Bacterial Challenge Stimulates Formation of Arachidonic Acid Metabolites by Human Keratinocytes and Neutrophils In Vitro

Abstract

Although the interactions of bacteria with keratinocytes induce the synthesis of various mediators, the capability of epithelial cells to form arachidonic acid mediators has not been studied, and therefore the first part of this study was initiated. The complex mixture of epithelium-derived mediators suggests that chemoattraction is not their only effect on neutrophils and that they may also affect neutrophil mediator synthesis. The effect of epithelium-derived mediators on neutrophil eicosanoide synthesis was evaluated in the second part of this study. We incubated human keratinocytes with human-pathogenic bacteria for 2 h and harvested the supernatants after 4, 6, 10, and 18 h of culture. Subsequently, the supernatants were coincubated for 5 min with human neutrophils with or without arachidonic acid. The formation of the arachidonic acid metabolites prostaglandin E₂ (PGE₂), leukotriene B₄ (LTB₄), 12-hydroxyeicosatetraenoic acid (12-HETE), and 15-HETE in keratinocytes and neutrophils was measured by reverse-phase high-pressure liquid chromatography. We demonstrated for the first time that keratinocytes produced significant amounts of LTB₄ and 12-HETE 4 to 6 h after bacterial challenge. Upon stimulation with epithelial supernatants, neutrophils produced significant amounts of PGE₂, LTB₄, 12-HETE, and 15-HETE throughout the observation period of 18 h, with a maximum synthesis by supernatants harvested 4 to 10 h after bacterial infection. The results of the study suggest that arachidonic acid mediator formation by epithelial cells following bacterial challenge may act as an early inflammatory signal for the initiation of the immune response. The epithelial supernatants were capable of inducing the formation of arachidonic acid mediators by neutrophils, which may have further regulatory effects on the immune response.

Although the mechanisms controlling inflammation are poorly understood, the release of pro- and anti-inflammatory mediators plays a central role in the initiation, perpetuation, and limitation of this process. Bacterial infection is an important stimulus for the synthesis of inflammatory mediators in human tissues. Epithelial cells are the first cell line that interacts with bacteria and causes the increased expression and secretion of a number of cytokines and proinflammatory mediators (24). These mediators are characterized by their ability to chemoattract and activate polymorphonuclear leukocytes (PMNs), suggesting that an important function of epithelial cells is to initiate the mucosal influx of PMNs. The arachidonic acid metabolites leukotriene B₄ (LTB₄), 12-hydroxyeicosatetraenoic acid (12-HETE), and 15-HETE regulate PMN functions, including chemotaxis (6, 17, 33, 42), degranulation (13, 38, 41), and adherence (7), which may indicate the critical role of these metabolites for early events in inflammation.

The synthesis of various inflammatory mediators by keratinocytes upon bacterial infection suggests that the induction of chemokinesis is not the only functional change that is induced in neutrophils. The capacity of neutrophils to synthesize arachidonic acid mediators upon stimulation is well documented (15, 16, 27, 37) and may have numerous effects on the propagation of inflammation because of their impact on various cells of the inflammatory reaction (18, 45). In the first part of this study we wished to assess the capacity of keratinocytes to synthesize arachidonic acid metabolites upon bacterial challenge. In the second part we investigated the synthesis of arachidonic acid metabolites in neutrophils upon induction by metabolites formed by keratinocytes after bacterial stimulation. These experiments address the events that occur during early stages of inflammation, from the interactions of bacteria with epithelial cells to the consequent effect on mediator synthesis by neutrophils with special respect to arachidonic acid mediators.

MATERIALS AND METHODS

Bacterial growth.

Actinobacillus actinomycetemcomitans MCCM 01881 and MCCM 03162 were used as the infectious agents for the experiments. This invasive microorganism was used because of its well-documented capacity to induce an cytokine response in epithelial cells (21). A. actinomycetemcomitans strains were grown in Trypticase soy broth containing 10% yeast extract. The bacteria were incubated at 37°C in a CO₂-enriched environment for 48 to 72 h until an optical density of 0.2 at 495 nm was reached. The bacteria were centrifuged at 1,500 × g for 15 min, washed in phosphate-buffered saline (PBS) buffer, and finally suspended in the antibiotic-free medium used to grow the epithelial cell line to reach a multiplicity of infection of 1,000:1, which had recently been determined as a ratio for maximum stimulation of epithelial cells (21).

