JOURNAL OF CELT,ULAR PHYSIOLOGY 142379-385 (1990)
Eicosanoid Release From Human Bronchial Epithelial Cells Upon Exposure to Toluene Diisocyanate In Vitro SABRINA MATTOLI,* M A U R O MASIERO, FRANCESCO CALABRO, M A U R l Z l O MEZZETTI, MARIO PLEBANI, AND LUlGl ALLEGRA The Interuniversity Center on Cellular and Molecular Mechanisms of Lurig lnjury (S.M., L A . ) and Departmen1 of Thoracic Surgery IM.Me.1, University of' Milan, 20100 Milan, Italy; lnstilule ol' Clinical Chemistry (M.Ma., M.P.) and Department of Surgery (F.C.1, Uriiversi1y of Padud, 35 100 fadua, ltaly Epithelial injury and inflammation are involved in airway hyperresponsivenessand asthma induced by toluene diisocyanate. In that isocyanates are insoluble and highly reactive c:ompounds, bronchial epithelial cells may represent the most important target cells of their toxic effect. We hypothesized that damage to airway epithelium by toluene diisocyanate may result in the release of metabolites of arachidonic acid, which are known to promote inflarnniation and to alter epithelial cell function and airway smooth muscle responsiveness.To test this hypothesis we examined eicosanoid products i n the culture media of bronchial epithelial cells exposed in vitro to 8 and 18 ppb toluene diisocyanate. Epithelial cells derived from human bronchi obtained at surgery were cultured to confluency on collagencoated microporous membrane5. Those cells, which expressed differentiated characteristics of epithelial cells (they showed keratin-containing filaments and had a cobblestone appearance), were alternatively exposed to toluene diisocyanate or air for 30 min in a specially designed in vitro chamber. The production of metabolites of arachidonic acid was assessed by measuring the release of immunoreactive products into the cell medium at the end of the exposure and during a 2 hr period after exposure. This mehod revealed a predominant isocyanate-induced release of immunoreactive 15-hydroxyeicosatetraenoic acid. Release rate of this compound tended to be dose-related and was associated wilh cell darnage as asessed by the release of lactate dehydrogenase in the medium.
Toluene diisocyanate (TDI) is the basic constituent of polyurethanes, and TDI exposure is one of the most important causes of occupational asthma (Karr et al., 1978). In animals and humans, experimental TDI exposure provokes bronchial epithelium injury and airway inflammation, and this is associated with an increase in nonspecific bronchial responsiveness (Gordon et al., 1985; Fabbri e t al., 1987). In that TDI is a lowmolecular-weight, insoluble, and highly reactive and irritant compound (Saunders and Fresch, 19781,part of its toxicity may be the result of adverse events within the epithelium. Ultrastructural studies have in fact documented alterations of cellular metabolism and cell damage after exposure of epithelial cell monolayers t o this compound in vitro (Hastie et al., 1989). Other stimuli that induce asthma and airway hyperresponsiveness have been demonstrated to stimulate eicosanoid release during periods when bronchial hyperresponsiveness is detectable in animals and humans (Hamberg et al., 1980; Murray e t al., 1986; Gravelyn et al., 1988; Leikauf et al., 1989), and the in vitro exposure of epithelial cells to some of those stimuli results in eicosanoid release (Leikauf et al., 1988; Doupnik and Leikauf, 1989). Similar events may occur
:C 1990 WILEY-LISS. INC
in human airways upon exposure to TDI, and the release of eicosanoids from epithelial cells may contribute to the development of some of the pathological alterations observed in isocyanate-induced asthma. The present study indicates that the in vitro exposure of human bronchial epithelial cells to TDI can result in the predominant release of a product of the 15-lipoxygenase pathway of arachidonic acid metabolism and that this is associated with cell damage.
MATERIALS AND METHODS Epithelial cell cultures Bronchial epithelial cells were isolated from grossly normal bronchi of 10 subjects undergoing lobectomy for standard clinical reasons. Epithelium was stripped from the bronchial mucosa, freed from the collagen stroma, dissected into small pieces, and incubated for 2 hr in CaiMg-free Hanks' balanced salt solution (HBSS)
Received June 28, 1989; accepted September 22, 1989.
"To whom reprint requestdcorrespondence should be addressed.
