that RNA encoding the inducible form of cyclooxygenase, COX2, was elevated 3-fold, demonstrating that TNFα stimulates cellular responses in these cells (not shown). The same assays revealed no evidence for the expression of the acetylcholine-synthesizing enzyme, choline acetyltransferase, or other human nAChR subunits (not shown).

Radioligand Binding—The binding of [³H]Eb to cell membrane preparations was done essentially as described (12, 16), with the following modifications. Cells were distributed into 100-mm culture dishes and treated (e.g. with nicotine) 48 h later. Cells were harvested 18–24 h after treatment, at which time the cultures were 50–75% confluent, into 50 mm Tris buffer (pH 7.4, 4 °C), pelleted, resuspended, and homogenized. Cellular debris and nuclei were removed by low speed centrifugation (100 × g; 5 min), and the supernatant was collected and centrifuged (20,000 × g for 10 min) to pellet remaining membranes for ligand-binding assays. For binding assays, 5 μgof membrane was incubated with 5 nm [³H]Eb for 2–4 h at room temperature (∼25 °C). Nonspecific binding was assessed by adding 500 μm nicotine hydrogen tartrate (Sigma) for 30 min before the addition of [³H]Eb to block specific binding. Samples were tested in triplicate. Bound ligand was separated from free ligand by vacuum filtration through Whatman GF/C filters and prepared for scintillation counting. Specific binding was defined by averaging the total binding minus the nonspecific (nicotine-blocked) binding. Data were analyzed using Prism 3 (GraphPad Software Inc., San Diego, CA) as described (12, 18, 19).

Calcium Imaging—Fura-2/AM calcium imaging was done essentially as before (20). Cells were grown on glass coverslips coated with poly-l-lysine and laminin and washed four times with Hanks' solution (with 1.3 mm CaCl₂ (Invitrogen)) over 20 min, and fresh buffer was added containing 5 μg/ml freshly prepared Fura-2/AM (Molecular Probes) in Me₂SO (Sigma). After a 30-min incubation, the cells on coverslips were again washed, and fresh Hanks' buffer was added before testing in a Zeiss Axiovert microscope equipped with the Attofluor calcium ratio system. Upon calibrating resting calcium-bound Fura-2 (340 nm) versus nonbound Fura-2 (380 nm) ratios, 1 μm nicotine was rapidly applied directly onto the cells using a fixed pipette or a picospritzer. The cell response (change in the 340/380 nm ratio) was then recorded. Nicotine was then added a second time, and the measurement was repeated before adding 1 mm acetylcholine to activate muscarinic receptors as a positive control for cell responsiveness (∼98% in each test group). Recordings were prepared from multiple coverslips in independent platings, which were scored for responsiveness. A responsive cell was defined as having a change in the peak intensity of the 340/380 nm ratio of 0.01 units from resting upon nicotine application and that failed or exhibited a very poor response to the second application (desensitization). Primary data were analyzed using Excel.

RESULTS

The subunit composition of nAChRs determines their pharmacological and functional response (1). Therefore, we first examined how nicotine, TNFα, and the combination of these agents impacts upon α4β2-nAChR, α4β4-nAChR, α3β2-nAChR, or α3β4-nAChR, respectively, stably expressed by 293 cells. For these experiments, cells were treated for 18–24 h with nicotine (1 μm), TNFα (25 ng/ml), or both agents, and specific binding of [³H]Eb to 5 μg of membrane protein was measured (see “Experimental Procedures”). Nicotine-induced up-regulation of [³H]Eb binding to α3 + β2 and α4 + β2 equaled 4–6-fold over controls, ∼2-fold for α4β4-nAChR, and no significant change for α3 + β4 (Fig. 1), as previously reported (1, 12). Although not as robust as nicotine, TNFα induced a small but significant increase in [³H]Eb binding by cells expressing α4β2-nAChR and α4β4-nAChR but not α3β2-nAChR or α3β4-nAChR. The most dramatic effect on ligand binding was in α4β2-nAChR-expressing cells simultaneously treated with TNFα and nicotine. In these cells, the increase in [³H]Eb sites exceeded 10-fold that of the control and often exceeded nicotine-treated cells by 2-fold. A smaller effect by TNFα on nicotine up-regulation was observed for α3β2-nAChR receptors and to a lesser but significant extent α4β4-nAChR. Again, cells expressing α3β4-nAChR exhibited only a slight trend toward enhanced ligand binding relative to controls that was not statistically significant.

