Capsaicin Inhibits Platelet-Activating Factor-Induced Cytosolic Ca²⁺ Rise and Superoxide Production¹

Platelat-activating factor (PAF)³ is one of several other important modulators of the inflammatory process (1). In neutrophils, PAF potently stimulates cell aggregation (2), the release of immunomodulators including histamine and leukotrienes (3), chemotaxis (4), and superoxide production (4, 5). In its course of action, PAF activates PAF-specific, G protein-coupled receptors (6). It also activates specific phospholipase C (PLC)-linked receptors and induces hydrolysis of phosphatidylinositol 4,5-bisphosphate and the production of inositol 1,4,5-trisphosphate (InsP₃) (7, 8). Its action thus leads subsequently to an increase in cytosolic Ca²⁺ and to superoxide production.

Because capsaicin transfers pain signals in afferent sensory neurons, it is used to alleviate symptoms such as pain and itching (9). In general, capsaicin acts on pain-linked neurons (10, 11), where its action is mediated by vanilloid receptors (receptors for capsaicin) that are expressed almost exclusively on primary sensory neurons (12). When capsaicin activates vanilloid receptors in afferent neurons, it stimulates the secretion of neurohormones, including substance P, tachykinin, calcitonin gene-related protein, and PAF into the space of peripheral nerve terminals (13, 14). The secreted neurohormones act as immunomodulators; they induce chemotaxis and superoxide production in immune cells in addition to vasodilation and increased vascular permeability. This process causes a unique type of inflammation, a so-called neurogenic inflammation (15, 16, 17). However, the prolonged presence of capsaicin leads to decreases in the secretion of neurohormones because the cells eventually become depleted of neurohormones. Up until now it was believed that the anti-inflammatory effect of capsaicin was based on desensitization (or death) of the afferent neurons that secreted the immunogenic neurohormones (18, 19, 20). However, several studies have observed that capsaicin could affect immune cells in the absence of afferent neurons. For example, it has been reported that capsaicin inhibited superoxide production (21) and the activation of NF-κB (22) in the absence of afferent neuron fibers. Ho et al. (23) reported that capsaicin regulated the expression of substance P and its receptor in monocytes without the involvement of afferent neurons. This suggested that there had to be another mechanism by which capsaicin acted on the immune cells directly, but such a mechanism was unknown.

Some immunomodulators, such as fMLP and PAF, mediate the activation of PLC in neutrophils or macrophages (24). We decided to investigate the effect of capsaicin on the inflammatory response induced by PAF. Finding capsaicin’s target in the PAF-induced signaling pathway, which is linked to PLC, might help us to better understand its anti-inflammatory action, because the increase in cytosolic Ca²⁺ and the activation of protein kinase C (PKC) subsequent to PLC activation are very important steps in the inflammatory process (4, 25).

Here we report that capsaicin inhibits the PAF-mediated Ca²⁺ rise and subsequent superoxide production by blocking the store-operated Ca²⁺ entry (SOCE) that follows PAF-induced PLC activation in HL-60 cells.

Capsaicin, PAF, ATP, fMLP, thapsigargin, SK&F96365, cytochrome c, and sulfinpyrazone were purchased from Sigma (St. Louis, MO). Fura-2 penta-acetoxymethyl ester (fura-2/AM) and 2′,7′-dichlorofluorescin diacetate (DCFH-DA) were obtained from Molecular Probes (Eugene, OR). [³H]Norepinephrine and [³H]InsP₃ were purchased from NEN (Boston, MA). RPMI 1640 and penicillin/streptomycin were obtained from Life Technologies (Grand Island, NY). Bovine calf serum and horse serum were obtained from HyClone (Logan, UT).

HL-60 cells were grown in RPMI 1640 supplemented with 10% (v/v) heat-inactivated bovine calf serum, 5% (v/v) heat-inactivated horse serum, and 1% (v/v) penicillin/streptomycin. The culture medium was changed daily. All cells were cultured in a humidified atmosphere of 95% air and 5% CO₂. We induced differentiation of the HL-60 cells by incubating them in 1 μM all-trans-retinoic acid for 5 days (26). We observed the morphology of the cells and monitored the fMLP-induced cytosolic calcium ion concentration ([Ca²⁺]_i) rise, which is only detectable in differentiated HL-60, as indicators of differentiation into neutrophil-like cells. We counted viable cells by the trypan blue exclusion method.