380
MATTOLI
(Sigma Chemical Co., St. Louis, MO) containing 2% Dispase (Collaborative Research Inc., Bedford, MA), 5 mM dithiothreitol, 1%penicillin-streptomycin solution (10.000 U-10 mgiml, Sigma Chemical Co.) (PSI,and 25 mM HEPES, pH 7.4, a t 37"C, 5% CO, in air, 100% humidity. The resulting cell suspension was centrifuged, and the cell pellet was treated with 144 mM ammonium chloride in 1 7 mM Tris HC1 (pH 7.2) to lyse erithrocytes. The cells were then depleted of fibroblasts, macrophage-monocytes, granulocytes, and lymphocytes by adherence to plastic surface and repeated panning (Wysocki and Sato, 1978) using the antiCD11, anti-CD16, anti-CD21, and anti-CD2 monoclonal antibodies (mAbs) (Ortho Diagnostic System SPA, Milan, Italy). Cell count and viability (trypan blue exclusion method) were assessed by using a Burker chamber. The enzymatic digestion procedure and the purification steps indicated above yelded a total number of epithelial cells per bronchus ranging from 3 to 5.5 x loG.Cells obtained from lobar and segmental bronchi from each donor were pooled to reach the adequate number of epithelial cells. The cells were finally resuspended in a complete medium (50:50 mixture of Dulbecco's modified Eagle's medium and Ham's F12 medium, supplemented with 1% PS and lo%, heat-inactivated fetal calf serum), and plated in replicate on transparent, collagen-treated, Transwell microporous membranes (Transwell-COL, 6.5 mm diameter, 3.0 pm pore size, Costar Corporate, Cambridge, MA) placed into the wells of a common 24-well cluster dish. Fluid compartments above and below those microporous membranes allow independent access to apical and basolateral domains. This permits removal of the supernatants above the cells for exposure to gas and still provides the cells with nutrients from below (basal feeding) (Whitcutt et al., 1988). Cells were seeded a t a density of 1 x 106/well and cultured to confluency (3-5 days) with both the apical and basolateral compartments filled with medium. Epithelial cell monolayers were characterized by phase-contrast microscopy and expression of keratin by using the antihuman cytokeratin mAb DAKO-CK1 (Dakopatts, Glostrup, Denmark) and tha APAAP (alkaline phosphatase-antialkaline phosphatase complex) immunohistochemical technique (Dako APAAP Kit, Dakopatts) (Cordell et al., 1984). For the immunohistochemical evaluation, the monolayers were briefly exposed to Dispase, and detached cells were transferred to frozen glass slides after collection onto Nucleopore cytodiagnostic membranes (Nucleopore Italia, Srl, Milan, Italy) by filtration. An irrelevant isotype-matched antibody was used as control. DAKO-CK1 mAb binds to a large number of human cytokeratins; fibroblasts and other mesenchimal cells are not labeled (Gatter et al., 1984). Monolayers were also tested for the presence of T cells and macrophage monocytes by using the antiCD2 and anti-CD11 mAbs (Ortho Diagnostic System SPA) and the APAAP immunoalkaline phosphatase technique. Only the cultures composed of more than 99% keratin-containing cells were used for evaluating the effect of TDI exposure.
In vitro exposure to TDI Cultures of epithelial cells were exposed to TDI or air, as control, employing a specially designed in vitro
m AL. chamber system. The system consisted of a 50 liter glass exposure chamber. To generate the desired concentration of TDI, a metered flow of filtered dry air was passed through the head space vapor of a flask containing 5 ml of a mixture of 80% 2,4-TDI and 20% 2,6-TDI. The gas mixture was diluted with filtered dry air in a 60 liter glass mixing chamber before it entered the exposure chamber. The gas mixture at the desired concentration of TDI was drawn into the exposure chamber by using a vacuum pump operating at 0.5 literimin. To minimize absorption of TDI vapor within the system, all tubing and connectors were made of glass or Teflon and were as short as possible. The entire apparatus was situated beneath a laboratory exhaust hood and inside a water bath, to maintain a constant temperature of 37°C in the chambers and along the tubing system. Experiments were performed with confluent cell monolayers. In each test, five or six cultures with epithelial cells obtained from each donor were exposed to air or two different concentrations of TDI. Cell counts were obtained before exposure in parallel duplicate cultures from each single donor, after detachment and dispersion on the cells by brief incubation with Dispase. For exposure of the cells to air or TDI, the culture medium was harvested and replaced by HBSS supplemented with 25 mM HEPES and 1%PS, pH 7.2. The cultures were incubated for 30 min, then the apical fluid was removed, and different cultures from each donor were exposed to air or two different concentrations of TDI, 8 and 18 ppb, for 30 min, in triplicate. The period of exposure was chosen on the basis of a preliminary study in which we determined the exposure time to air and 18 ppb TDI during which optimal release of eicosanoids occurred with the least change in cell viability. TI11 concentration was monitored continuously by a n MDA model 7005 isocyanate detection equipment (MIIA Scientific Inc., Glenview. IL) and by MarCali's method (Marcali, 1957). After the exposure, the basolateral medium was removed and processed a s indicated below. Parallel cultures of nonexposed epithelial cells were incubated for 30 min with supplemented HBSS in the lower compartment only. The medium was harvested at the end of the incubation for the subsequent evaluation of the level of eicosanoid production in unstimulated epithelial cells. To test the time course of the release of eicosanoid products, in some experiments fresh supplemented HBSS was added back to the cultures in the basolateral compartments and harvested after 2 hr.
