The primary characteristic features of any inflammatory or infectious lesions are immune cell infiltration, cellular proliferation, and the generation of proinflammatory mediators. TNF-α is a potent proinflammatory and immuno-regulatory cytokine. Decades of research have been focused on the physiological/pathophysiological events triggered by TNF-α. However, the signaling network initiated by TNF-α in human leukocytes is still poorly understood. In this study, we report that TNF-α activates phospholipase D1 (PLD1), in a dose-dependent manner, and PLD1 is required for the activation of sphingosine kinase and cytosolic calcium signals. PLD1 is also required for NFκB and ERK1/2 activation in human monocytic cells. Using antisense oligonucleotides to reduce specifically the expression of PLD isozymes showed PLD1, but not PLD2, to be coupled to TNF-α signaling and that PLD1 is required to mediate receptor activation of sphingosine kinase and calcium transients. In addition, the coupling of TNF-α to activation of the phosphorylation of ERK1/2 and the activation of NFκB were inhibited by pretreating cells with antisense to PLD1, but not to PLD2; thus, demonstrating a specific requirement for PLD1. Furthermore, use of antisense oligonucleotides to reduce expression of PLD1 or PLD2 demonstrated that PLD1 is required for TNF-α-induced production of several important cytokines, such as IL-1β, IL-5, IL-6, and IL-13, in human monocytes. These studies demonstrate the critical role of PLD1 in the intracellular signaling cascades initiated by TNF-α and its functional role for coordinating the signals to inflammatory responses.
Tumor necrosis factor-α, is one of the most potent and pleiotropic proinflammatory cytokines, and has been associated with a wide range of diseases, including acute and chronic infections, as well as in inflammatory, allergic, and autoimmune diseases (1, 2, 3, 4, 5, 6). TNF-α is produced by many cells, but in higher amounts by phagocytes and mast cells; it is also produced by lymphocytes, NK cells, Kupffer cells, glial cells, and adipocytes (1). TNF-α binding to its transmembrane receptors (TNFR I and II), expressed on a wide variety of cell-types, leads to the stimulation of multiple signaling cascades (7), including the MAPK and NFκB cascades, leading to the generation of proinflammatory cytokines and other proinflammatory molecules. However, the mechanisms that regulate TNF-α-induced MAPK and NFκB activation in human monocytes are still poorly understood.
Receptor-coupled signaling transduction mechanisms are tightly regulated by a series of protein phosphorylation and dephosphorylation events, and MAPK is a key one among them. MAPKs play a crucial role in amplifying signals and regulating signal transduction cascades leading to cell proliferation, chemotaxis, and the translation of diverse extracellular stimuli to the nucleus, resulting in cellular functions like gene expression, mitosis, differentiation, survival, and apoptosis (8). Five distinct MAPK subfamilies have been identified to date: these are the ERKs (p44/42-ERK1/2), the 38-kilodalton MAPKs (p38 α, β, γ, and δ), the JNKs (JNK1, 2, and 3), and the ERK 3, 4, and 5 kinases (8). ERK1/2 are activated by growth factors, phorbol esters, serum, cytokines, certain stresses, and ligands of G-protein-coupled receptors (9, 10); MAPKs are found to be key regulators of cell proliferation (8). p38 kinase activity is triggered by osmotic, physical, and chemical stress, UV irradiation, proinflammatory cytokines, and hormones (9, 11, 12) and is found to be vital in regulating immune and inflammatory responses (8).
Cell surface receptor engagement by a ligand leads to phospholipase-mediated hydrolysis of cellular phospholipids, leading to the generation of lipid-derived products, which in turn play critical roles in a wide range of intracellular signaling pathways (13, 14). Intracellular signaling mediated by phospholipase D (PLD)3 has been a target of interest in inflammation and tumor metastases for more than two decades (14). It is well known that PLD is a major source of second messengers, including phosphatidic acid (PA) (15), that can be converted to another important messenger, diacylglycerol (16). PLD is considered to regulate cellular responses that contribute to inflammation and tumor cell migration (14). In mammalian cells, activation of phosphocholine-specific PLD has been proposed to control signal transduction pathways regulating a wide range of physiological processes, including membrane trafficking (17) and cytoskeletal reorganization (18), phagocytosis (19), phagocytic cell NADPH-oxidase-respiratory burst (20, 21), and mast cell degranulation (22).