Superoxide generation was determined based on the change in absorbance of cytochrome c using a previously published procedure with slight modification (27). Briefly, 5 × 10⁶ cells were washed, resuspended with Locke’s solution (154 mM NaCl, 5.6 mM KCl, 1.2 mM MgCl₂, 2.2 mM CaCl₂, 5 mM HEPES, and 10 mM glucose, pH 7.3), and placed into a cuvette, and 40 μM cytochrome c was added. After a 1-min incubation, stimulants were added, and the change in absorbance at 550 nm was monitored. Superoxide dismutase was used as the control, setting the maximal value of the superoxide-mediated absorbance change. Calibration of the change in absorbance in terms of superoxide production was performed using the following equation: [superoxide] = ΔA v/t/K/l/cells, where ΔA is the change in absorbance, v is the reaction volume, t is the time, K is the extinction coefficient for the difference between the light absorption of reduced cytochrome c and that of oxidized cytochrome c (21 × 10³ cm⁻¹ M⁻¹), and l is the length of the cuvette.

The production of intracellular superoxides was determined based on the changes in fluorescence of DCFH-DA, an oxidation-sensitive fluorescence probe, with a slight modification of a previously published procedure (28, 29). Briefly, the cell suspension was incubated in fresh serum-free RPMI 1640 medium with 2 μM DCFH-DA at 37°C for 40 min under continuous stirring. The loaded cells were then washed twice with Locke’s solution. Then 2 × 10⁶ cells were placed into a cuvette and in a thermostatically controlled cell holder at 37°C and continuously stirred. Fluorescence was excited at 488 nm, and emission was recorded at 530 nm. The change in fluorescence intensity was monitored.

[Ca²⁺]_i was determined using the fluorescent Ca²⁺ indicator fura-2 as reported previously (30). Briefly, the cell suspension was incubated in fresh serum-free RPMI 1640 medium with 3 μM fura-2/AM at 37°C for 60 min with continuous stirring. The loaded cells were then washed twice with Locke’s solution. Sulfinpyrazone (250 μM) was added to all solutions to prevent dye leakage. For the fluorometric measurement of [Ca²⁺]_i, 1 × 10⁶ cells/ml were placed into a quartz cuvette in a thermostatically controlled cell holder at 37°C and continuously stirred. Fluorescence ratios were monitored, with dual excitation at 340 and 380 nm and emission at 500 nm. Calibration of the fluorescent signal in terms of [Ca²⁺]_i was performed as described by Grynkiewicz et al. (31) using the following equation: [Ca²⁺]_i = K_d[R − R_min)/(R_max − R)](S_f2/S_b2), where R is the ratio of fluorescence emitted by excitation at 340 and 380 nm. S_f2 and S_b2 are the proportionality coefficients at 380 nm excitation of Ca²⁺-free fura-2 and Ca²⁺-saturated fura-2, respectively. To obtain R_min, the fluorescence ratios of the cell suspension were measured successively at final concentrations of 4 mM EGTA, 30 mM Trizma base, and 0.1% Triton X-100. The cell suspension was then treated with CaCl₂ at a final concentration of 4 mM Ca²⁺, and the fluorescence ratios were measured to obtain the R_max.

The Mn²⁺ quenching assay was performed as described by Lee et al. (32) to measure the influx of Ca²⁺ from the extracellular space. Briefly, fura-2-loaded cells (5 × 10⁶ cells/ml; described above) were placed into a quartz cuvette in a thermostatically controlled cell holder at 37°C under continuous stirring. Fluorescence was excited at 360 nm, i.e., the isosbestic wavelength at which Ca²⁺ does not affect fura-2 fluorescence and at which, therefore, changes are caused by Mn²⁺ quenching. Emission was recorded at 500 nm. The potency and slope of the change in fluorescence intensity were recorded after applying 2 mM MnCl₂ and the drugs to be tested.

