Archs oral Bid.
Vol. 35,lVo.5,pp. 387-395,1990
0003-9969/90 53.00+ 0.00
Printed in GreatBritain. All rightsreserved
Copyright 0 1990Pergamon Pressplc
EFFECTS OF CYTOKINES ON PROSTAGLANDIN E AND CAMP LEVELS IN HUMAN PERIODONTAL LIGAMENT FIBROBLASTS IN VITRO S. SAITO,’ P. NGAN,~* M. SAITO,’ K. KIM,] R. LANESE,~J. SHANFELD’and
Z. DAVIDOVITCH’
‘Department of Orthodontics, School of Dentistry, Showa University, Tokyo, Japan, Departments of *Orthodontics, 4Preventive Medicine and ‘Oral Biology, The Ohio State University, Columbus, OH 43210, U.S.A., I’Department of Orthodontics, College of Dentistry, Yon Sei University, Seoul, Korea (Accepted 31 October 1989)
Summary-The stimulation of PGE synthesis and CAMP production by cytokines have important physiological effects in many target tissues. The effects of interleukin-la and -11. tumour necrosis factor-a and interferon-y on PGE and CAMP production by periodontal ligament fibroblasts were studied. Fibroblasts in the 4th-dth passage, grown and maintained in DMEM supplemented with 10% equine serum, were incubated with graded doses of the various cytokines for 0.25, 0.5, 1, 2, 4, 24, 48 or 72 h. At the end of each incubation, PGE in the medium and the cellular content of CAMP were evaluated by a combined immunohistochemical microphotometric procedure, and conventional radiometric assays. The fibroblasts responded to all the cytokines with a dose- and time-related increase in the levels of PGE and CAMP. Such increases were inhibited by the inclusion of indomethacin in the medium. The addition of exogenous PiGE reversed that inhibition in respect of CAMP production. Immunohistochemical localization showed PGE predominantly in the cytoplasm and CAMP in the nucleus. These findings indicate that: (1) human periodontal ligament fibroblasts respond to these cytokines by increased synthesis of PGE and the production of CAMP; and (2) the CAMP production is secondary to the PGE synthesis. They suggest that these cytokines may regulate the function of these fibroblasts in physiological remodelling of the periodontium, as well as in inflammatory reactions. Key words: ‘cytokines, inflammation, growth, periodontium, fibroblasts.
INTRODUCTION
Periodontal ligament fibroblasts are a unique population of cells whose function is the maintenance and remodelling of the ligament itself, and also to some extent that of the associated alveolar bone and
cementurn. Ten Gate, Deporter and Freeman (1976) observed morphological features which suggested that these fibroblasts are capable of synthesizing as well as degrading collagen during physiological tooth movement. Bippin (1976) examined the rate of protein turnover in young rat molars subjected to altered functional forces and found it increased throughout the whole width of the ligament, which would facilitate the tissue remodelling associated with tooth movement. Boberts and Chase (1981) reported that the osteogenic component of the periodontal ligament cells responds to orthodontic force by increased proliferation as well as by differentiation of such cells into osteoblasts. *Address correspondence to: Dr P. Ngan, Department of Orthodontics, The Ohio State University, College of Dentistry, 305 W. 12th Avenue, Columbus, OH 43210, U.S.A. Abbreoiarions: ANOVA, analysis of variance; BSA, bovine serum albumin; DMEM, Dulbecco’s modified Eagle medium; HBSS, Hank’s balanced salt solution; IBMX, 3-isobutyl-l-meth yl xanthine; PG(E,, E2), prostaglandin (E,, E2).