Effect of TDI exposure on cell viability To evaluate the degree of cell damage in epithelial cell monolayers exposed to air or TDI, the percent release of the cytoplasmic enzyme lactate dehydrogenase (LDH) in the culture medium was assessed a t the end of the exposure and 2 h r thereafter, according to published methods (Hergmeyer et al., 1974). LDH measurements were performed in the culture media of duplicate cultures. Residual LDH cell content was assayed after disruption of the cells from the same monolayers by treatment with Triton X-100. The amount of released enzyme was expressed a s a percentage of the total (cell content plus the amount measured in the medium). Viability of the cells was also deter-
EICOSANOID RELEASE FROM B8RONCHIALEPITHELIAL CELLS
mined by the trypan blue exclusion method after cell dispersion by brief exposure to Dispase. Arachidonic acid metabolites analysis The cyclooxygenase and lipoxygenase products were extracted from the media of epithelial cells by a nonpolar, solid-phase silica sorbent extraction procedure (Johnson et al., 1985). The aqueous samples were acidified, and the lipid was extracted by passage over Amprep C2-ethyl minicolums (Amersham International Plc, Amersham, England), previously dampened with 100% methanol and water. The minicolums were then washed with 5 ml water, 5 ml 10% ethanol, and 5 ml hexane. The lipid fraction containing prostaglandins and lipoxygenase products was eluted with 5 ml methylformate and evaporated to dryness. The dry samples were reconstituted in 500 pl buffer, aliquoted, and stored at -80°C until analysis. By using the same radioimmunological assays indicated below, the recovery of known amounts of standard 15-hydroxyeicosatetraenoic acid (15-HETE), prostaglandin E, (PGE,), prostaglandin F,, (PGF,,), and leukotriene B, (LTB,) from the buffer used for cell exposure, extracted a s indicated above, ranged from 89% to 95% (five experiments). The amounts of cyclooxygenase and lipoxygenase metabolites of arachidonic acid produced by bronchial epithelial cells exposed to air or TDI were measured a s the release of radioimmunoreactive products. Extracted medium from parallel, nonexposed cultures was assayed in the same manner to determine baseline release of eicosanoid products. The following assays were employed: 15-HETE (6012 Advanced Magnetics Inc., Cambridge, MA) LTB, (36-104-6020 Paesel, GMBH & Co., Frankfurt, Federal Republic of Germany), PGE, (6223 Advanced Magnetics Inc.), and PGF,, (Paesel, GMBH & Go.) (Lingren et al., 1974; Bryant and Hwang, 1983). The antisera crossreactivity for the assays was 0.5% (except PGE,, which had 10% cross reactivity with PGE,) Values of immunoreactive products were expressed as nanograms released per l o 6 cells after correction for the mean recovery of each known standard during the extraction procedure.
Statistical analysis Results from epithelial cells exposed to air o r different concentrations of TDI were compared by a nonparametric equivalent of analysis of variance for repeated measurements and a multiple comparison test.
RESULTS Characterization of epithelial cells The monolayers of epithelial cells were characterized by phase-contrast light microscopy and immunohistochemical techniques. The monolayers invariably showed a cobblestone appearance under phase-contrast light microscopy, and more than 99% of the cells had keratin filaments, as revealed by immunospecific labeling (Fig. 1).Labeling of the cells with mAbs to the surface markers of lymphocytes, granulocytes, and monocytes showed negative results.
381
Effect of TDI exposure on cell viability There was a n appreciable decrease in the number of viable cells in the epithelial cell cultures exposed to the two concentrations of TDI compared to the cultures exposed t o air (Fig. 2). Cell damage due to TDI exposure was confirmed by the measurements of the release of a cytoplasmic enzyme in the culture media soon after exposure and in the 2 h r period following exposure (Fig. 2): LDH release from the cells exposed to TDI was significantly higher than t h a t from cells exposed to air or from nonexposed cells and was maximal 2 h r after exposure. Although the effect tended to be dose related, this did not reach statistical significance (P > 0.1). Analysis of the metabolites of arachidonic acid The variability of the assays in the samples from duplicate cultures of epithelial cells from each donor, exposed to the highest concentration of TDI, ranged from 3% (PGE, assay) to 19% (15-HETE assay). The medium from cultures of epithelial cells not exposed to air or TDI, and incubated for 30 min with HBSS buffer in the lower compartment only, did not contain any detectable amounts of arachidonic acid metabolites (sensitivity of the assays: 15-HETE, 82 pgiml; PGE,, 41 pgiml; PGF,,, 48 pgiml; LTB,, 4.6 pgiml, which in our system corresponded to 82, 41, 48, and 4.6 pg/106 cells, respectively). When epithelial cell monolayers were exposed to air, they released appreciable amounts of immunoreactive PGE, (mean ng/106 cells 5 SE: 1.18 2 0.83) and very low amounts of immunoreactive PGF,, (0.08 2 0.01) and 15-HETE (0.14 ? 0.07) to the culture media (Fig. 3). TDI significantly enhanced the release of immunoreactive 15-HETE from these cells, and the effect tended to be related to the concentration of TDI used for exposure (Fig. 3). The release of immunoreactive PGF,, also increased slightly (P = 0.049) in the medium of cells exposed to the highest concentration of TDI (Fig. 3). Immunoreactive PGE, in the medium of TDI-exposed cells did not differ significantly from t h a t detected in the medium from cultures exposed t o air (Fig. 3). No detectable amount of immunoreactive LTB, was recorded in the culture media of cells exposed to air or TDI. After cessation of exposure, the release of immunoreactive 15-HETE and PGF,,, began to decrease. 15HETE remained elevated for a t least 2 h r in the medium of cells exposed to the two concentrations of TDI (Figs. 4, 51, whereas PGF,, returned to baseline over the same period of time (Fig. 4). Surprisingly, during the 2 h r period following exposure, PGE, release fell more quickly in the medium of cultures exposed to TDI 18 ppb (mean ng/lOGcells 2 SE: 0.238 2 0.01) than in cultures exposed t o air (0.925 2 0.09) (P < 0.05).