Epithelial growth.

Human immortalized keratinocytes were routinely cultured (passages 30 to 90) on Falcon plastic dishes in Dulbecco’s modified Eagle medium (DMEM)-10% fetal calf serum-glucose (4.5 g/liter)-MEM nonessential amino acids-penicillin (100 U/ml)-streptomycin (100 μg/ml) and subcultured every 7 to 10 days. These cells are a nontransformed human keratinocyte line derived from skin and maintain a substantial differentiation potential in culture (2). Monolayers of human keratinocytes were treated with trypsin (0.2% in 0.05% EDTA) and washed in magnesium- and calcium-free PBS, and cells were seeded into 24-well Costar tissue plates at a density of 10⁵ in volumes of 0.5 ml. Cells were grown in a 5% CO₂ atmosphere at 37°C to confluent monolayers before infection with bacteria.

Infection of epithelial cells.

The cell growth medium was removed from the confluent epithelial cells, and 200 μl of DMEM with or without bacteria was added. Bacteria were cocultured with epithelial cells at 37°C for 2 h to allow interactions between bacteria and epithelial cells. After incubation the supernatant was carefully removed. The monolayers were washed three times with 500 μl of PBS to remove all extracellular bacteria. The cultures were further incubated for 4, 6, 10, and 18 h with 500 μl of DMEM supplemented with 0.1 mg of gentamicin per ml to kill the remaining extracellular bacteria. The supernatants were carefully removed after individual time points and stored at −50°C for the following procedures.

Isolation of human PMNs.

Blood was obtained from fasting healthy human volunteers, who had not taken nonsteroidal anti-inflammatory drugs during the previous 2 weeks. Five milliliters of anticoagulant stock solution (7.5 g of trisodium hydrate, 2.0 g of glucose, and 66 μl of heparin in a volume of 100 ml of 0.9% NaCl) and dextran stock solution (6 g of dextran in 100 ml of 0.9% NaCl) were added. After 45 min the mixture was centrifuged and the pellet was resuspended in 10 ml of washing buffer (0.274 g of EDTA in 1,000 ml of 0.9% NaCl) and covered with a layer of Ficoll (Biocoll separation solution; Biochrom). The mixture was centrifuged (45 min), and the pellet was resuspended in 10 ml of washing buffer. The erythrocytes were hemolyzed by the addition of ice-cold distilled water followed by the addition of 3 ml of NaCl solution (2.7 g of NaCl and 0.11 g of EDTA in 100 ml of 0.9% NaCl). The samples were centrifuged again, and the pellet was resuspended in wash buffer. The purity of cells was indicated by morphological criteria. After isolation the cells were suspended at a concentration of 10⁷/ml and were kept at 4°C until analysis.

Stimulation of neutrophils.

Neutrophils (500 μl, 10⁷ cells/ml) were preincubated for 10 min at 37°C with or without exogenous arachidonic acid (Sigma-Aldrich, Steinheim, Germany) at a concentration of 50 μM. Subsequently, 420 μl of the epithelial supernatants was added together with 20 μl of MgCl₂ (final concentration of 1.9 mM) and 20 μl of CaCl₂ (final concentration of 3.6 mM). The ionophore A23187 (Sigma-Aldrich) was added in controls at a final concentration of 25μM in 1% ethanol. A23187 causes an increase in cytoplasmatic Ca²⁺, leading to an activation of Ca²⁺-dependent phospholipase and resulting in the liberation of arachidonic acid from the cell membrane (1). Incubations were carried out for 5 min at 37°C in a shaking water bath. Incubations were terminated by cooling the reaction mixtures to 0°C in an ice water bath.

RP-HPLC analysis of lipid mediators.