Interleukin- 1, tumour necrosis factor-a and interferon-y are cytokines known to mediate a variety of actions important in host defence, inflammation and autoimmune responses. The interleukin and the necrosis factor are largely products of monocytes/ macrophages, whereas the interferon is produced primarily by lymphocytes. The pleiotropic biological effects of these cytokines on a variety of target tissues have been reviewed in detail (Dinarello, 1988; Le and Vilcek, 1987; Hughes and Baron, 1987). Augmented production of PGE, by fibroblastic cells in response to stimulation by cytokines is a prominent feature of inflammatory reactions (Whiteley and Needleman, 1984). The sequelae of PGE, release include vasodilatation, increased vascular permeability and altered immune cell function. In addition, PGE, acts in an autocrine manner to regulate fibroblast proliferation, collagen synthesis and generation of tissue factor such as transforming growth factor. Interleukin-1 is involved in the process of alveolar bone remodelling during orthodontic tooth movement, acting via PGE synthesis by periodontal ligament fibroblasts (Davidovitch er al., 1988). Human periodontal ligament fibroblasts respond to the administration of parathyroid hormone and interleukin-l/l with an increase in the synthesis of PGE and CAMP, indicating an interaction between hormone and paracrine/autocrine factors in their regulation (Ngan et al., 1988).
S. S~rro et al.
388
The mechanism of action of cytokines on PGE production by fibroblasts is not completely clear. Receptors have been identified for interleukin-1 and tumour necrosis factor (Vilcek et al., 1987; Dinarello and Savage, 1989) in other cell types. Interleukin-1 apparently activates arachidonic acid metabolism in fibroblasts by release of arachidonic acid from cellular phospholipids by specific phospholipases (Burch and Axelrod, 1987; Parker, Daniel and Waite, 1987), and through conversion by cyclooxygenase of released arachidonate into PGs (Raz et al., 1988). Cyclic AMP mediates the effects of various hormonal stimuli on the activity of target cells as second messengers (Kuehl, 1974). Interleukin-1 can induce the release of neutrophils from bone marrow in rats (Kapschmidt and Upchurch, 1982); the induction of synovial-cell plasminogen activator (Mochan, Uhl and Newton, 1986) is mediated by CAMP. In fibroblasts, the induction of interleukin-6 is mediated by CAMP-dependent protein kinase upon stimulation by interleukin-1 and tumour necrosis factor-cc (Zhang et al., 1988). Our objective was to determine whether human periodontal ligament fibroblasts respond to cytokines by altered production of PGE and CAMP.
MATERIALS AND METHODS
Collection of human periodontal ligament jibroblasts
Fibroblasts were prepared according to the method of Ragnarsson, Carr and Daniel (1985), with slight modifications, Premolars extracted in the course of orthodontic treatment were obtained fresh from the Oral Surgery department. The teeth were washed in HBSS, and the gingival attachments were carefully removed with a sharp scalpel. The crowns were dipped in a 5.25% sodium hypochlorite solution for 2 min to kill bacteria and any remaining gingival cells. The teeth were then rinsed in three changes of 0.1% collagenase (Cooper Biomedical, West Chester, Pa, U.S.A.) and incubated at 37°C for 6 h. After incubation, the tube was centrifuged at IOOOg for lOmin, the tooth was removed and the cell pellet collected. The pellet was washed in complete medium consisting of DMEM (GIBCO, Grand Island, N.Y., U.S.A.) containing 10% decomplemented equine serum (Hyclone Laboratories, Logan, Utah, U.S.A.; heat inactivated at 56°C for 30min), 50pg/ml gentamycin (GIBCO) and 50 pg/ml nystatin (GIBCO). It was then plated onto 60 mm culture dishes, which were incubated in a humid environment of 95% air, 5% CO,, at 37°C for 48 h, to permit attachment to occur, then washed free of debris and refilled daily with complete medium. To separate fibroblasts from epithelial cells, the cell masses were subcultured after differential trypsinization as follows. The dishes were washed with HBSS and the cells removed by incubation in 0.15% trypsin (GIBCO) for 7 min at 37°C; cell removal was closely monitored by phase contrast microscopy, which revealed that the trypsin removed the fibroblasts more rapidly than the epithelial cells (Owens, 1974). Cells in the 4th-6th passage were used for the experiments.
