Interactive Effects Between Cytokines on PGE Production by Human Periodontal Ligament Fibroblasts in vitro S. SAITO, P. NGAN1"4, M. SAITO, R. LANESE2, J. SHANFELD3, and Z. DAVIDOVITCH' Department of Orthodontics, Showa University, Tokyo, 145, Japan; and Departments of 'Orthodontics, 2Preventive Medicine, and 3Oral Biology, The Ohio State University, Columbus, Ohio 43210
Mononuclear cell production of cytokines that stimulate fibroblast production of prostaglandin E (PGE) is an important mechanism by which mononuclear cells regulate fibroblast function. The objective of this investigation was to determine the effects of the cytokines interleukin 11 (IL-113), interleukin lot (IL-lo), tumor necrosis factor c (TNF-ot), and interferon y (IFN-y), alone or in paired combinations, on PGE production by near-confluent human periodontal ligament (PDL) fibroblasts in vitro. Premolars extracted in the course of orthodontic treatment were used for this study. Fibroblast cultures, free of epithelial cells, were obtained after the fourth subculture by the use of accurately-timed trypsin treatment. Cells in the fourth to sixth passage, incubated in DMEM supplemented with 10% equine serum, were used for these experiments. Cells (1 x 105) were seeded in 12- x -75-mm tissue culture tubes and incubated with various doses of IL-1r, IL-lx, TNF-cx, and IFNy, alone or in specific combinations, for 15 min, two, 12, 24, and 72 h. PGE concentrations in the media were measured by radio-immunoassay. The results showed that human PDL fibroblasts responded to the administration of cytokines by an elevation in the synthesis of PGE in a dose- and time-related fashion. The increase in PGE production was inhibited by the addition of indomethacin. The interactions between these cytokines varied in degree, depending on the particular combinations of cytokines. In addition, the administration of cytokine combinations was found to be additive, synergistic, subtractive, or suppressive on the production of PGE by PDL fibroblasts, depending on the duration of incubation. These experiments demonstrate the importance of the consideration of the interplay between cytokines produced by mononuclear cells on the mechanisms that regulate the functions of PDL fibroblasts. J Dent Res 69(8):1456-1462, August, 1990
Introduction. The main function of periodontal ligament (PDL) fibroblasts appears to be the maintenance and remodeling of the ligament itself and the establishment of a viable connection between alveolar bone and dental cementum. Fibroblasts and epithelial cells have been cultured from human PDL (Litwin et al., 1971; Blomlof and Otteskog, 1981; Ragnarsson et al., 1985). Ten Cate et al. (1976) have shown that PDL fibroblasts are capable of synthesizing, as well as degrading, collagen during physiological tooth movement. Rippin (1976) examined the rate of collagen turnover in young rat molars subjected to altered functional forces, and found an increased turnover of collagen throughout the width of the ligament, facilitating the Received for publication June 7, 1989 Accepted for publication December 4, 1989 4To whom correspondence and reprint requests should be addressed This investigation was supported by The Ohio State University College of Dentistry. 1456
tissue remodeling associated with tooth movement. Davidovitch et al. (1988) studied the localization of the cytokines interleukin 11 (IL-1p) and IL-lot in dental and paradental tissues surrounding orthodontically treated teeth in cats by an immunohistochemical procedure. They found a marked increase in staining intensity for IL-lot and 11 in cells around the tooth, particularly at compression sites. More direct evidence on the activity of PDL cells comes from in vitro work that demonstrates increased production of PGE and cAMP by fibroblasts undergoing mechanical stretch (Ngan et al., 1988). PDL fibroblasts have also been implicated in the process of tooth eruption; cAMP and PGE have been shown to be localized around the paradental tissues of erupting teeth in kittens (Gustafson et al., 1977; Shanfeld et al., 1986). The source of PGE around these teeth is not clear. However, a potential mechanism could be the production of PGE in response to IL-1. The cytokines, IL-1, tumor necrosis factor (TNF), and interferon (IFN) are released by cells of the immune system upon stimulation by a variety of agents. IL-1 and TNF are largely products of monocytes, whereas IFN-y is produced primarily by lymphocytes. Despite the fact that little, if any, homology exists among these molecules, which seem to bind to distinct receptors, they share some common functions. Fortunately, all of these cytokines are known as biological response modifiers that demonstrate pleiotropic effects. All are available as products of recombinant DNA technology, so that examination of their actions is