Eicosanoids and prostanoids were extracted from cell supernatants, separated, and quantified by reverse-phase high-pressure liquid chromatography (RP-HPLC) as described by Eberhard et al. (10). The reaction mixtures were centrifuged at 1,500 × g at 0°C for 10 min. The lipids were extracted by the addition of 200 μl of ice-cold methanol and stored for 10 min at −20°C. The reaction mixtures were centrifuged at 1,500 × g at 0°C for 10 min, and 200 μl of the supernatants was acidified with trifluoracetic acid (TFA) and loaded onto an RP-HPLC system as reported previously. Briefly, the HPLC consisted of a tertiary gradient system equipped with a photodiode array detector (Gynkotek, Germering, Germany). Chromatography was performed at room temperature, and the flow rate was 0.5 ml/min. RP-HPLC was carried out on an analytical column (125.0 by 4.6 mm) of HyPurity Elite C₁₈ (5-μm particle size). A guard column (20 by 4.5 mm) of the same material, HyPurity Elite-HCT Guards C₁₈ (5-μm particle size; ThermoHypersil; Cheshire, England) was used. Two continually degassed solvents were used: (i) 0.002% (vol/vol) TFA in 25% water and 75% acetonitrile (vol/vol) and (ii) 0.002% (vol/vol) TFA in 7.5% water, 38.5% acetonitrile, and 54% methanol (vol/vol). The eluate was monitored for prostaglandin E₂ (PGE₂) at 220 nm, for LTB₄ at 280 nm, and for 15(S)- and 12(R)-HETE at 237-nm absorption. For calibration, standards of LTB₄, PGE₂, 12(R)-HETE, and 15(S)-HETE were diluted in 0.002% (vol/vol) TFA in 25% water and 75% acetonitrile and analyzed by RP-HPLC. Their purity was verified by HPLC (>95%, dry weight). We analyzed serial dilutions of PGE₂ between 1 and 300 ng and of LTB₄, 12-HETE, and 15-HETE between 1 and 250 pg as external standards to demonstrate their linear responses and recoveries.

Statistical methods.

Each experiment was performed at least with five repeats, and data were read twice. The median and interquartile range were computed. All analyses were computed with the program SPSS 9.0 (SPSS Inc., Chicago, Ill.). Normal distribution of data was tested by the Kolmogorov-Smirnov test. Data were analyzed by Friedman analysis of variance on ranks followed by a multiple pairwise comparison procedure. The significance level was set at 0.05 (P < 0.05).

RESULTS

Arachidonic acid mediator synthesis by human keratinocytes after bacterial challenge.

Figure 1 shows a typical chromatogram of epithelial supernatants analyzed by RP-HPLC. We did not find significant differences in mediator synthesis between the two bacterial strains and in consequence did not analyze the data separately. Table 1 summarizes the eicosanoid and prostanoid synthesis of keratinocytes after bacterial challenge for 2 h. We did not detect PGE₂ in the supernatants at any time point. LTB₄, 12-HETE, and 15-HETE were detected as products of keratinocytes, but only LTB₄ and 12-HETE formation increased statistically significantly in relation to that in unstimulated cells. The levels of LTB₄ and 12-HETE increased to a maximum 4 to 6 h after bacterial stimulation and decreased until the end of observation at 18 h.

FIG. 1.

(A) Typical chromatogram produced by RP-HPLC analysis of epithelial supernatants recorded at 220, 237, and 280 nm. (B) The metabolite was identified by the retention time of controls. Metabolites and retention times are indicated.

TABLE 1.

Synthesis of eicosanoids by human immortalized keratinocytes after bacterial challenge for 2 h

Metabolite	Median concn (ng/ml)a
	In unstimulated control	At time (h) after bacterial challenge
		4	6	10	18
PGE2	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)
LTB4	0.00 (0.04)	0.05 (0.03)*	0.07 (0.06)*	0.04 (0.03)	0.05 (0.03)
12-HETE	0.00 (0.00)	0.06 (0.15)*	0.13 (0.26)*	0.00 (0.01)	0.02 (0.07)
15-HETE	0.00 (0.02)	0.00 (0.00)	0.03 (0.05)	0.04 (0.05)	0.02 (0.04)

^aValues in parentheses are interquartile ranges. *, statistically significantly different from the value for unstimulated epithelial cells (Friedman analysis of variance on ranks followed by multiple pairwise comparisons; P < 0.05, n = 5).

Arachidonic acid mediator synthesis by neutrophils.