Incubation and assay for PGE
Cells (1 x 105; counted by haemocytometer) were seeded into 12 x 75 mm plastic tissue culture tubes. Satisfactory attachment of fibroblasts was usually obtained in 6-12 h. The doses tested for recombinant human interleukin-la (a gift from Dr S. Gillis; Immunex, Seattle, Wash., U.S.A.) were 0.3, 1.0, 3.0 and 10.0 ng/ml; those for recombinant human interleukin-l/I (a gift from Dr S. Gillis, Immunex) were 0.1, 0.3, 1.0 and 3.0 ng/ml; those for recombinant human tumour necrosis factor-u (Genentech, San Francisco, Calif., U.S.A.) were 0.1, 0.3, 1.0 and 3.0 nM; and those for recombinant human interferon-y (Genzyme, Boston, Mass., U.S.A.) were 10, 30, 100 and 300 U/ml. Indomethacin (a gift from Dr C. A. Stone, Merck Sharp & Dohme Research Laboratories, West Point, Pa., U.S.A.), when used, was added to a final concentration of 1 x 10m6M. These agents were added to the test tubes with 0.5% BSA (RIA grade, Sigma, St Louis, MO., U.S.A.) for 0.25, 0.5, 1, 2, 4, 24, 48 or 72 h. The PGE content of sample media was measured in duplicate by radioimmunoassay. Subsamples of media were incubated with diluted rabbit anti-PGE serum (ICN Immuno Biologicals, Lisle, Ill., U.S.A.tthe term PGE instead of PGE2 is used here because the antibodies cross-react with PGE,-and [3H]-PGE, (10,000 dis/min, New England Nuclear, Boston, Mass., U.S.A.) for 90min at 4°C. After incubation, cold dextran-coated charcoal suspension (0.5% Norit A, Fisher Scientific, Fair Lawn, N.J., U.S.A.; 0.05% dextran T 70, Pharmacia, Uppsala, Sweden; in 0.1 M potassium phosphate buffer, pH 7.4, plus 0.1% BSA) was added to all assay tubes simultaneously, 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 2, Fisher Scientific) were added, then the mixture was vortexed and measured in a scintillation counter (Packard, Model B4430, Downers Grove, Ill., U.S.A.). The sensitivity of the assay was lOpg, and the average absolute binding (zero standard) was 3540%. In order to determine if cytokines in the media have any effect on the PGE assay, 10.0 ng/ml of interleukin-1 ct, 3.0 ng/ml of interleukin-lp, 3.0 nM of tumour necrosis factor-cc or 300 U/ml of interferon-y were added; these additions had no effect on the quantification of PGE. Incubation and assay for CAMP
For the assay of CAMP, cells (1.5 x 105) were pre-incubated for 20 min in 1.Oml of DMEM containing 2.0mM IBMX (Sigma) and 0.5% BSA, pH 7.4, at 37°C. Interleukin-lcr at doses of 0.3, 1.0, 3.0 and 10.0 ng/ml; interleukin-l/3 at doses of 0.1,0.3, 1.O and 3.0 ng/ml; tumour necrosis factor-cc at doses of 0.3, 1.0, 3.0 and 10.0 nM and interferon-y at doses of 30, 100, 300 and 1000 U/ml were added to the test tubes for either 5, 15, 30 or 60 min. The reaction was terminated with 1.0 ml of cold buffer (50 mM tris-HCl, 5 mM EDTA and 0.2% BSA; pH 7.6). Cells were sonicated for 30 s and centrifuged 3000g for 10min. Subsamples of the supernatants were used to measure CAMP content by a modification of the protein binding method of Gilman (1970).