not limited by availability or degree of purity. There are two distinct IL-1 gene products, IL-lx and IL-1P (March et al., 1985), with the latter being the major physiological form, which has been shown to be synonymous with osteoclast activating factor (Dewhirst, 1985). IL-1 and TNF-a have been shown to have synergistic effects on the production of PGE2 (Elias et al., 1987; Last-Barney et al., 1988), induction of bone resorption (Stashenko et al., 1987), and on growth arrest of melanoma cells in vitro (Ruggiero and Baglioni, 1987). In addition, IL-1 and TNF-cx have overlapping activities, including the regulation of phospholipid metabolism, induction of acute phase proteins, induction of prostacyclin synthesis, synthesis of hemopoietic growth factors, growth-promoting effects on fibroblasts, induction of fever, and regulation of the expression of human major histocompatibility antigens (Ruggiero and Baglioni, 1987; Le and Vilcek, 1987; Lovett et al., 1986). The chemical similarities and distinctions, as well as studies on the synergistic and overlapping activities of these two molecules, have been reviewed recently by Le and Vilcek (1987). TNF-at and IFN-y also share some functional activities. Both cytokines induce the re-organization of endothelial cells in monolayer culture (Stolpen et al., 1986) and activate polymorphonuclear neutrophil functions (Shalaby et al., 1985). IFNy also stimulates the production of TNF by cultured human peripheral blood monocytes, in contrast to IFN-at, which does not stimulate TNF production (Scuderi et al., 1987). However, IFN-y has been shown to inhibit certain inflammatory reactions, especially those associated with PGE2 synthesis and bone resorption (Boraschi et al., 1985; Gowen and Mundy, 1986).
Downloaded from jdr.sagepub.com at UNIV CALIFORNIA SANTA BARBARA on July 8, 2015 For personal use only. No other uses without permission.
CYTOKINE EFFECTS ON PDL FIBROBLAST PGE PRODUCTION
Vol. 69 No. 8
IFN-y has been shown to suppress IL-1-stimulated PGE production by human monocytes (Browing and Ribolini, 1987; Friteau et al., 1988), but not by human fibroblasts (Friteau et al., 1988). The objective of this study was to investigate the interactive effects of simultaneous administrations of paired cytokines, i.e., IL-lot, IL-1r, TNF-at, and IFN-y, on the production of PGE by human PDL fibroblasts in vitro.
Materials and methods. Collection and culture of human PDL fibroblasts. -PDL fibroblasts were prepared according to the method of Ragnarsson et al. (1985). Premolars, extracted from healthy male patients 16 to 25 years old in the course of orthodontic treatment, were obtained fresh from the Oral Surgery Department. The teeth were washed in Hanks' balanced salt solution (HBSS), and the gingival attachments were carefully removed with a sharp scalpel. The crowns (plus 2 mm past the gingival margins) were dipped in a 5.25% sodium hypochlorite solution for two min so that bacteria and any remaining gingival cells would be killed. The teeth were rinsed in three changes of HBSS. Each tooth was placed in a sterile 15-mL centrifuge tube with 5 mL of 0.1% collagenase (Cooper Biomedical, West Chester, PA) and incubated at 370C for six h. After incubation, the tube was centrifuged at 1000 g for ten min, the tooth was removed, and the cell pellet collected. The pellet was washed in complete medium consisting of Dulbecco's Modified Eagle Medium (DMEM, GIBCO, Grand Island, NY) containing 10% decomplemented equine serum (Hyclone Laboratories, Logan, UT) heat-inactivated at 560C for 30 min, 50 pLg/mL gentamycin (GIBCO), 50 ig/mL nystatin (GIBCO), and plated in 60-mm culture dishes. The dishes were incubated in a humid environment of 95% air, 5% C02, at 370C for 48 h, so that attachment would occur; the dishes were then washed free of debris and re-fed daily with complete medium. Culture characteristics and separation of cell type. -One human premolar or molar yielded enough cells to seed two 60mm dishes. Two populations of cells could be clearly distinguished after 48 h in culture, one having characteristic fibroblast morphology and the other epithelial morphology. In five to seven days, the cells grew to confluence, with the fibroblasts showing a greater rate of proliferation than the epithelial cells. When confluence was obtained, separation and subculturing of the cells were attempted by means of differential trypsinization. The dishes were washed with HBSS, and the cells were removed by incubation in 0.15% trypsin (GIBCO) for seven min at 37°C. The removal of cells was closely monitored by phase-contrast microscopy. Trypsin removes the fibroblasts more rapidly than the epithelial cells (Owens, 1974). In this manner, it was possible for a pure population of fibroblasts free of any epithelial cell contamination to be obtained (Ragnarsson et al.,
1985).