DISCUSSION Recent evidence indicates that the airway epithelium may have a n important role in airway hyperresponsiveness (Cuss and Barnes, 1987; Sigal and Nadel, 19881, a condition inducible by TDI (Gordon et al., 1985; Fabbri et al., 1987) and one of the most important diagnostic features of asthma (Laitinen et al., 1985). This hypothesis may be particularly relevant to the environmental exposure in that airway epithelium rep-
382
MATTOLI ET AL
Fig. 1. Human bronchial epithelial cells grown on collagen-treated Costar Transwell microporous membranes and labeled with the anticytokeratin mAb DAKO-CK1. The indirect APAAP immunoenzymatic technique was used to visualize the keratin-containing cells,
which stain red by this method. Left: Monolayer of epithelial cells. Right: epithelial cells labeled after detachment from the culture membranes and transfer onto glass slides. x 400.
resents the first respiratory cell barrier to interact with inhaled substances such a s TDI. In addition, in that TDI is a highly reactive compound and the epithelial cell membrane is the first point of contact, initiating events within the epithelial membrane or inside the epithelial cells may be sufficient to explain much of its toxicity (Sheppard et al., 1988; Hastie et al., 1989). Many other irritating asthmogenic substances known to induce epithelial cell damage, inflammation, and increased airway responsiveness, such as ozone and aldehydes (Murlas and Roum, 1985; Seltzer et al., 1986; Leikauf et al., 1989), cause epithelial injury in vitro and the release of metabolites of arachidonic acid, with inflammatory properties from bronchial epithelial cells (Leikauf et al., 1988; Doupnik and Leikauf, 19891, particularly 15-HETE, PGE,, and PGF,,,. The same mediators have been demonstrated previously to increase in the lung of asthmatic subjects, either spontaneously or after exposure to asthmogenic stimuli (Hamberg et al., 1980; Murray et al., 1986; Gravelyn et al., 19881, suggesting that they are relevant to the pathogenesis of the disease. Because TDI induces epithelial damage in vivo and in vitro (Gordon et al., 1985; Hastie et al., 19891, the objective of this study was to determine whether TDI could alter arachidonic acid metabolism in the epithelium. To test this hypothesis, we cultured epithelial cells onto transparent, collagen-treated microporous membranes, which are known to promote the growth
and the polarity of differentiation of cells (Whitcutt et al., 1988) and allow direct exposure t o gas after the removal of the apical fluid. Pure epithelial cells were obtained by depletion of granulocytes, monocytes, and lymphocytes with monoclonal antibodies against the surface markers of those cells and immunoadherence, eliminating possible contributions of arachidonic acid metabolites from nonepithelial cells. We also used primary cultures rather than cells that have been repeatedly passaged, and this, together with the use of collagen-coated microporous filters, should have limited the effects of cellular dedifferentiation known to occur in classic cultures. Epithelial cell monolayers were composed of more than 99% keratin-containing cells, as assessed by immunohistochemical techniques, and showed a cobblestone appearance under phase-contrast microscopy. Both keratin-containing filaments and the formation of domes in cultures are characteristic features of differentiated, functional airway epithelial cells. The cells were exposed to air or TDI under controlled physiologic conditions. In this manner, we demonstrated that the exposure of human bronchial epithelial cells to concentrations of TDI normally encountered in the environment of naturally exposed subjects (Fabbri et al., 1987) did result in the predominant release of immunoreactive 15HETE and release of appreciable amounts of immunoreactive PGF,,,. Our findings confirm the results of previous studies
383
EICOSANOID RELEASE FROM BRONCHIAL EPITHELIAL CELLS
50 m 60 -I
70 W
5
> aa
80
1 90 lJ
100
0
-
10
lA
20
W
2 2 -I
30
ae 1
z
40
D
soon a f t e r exposure 2 h a f t e r exposure
I n
J*q
**
**
J
50
CONTROL
A1 R
8.0
t
18.0
TDI (ppb) Fig. 2. Top: Viability of epithelial cells soon after exposure t o air or TDI, 8 and 18 ppb, and 2 hr therafter. Nonexposed monolayers from the same donors were used as control. Results are expressed as the mean percent number of epithelial cells excluding trypan blue (ZSE). Bottom: Release of LDH from epithelial cell8 upon exposure to air or TDI and over a 2 hr period following exposure. Mean percent release: lo6 cells ( t S E ) from five experiments in duplicate. *P < 0.05, ""P