InsP3 mobilization was determined by competition assay of [³H]InsP₃ for binding protein as described previously (33). To determine InsP₃ production, 1 × 10⁶ cells were stimulated with the drugs to be tested. The reaction was terminated by the addition of ice-cold 5% TCA containing 10 mM EGTA. The supernatant of the lysate was then saved and extracted with diethyl ether to remove TCA. The aqueous fraction after a final extraction was neutralized with 200 mM Trizma base to adjust it to pH 7.4. Twenty milliliters of extract was added to 20 ml of assay buffer (0.1 M Tris buffer containing 4 mM EDTA) and 20 ml of [³H]InsP₃ (100 nCi/ml). The mixture was incubated for 15 min on ice and then centrifuged at 2000 × g for 10 min. Water (100 ml) and 1 ml of liquid scintillation cocktail were added to the pellet to measure the radioactivity. The InsP₃ concentration of the sample was determined by comparison to a standard curve and expressed as picomoles per milligram of protein. The total cellular protein concentration was measured using the Bradford method after sonication of 1 × 10⁶ cells.

All quantitative data are expressed as the mean ± SEM. We calculated the IC₅₀ with the Microcal Origin for Windows program. Differences were considered significant only for p < 0.05.

We studied the effect of capsaicin on PAF-induced superoxide production and the increase in [Ca²⁺]_i in human promyelocyte HL-60 cells. HL-60 cells have served as a good model in which to study signal transduction of various receptors involved in the inflammatory processes of leukocytes (34). HL-60 cells express the PAF receptor, and its level of expression increases during granulocytic differentiation (35). As shown in Fig. 1,A, PAF triggered differentiated HL-60 cells to secrete superoxide into the extracellular space. Under the above conditions, addition of capsaicin attenuated the production of superoxide in a concentration-dependent manner. It has been reported that PAF activates PLC and increases cytosolic Ca²⁺ levels. We, therefore, tested the effect of capsaicin on the responses mediated by other PLC-coupled receptors in HL-60 cells. Capsaicin also inhibited superoxide production induced by fMLP (Fig. 1,B) and extracellular ATP (Fig. 1,C) with a similar inhibitory potency as that seen in the PAF response. However, PMA-induced superoxide production was not affected by capsaicin (Fig. 1 D). The results suggest that capsaicin acts on the Ca²⁺ response in the PLC signaling pathway and not on PKC.

It has been reported that the incubation of HL-60 cells with >100 μM capsaicin induces changes in the cytoskeleton and apoptosis (36). When we challenged cells with 300 μM capsaicin, we observed a slow, but prolonged, Ca²⁺ rise, even though the increase was small, i.e., <100 nM (data not shown). The capsaicin-induced elevation of cytosolic Ca²⁺ under the above conditions resulted from influx of Ca²⁺ from the extracellular space and the release of Ca²⁺ from the intracellular Ca²⁺ pool. We suggest that this is not due to the activation of vanilloid receptors but, rather, to plasma and microsomal membrane perturbations that may occur and, upon prolonged Ca²⁺ influx, drive the cell to apoptosis. Up to 100 μM capsaicin by itself did not trigger any cytosolic Ca²⁺ rise.

When we added capsaicin, PAF-induced increases in cytosolic Ca²⁺ were inhibited in undifferentiated HL-60 cells (Fig. 2,A). The inhibition was more obvious in the sustained Ca²⁺ level rather than the peak level. Capsaicin’s inhibitory effect disappeared in the absence of extracellular Ca²⁺ using Ca²⁺-free Locke’s solution (156.2 mM NaCl, 5.6 mM KCl, 1.2 mM MgCl₂, 5 mM HEPES, and 10 mM glucose, pH 7.3), but became prominent again when extracellular 2.2 mM Ca²⁺ was reintroduced (Fig. 2 B). The results, therefore, suggest that capsaicin inhibits SOCE-mediated by PAF. The Ca²⁺ responses to PAF were more dramatic in granulocytic HL-60 cells differentiated by incubation with 1 μM all-trans-retinoic acid for 5 days, but the characteristics of the signal transduction were same (data not shown).