The stimulation of neutrophils with epithelial supernatants harvested directly after bacterial infection (time zero) and 4, 6, 10, and 18 h later resulted in significant increased levels of PGE₂, LTB₄, 12-HETE, and 15-HETE compared to neutrophils stimulated with epithelial supernatants that were not challenged by bacteria (Table 2). The slope for the concentration of all four arachidonic acid metabolites was characterized by a rapid increase directly after stimulation of epithelial cells by bacteria. The mediator formation reached maximum intensity 4 to 10 h after bacterial challenge for PGE₂, LTB₄, 12-HETE, and 15-HETE. Eighteen hours after bacterial challenge, the levels of PGE₂, LTB₄, 12-HETE, and 15-HETE had not returned to baseline levels. The presence of exogenous arachidonic acid had no significant influence on the PGE₂, LTB₄, 12-HETE, and 15-HETE levels during coincubation of epithelial supernatants with neutrophils (Table 2). The stimulation of neutrophils with the calcium ionophore A23187 induced the formation of PGE₂ (median [interquartile range]; 80.37 [2.44] ng/10⁷ cells), LTB₄ (1.93 [1.44] ng/10⁷ cells), 12-HETE (0.59 [1.18] ng/10⁷ cells), and 15-HETE (2.00 [1.28] ng/10⁷ cells). The formation of LTB₄ and 15-HETE significantly increased (5.04 [1.70] ng/10⁷ cells and 5.35 [1.83] ng/10⁷ cells, respectively) in the presence of exogenous arachidonic acid, while the formation of PGE₂ and 12-HETE were not significantly different in the presence of arachidonic acid (Table 3).

TABLE 2.

Synthesis of eicosanoids by PMNs after stimulation with epithelial supernatants

Metabolitea	Median concn (ng/ml)b
	In unstimulated control	At time (h) after bacterial challenge
		0	4	6	10	18
Without AA
PGE2	0.00 (4.04)	90.46 (105.18)	110.97 (41.24)*	159.59 (80.42)*	99.00 (55.12)	128.16 (29.04)*
LTB4	2.49 (0.57)	13.35 (1.54)	2.51 (3.25)	4.32 (1.08)*	4.00 (3.59)*	3.15 (3.56)
12-HETE	0.00 (0.60)	1.67 (1.50)	1.56 (0.73)	2.65 (1.63)*	1.62 (1.89)	0.00 (1.21)
15-HETE	4.28 (1.18)	7.87 (2.10)	5.73 (2.74)	7.24 (2.79)	5.82 (2.49)	6.86 (1.84)
With AA
PGE2	0.00 (0.00)	11.53 (19.52)*	254.04 (149.27)*	113.85 (25.47)*	160.75 (100.67)*	12.42 (47.17)*
LTB4	2.08 (1.09)	2.66 (1.32)	2.85 (2.16)	2.94 (1.51)*	3.45 (2.98)*	2.54 (1.05)
12-HETE	0.00 (1.84)	2.52 (0.71)*	2.43 (1.41)*	0.00 (0.69)	1.84 (1.61)	2.06 (2.49)
15-HETE	3.30 (0.62)	5.33 (2.17)	7.15 (3.86)*	5.82 (1.82)*	7.46 (1.33)*	5.02 (2.62)

^aAA, arachidonic acid.

^bValues in parentheses are interquartile ranges. *, statistically significantly different from value for unstimulated polymorphnuclear cells (Friedman analysis of variance on ranks followed by multiple pairwise comparisons; P < 0.05, n = 5).

TABLE 3.

Synthesis of eicosanoids by PMNs after stimulation with A23187 or epithelial supernatants

Metabolitea	Median concn (ng/ml) after stimulation withb:
	A23187	Epithelial supernatantc
Without AA
PGE2	80.37 (2.44)	159.59 (80.42)*
LTB4	1.93 (1.44)	4.32 (1.08)
12-HETE	0.59 (1.18)	2.65 (1.63)*
15-HETE	2.00 (1.28)	7.87 (2.10)*
With AA
PGE2	60.70 (102.71)	254.04 (149.27)*
LTB4	5.04 (1.70)**	2.94 (1.51)*
12-HETE	0.00 (1.03)	2.52 (0.71)*
15-HETE	5.35 (1.83)**	7.46 (1.33)

^aAA, arachidonic acid.