Characteristics of PDL cells.-Morphologically, as examined by phase-contrast microscopy, PDL cells appeared to be spindle-shaped and elongated in appearance, characteristics of "fibroblast-like" cells. Their doubling time as measured by Reichert's hemocytometer was approximately 36 to 48 h, reaching confluence in five days. Comparison of biosynthetic activity between bone cells (MC3T3-E1 cells) and PDL cells showed that the basal level of alkaline phosphatase activity (determined by a modified kinetic method, Sigma #245) was two-fold higher in bone cells (160 x 10-3 units ALP/105 cells vs. 78 x 10 -5units). PTH induces a 40-fold increase in cAMP activities in bone cells and a five-fold increase in PDL cells. However, both cells showed decreased alkaline phosphatase activity (25%) with PTH administration.
1457
Incubation and assay for PGE2. -Cells (1 x 105) were seeded on 12- x -75-mm plastic tissue-culture tubes. Satisfactory attachment of fibroblasts was usually obtained in six to 12 h. Various doses of recombinant human IL-13 (Immunex, Seattle, WA), recombinant human IL-lat (Immunex), recombinant human TNF-at (Genentech, San Francisco, CA), and recombinant human IFN--y (Genzyme, Boston, MA) were added to the test tubes for various time periods. The doses tested for IL-11 were 0.1 ng/mL (= 5.83 x 10-12 mol/L), 0.3 ng/mL (= 1.75 x 10-11 mol/L), and 1.0 ng/mL (= 5.83 x 10-1 mol/L); for IL-la, the doses were 0.3 ng/mL (= 1.75 x 10- l mol/L), 1.0 ng/mL (= 5.83 x 10-11 mol/L), and 3.0 ng/mL (= 1.75 x 10-10 mol/L); for TNF-at, the doses were 0.1 nmol/L (= 10-10 mol/L), 0.3 nmol/L (= 3 x 10-10 mol/L), and 1.0 nmol/L (= 10-9 mol/L); and for IFN-y, the doses were 10 U/mL (= 2.33 x 10-11 mol/L), 30 U/mL (= 7.0 x 10-11 mol/L), and 100 U/mL (= 2.33 x 10-10 mol/L). These recombinant cytokine preparations were relatively pure and free of contamination by lipopolysaccharide (LPS). The amount of LPS as determined by the limulus assay was in the order of pg/mg protein. In addition, the cytokine preparations were relatively stable in tissue culture medium, due to the hydrophobic core, and were resistant to proteolysis. The halflife of these proteins was in the order of days. PGE content in sample media was measured in duplicate by radio-immunoassay (RIA). Media aliquots were incubated with diluted rabbit anti-PGE serum (ICN ImmunoBiologicals, Lisle, IL; the term PGE instead of PGE2 is used here because the antibodies crossreact with PGE1) and 3H-PGE2 (10,000 dpm, New England Nuclear, Boston, MA) for 90 min at 4°C. After incubation, cold Dextran-coated charcoal suspension (0.5% Norit A, Fisher Scientific, Fair Lawn, NJ; 0.05% Dextran T 70, Pharmacia, Uppsala, Sweden; in 0.1 mol/L potassium phosphate buffer, pH 7.4, plus 0.1% bovine serum albumin, RIA grade BSA, Sigma Chemical, St. Louis, MO) was added to all assay tubes simultaneously. The tubes were vortexed and then centrifuged at 4500 g for 20 min at 4°C. The supernatants were decanted into mini-vials, 4 mL of scintillation fluid (Scinti Verse E, Fisher Scientific) was added, vortexed, and counted in a scintillation counter (Packard, Model B4430, Downers Grove, IL). The sensitivity of the assay was 10 pg, and the average absolute binding (zero standard) was from 35 to 40%. Statistical analysis of data. -Data were analyzed by a twoway analysis of variance (ANOVA), where both factors, dose and time, were the repeated measures. In all of the analyses, there were four samples in each cell of the experimental design. The variability of PGE responses required transformations to natural logarithms (Sokal and Rohlf, 1981). Degrees of freedom for the main effects of dose, time, and their interaction in the analysis of variance were adjusted by the Greenhouse-Geisser method. All dose comparisons were made with reference to control values by use of Dunnett's t test and the appropriate error term from the analysis of variance.