When cytosolic Ca²⁺ was elevated by the activation of other PLC-coupled receptors with fMLP or extracellular ATP, capsaicin inhibited the Ca²⁺ increase in a similar manner as the PAF-induced Ca²⁺ increase (data not shown), suggesting that capsaicin acted as a PLC inhibitor. We therefore tested whether capsaicin directly inhibited PLC by measuring the capsaicin effect on InsP₃ production induced by PAF, fMLP, and extracellular ATP. Capsaicin did not inhibit InsP₃ production induced by the above agents in undifferentiated (Fig. 3,A) and differentiated granulocytic HL-60 cells (Fig. 3 B). The results, therefore, reveal that capsaicin is not a PLC inhibitor.

Thapsigargin inhibits microsomal Ca²⁺-ATPase, depletes intracellular Ca²⁺ pools, and induces Ca²⁺ influx of the SOCE kind. Capsaicin inhibited the thapsigargin-induced increase in cytosolic Ca²⁺ (Fig. 4, A and B) in a concentration-dependent manner with an IC₅₀ of 24.8 ± 2.4 μM (Fig. 4,C). To confirm the inhibitory effect of capsaicin on SOCE, we tested the thapsigargin-evoked influx of Ba²⁺ and Mn²⁺ ions that are experimentally added to the extracellular space to monitor the influx of Ca²⁺ separate from the store-operated release. Capsaicin inhibited the fluorescence changes induced by the influx of Ba²⁺ (Fig. 5,A). Capsaicin also decreased the rate of fluorescence quenching caused by the binding of cytosolic fura-2 to Mn²⁺ entering from the extracellular space (Fig. 5 B). The data thus consistently indicated that the target site of capsaicin was the Ca²⁺ influx through store-operated channels.

To further assess the target site of capsaicin, we compared the actions of capsaicin and SK&F96365, which is a known antagonist of SOCE (37). Fig. 6 shows that SK&F96365 decreased the thapsigargin-induced elevation of [Ca²⁺]_i and that the successive addition of capsaicin did not add to the inhibition of the sustained Ca²⁺ level (Fig. 6,A) and vice versa (Fig. 6 B). The results suggest that capsaicin and SK&F96365 possibly act on the same site.

To study the capsaicin effect on inflammatory reactions, we monitored the effect of SOCE on superoxide formation in granulocytic differentiated HL-60 cells. PAF prominently evoked the formation of cytosolic superoxide; however, removal of extracellular Ca²⁺ dramatically reduced the production of cytosolic superoxide (Fig. 7,A) as previously reported by Gallois et al. (38). This suggests that external Ca²⁺ entry significantly contributes to the formation of superoxides. When cells were treated with capsaicin, the PAF-induced cytosolic superoxide production was clearly diminished (Fig. 7 B). SK&F96365 decreased the PAF-induced superoxide formation just a little more potently than capsaicin. The results thus demonstrate that SK&F96365 and capsaicin exert the same effect on superoxide production.

There still remained the possibility that capsaicin acted on vanilloid receptors. We therefore tested capsaicin in combination with an agonist and an antagonist of vanilloid receptors. Resiniferatoxin is a known vanilloid agonist that is 100 times more potent than capsaicin (39). When cells were treated with 1 μM resiniferatoxin, which is a concentration 100 times higher than that generally used, an inhibitory effect on the thapsigargin-induced SOCE was not detectable (Fig. 8,A). It has been reported that ruthenium red and capsazepine act on the vanilloid receptor as antagonists (40). Pretreatment of cells with 10 μM ruthenium red did not block the capsaicin-induced inhibition of the thapsigargin-induced SOCE (Fig. 8,B). Finally, capsazepine had a similar effect on the inhibition of the thapsigargin-induced SOCE as capsaicin (Fig. 8 C).