Some studies have suggested a role for PLD in TNF-α-mediated cellular cytotoxicity (23, 24). However, these early studies have not been followed or validated, and the functional significance of these observations remains controversial (25). The role of PLD in TNF-α-induced intracellular signaling events remains largely unknown.
In this study, we show that PLD1 is activated by TNF-α, plays a key role in the signaling hierarchy, and is essential for TNF-α-triggered sphingosine kinase, NFκB, and ERK1/2 activation, but not for p38 phosphorylation. Furthermore, the physiological importance of PLD1 in the TNF-α-mediated signaling pathways is emphasized by its role in mediating proinflammatory cytokine production.
Materials and Methods
U937 cells were cultured in RPMI 1640 (Life Technologies), supplemented with 10% FCS, 2 mM glutamine, 10 U/ml penicillin, and 10 mg/ml streptomycin at 37°C, 6.8% CO2 in a water-saturated atmosphere. The cells were treated with IFN-γ (Bender Wien Ltd; 200 ng/ml) for 16 h for them to be differentiated into a monocytic phenotype.
Mononuclear cells were isolated from heparinized fasting venous blood by Ficoll-Hypaque centrifugation as previously described (27). The 20 ml of blood (anticoagulated with 10 U/ml heparin) was layered carefully on 15 ml of Ficoll-Hypaque gradient and centrifuged at 500g, without brakes, at room temperature for 30 min. The mixed mononuclear band was aspirated and the cells were washed three times in phenol red RPMI 1640 medium containing 100 U/ml penicillin, 100 × g/ml streptomycin, and 2 mM glutamine and suspended in a known volume. Leukocyte count was performed on a Coulter counter and then cells were plated (5–7 × 106 cells) in 6-well culture plates in RPMI 1640 medium. Incubation was conducted at 37°C for 2 h in 5% CO2/95% air, after which, nonadherent cells were removed by washing the wells twice with RPMI 1640, and the remaining adherent cells were grown in the culture medium supplemented with 10% FCS and 2 mM glutamine for 2 days. The cells were used after 2 days of culture. Cell viability, determined by trypan blue exclusion, was ∼94% in all experiments.
Antisense oligonucleotides were purchased from 1st Base (Singapore); 24-mers were synthesized, capped at either end by the phosphothiorate linkages (first two and last two linkages), and corresponded to the reverse complement of the first 8 amino acids for either PLD1 or PLD2. The sequences of the oligonucleotides were:
5‘CCGTGGCTCGTTTTTCAGTGACAT 3‘ for PLD1 and 5‘GAGGCTCTCAGGGGTCGCCGTCAT 3‘ for PLD2.
U937 cells were incubated in 10 μM oligonucleotide for a total of 36 h (20 h before and then for the duration of culture with IFN-γ). Primary monocytes, after 12 h in culture, were incubated in 10 μM oligonucleotide for a total of 36 h.
Measurement of PLD activity
PLD activity was measured as previously described (21, 26), using the transphosphatidylation assay. Briefly, U937 cells were labeled (106 cells/ml) with [3H]palmitic acid (5 μCi/ml; Amersham Biosciences) in the cell culture medium for 16 h. Following washing, the cells were incubated at 37°C for 15 min in RPMI 1640 medium containing ethanol (0.3% final). Following TNF-α (PeproTech) 10 ng/ml stimulation for 0, 2, 5, 10, 20, 30, and 60 min at 37°C, cells were then lysed and lipids extracted by Bligh-Dyer phase separation. The accumulated phosphatidylethanol was assayed as described previously (21).