Results. Fig. 1 shows the dose-response curves of PGE production by human PDL fibroblasts incubated with IL-1 3, IL-1a, TNF-
a, or IFN-y for 24 h. All of the doses tested for each of the cytokines showed significant PGE elevations, as compared with control (p
Mononuclear cell production of cytokines that stimulate fibroblast production of prostaglandin E (PGE) is an important mechanism by which mononuclear cells regulate fibroblast function. The objective of this investigation was to determine the effects of the cytokines interleukin 11 (IL-113), interleukin lot (IL-lo), tumor necrosis factor c (TNF-ot), and interferon y (IFN-y), alone or in paired combinations, on PGE production by near-confluent human periodontal ligament (PDL) fibroblasts in vitro. Premolars extracted in the course of orthodontic treatment were used for this study. Fibroblast cultures, free of epithelial cells, were obtained after the fourth subculture by the use of accurately-timed trypsin treatment. Cells in the fourth to sixth passage, incubated in DMEM supplemented with 10% equine serum, were used for these experiments. Cells (1 x 105) were seeded in 12- x -75-mm tissue culture tubes and incubated with various doses of IL-1r, IL-lx, TNF-cx, and IFNy, alone or in specific combinations, for 15 min, two, 12, 24, and 72 h. PGE concentrations in the media were measured by radio-immunoassay. The results showed that human PDL fibroblasts responded to the administration of cytokines by an elevation in the synthesis of PGE in a dose- and time-related fashion. The increase in PGE production was inhibited by the addition of indomethacin. The interactions between these cytokines varied in degree, depending on the particular combinations of cytokines. In addition, the administration of cytokine combinations was found to be additive, synergistic, subtractive, or suppressive on the production of PGE by PDL fibroblasts, depending on the duration of incubation. These experiments demonstrate the importance of the consideration of the interplay between cytokines produced by mononuclear cells on the mechanisms that regulate the functions of PDL fibroblasts. J Dent Res 69(8):1456-1462, August, 1990
Introduction. The main function of periodontal ligament (PDL) fibroblasts appears to be the maintenance and remodeling of the ligament itself and the establishment of a viable connection between alveolar bone and dental cementum. Fibroblasts and epithelial cells have been cultured from human PDL (Litwin et al., 1971; Blomlof and Otteskog, 1981; Ragnarsson et al., 1985). Ten Cate et al. (1976) have shown that PDL fibroblasts are capable of synthesizing, as well as degrading, collagen during physiological tooth movement. Rippin (1976) examined the rate of collagen turnover in young rat molars subjected to altered functional forces, and found an increased turnover of collagen throughout the width of the ligament, facilitating the Received for publication June 7, 1989 Accepted for publication December 4, 1989 4To whom correspondence and reprint requests should be addressed This investigation was supported by The Ohio State University College of Dentistry. 1456
tissue remodeling associated with tooth movement. Davidovitch et al. (1988) studied the localization of the cytokines interleukin 11 (IL-1p) and IL-lot in dental and paradental tissues surrounding orthodontically treated teeth in cats by an immunohistochemical procedure. They found a marked increase in staining intensity for IL-lot and 11 in cells around the tooth, particularly at compression sites. More direct evidence on the activity of PDL cells comes from in vitro work that demonstrates increased production of PGE and cAMP by fibroblasts undergoing mechanical stretch (Ngan et al., 1988). PDL fibroblasts have also been implicated in the process of tooth eruption; cAMP and PGE have been shown to be localized around the paradental tissues of erupting teeth in kittens (Gustafson et al., 1977; Shanfeld et al., 1986). The source of PGE around these teeth is not clear. However, a potential mechanism could be the production of PGE in response to IL-1. The cytokines, IL-1, tumor necrosis factor (TNF), and interferon (IFN) are released by cells of the immune system upon stimulation by a variety of agents. IL-1 and TNF are largely products of monocytes, whereas IFN-y is produced primarily by lymphocytes. Despite the fact that little, if any, homology exists among these molecules, which seem to bind to distinct receptors, they share some common functions. Fortunately, all of these cytokines are known as biological response modifiers that demonstrate pleiotropic effects. All are available as products of recombinant DNA technology, so that examination of their actions is not limited by availability or degree of purity. There