Cytokines and fibroblast PGE and CAMP Each subsample was incubated at 4°C for 120 min with CAMP-binding protein and [)H]-CAMP (10,000 dis/min, sp. act. 31.2 Ci/mmol, New England Nuclear). The binding protein was prepared from rabbit skeletal muscle in 5 mM EDTA (Sigma) according to the method of Miyamoto, Kuo and Greengard (1969), without purification, and diluted with 50 mM potassium phosphate containing 0.1% BSA, pH 7.4, to p.rovide 40-50% binding. After incubation, dextran-coated charcoal suspension was added to all assay tubes at the same time, and the assay completed in manner described above for PGE. The assay sensitivity was 0.1 pmol. Immunohistochemical staining and microphotometric measurement of cellular staining intensity Cells (2 x 104) were seeded on tissue-culture chamber/slides (Miles Scientific, Naperville, Ill., U.S.A.) and incubated with cytokines as described above. After freeze-drying, cells were incubated with normal goat serum (Cappel Laboratories, Malvern, Pa, U.S.A.), 2 mg total protein/ml, for 20 min, followed by overnight incubation with either anti-PGE (diluted 1:20) or anti-CAMP (diluted 1:20) mouse monoclonal antibod.ies at 4°C. The preparation of these antibodies in ascites fluid has been described by Davidovitch, Shanfeld and Lally (1982). After a 10 min rinse in tris buffer, the cells were incubated for 20 min at 22°C with peroxidase-conjugated goat antimouse IgG (Cappel Laboratories), 26.1 mg/ml total protein, diluted 1: 60. The presence of peroxidase was determined by the substrate 3,3’-diaminobenzidine (97-99% pure, Sigma). Stained cells were examined with a Zeiss Universal microscope fitted with a model 03 microphotometer, interfaced with a Zonax microcomputer (Carl Zeiss, Oberkochen, F.R.G.). Part of the light that reached the specimen was absorbed by the stained cells; the extent of absorption was dependent on the density of the 3,3’-diaminobenzidine reaction products in the field. A fixed aperture (5 pm), positioned in front of the sensor, controlled the diameter of the field to be screened. The instrument was adjusted to 100% transmission by focusing on the area of the slide devoid of cells. All measurements were performed with a 40x objective. Fifteen cells were examined microphotometrical1.y at 600 nm wavelength in the visible light range, which is optima1 for 3,3’diaminobenzidine. SpeciJicity tests for PGE and CAMP staining Cultured cells were incubated with 3,3’diaminobenzidine only, or with goat anti-mouse IgG conjugated to horseradish peroxidase, without prior incubation with the monoclonal antibodies to PGE or CAMP. Cellular staining was absent in both cases. In other control experiments, the monoclonal antibodies to PGE were incubated at 4°C with each of the following PGs: PGB, , PGEz, PGF,,, PGF,,, PGAz, PGB,-all at a concentration of 1 x 10m4M. After 12 h of incubation, the antibody-antigen mixture was then centrifuged (13,OOOg, 30 min) and the supernatant was placed over the cells as the first step of the immunohistochemical procedure. Incubation of the anti-PGE antibodies with PGs other than PGE, and PGE, did not prevent cellular staining. Similarly, the
389
monoclonal antibody to CAMP was incubated at 4°C with the following compounds: CAMP, cGMP, ATP, GTP, 5’-GMP, adenosine, guanosine, DNA and RNA at concentrations of 1 x 10m4, 10e5, and 10e6M. Such incubation of the anti-CAMP antibodies with compounds other than CAMP did not prevent cellular staining. Statistical analysis of data Data were analysed using a two-way ANOVA, where both factors, dose and time, were repeated measures. In all of the analysis, there were 4 samples in each cell of the experimental design. The variability of PGE and CAMP responses required transformations to natural logarithms. Degrees of freedom for the main effects of dose, time and their combination in the ANOVA were adjusted by the method of Geisser and Greenhouse (1958). All dose comparisons were made with reference to control values using Dunnett’s t-test and the appropriate error term from the ANOVA. RESULTS