To study re-localization patterns of PLD isoforms, normal resting cells were stimulated by TNF-α over a time scale (0, 2, 5, and 10 min). After TNF-α stimulation, the suspended cells were fixed in 4% paraformaldehyde and deposited on microscope slides using a cytospin centrifuge; they were then permeabilized for 5 min in 0.1% Triton X-100 in PBS. The permeabilized cells were blocked for nonspecific binding with 5% FCS for 10 min at room temperature. Fluorescent labeling was done by incubating the cells with goat-polyclonal PLD1 (Santa Cruz Biotechnology), goat-polyclonal PLD2 (Santa Cruz Biotechnology), and an irrelevant goat-polyclonal (Santa Cruz Biotechnology), primary Abs for 1 h at room temperature. The cells were washed with PBS and then incubated with appropriate secondary Abs (anti-goat IgG-FITC conjugate) for an hour. The cell-laden slides were then washed, cover slips were mounted, and the cells preserved using Fluorsave reagent (Calbiochem). To a set of cells, irrelevant control Abs plus the secondary Abs were added as control. Staining was visualized with an inverted fluorescence Leica DM IRB microscope and recorded by a Leica DC 300F digital camera; pictures were analyzed with the Leica IM500 Image Manager software.
Immunoprecipitation of PLD
PLD1 and PLD2 were immunoprecipitated from cell lysates before Western blot analysis of the desired proteins. Goat polyclonal Ab (2 μg), either anti-PLD1 or anti-PLD2 (Santa Cruz Biotechnology), were incubated with 50 μl of 50% Protein A-agarose and 450 μl of PBS, for 2 h on a rocking platform at 4°C to form precipitating complexes, a goat-polyclonal against ARF-1 was used as an irrelevant control for immunoprecipitation. Then, the Ab and Protein A-agarose mix was washed to remove unbound Ab. Following this, 500 μl of cell lysate containing 200 μg protein was mixed with the precipitating (Ab:protein A-agarose) complex and placed in a tumbler at 4°C for 4 h. Following incubation, the precipitating complex was centrifuged and washed, before addition of Lamelli buffer for loading on to 8% polyacrylamide gels (SDS-PAGE).
Following TNF-α stimulation at various time points (0, 5, 10, and 30 min), the cell lysates were prepared using RIPA lysis buffer and then protein concentrations were estimated. Proteins were then resolved on 12% polyacrylamide gels (SDS-PAGE) under denaturing conditions and then transferred to polyvinylidene difluoride membranes (Millipore). The polyvinylidene difluoride membranes were handled as per the manufacturer’s instructions. The membranes were incubated with the relevant primary Abs (p-p44/42, p-p38 – Cell Signaling; GAPDH – Santa Cruz Biotechnologies) and appropriate HRP-conjugated secondary Abs (Sigma-Aldrich) at room temperature. The membranes were washed with washing buffer (1% Tween 1× PBS) and bands were visualized by autoradiography using ECL Western blotting detection system (Amersham Biosciences).
Cytosolic calcium measurement
Cytosolic calcium was measured as described previously (27). Briefly, cells were loaded with 1 μg/ml Fura2-AM in PBS, 1.5 mM Ca2+, and 1% BSA. After removal of excess reagents by dilution and centrifugation, the cells were resuspended in 1.5 mM Ca2+-supplemented PBS and warmed to 37°C in the cuvette. After the basal line was obtained, the cells were stimulated by the addition of TNF-α. Fluorescence was measured at 340 and 380 nm and the background-corrected 340:380 ratio was calibrated as previously described (27).
Sphingosine kinase (SphK) activity in cell extracts
Cells (2 × 106) per sample were stimulated with TNF-α, as above. Following TNF-α stimulation the cells were lysed and the cell extracts were assayed for SphK activity. SphK activity was measured as described previously (27). Briefly, the system is based upon the SphK-catalyzed transfer of the γ-phosphate group of ATP (using a mixture of cold ATP and [γ32P] ATP (1 μCi/sample; Amersham Biosciences) to a specific substrate, then the products were separated and analyzed as previously described (27).
NFκB activity assay
NFκB activity was analyzed from cells, which were either pretreated or not with the antisense against PLD1 or PLD2 and stimulated with TNF-α (10 ng/ml). NFκB activity was analyzed using the EZ-Detect transcription factor kit (Pierce) following the manufacturer’s instructions. Briefly, this kit is based on an ELISA format, provided in a 96-well format with oligonucleotides containing the consensus binding sequences for the transcription factor coated on the wells. Cell extracts are incubated in the wells, and bound transcription factors are then detected by a specific primary Ab; a HRP-conjugated secondary Ab is then used to detect the bound primary Ab. The enzymatic product can be measured with any standard plate reader.