Eicosanoid Release From Human Bronchial Epithelial Cells Upon Exposure to Toluene Diisocyanate In Vitro SABRINA MATTOLI,* M A U R O MASIERO, FRANCESCO CALABRO, M A U R l Z l O MEZZETTI, MARIO PLEBANI, AND LUlGl ALLEGRA The Interuniversity Center on Cellular and Molecular Mechanisms of Lurig lnjury (S.M., L A . ) and Departmen1 of Thoracic Surgery IM.Me.1, University of' Milan, 20100 Milan, Italy; lnstilule ol' Clinical Chemistry (M.Ma., M.P.) and Department of Surgery (F.C.1, Uriiversi1y of Padud, 35 100 fadua, ltaly Epithelial injury and inflammation are involved in airway hyperresponsivenessand asthma induced by toluene diisocyanate. In that isocyanates are insoluble and highly reactive c:ompounds, bronchial epithelial cells may represent the most important target cells of their toxic effect. We hypothesized that damage to airway epithelium by toluene diisocyanate may result in the release of metabolites of arachidonic acid, which are known to promote inflarnniation and to alter epithelial cell function and airway smooth muscle responsiveness.To test this hypothesis we examined eicosanoid products i n the culture media of bronchial epithelial cells exposed in vitro to 8 and 18 ppb toluene diisocyanate. Epithelial cells derived from human bronchi obtained at surgery were cultured to confluency on collagencoated microporous membrane5. Those cells, which expressed differentiated characteristics of epithelial cells (they showed keratin-containing filaments and had a cobblestone appearance), were alternatively exposed to toluene diisocyanate or air for 30 min in a specially designed in vitro chamber. The production of metabolites of arachidonic acid was assessed by measuring the release of immunoreactive products into the cell medium at the end of the exposure and during a 2 hr period after exposure. This mehod revealed a predominant isocyanate-induced release of immunoreactive 15-hydroxyeicosatetraenoic acid. Release rate of this compound tended to be dose-related and was associated wilh cell darnage as asessed by the release of lactate dehydrogenase in the medium.
Toluene diisocyanate (TDI) is the basic constituent of polyurethanes, and TDI exposure is one of the most important causes of occupational asthma (Karr et al., 1978). In animals and humans, experimental TDI exposure provokes bronchial epithelium injury and airway inflammation, and this is associated with an increase in nonspecific bronchial responsiveness (Gordon et al., 1985; Fabbri e t al., 1987). In that TDI is a lowmolecular-weight, insoluble, and highly reactive and irritant compound (Saunders and Fresch, 19781,part of its toxicity may be the result of adverse events within the epithelium. Ultrastructural studies have in fact documented alterations of cellular metabolism and cell damage after exposure of epithelial cell monolayers t o this compound in vitro (Hastie et al., 1989). Other stimuli that induce asthma and airway hyperresponsiveness have been demonstrated to stimulate eicosanoid release during periods when bronchial hyperresponsiveness is detectable in animals and humans (Hamberg et al., 1980; Murray e t al., 1986; Gravelyn et al., 1988; Leikauf et al., 1989), and the in vitro exposure of epithelial cells to some of those stimuli results in eicosanoid release (Leikauf et al., 1988; Doupnik and Leikauf, 1989). Similar events may occur
:C 1990 WILEY-LISS. INC
in human airways upon exposure to TDI, and the release of eicosanoids from epithelial cells may contribute to the development of some of the pathological alterations observed in isocyanate-induced asthma. The present study indicates that the in vitro exposure of human bronchial epithelial cells to TDI can result in the predominant release of a product of the 15-lipoxygenase pathway of arachidonic acid metabolism and that this is associated with cell damage.
MATERIALS AND METHODS Epithelial cell cultures Bronchial epithelial cells were isolated from grossly normal bronchi of 10 subjects undergoing lobectomy for standard clinical reasons. Epithelium was stripped from the bronchial mucosa, freed from the collagen stroma, dissected into small pieces, and incubated for 2 hr in CaiMg-free Hanks' balanced salt solution (HBSS)
Received June 28, 1989; accepted September 22, 1989.
"To whom reprint requestdcorrespondence should be addressed.