In the present study we demonstrate that capsaicin inhibits PAF-mediated superoxide production. Many inflammatory signals, including PAF, trigger receptor-mediated PLC activation as part of their signaling mechanism. Our experiments with fMLP and extracellular ATP, which activate PLC-coupled receptors (Fig. 1), suggested that the capsaicin-mediated inhibition of PAF-induced superoxide production was mediated by blockage of a common site in the PLC signaling pathway. Activation of PLC leads to an increase in intracellular Ca²⁺ and activates PKC. It is generally accepted that the Ca²⁺ increase and PKC activation are synergistically involved in superoxide production (41). Interestingly, capsaicin did not inhibit PMA-induced superoxide production. This indicates that capsaicin does not directly inhibit PKC.

PLC-mediated cytosolic Ca²⁺ elevation is achieved by Ca²⁺ release from InsP₃-sensitive stores and subsequent Ca²⁺ influx from the extracellular space, the so-called SOCE, which is activated by the depletion of intracellular Ca²⁺ stores (42). SOCE is thought to be a major regulator of immune responses, including O₂⁻ production in granulocytic differentiated HL-60 cells (37) and neutrophils (43), IL-8 release in neutrophils (44), histamine release in mast cells (45), and platelet aggregation (46). It has been reported that the PAF-induced priming of neutrophils requires Ca²⁺ influx (47), which has been found to be SOCE (24). Our results reveal that capsaicin inhibited Ca²⁺-sensitive superoxide production (Figs. 1 and 7) and Ca²⁺ influx, which is activated subsequent to the depletion of Ca²⁺ stores (Figs. 4 and 5) and is affected by SK&F 96365, a SOCE inhibitor (Fig. 6), whereas capsaicin did not directly inhibit PLC or the InsP₃-sensitive Ca²⁺ release (Figs. 2 and 3). We thus can conclude that capsaicin-induced inhibition of SOCE will result in a reduction of PAF-induced superoxide formation in HL-60 cells.

In this report we demonstrate that capsaicin directly acts on immune cells attenuating their inflammatory responses in addition to their effect on afferent neurons. There was evidence presented in a previous report that capsaicin could inhibit the production of superoxides by macrophages in the absence of afferent neuron fibers (21). Capsaicin had been thought to exclusively have an effect on the desensitization of neurogenic inflammation. However, we carefully suggest the possibility that the direct inhibition of superoxide formation by capsaicin could be another important aspect in the alleviation of inflammation. The capsaicin concentrations used for the blockage of neurogenic inflammation are 25–100 mg/kg (17, 18, 19), which roughly equals 1–5 mM and therefore could be enough to directly block the inflammatory action of immune cells.

It is known that vanilloid receptors are exclusively expressed on afferent neurons. Although vanilloid receptors were detected on murine mast cells (48), there is no evidence of expression of vanilloid receptors on other immune cells such as monocytes, macrophages, or neutrophils. The capsaicin-induced effects could be classified into two different categories: vanilloid receptor-mediated responses and nonvanilloid type responses. Our findings suggest that the capsaicin effect is of the nonvanilloid type because of 1) its high effective concentration needed for the effect (Fig. 1), 2) the lack of antagonistic effect of a classical vanilloid agonist, such as resiniferatoxin (Fig. 8,A), and 3) no detectable antagonistic effect of classical vanilloid antagonists, such as capsazepine and ruthenium red (Fig. 8, B and C). Interestingly, capsazepine also inhibited SOCE just as capsaicin. Our results correlate with other studies that have seen effective concentrations of capsaicin in the micromolar range (17, 18), whereas vanilloid receptor can be activated with concentrations in the nanomolar range. Our results also agree with reports of some capsaicin-mediated effects not correlating with typical features of the unusual responses of vanilloid receptors to vanilloid antagonists (49).

We studied the inhibitory effect of capsaicin on PAF-induced superoxide formation. Because PAF is a potent inducer of inflammation, a PAF antagonist could be a promising agent for therapeutic applications that treat inflammation, although clinically available drugs are still at the developmental stage. Our results provide a first clue toward understanding of the capsaicin anti-inflammatory effect as it inhibits PAF-mediated reactions.

We thank G. Hoschek for editing this manuscript.