We focused the subsequent experiments toward examining the dramatic enhancement by TNFα on nicotine up-regulation of α4β2-nAChR. To begin, dose-response assays for up-regulation of [³H]Eb binding in response to 18–24 h of exposure to nicotine, TNFα, or nicotine plus TNFα were performed. Up-regulation was observed in cells treated with as little as 100 nm nicotine (∼2-fold) and saturated at 1 μm nicotine (Fig. 2A). TNFα (Fig. 2B) induced weak up-regulation at all doses tested (optimal dose, 12.5–25 ng/ml). The results of experiments where nicotine was held constant at 1 μm and the concentration of TNFα varied are shown in Fig. 2C. The optimal enhancement of up-regulation (12-fold relative to 4-fold for nicotine alone in this experiment) occurred at 1 μm nicotine and 25 ng/ml TNFα, which was also true in experiments where constant TNFα (25 ng/ml) and varied nicotine (10 nm to 10 μm) were done (not shown). The order of addition (TNFα versus nicotine), including in experiments where they were added 24 h in advance of the other, had no significant influence on the results (not shown). Therefore, 1 μm nicotine and 25 ng/ml TNFα were used in subsequent experiments unless otherwise noted.

An important variable in tissue culture experiments is cell culture density. This is particularly true when variables such as cell cycle, nutrients, pH, and cell-cell contact (that may also affect accessibility by ligands to receptors) influence the experimental results. The influence of cell density on TNFα-enhancement of nicotine-mediated up-regulation was examined in cultures that were plated to reach confluence of 100, 75, 50, and 25%, respectively, following a total of 36 h in culture and 24 h after treatment with nicotine, TNFα, or both agents as above. The results show (Fig. 2D) that cell cultures at ∼50% confluence display the greatest enhancement by TNFα of up-regulation (1.8-fold). There was, however, a decline in the amount of TNFα enhancement of up-regulation with increasing cell density (∼1.4-fold). This result indicates that cell culture conditions influence the ability of TNFα to enhance nicotine-mediated up-regulation. This finding is consistent with cell density influencing nAChR expression by transfected cultured cells. For all experiments reported, culture density was ∼50% confluence at the time of harvest.

TNFα Enhancement of Nicotine-mediated Up-regulation Requires New RNA Transcription and Protein Synthesis—The mechanism of nicotine plus TNFα-enhanced up-regulation was examined next. To begin, saturation binding of [³H]Eb over a concentration range of 1–5000 pm for each treatment group was measured (Fig. 3A). For control cells, the [³H]Eb-specific binding was 537 ± 140 fmol/mg, and the saturable best fit was to a single high affinity site (control mean K_d of 53 ± 14 pm) as calculated using Prism 3.0 software (12). Nicotine competition curves averaged from three independent experiments (Fig. 3B; nicotine binding competition against 500 pm [³H]Eb) show that at 10^–4 m to ∼5 × 10^–5 m, nicotine competed for >85% of the specific [³H]Eb binding. There was no significant difference in the affinity of this compound measured in these experiments (nicotine IC₅₀ of 12 ± 4 nm; see Ref. 12) among the different treatment groups. Collectively, these data support a model where combining nicotine and TNFα up-regulation by co-exposure is related to changes in receptor number.

How TNFα enhances nicotine-mediated up-regulation of α4β2-nAChR [³H]Eb binding sites was examined further. Although previous reports show ligand-mediated up-regulation of nAChRs to be independent of transcription (1), how TNFα enhancement of up-regulation occurs is unknown. The first measurement determined if intact cells were required for TNFα to enhance nicotine-mediated up-regulation. For this experiment, a crude membrane fraction of untreated cells was prepared as described for binding assays. Equal aliquots of these preparations were treated with saline, 1 μm nicotine, 25 ng TNFα, or both nicotine and TNFα and incubated at 37 °C for 24 h with gentle rocking. Membranes were then collected by centrifugation and washed, and radioligand-binding measurements were performed. The results (Fig. 3C) show that incubation of membranes with TNFα alone had no effect on [³H]Eb binding, although a small but significant increase was measured in nicotine-treated samples and those with nicotine plus TNFα. This persistent increase is consistent with reports of up-regulation through mechanisms that influence receptor affinity for ligand (9). Nevertheless, the changes in ligand binding were well below the 4–6-fold change expected from whole cell preparations, and no TNFα enhancement of this up-regulation effect was present, indicating that TNFα acts on up-regulation through mechanisms requiring intact cells.