380
MATTOLI
(Sigma Chemical Co., St. Louis, MO) containing 2% Dispase (Collaborative Research Inc., Bedford, MA), 5 mM dithiothreitol, 1%penicillin-streptomycin solution (10.000 U-10 mgiml, Sigma Chemical Co.) (PSI,and 25 mM HEPES, pH 7.4, a t 37"C, 5% CO, in air, 100% humidity. The resulting cell suspension was centrifuged, and the cell pellet was treated with 144 mM ammonium chloride in 1 7 mM Tris HC1 (pH 7.2) to lyse erithrocytes. The cells were then depleted of fibroblasts, macrophage-monocytes, granulocytes, and lymphocytes by adherence to plastic surface and repeated panning (Wysocki and Sato, 1978) using the antiCD11, anti-CD16, anti-CD21, and anti-CD2 monoclonal antibodies (mAbs) (Ortho Diagnostic System SPA, Milan, Italy). Cell count and viability (trypan blue exclusion method) were assessed by using a Burker chamber. The enzymatic digestion procedure and the purification steps indicated above yelded a total number of epithelial cells per bronchus ranging from 3 to 5.5 x loG.Cells obtained from lobar and segmental bronchi from each donor were pooled to reach the adequate number of epithelial cells. The cells were finally resuspended in a complete medium (50:50 mixture of Dulbecco's modified Eagle's medium and Ham's F12 medium, supplemented with 1% PS and lo%, heat-inactivated fetal calf serum), and plated in replicate on transparent, collagen-treated, Transwell microporous membranes (Transwell-COL, 6.5 mm diameter, 3.0 pm pore size, Costar Corporate, Cambridge, MA) placed into the wells of a common 24-well cluster dish. Fluid compartments above and below those microporous membranes allow independent access to apical and basolateral domains. This permits removal of the supernatants above the cells for exposure to gas and still provides the cells with nutrients from below (basal feeding) (Whitcutt et al., 1988). Cells were seeded a t a density of 1 x 106/well and cultured to confluency (3-5 days) with both the apical and basolateral compartments filled with medium. Epithelial cell monolayers were characterized by phase-contrast microscopy and expression of keratin by using the antihuman cytokeratin mAb DAKO-CK1 (Dakopatts, Glostrup, Denmark) and tha APAAP (alkaline phosphatase-antialkaline phosphatase complex) immunohistochemical technique (Dako APAAP Kit, Dakopatts) (Cordell et al., 1984). For the immunohistochemical evaluation, the monolayers were briefly exposed to Dispase, and detached cells were transferred to frozen glass slides after collection onto Nucleopore cytodiagnostic membranes (Nucleopore Italia, Srl, Milan, Italy) by filtration. An irrelevant isotype-matched antibody was used as control. DAKO-CK1 mAb binds to a large number of human cytokeratins; fibroblasts and other mesenchimal cells are not labeled (Gatter et al., 1984). Monolayers were also tested for the presence of T cells and macrophage monocytes by using the antiCD2 and anti-CD11 mAbs (Ortho Diagnostic System SPA) and the APAAP immunoalkaline phosphatase technique. Only the cultures composed of more than 99% keratin-containing cells were used for evaluating the effect of TDI exposure.
In vitro exposure to TDI Cultures of epithelial cells were exposed to TDI or air, as control, employing a specially designed in vitro
m AL. chamber system. The system consisted of a 50 liter glass exposure chamber. To generate the desired concentration of TDI, a metered flow of filtered dry air was passed through the head space vapor of a flask containing 5 ml of a mixture of 80% 2,4-TDI and 20% 2,6-TDI. The gas mixture was diluted with filtered dry air in a 60 liter glass mixing chamber before it entered the exposure chamber. The gas mixture at the desired concentration of TDI was drawn into the exposure chamber by using a vacuum pump operating at 0.5 literimin. To minimize absorption of TDI vapor within the system, all tubing and connectors were made of glass or Teflon and were as short as possible. The entire apparatus was situated beneath a laboratory exhaust hood and inside a water bath, to maintain a constant temperature of 37°C in the chambers and along the tubing system. Experiments were performed with confluent cell monolayers. In each test, five or six cultures with epithelial cells obtained from each donor were exposed to air or two different concentrations of TDI. Cell counts were obtained before exposure in parallel duplicate cultures from each single donor, after detachment and dispersion on the cells by brief incubation with Dispase. For exposure of the cells to air or TDI, the culture medium was harvested and replaced by HBSS supplemented with 25 mM HEPES and 1%PS, pH 7.2. The cultures were incubated for 30 min, then the apical fluid was removed, and different cultures from each donor were exposed to air or two different concentrations of TDI, 8 and 18 ppb, for 30 min, in triplicate. The period of exposure was chosen on the basis of a preliminary study in which we determined the exposure time to air and 18 ppb TDI during which optimal release of eicosanoids occurred with the least change in cell viability. TI11 concentration was monitored continuously by a n MDA model 7005 isocyanate detection equipment (MIIA Scientific Inc., Glenview. IL) and by MarCali's method (Marcali, 1957). After the exposure, the basolateral medium was removed and processed a s indicated below. Parallel cultures of nonexposed epithelial cells were incubated for 30 min with supplemented HBSS in the lower compartment only. The medium was harvested at the end of the incubation for the subsequent evaluation of the level of eicosanoid production in unstimulated epithelial cells. To test the time course of the release of eicosanoid products, in some experiments fresh supplemented HBSS was added back to the cultures in the basolateral compartments and harvested after 2 hr.