^bValues in parentheses are interquartile ranges. *, statistically significantly different from the value for A23187-stimulated PMNs (Friedman analysis of variance on ranks followed by multiple pairwise comparisons; P < 0.05, n = 5). **, statistically significantly different from the value for stimulation without arachidonic acid (Friedman analysis of variance on ranks followed by multiple pairwise comparisons; P < 0.05, n = 5).

^cValues are maximum observed stimulation.

The formation of PGE₂ and 12-HETE in neutrophils was significantly increased by stimulation with epithelial supernatants compared to A23187 with or without arachidonic acid (Table 3). A significantly reduced stimulation of LTB₄ formation was observed with epithelial supernatants in contrast to A23187 in the presence of arachidonic acid. The synthesis of 15-HETE was not different after stimulation with A23187 or epithelial supernatants in the presence of arachidonic acid, while without arachidonic acid the stimulation with epithelial supernatants resulted in the synthesis of significantly more 15-HETE than those with stimulation with A23187.

DISCUSSION

We incubated strains of the human pathogen A. actinomycetemcomitans with keratinocytes to induce epithelial mediator synthesis and measured PGE₂, LTB₄, 12-HETE, and 15-HETE formation by RP-HPLC. We did not find significant differences in mediator synthesis between the laboratory strain and the clinical isolate. It may be that we were not able to detect differences due to the analytical system we used or that the different stimulation did not affect the formation of arachidonic acid metabolites by epithelial cells. We have shown that the stimulation of epithelial cells by human pathogenic bacteria leads to the synthesis of arachidonic acid metabolites LTB₄ and 12-HETE. An increasing number of studies reported the release of cytokines and proinflammatory mediators by epithelial cells upon stimulation with bacteria (8, 11, 23, 28, 36). The present study is to our knowledge the first report that demonstrates the synthesis of arachidonic acid metabolites by epithelial cells after bacterial stimulation. This study shows that keratinocytes have the capacity to form LTB₄ and 12-HETE, which is in agreement with other studies using cell cultures (22, 39), although other experiments failed to show 5-lipoxygenase activities in keratinocytes (3). In contrast to canine airway epithelial cells (29) and cultured human bronchial epithelial cells (35), which release PGE₂ after electrical and phorbol myristate acetate stimulation, respectively, we did not observe PGE₂ formation after stimulation by bacteria. This observation suggests that a bacterial stimulus does not trigger PGE₂ synthesis in keratinocytes. The formation of LTB₄ and 12-HETE significantly increased with a maximum 4 to 6 h after bacterial challenge and decreased to baseline levels after 18 h. This time course was in agreement with similar experimental approaches that showed a predominant increase of mediator synthesis during the first 5 to 10 h after bacterial stimulation (11, 21). The simultaneous nature of increased levels of the potent chemoattractant mediators LTB₄ and 12-HETE after bacterial stimulation suggests that they may be important stimuli for the neutrophil infiltration at the site of bacterial-epithelial interaction. This is in agreement with animal experiments that demonstrated the participation of LTB₄ in neutrophil influx (4, 5).

In the second part of our study we observed the increased formation of PGE₂, LTB₄, 12-HETE, and 15-HETE by human neutrophils after stimulation with epithelial supernatants, which were synthesized in response to a bacterial challenge. With respect to the volumes incubated during neutrophil coincubation, the portions of PGE₂, LTB₄, 12-HETE, and 15-HETE formed in keratinocytes were low compared to the total amount measured after neutrophil coincubations. In contrast to keratinocytes, neutrophils produced large amounts of PGE₂ upon stimulation, which is in agreement with other in vitro studies (12, 32, 47). The immunohistochemistry of inflamed tissues demonstrated cyclooxygenase-2 activity mainly within infiltrating neutrophils and emphasizes that these cells are an important source for PGE₂ in vivo as well (27). The generation of LTB₄ by neutrophils is a well-documented mechanism that is induced by mediators such as granulocyte-macrophage colony-stimulating factor (40), interleukin 8 (IL-8) (37), and IL-4 and IL-13 (48). Although the capability of human neutrophils to synthesize 12-HETE has been reported (16), the description of mediators that induce 12-HETE generation is limited. Although we did not identify the components that caused 12-HETE release, we demonstrated that epithelial supernatants harbor a potent stimulus for 12-HETE generation in neutrophils. The 15-HETE generation by neutrophils is induced by the proinflammatory mediators IL-1, IL-8 (15), and IL-13 (31).