The time course of TNFα-enhanced up-regulation was examined to determine when TNFα-enhanced up-regulation begins. These experiments were done as above, only cells were harvested at various times after the addition of nicotine and/or TNFα. As shown in Fig. 3D, a small enhancement by TNFα of nicotine up-regulation was observed with in 2 h of treatment; however, significant enhancement of up-regulation was not observed until ∼8 h post-treatment. This result indicates that the impact of TNFα on up-regulation is not likely to be through modifications of preexisting protein pools. To test this, we next examined the influence of the transcriptional inhibitor, actinomycin D, on TNFα-enhanced up-regulation. In these experiments, [³H]Eb binding of membranes isolated from cells treated with nicotine, TNFα, nicotine plus TNFα, or saline for 18 h, either in the presence or absence of 10 μm actinomycin D (a dose tolerated by the cells; not shown), was measured. Upon normalizing the results to the control cells as in Fig. 3E, it was found that consistent with the previous reports noted above, there was no effect by transcription inhibition with actinomycin D on the up-regulation of [³H]Eb binding sites treated with nicotine relative to control values. However, TNFα enhancement of the nicotine-induced up-regulation process was abolished. In a similar experiment, inhibition of new protein synthesis with 10 μm cycloheximide was done as described above for the actinomycin D experiments. The results of these experiments, shown in Fig. 3F, demonstrate that cycloheximide abolishes essentially all up-regulation as well as TNFα enhancement of up-regulation relative to control cells. Notably, a small amount of up-regulation in cells treated with nicotine did persist, similar to that seen in membrane incubation experiments (Fig. 3C). However, we did not explicitly rule out the possibility that some residual protein synthesis could have occurred. What is important is the substantial reduction of up-regulation and enhancement by TNFα (>80% relative to controls). Together with the results of actinomycin D studies and the timing of the onset of enhanced up-regulation, these findings argue that the mechanism(s) of TNFα enhancement of [³H]Eb binding sites requires both new transcription and protein synthesis.

Up-regulation Produced by Ligands Other than Nicotine Is Enhanced by TNFα—Most agonists that bind nAChRs (e.g. nicotine) and some antagonists (e.g. dihydro-β-erythroidine) produce up-regulation (21). In this context, to examine the specificity of TNFα enhancement of up-regulation, the change in [³H]Eb binding sites by α4β2-nAChR-expressing cells in the presence or absence of TNFα for 24 h and either 1 mm carbachol, 10 μm dihydro-β-erythroidine, 10 μm cytisine (a competitive partial agonist), or mecamylamine a noncompetitive antagonist not generally reported to produce up-regulation). All agents, except mecamylamine, consistently produce significant up-regulation of α4β2-nAChR [³H]Eb binding (Fig. 4). TNFα enhanced up-regulation induced by these agents proportionately to that seen with nicotine except for mecamylamine, where again no significant effect was observed. The greatest up-regulation was observed with carbachol, although enhancement of this effect by TNFα was proportional to that seen with other receptor ligands. Because carbachol is also an agonist of muscarinic receptors, the same experiment (Fig. 4A) was done in the presence of atropine (which by itself had no effect on [³H]Eb binding; not shown). Atropine had no measurable effect on carbachol up-regulation of [³H]Eb binding or TNFα enhancement. Also, because the concentration of carbachol used was relatively high, a dose-response assay was done (Fig. 4B). At concentrations to 60 μm, carbachol produced significant α4β2-nAChR up-regulation of [³H]Eb, and TNFα (25 ng) enhanced this up-regulation. As reported (8), the up-regulation of α4β2-nAChR by the competitive antagonist dihydro-β-erythroidine was approximately one-third that of nicotine, yet TNFα also enhanced [³H]Eb binding proportionately. Collectively, these results indicate that in all cases tested where ligand binding by the nAChR initiates the up-regulation mechanism, proportional enhancement by TNFα occurs.

Carbachol-mediated up-regulation is particularly useful, since this method permits conditions for Fura-2 calcium imaging measurements. This is because nicotine up-regulation results in receptors that are also deeply desensitized, and the extensive amounts of washing required to restore function are often impractical. In contrast, nAChR function is restored in carbachol-treated cells within 20–30 min of washing (not shown). Following 18–24 h of 500 μm carbachol, rapid application of 1 μm nicotine produces an easily measurable change in the Fura-2 340/380 ratio in ∼60% of the cells (193 of 315 cells for the experiment shown; Fig. 4D), which is in contrast to control cells that are rarely able to produce a convincing response (Fig. 4C). In cells treated with carbachol + TNFα, the peak response measured was on average 30% greater than the response measured (Fig. 4E) in carbachol only-treated cells, although there was only a small increase in the percentage of cells responding that was not seen in all experiments (236 of 350 cells). This result indicates that TNFα, in addition to increasing the number of [³H]Eb sites, also increases the number of functional receptors (or their duration of opening) but does not necessarily alter the number of cells able to produce a measurable functional response.