Effect of TDI exposure on cell viability To evaluate the degree of cell damage in epithelial cell monolayers exposed to air or TDI, the percent release of the cytoplasmic enzyme lactate dehydrogenase (LDH) in the culture medium was assessed a t the end of the exposure and 2 h r thereafter, according to published methods (Hergmeyer et al., 1974). LDH measurements were performed in the culture media of duplicate cultures. Residual LDH cell content was assayed after disruption of the cells from the same monolayers by treatment with Triton X-100. The amount of released enzyme was expressed a s a percentage of the total (cell content plus the amount measured in the medium). Viability of the cells was also deter-
EICOSANOID RELEASE FROM B8RONCHIALEPITHELIAL CELLS
mined by the trypan blue exclusion method after cell dispersion by brief exposure to Dispase. Arachidonic acid metabolites analysis The cyclooxygenase and lipoxygenase products were extracted from the media of epithelial cells by a nonpolar, solid-phase silica sorbent extraction procedure (Johnson et al., 1985). The aqueous samples were acidified, and the lipid was extracted by passage over Amprep C2-ethyl minicolums (Amersham International Plc, Amersham, England), previously dampened with 100% methanol and water. The minicolums were then washed with 5 ml water, 5 ml 10% ethanol, and 5 ml hexane. The lipid fraction containing prostaglandins and lipoxygenase products was eluted with 5 ml methylformate and evaporated to dryness. The dry samples were reconstituted in 500 pl buffer, aliquoted, and stored at -80°C until analysis. By using the same radioimmunological assays indicated below, the recovery of known amounts of standard 15-hydroxyeicosatetraenoic acid (15-HETE), prostaglandin E, (PGE,), prostaglandin F,, (PGF,,), and leukotriene B, (LTB,) from the buffer used for cell exposure, extracted a s indicated above, ranged from 89% to 95% (five experiments). The amounts of cyclooxygenase and lipoxygenase metabolites of arachidonic acid produced by bronchial epithelial cells exposed to air or TDI were measured a s the release of radioimmunoreactive products. Extracted medium from parallel, nonexposed cultures was assayed in the same manner to determine baseline release of eicosanoid products. The following assays were employed: 15-HETE (6012 Advanced Magnetics Inc., Cambridge, MA) LTB, (36-104-6020 Paesel, GMBH & Co., Frankfurt, Federal Republic of Germany), PGE, (6223 Advanced Magnetics Inc.), and PGF,, (Paesel, GMBH & Go.) (Lingren et al., 1974; Bryant and Hwang, 1983). The antisera crossreactivity for the assays was 0.5% (except PGE,, which had 10% cross reactivity with PGE,) Values of immunoreactive products were expressed as nanograms released per l o 6 cells after correction for the mean recovery of each known standard during the extraction procedure.
Statistical analysis Results from epithelial cells exposed to air o r different concentrations of TDI were compared by a nonparametric equivalent of analysis of variance for repeated measurements and a multiple comparison test.
RESULTS Characterization of epithelial cells The monolayers of epithelial cells were characterized by phase-contrast light microscopy and immunohistochemical techniques. The monolayers invariably showed a cobblestone appearance under phase-contrast light microscopy, and more than 99% of the cells had keratin filaments, as revealed by immunospecific labeling (Fig. 1).Labeling of the cells with mAbs to the surface markers of lymphocytes, granulocytes, and monocytes showed negative results.
381
Effect of TDI exposure on cell viability There was a n appreciable decrease in the number of viable cells in the epithelial cell cultures exposed to the two concentrations of TDI compared to the cultures exposed t o air (Fig. 2). Cell damage due to TDI exposure was confirmed by the measurements of the release of a cytoplasmic enzyme in the culture media soon after exposure and in the 2 h r period following exposure (Fig. 2): LDH release from the cells exposed to TDI was significantly higher than t h a t from cells exposed to air or from nonexposed cells and was maximal 2 h r after exposure. Although the effect tended to be dose related, this did not reach statistical significance (P > 0.1). Analysis of the metabolites of arachidonic acid The variability of the assays in the samples from duplicate cultures of epithelial cells from each donor, exposed to the highest concentration of TDI, ranged from 3% (PGE, assay) to 19% (15-HETE assay). The medium from cultures of epithelial cells not exposed to air or TDI, and incubated for 30 min with HBSS buffer in the lower compartment only, did not contain any detectable amounts of arachidonic acid metabolites (sensitivity of the assays: 15-HETE, 82 pgiml; PGE,, 41 pgiml; PGF,,, 48 pgiml; LTB,, 4.6 pgiml, which in our system corresponded to 82, 41, 48, and 4.6 pg/106 cells, respectively). When epithelial cell monolayers were exposed to air, they released appreciable amounts of immunoreactive PGE, (mean ng/106 cells 5 SE: 1.18 2 0.83) and very low amounts of immunoreactive PGF,, (0.08 2 0.01) and 15-HETE (0.14 ? 0.07) to the culture media (Fig. 3). TDI significantly enhanced the release of immunoreactive 15-HETE from these cells, and the effect tended to be related to the concentration of TDI used for exposure (Fig. 3). The release of immunoreactive PGF,, also increased slightly (P = 0.049) in the medium of cells exposed to the highest concentration of TDI (Fig. 3). Immunoreactive PGE, in the medium of TDI-exposed cells did not differ significantly from t h a t detected in the medium from cultures exposed t o air (Fig. 3). No detectable amount of immunoreactive LTB, was recorded in the culture media of cells exposed to air or TDI. After cessation of exposure, the release of immunoreactive 15-HETE and PGF,,, began to decrease. 15HETE remained elevated for a t least 2 h r in the medium of cells exposed to the two concentrations of TDI (Figs. 4, 51, whereas PGF,, returned to baseline over the same period of time (Fig. 4). Surprisingly, during the 2 h r period following exposure, PGE, release fell more quickly in the medium of cultures exposed to TDI 18 ppb (mean ng/lOGcells 2 SE: 0.238 2 0.01) than in cultures exposed t o air (0.925 2 0.09) (P < 0.05).