Cells (2 × 106) pretreated or not with antisense oligos were stimulated by the addition of TNF-α (10 ng/ml) for 24 h. Following stimulation, the supernatants were collected at the indicated time points and stored at −20°C until used. IL-1β, IL-6, IL-5, and IL-13 levels in the supernatants were evaluated using ELISA (BD Pharmingen), following the manufacturer’s instructions with lower detection limits of ≤10 pg/ml.
Statistical differences between control and treated cells were calculated using Student’s t test. A statistical difference of at least 95% (p < 0.05) was considered significant.
TNF-α induces PLD activity in human monocytic cells and subcellular re-localization of PLD1
We first investigated whether TNF-α would induce the activity of PLD in the human monocytic cell line used. TNF-α stimulated a rapid increase in PLD activity, which was detectable even after 2 min of TNF-α stimulation (Fig. 1,A), and this was a dose-dependent response with maximal activity at 10 ng of TNF-α; however, activity at 2 and 5 ng of TNF-α was quite robust (Fig. 1,B). Since both PLD1 and PLD2 (PLD isoforms) have been found to be expressed in the human monocytic cells used (21), we decided to look at the subcellular localization of each of the two isoforms and whether TNF-α stimulated their relocalization, as a means to start dissecting the specific isoform(s) activated by this cytokine. Fluorescent microscopy-derived results, in our study, reveal that in resting cells, both PLD isoforms have a general cytosolic localization and that TNF-α induces the re-localization of PLD1, but not PLD2, to the cells’ periphery (Fig. 1, B and C). This would suggest that the PLD isoform activated by TNF-α is, potentially, PLD1.
TNF-α stimulates PLD1
As both isozymes for PLD are expressed in U937 cells, experiments were performed to examine their respective roles; in particular, their activities following TNF-α stimulation. To do this, specific antisense oligonucleotides were designed against each of the PLD isozymes to specifically knock-down the expression of each enzyme (i.e., antisense to PLD1 and antisense to PLD2). We have previously shown that U937 cells are sensitive to antisense manipulation (21). IFN-γ primed cells were treated with antisense oligonucleotides, and PLD activity was assayed in unstimulated cells to measure basal levels of activity, or after stimulation with TNF-α. The specificity of the antisense oligonucleotides on relative PLD isozyme expression was checked by Western blot analysis (Fig. 2 A). In this study, it was found that, in cells treated with antisense to PLD1, there was a reduction in PLD1 immunoreactivity (a reduction of 81 ± 5% was quantified by densitometry, from three separate experiments), whereas PLD2 immunoreactivity was unaffected. Conversely, in cells treated with antisense to PLD2, there was a reduction in PLD2 immunoreactivity (a reduction of 85 ± 10% was quantified by densitometry, from three separate experiments), whereas PLD1 immunoreactivity remained unchanged. Each antisense oligonucleotide, therefore, acted as an internal control for the other. Further controls were performed by the immunoprecipitation of ARF1, with a goat primary Ab, as control of immunoprecipitation experiments, as well as of antisense specificity.
Treatment of cells with the antisense oligonucleotide to PLD1 resulted in no change in basal PLD activity. However, following TNF-α stimulation, the increase in PLD activity was significantly reduced, compared with the control cells (p < 0.01) (Fig. 2,B). The reduction in the increase after TNF-α activation was 80 ± 5% in cells treated with antisense PLD1, compared with control cells, and was proportional to the observed reduction in protein expression by Western blot analysis. In contrast, treatment of cells with the antisense oligonucleotide to PLD2 significantly reduced basal PLD activity (p < 0.01). TNF-α-mediated activation of PLD was marginally reduced in cells treated with the antisense to PLD2, but this reduction was entirely accounted for by the reduction in basal levels; the increment over basal was identical in control (untreated) cells and those pretreated with PLD2 antisense oligonucleotide (Fig. 2 B), demonstrating that, at least in this system, TNF-α stimulation specifically activates PLD1.
PLD1, and not PLD2, couples TNF-α to the activation of SphK, cytosolic calcium transients, and NF-κB activation
Previously, we have shown that TNF-α, in human monocytes, results in a SphK-dependent release of calcium from intracellular stores (27). In this study, we show that SphK activity and cytosolic calcium responses triggered by TNF-α depend on PLD1.