are two distinct IL-1 gene products, IL-lx and IL-1P (March et al., 1985), with the latter being the major physiological form, which has been shown to be synonymous with osteoclast activating factor (Dewhirst, 1985). IL-1 and TNF-a have been shown to have synergistic effects on the production of PGE2 (Elias et al., 1987; Last-Barney et al., 1988), induction of bone resorption (Stashenko et al., 1987), and on growth arrest of melanoma cells in vitro (Ruggiero and Baglioni, 1987). In addition, IL-1 and TNF-cx have overlapping activities, including the regulation of phospholipid metabolism, induction of acute phase proteins, induction of prostacyclin synthesis, synthesis of hemopoietic growth factors, growth-promoting effects on fibroblasts, induction of fever, and regulation of the expression of human major histocompatibility antigens (Ruggiero and Baglioni, 1987; Le and Vilcek, 1987; Lovett et al., 1986). The chemical similarities and distinctions, as well as studies on the synergistic and overlapping activities of these two molecules, have been reviewed recently by Le and Vilcek (1987). TNF-at and IFN-y also share some functional activities. Both cytokines induce the re-organization of endothelial cells in monolayer culture (Stolpen et al., 1986) and activate polymorphonuclear neutrophil functions (Shalaby et al., 1985). IFNy also stimulates the production of TNF by cultured human peripheral blood monocytes, in contrast to IFN-at, which does not stimulate TNF production (Scuderi et al., 1987). However, IFN-y has been shown to inhibit certain inflammatory reactions, especially those associated with PGE2 synthesis and bone resorption (Boraschi et al., 1985; Gowen and Mundy, 1986).
Downloaded from jdr.sagepub.com at UNIV CALIFORNIA SANTA BARBARA on July 8, 2015 For personal use only. No other uses without permission.
CYTOKINE EFFECTS ON PDL FIBROBLAST PGE PRODUCTION
Vol. 69 No. 8
IFN-y has been shown to suppress IL-1-stimulated PGE production by human monocytes (Browing and Ribolini, 1987; Friteau et al., 1988), but not by human fibroblasts (Friteau et al., 1988). The objective of this study was to investigate the interactive effects of simultaneous administrations of paired cytokines, i.e., IL-lot, IL-1r, TNF-at, and IFN-y, on the production of PGE by human PDL fibroblasts in vitro.
Materials and methods. Collection and culture of human PDL fibroblasts. -PDL fibroblasts were prepared according to the method of Ragnarsson et al. (1985). Premolars, extracted from healthy male patients 16 to 25 years old in the course of orthodontic treatment, were obtained fresh from the Oral Surgery Department. The teeth were washed in Hanks' balanced salt solution (HBSS), and the gingival attachments were carefully removed with a sharp scalpel. The crowns (plus 2 mm past the gingival margins) were dipped in a 5.25% sodium hypochlorite solution for two min so that bacteria and any remaining gingival cells would be killed. The teeth were rinsed in three changes of HBSS. Each tooth was placed in a sterile 15-mL centrifuge tube with 5 mL of 0.1% collagenase (Cooper Biomedical, West Chester, PA) and incubated at 370C for six h. After incubation, the tube was centrifuged at 1000 g for ten min, the tooth was removed, and the cell pellet collected. The pellet was washed in complete medium consisting of Dulbecco's Modified Eagle Medium (DMEM, GIBCO, Grand Island, NY) containing 10% decomplemented equine serum (Hyclone Laboratories, Logan, UT) heat-inactivated at 560C for 30 min, 50 pLg/mL gentamycin (GIBCO), 50 ig/mL nystatin (GIBCO), and plated in 60-mm culture dishes. The dishes were incubated in a humid environment of 95% air, 5% C02, at 370C for 48 h, so that attachment would occur; the dishes were then washed free of debris and re-fed daily with complete medium. Culture characteristics and separation of cell type. -One human premolar or molar yielded enough cells to seed two 60mm dishes. Two populations of cells could be clearly distinguished after 48 h in culture, one having characteristic fibroblast morphology and the other epithelial morphology. In five to seven days, the cells grew to confluence, with the fibroblasts showing a greater rate of proliferation than the epithelial cells. When confluence was obtained, separation and subculturing of the cells were attempted by means of differential trypsinization. The dishes were washed with HBSS, and the cells were removed by incubation in 0.15% trypsin (GIBCO) for seven min at 37°C. The removal of cells was closely monitored by phase-contrast microscopy. Trypsin removes the fibroblasts more rapidly than the epithelial cells (Owens, 1974). In this manner, it was possible for a pure population of fibroblasts free of any epithelial cell contamination to be obtained (Ragnarsson et al.,
1985).