DISCUSSION Recent evidence indicates that the airway epithelium may have a n important role in airway hyperresponsiveness (Cuss and Barnes, 1987; Sigal and Nadel, 19881, a condition inducible by TDI (Gordon et al., 1985; Fabbri et al., 1987) and one of the most important diagnostic features of asthma (Laitinen et al., 1985). This hypothesis may be particularly relevant to the environmental exposure in that airway epithelium rep-
382
MATTOLI ET AL
Fig. 1. Human bronchial epithelial cells grown on collagen-treated Costar Transwell microporous membranes and labeled with the anticytokeratin mAb DAKO-CK1. The indirect APAAP immunoenzymatic technique was used to visualize the keratin-containing cells,
which stain red by this method. Left: Monolayer of epithelial cells. Right: epithelial cells labeled after detachment from the culture membranes and transfer onto glass slides. x 400.
resents the first respiratory cell barrier to interact with inhaled substances such a s TDI. In addition, in that TDI is a highly reactive compound and the epithelial cell membrane is the first point of contact, initiating events within the epithelial membrane or inside the epithelial cells may be sufficient to explain much of its toxicity (Sheppard et al., 1988; Hastie et al., 1989). Many other irritating asthmogenic substances known to induce epithelial cell damage, inflammation, and increased airway responsiveness, such as ozone and aldehydes (Murlas and Roum, 1985; Seltzer et al., 1986; Leikauf et al., 1989), cause epithelial injury in vitro and the release of metabolites of arachidonic acid, with inflammatory properties from bronchial epithelial cells (Leikauf et al., 1988; Doupnik and Leikauf, 19891, particularly 15-HETE, PGE,, and PGF,,,. The same mediators have been demonstrated previously to increase in the lung of asthmatic subjects, either spontaneously or after exposure to asthmogenic stimuli (Hamberg et al., 1980; Murray et al., 1986; Gravelyn et al., 19881, suggesting that they are relevant to the pathogenesis of the disease. Because TDI induces epithelial damage in vivo and in vitro (Gordon et al., 1985; Hastie et al., 19891, the objective of this study was to determine whether TDI could alter arachidonic acid metabolism in the epithelium. To test this hypothesis, we cultured epithelial cells onto transparent, collagen-treated microporous membranes, which are known to promote the growth
and the polarity of differentiation of cells (Whitcutt et al., 1988) and allow direct exposure t o gas after the removal of the apical fluid. Pure epithelial cells were obtained by depletion of granulocytes, monocytes, and lymphocytes with monoclonal antibodies against the surface markers of those cells and immunoadherence, eliminating possible contributions of arachidonic acid metabolites from nonepithelial cells. We also used primary cultures rather than cells that have been repeatedly passaged, and this, together with the use of collagen-coated microporous filters, should have limited the effects of cellular dedifferentiation known to occur in classic cultures. Epithelial cell monolayers were composed of more than 99% keratin-containing cells, as assessed by immunohistochemical techniques, and showed a cobblestone appearance under phase-contrast microscopy. Both keratin-containing filaments and the formation of domes in cultures are characteristic features of differentiated, functional airway epithelial cells. The cells were exposed to air or TDI under controlled physiologic conditions. In this manner, we demonstrated that the exposure of human bronchial epithelial cells to concentrations of TDI normally encountered in the environment of naturally exposed subjects (Fabbri et al., 1987) did result in the predominant release of immunoreactive 15HETE and release of appreciable amounts of immunoreactive PGF,,,. Our findings confirm the results of previous studies
383
EICOSANOID RELEASE FROM BRONCHIAL EPITHELIAL CELLS
50 m 60 -I
70 W
5
> aa
80
1 90 lJ
100
0
-
10
lA
20
W
2 2 -I
30
ae 1
z
40
D
soon a f t e r exposure 2 h a f t e r exposure
I n
J*q
**
**
J
50
CONTROL
A1 R
8.0
t
18.0
TDI (ppb) Fig. 2. Top: Viability of epithelial cells soon after exposure t o air or TDI, 8 and 18 ppb, and 2 hr therafter. Nonexposed monolayers from the same donors were used as control. Results are expressed as the mean percent number of epithelial cells excluding trypan blue (ZSE). Bottom: Release of LDH from epithelial cell8 upon exposure to air or TDI and over a 2 hr period following exposure. Mean percent release: lo6 cells ( t S E ) from five experiments in duplicate. *P < 0.05, ""P