Pretreating cells with the antisense oligonucleotide to PLD1 to knock-down isozyme expression significantly reduced the activation of sphingosine kinase, following TNF-α stimulation, by 78 ± 6% (p < 0.01). The reduction in peak activation was proportional to the loss of PLD1 enzyme expressed in these cells, as assessed by Western blot analysis. Reduction in expression of PLD2 had no effect on the ability of TNF-α to trigger sphingosine kinase activation (Fig. 3,A). Similarly, reduction in expression of PLD1 by pretreatment of cells with antisense PLD1 oligonucleotide resulted in the attenuation of the cytosolic calcium response observed, following TNF-α stimulation (Fig. 3 B). Reducing expression of PLD2 had no effect on the calcium transients, compared with controls.
One of the key transcription factors that trigger the generation of many proinflammatory molecules is NFκB. In this study, we show that the TNF-α-triggered activation of p50 and p65-NFκB in human monocytes is substantially reduced in cells pretreated with the antisense oligonucleotide against PLD1, whereas pretreatment with a PLD2 antisense oligonucleotide had no observed effect on NFκB activity (Fig. 3 C).
These results suggest that PLD1 is upstream of SphK activity and its subsequent coupling to cytosolic calcium signals and NFκB activation.
TNF-α is functionally coupled to ERK1/2 phosphorylation through PLD1 activation, but TNF-α-induced p38 phosphorylation is independent of PLD activity
TNF-α-triggered activation of MAPKs has important functional consequences for several cellular responses, including for the generation of cytokines. Our study demonstrates a critical role for PLD1 in mediating the TNF-α-induced phosphorylation of ERK1/2. Treatment of cells with the antisense oligonucleotide to PLD1 resulted in the inhibition of the phosphorylation levels of ERK1/2 triggered by TNF-α (Fig. 4,A). In contrast, treatment of cells with the antisense oligonucleotide to PLD2 did not alter TNF-α-induced ERK1/2 phosphorylation, compared with control cells (Fig. 4,A). To evaluate the relevance of PLD1 in p38 kinase-activity in the TNF-α-triggered signaling pathway, we looked at the phosphorylation of p38 induced by TNF-α. We found that TNF-α did indeed trigger p38 phosphorylation in the human monocytic cells; however, in contrast to ERK1/2 phosphorylation, the phosphorylation of p38 was not inhibited when the cells were pretreated with antisense oligonucleotide to PLD1 (Fig. 4,B). Similarly, pretreatment of cells with the antisense oligonucleotide to PLD2 did not alter the TNF-α-induced p38 phosphorylation (Fig. 4 B).
Taken together these results suggest that PLD1 is upstream of ERK1/2, but TNF-α-mediated p38 phosphorylation is independent of PLD activity. It has been suggested that PLD activity may be downstream of ERK1/2 activation (28, 29, 30, 31). To investigate this, experiments were conducted in cells pretreated with MEK and p38 inhibitors, and the TNF-α-triggered PLD activity was measured. The results shown in Fig. 4,C demonstrate that, at least in this system, the TNF-α-induced PLD activation is independent of MAPK activity, as both inhibitors failed to moderate its activity. The inhibitors were shown to be working properly as both of them inhibited the TNF-α-induced phosphorylation of their target proteins (Fig. 4, A and B).
PLD1 is required for TNF-α-triggered cytokine generation
TNF-α is capable of amplifying the inflammatory response by promoting the generation and release of several proinflammatory cytokines and chemokines. In this study, we show that TNF-α stimulates IL-1β, IL-5, IL-6, and IL-13 production and release from human monocytes (Fig. 5), and that in cells pretreated with the antisense-PLD1 oligonucleotide, TNF-α-triggered cytokine release was inhibited (Fig. 5). In contrast, pretreatment with a PLD2 antisense had no inhibitory effect on cytokine production (Fig. 5), suggesting that PLD1 mediates the TNF-α-triggered cytokine production in human monocytes.
TNF-α activated the PLD1 pathway in primary human monocytes
All the previous experiments were conducted in a human monocytic cell line (U937 cells) differentiated to a mature monocyte phenotype with IFN-γ; we then decided to investigate whether TNF-α is also capable of activating the PLD1 pathway in primary human monocytes.