Characteristics of PDL cells.-Morphologically, as examined by phase-contrast microscopy, PDL cells appeared to be spindle-shaped and elongated in appearance, characteristics of "fibroblast-like" cells. Their doubling time as measured by Reichert's hemocytometer was approximately 36 to 48 h, reaching confluence in five days. Comparison of biosynthetic activity between bone cells (MC3T3-E1 cells) and PDL cells showed that the basal level of alkaline phosphatase activity (determined by a modified kinetic method, Sigma #245) was two-fold higher in bone cells (160 x 10-3 units ALP/105 cells vs. 78 x 10 -5units). PTH induces a 40-fold increase in cAMP activities in bone cells and a five-fold increase in PDL cells. However, both cells showed decreased alkaline phosphatase activity (25%) with PTH administration.
1457
Incubation and assay for PGE2. -Cells (1 x 105) were seeded on 12- x -75-mm plastic tissue-culture tubes. Satisfactory attachment of fibroblasts was usually obtained in six to 12 h. Various doses of recombinant human IL-13 (Immunex, Seattle, WA), recombinant human IL-lat (Immunex), recombinant human TNF-at (Genentech, San Francisco, CA), and recombinant human IFN--y (Genzyme, Boston, MA) were added to the test tubes for various time periods. The doses tested for IL-11 were 0.1 ng/mL (= 5.83 x 10-12 mol/L), 0.3 ng/mL (= 1.75 x 10-11 mol/L), and 1.0 ng/mL (= 5.83 x 10-1 mol/L); for IL-la, the doses were 0.3 ng/mL (= 1.75 x 10- l mol/L), 1.0 ng/mL (= 5.83 x 10-11 mol/L), and 3.0 ng/mL (= 1.75 x 10-10 mol/L); for TNF-at, the doses were 0.1 nmol/L (= 10-10 mol/L), 0.3 nmol/L (= 3 x 10-10 mol/L), and 1.0 nmol/L (= 10-9 mol/L); and for IFN-y, the doses were 10 U/mL (= 2.33 x 10-11 mol/L), 30 U/mL (= 7.0 x 10-11 mol/L), and 100 U/mL (= 2.33 x 10-10 mol/L). These recombinant cytokine preparations were relatively pure and free of contamination by lipopolysaccharide (LPS). The amount of LPS as determined by the limulus assay was in the order of pg/mg protein. In addition, the cytokine preparations were relatively stable in tissue culture medium, due to the hydrophobic core, and were resistant to proteolysis. The halflife of these proteins was in the order of days. PGE content in sample media was measured in duplicate by radio-immunoassay (RIA). Media aliquots were incubated with diluted rabbit anti-PGE serum (ICN ImmunoBiologicals, Lisle, IL; the term PGE instead of PGE2 is used here because the antibodies crossreact with PGE1) and 3H-PGE2 (10,000 dpm, New England Nuclear, Boston, MA) for 90 min at 4°C. After incubation, cold Dextran-coated charcoal suspension (0.5% Norit A, Fisher Scientific, Fair Lawn, NJ; 0.05% Dextran T 70, Pharmacia, Uppsala, Sweden; in 0.1 mol/L potassium phosphate buffer, pH 7.4, plus 0.1% bovine serum albumin, RIA grade BSA, Sigma Chemical, St. Louis, MO) was added to all assay tubes simultaneously. The tubes were vortexed and then centrifuged at 4500 g for 20 min at 4°C. The supernatants were decanted into mini-vials, 4 mL of scintillation fluid (Scinti Verse E, Fisher Scientific) was added, vortexed, and counted in a scintillation counter (Packard, Model B4430, Downers Grove, IL). The sensitivity of the assay was 10 pg, and the average absolute binding (zero standard) was from 35 to 40%. Statistical analysis of data. -Data were analyzed by a twoway analysis of variance (ANOVA), where both factors, dose and time, were the repeated measures. In all of the analyses, there were four samples in each cell of the experimental design. The variability of PGE responses required transformations to natural logarithms (Sokal and Rohlf, 1981). Degrees of freedom for the main effects of dose, time, and their interaction in the analysis of variance were adjusted by the Greenhouse-Geisser method. All dose comparisons were made with reference to control values by use of Dunnett's t test and the appropriate error term from the analysis of variance.
Results. Fig. 1 shows the dose-response curves of PGE production by human PDL fibroblasts incubated with IL-1 3, IL-1a, TNF-
a, or IFN-y for 24 h. All of the doses tested for each of the cytokines showed significant PGE elevations, as compared with control (p