Mechanism of TNFα Enhancement of α4β2-nAChR Up-regulation—TNFα signaling through its receptors impacts on numerous intracellular pathways (22, 23). Of interest to us has been the activation of p38 mitogen-activated protein kinase (p38^MAPK) pathways. We have examined the possible contribution of this intracellular pathway to enhancement of up-regulation using various inhibitors, including those affecting phosphatidylinositol 3-kinase/protein kinase B, extracellular signal-regulated kinase, and p38^MAPK. For these experiments, all treatments with saline, nicotine, TNFα, or both agents were done as described above. Me₂SO, which is used as the carrier for inhibitor stock solutions, has no effect on ligand binding. Representative results for these inhibitors on the TNFα enhancement of nicotine-mediated up-regulation are shown in Fig. 5. Two inhibitors affected the TNFα enhancement of nicotine-induced up-regulation. The MEK1/2 inhibitor, U0126, inhibits ∼50% of TNFα-enhanced up-regulation, but the MEK1-preferring inhibitor (24), PD98059, has no effect, suggesting that signaling through at least the MEK2 pathway is important. However, effectively all enhanced up-regulation was blocked by the p38^MAPK inhibitor, SB 202190 (Fig. 5). None of these agents increased or inhibited nicotine-induced up-regulation (not shown).

DISCUSSION

Nicotine has long been associated with its ability to up-regulate the expression of α4β2-nAChRs, and this correlates with phenotypes related to addiction (1). More recently, this compound has also been found to directly impact upon inflammatory and immunologic responses (25). In a reciprocal manner, inflammatory mediators can affect the expression of nicotinic receptors (16). In this report, using stably transfected cell lines, we find that TNFα strongly enhances α4β2-nAChR up-regulation, independent of other inflammatory conditions or nAChR subtypes, through an actinomycin D and cycloheximide-sensitive p38^MAPK intracellular signaling pathway. A diagram reflecting these pathways is shown in Fig. 6. The p38^MAPK cascade has a central role in regulating many immunologic and inflammatory responses as well as physiological functions related to cell growth, cell differentiation, and apoptosis. Partial inhibition of TNFα-enhanced up-regulation of α4β2-nAChR by the MEK1/2 inhibitor U0126, but not the MEK1-preferring PD98059 (24), implies participation of an additional pathway that leads to p38^MAPK activation. Collectively, the delay in the onset of TNFα-mediated enhancement of up-regulation, the selective diminishment of this effect by both transcription and translation inhibitors, and the general role of p38^MAPK in activating stress-related genes combine to suggest that TNFα-mediated enhancement of up-regulation occurs through induction of a protein that favors nicotinic receptor expression by a post-translational mechanism(s). Particularly attractive candidates are chaperone-like proteins, since these would promote more efficient folding, assembly, and/or transport from the endoplasmic reticulum to the cell surface. Also, the impact of these signals appears to be based upon subunit composition. In particular, enhancement of ligand-mediated up-regulation is most notable in receptors harboring the β2 subunit. However, we observed a small but significant and direct influence by TNFα on promoting increased ligand-binding sites for α4β2-nAChR and α4β4-nAChR but not for α3-containing receptors. This indicates that the α4 subunit is also likely to harbor structures that are targeted directly by the TNFα enhancement up-regulation signal.

This study and others (13–16) also indicate that an interaction between inflammation and various nAChR subtypes occurs through multiple mechanisms. For example, α7-nAChR has been demonstrated to reduce the release of TNFα (14), which would dampen inflammation (and inflammatory cytokines), and indirectly regulate the amount of α4β2-nAChR up-regulation. However, the relationship between α7-nAChR, TNFα, and enhanced up-regulation of α4β2-nAChR may not be straightforward, since TNFα can originate from multiple sources that may or may not be regulated by α7-nAChR.

A final point is to speculate regarding the translational relevance of this observation in the animal. One place where this is possible is related to the observation that when individuals are ill, they often experience a diminished desire to continue nicotine administration. In fact this has been used as an indicator of nicotine dependence (26). Because illness is often associated with elevated TNFα, in the presence of chronic nicotine administration, enhanced up-regulation might actually exceed the comfort level and possibly even become toxic. A second point involves changes in nicotinic cholinergic receptor expression and inflammatory dysfunction that is characteristic of aging animals. There is a dramatic and selective loss of α4β2-nAChR expression in the aged brain of both rodents (27–30) and humans (especially those with Alzheimer's disease; e.g. see Refs. 31 and 32). Coincident with the onset of old age and pathology is also an unregulated increase in TNFα and other proinflammatory cytokines. At present, the correlation between the loss of α4β2-nAChR expression and onset (or progression) of pathology is unknown. However, the possibility of a mechanistic interaction contributing to this coincidental dysregulation in both systems is intriguing. These results and other possible interactions between inflammatory cytokines and nAChR expression extend our understanding of the physiologic relevance of up-regulation as a normal contributor to local and conditional regulation of nAChR function.