TNF-α stimulated PLD activity in a dose-dependent manner (Fig. 6,A), similar to that observed for U937 cells (Fig. 1,B). Both PLD1 and PLD2 are also present in primary monocytes and have a general cytosolic localization in resting cells; however, TNF-α triggers the re-localization of PLD1, but not PLD2, to the cells’ periphery (Fig. 6, B and C). We then pretreated the primary monocytes with antisense oligonucleotides against the two PLD isoforms. The levels of PLD1 can be reduced by 85 ± 5% (quantified by densitometry, from three separate experiments), and PLD2 had a reduction by 82 ± 8% (quantified by densitometry, from three separate experiments), (Fig. 6,D). Furthermore, controls were performed by the immunoprecipitation of ARF1, with a goat primary Ab, as control of immunoprecipitation and antisense specificity (Fig. 6 D).
Treatment of cells with the antisense oligonucleotide to PLD1 resulted in no change in basal PLD activity. However, following TNF-α stimulation, the increase in PLD activity was significantly reduced, compared with the control cells (Fig. 6,E). The reduction in the increase after TNF-α activation was 83 ± 7% in cells treated with antisense PLD1, compared with control cells, and was proportional to the observed reduction in protein expression by Western blot analysis. In contrast, treatment of cells with the antisense oligonucleotide to PLD2 only had an effect on the basal PLD activity (Fig. 6 E).
Furthermore, we show here that, as in the U937 cells, in primary human monocytes, TNF-α also couples PLD1 to the activation of SphK and to cytosolic calcium responses (Fig. 6, F and G). Moreover, in the primary human monocytes PLD1 was also essential for the TNF-α-mediated release of cytokines (Fig. 6 H).
Taken together, the data presented in this report strongly suggest that PLD1 is a key signaling molecule that mediates TNF-α-triggered proinflammatory responses in human monocytic cells.
In this study, we have shown that TNF-α is functionally coupled to PLD1, but not PLD2, in IFN-γ primed U937 cells, as well as in human primary monocytes, even though both enzymes are expressed in these cells. We further show that PLD1, and not PLD2, is required for TNF-α-mediated activation of the sphingosine kinase-mediated intracellular calcium responses, the activation of key transcription factors and cytokine generation.
Understanding the intracellular signal transduction mechanisms that regulate TNF-α-mediated responses has profound implications, not the least of which is to identify novel molecules as potential therapeutic targets. There is overwhelming evidence for believing that TNF-α is associated with a variety of inflammatory conditions in several diseases (32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47). Efforts are being made to find novel ways to balance TNF-α levels or TNF-α-induced responses in several diseases, including rheumatoid arthritis and systemic lupus erythematosus. Regulation of molecules involved in the TNF-α-triggered signaling pathways has gained attention in this regard, as a variety of signaling molecules such as PDE4, p38 MAP-kinase, and NFκB inhibitors are being studied in clinical trials (4), but have met with undesired side effects so far. More efficient therapies may become available by elucidation of the molecular mechanisms, and the role of key molecules, involved in the TNF-α-triggered signaling events. Thus, we investigated the role of PLD isoforms in the TNF-α-mediated intracellular and effector responses in human monocytes.
Our results show that TNF-α triggers PLD activity and that PLD1 (but not PLD2) is rapidly translocated from a cytosolic distribution to the plasma-membrane periphery. We and others have previously shown that PLD plays a role in inflammatory signals (30) and may stimulate MAPK phosphorylation events (48, 49, 50, 51), or prevent protein de-phosphorylation (51).
The cells used in our study expressed both PLD isoforms (PLD1 and PLD2) (21); however, this study shows a selective re-localization of PLD1 to the cell membrane after TNF-α stimulation. This would suggest that TNF-α induces the selective activation of PLD1. However, as there are many reports indicating the stimulation of isoform-specific PLDs in different cells, responding to distinct stimuli (21, 28, 29, 30, 31), we decided to further investigate the specific isoform activated by TNF-α. We used antisense oligonucleotides to specifically knockdown the expression of either PLD1 or PLD2. Our results using the antisense demonstrate that TNF-α specifically activates PLD1 but not PLD2.