The generation of PGE₂ and 12-HETE by neutrophils was significantly increased by stimulation with epithelial supernatants compared to A23187 with or without arachidonic acid. These results indicate that epithelial supernatants were a very strong stimulus for the upregulation of cyclooxygenase and lipoxygenase enzymes and the consequent synthesis of these arachidonic acid metabolites in neutrophils. The relatively short incubation time suggests that the upregulation of mediator synthesis was not related to increased protein synthesis. In the presence of arachidonic acid, the formation of LTB₄ by the stimulation with supernatants was about 75% lower than that with A23187. For comparison, Fogh et al. (15) observed an approximately 65% decrease in LTB₄ formation and a 40% increase in 15-HETE formation by IL-8 in the presence of arachidonic acid compared to stimulation with A23187. This comparison showed that IL-8 and epithelial supernatants were not as effective in LTB₄ stimulation as A23187 and suggests additional regulatory capacities of IL-8 or epithelial supernatants for the 5-lipoxygenase pathway. This proposed additional regulation of cyclooxygenase and lipoxygenase pathways could be of epithelial origin or by the paracrine action of neutrophil derived arachidonic acid mediators itself (7, 14, 19, 26, 43, 44).

During stimulation of neutrophils with epithelial supernatants we did not observe any significant effect of arachidonic acid on PGE₂, LTB₄, 12-HETE, or 15-HETE formation, while the effect of arachidonic acid on A23187-stimulated neutrophils was in agreement with data in other reports (15, 37). This may indicate that the synthesis of cyclooxygenase and lipoxygenase enzymes was not limited by substrate availability in the present study. Enzyme saturation may be one possibility; additional regulatory mechanisms that limited exogenous arachidonic acid metabolism also have to be taken into account. In contrast to IL-8, which induced the release of 15-HETE only in the presence of exogenous arachidonic acid (37), we observed elevated levels of PGE₂, 12-HETE, and 15-HETE after stimulation with epithelial supernatants without exogenous arachidonic acid, which suggests that this stimulus may also activate endogenous arachidonic acid release.

The expression of the strong neutrophil chemoattractant mediators LTB₄ and 12-HETE in keratinocytes suggests that bacterial-epithelial cell interactions play a key role in the induction of neutrophil migration during infection. These findings indicate that epithelial cells are an integral part of the immune system, eliciting proinflammatory responses to a bacterial challenge. In contrast to experiments that demonstrated the synthesis of anti-inflammatory mediators by airway epithelial cells upon antigen stimulation in animal experiments and in vitro (25, 30, 35), the present investigation clearly indicates that the reaction of keratinocytes to a bacterial challenge was dominated by the synthesis of proinflammatory arachidonic acid mediators. The released arachidonic acid mediator PGE₂ is able to affect other inflammatory cells, like the modulation of T-lymphocyte lymphokine secretion and cytotoxicity (18) or the stimulation of immunoglobulin production by B cells (20, 32). T lymphocytes are also equipped with LTB₄ receptors by which LTB₄ exerts the induction of the IL-5 production that in turn modulates eosinophil function (9, 46). LTB₄ also activates B cells and enhances IL-4-mediated immunoglobulin G, M, and E synthesis (45).

In conclusion, this study demonstrated for the first time the possibility of arachidonic acid mediator synthesis in keratinocytes induced by infection with human-pathogenic bacteria. The epithelial supernatants harvested after infection were able to induce the simultaneous formation of the pro- and anti-inflammatory eicosanoid mediators PGE₂, LTB₄, 12-HETE, and 15-HETE in human neutrophils. Even though the present study was an in vitro experiment, it could outline a possible cascade of events during early phases of the inflammatory reaction in vivo that involves the interaction of bacteria with epithelial cells and the consequent release of eicosanoid mediators within neutrophils.