To further our studies of the role of PLD1 in TNF-α-mediated intracellular signaling events, we looked at the role of PLD in signals triggered by TNF-α. We have previously shown that, in these cells, TNF-α activates and uses sphingosine kinase to mediate calcium release from internal stores (27). In this study, we show that in cells pretreated with the antisense against PLD1, the TNF-α-mediated SphK activity and cytosolic calcium responses are substantially inhibited, whereas in the cells pretreated with the antisense against PLD2, these TNF-α-mediated responses remain intact. This is further proof for the specificity of PLD1 in the TNF-α-mediated responses.
MAPKs are vital regulators and/or amplifiers of extracellular stimuli leading to cellular functions; they are regulated by sequential phosphorylation of their preceding kinase-family members (8). Our study shows that the TNF-α-stimulated ERK1/2 phosphorylation is dependent on PLD1; however, the phosphorylation of p38 is not. Interestingly, several studies have reported receptor-coupled activation of ERK1/2 to be not only independent of PLD activity, but that PLD activity was actually dependent on ERK1/2 activity (29). We show here that, at least in human monocytic cells, TNF-α signals in a very different way, i.e., TNF-α-stimulated PLD activity is upstream of ERK1/2 phosphorylation, whereas p38 phosphorylation is independent of PLD activity. This contrasts with an elegant report by Bechoua and Daniel (28), that showed that fMLP-triggered p38 activation in HL-60 cells was dependent on PLD activity, whereas ERK1/2 phosphorylation was not. To clarify these potential contradictions, and to establish the molecular specificity in our system, we used MEK and p38 inhibitors and looked at PLD activity triggered by TNF-α. Our results showed that neither inhibition of MEK nor p38 had any effect on TNF-α-mediated PLD activity; the inhibitors were shown to be working correctly as they inhibited their respective targets.
As the NFκB and ERK1/2 pathways appear to be dependent on PLD activity, we went on to measure the TNF-α-triggered proinflammatory cytokine production, and showed that the antisense to PLD1 substantially blocks the various cytokines measured.
Taken together, the data presented here suggest that, in human monocytes, TNF-α mobilizes intracellular calcium through the coupling of PLD1 to SphK1; so far this is the first example of this pathway to be shown in cytokine signaling. The fact that TNF-α needs to couple to PLD1 to trigger calcium release from internal stores may, at least in part, explain the role of PLD1 in TNF-α-mediated NFκB activation, as it has been shown that calcium amplitude and/or modulation is required for NFκB activation (52, 53, 54); moreover, it has also been shown that the product of SphK-activity (S1P) can activate NFκB in U937 cells (55). In this study, we also demonstrate that PLD1 is required for the phosphorylation/activation of ERK1/2 MAPKs. The immediate product of PLD1 is phosphatidic acid. Previous studies have shown that Raf-1 can be activated by phosphatidic acid (48, 49). Our finding that ERK1/2 is downstream of PLD is, therefore, consistent with this in vitro work (48, 49). Interestingly, it has long been established that members of the MAPK family can phosphorylate the IκB α kinase complex, which in turn leads to the activation of NFκB (56 , reviewed in Ref. 54).
TNF-α stimulation in myeloid cells triggers a number of effector functions, including the generation of several cytokines. The novel intracellular signaling pathway demonstrated here appears to be functionally interactive/associated with these. In the study reported here, silencing PLD1 reduced or abolished the ability of TNF-α to mobilize calcium from intracellular stores. In addition, the knockdown of PLD1 significantly reduced the activation of key transcription factors, such as NFκB and ERK1/2 MAPKs. It is of interest that these transcription factors play major roles in the inflammatory responses, by triggering cytokine and chemokine genes, as well as cyclooxygenases and other genes involved in the inflammatory responses (57). The finding that TNF-α-triggers the rise in cytosolic calcium and cytokine production, via a novel pathway that uses the sequential activation of PLD1 and SphK, has profound implications for the development of strategies for therapeutic intervention against differential myeloid responses to inflammation.
We thank A.-K. Fraser-Andrews for proofreading the manuscript.
The authors have no financial conflict of interest.
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 was supported by a grant from the Biomedical Research Council, Singapore.
Abbreviations used in this paper: PLD, phospholipase D; SphK, sphingosine kinase.