The production of PGE by human periodontal ligament fibroblasts in response to the administration of each of the cytokines is shown in Text Figs l-4. In each illustration, the top figure (A) depicts the results of short incubation times (within 4 h) and the bottom figure (B) the results of long incubation times (24-72 h). The effect of interleukin- ICCis shown in Text Fig. 1. A significant increase in PGE production was observed after 1 h of incubation. The dose of 3.0 ng/ml produced peak PGE stimulation between 1 and 4 h of incubation. Long-term incubation showed a timerelated increase in PGE production with 0.3 ng/ml of interleukin-or. However, at higher concentrations (1 .O, 3.0 ng/ml), PGE production reached a maximum and started to form a plateau at 72 h of incubation. The effect of interleukin-l/l is shown in Text Fig. 2. At the concentrations tested (0.1-3.0 ng/ml), the cells responded in a dose- and time-dependent fashion; this was particularly prominent between 1 and 4 h of incubation. The dose of 1.0 ng/ml of interleukin-lg produced peak PGE stimulation; the dose of 3.0 ng/ml caused less response. In contrast to the time course with interleukin-lee, there was a small but significant increase in PGE production for the first 15 and 30 min, and then production increased rapidly for the next 4 h of incubation. Long-term administration of interleukin-lfl was associated with PGE production that reached a maximum after 24 h of incubation for concentrations of 1.0 ng/ml, 48 h for 0.3 ng/ml, and at 72 h for 0.1 ng/ml. The dose- and time-dependent experiments with tumour necrosis factor-cc (Text Fig. 3) showed a pattern similar to that of interleukin-1. This effect was dose-dependent, was most prominent between 1 and 4 h of incubation, and maximal stimulation was noted at 1.0 nM. PGE production in response to the administration of this factor was increased significantly (P < 0.01) as compared to control after 30 min of incubation, then reached a maximum after 48 h. The production of PGE by fibroblasts in response to interferon-y (Text Fig. 4) at the concentrations
390
S. 1.0.
A-O c---4
0.8.
e........ e...-..
--+
0.3 1.0 3.0 10.0
1
u-lo
..’
sAIlI
et al.
;5
,..”
(w/ml)
,,./’ (./
0.6.
3
”
0.25 0.6
”
I
24
i
i
i
42
72
Incubation
I
Time
46
(h)
Fig. 1. Dose- and time-dependent effects of interleukin-la (IL-la) on the synthesis of PGE by human periodontal ligament fibroblasts. Cells (1 x 105) were plated into tubes and incubated for 15min to 4 h (A) or 24-72 h (B). Values are mean & SE for 4 cultures. Note the scales of X and Y axis in (A) and (B) are different. Absence of error bar indicates that SE is less than the size of the symbol. Individual data that were too close together are presented by only one symbol instead of two or more symbols. Open circle and closed circle represent significant (p i 0.01) and no significant differences from control. respectively.
Incubation
Time
(h)
Fig. 3. Dose- and time-dependent effects of tumour necrosis factor-a (TNF-a) on the synthesis of PGE by human periodontal ligament fibroblasts; all information as in legend to Fig. 1. tested (l&300 U/ml) was dose- and time-dependent. The earliest incubation time to cause a significant PGE elevation was 1 h. Maximum stimulation was noted with 100 U/ml of interferon-y. Long-term incubations showed a lower response in PGE production after 48 h of incubation. At 72 h, such production for all the doses tested was not significantly different from that of control cells.
t
n
2c 1
I-
- ,.,
24
2.
46
Incubation
Time
(h) .’
Fig. 2. Dose- and time-dependent effects of interleukin-lp (IL-1B) on the synthesis of PGE by human periodontal ligament fibroblasts. Open circle represents significant (p < 0.01) difference from control; all other information as in legend to Fig. 1.
I
24
44 Incubation
Fig. 4. Dose- and time-dependent (IFN-y) on the synthesis of PGE . hgament tibroblasts; all information
72
Time
(h)
effects of interferon-y by human periodontal as in legend to Fig. 1.
391
Cytokines and fibroblast PGE and CAMP Table 1. Effect of cytokines on the synthesis of PGE in the presence or absence of indomethacin by human periodontal ligament fibroblasts Cytokine* IL-la IL-la IL-l/I IL-l/l TNF-a TNF-a IFN-y IFN-y
Table 3. Dose-dependent effect of cytokines (abbreviations as in Table 1) on the production of CAMP by human periodontal ligament fibroblasts Cvtokine
Indomethacin (1O-6 M) + + + +
PGE production 78 + 15.4t 37 * 4.53 245 k 38.0 42 + 6.8$ 398 + 42.3 56 & 9.2f 215 + 19.8 38 5 6.51 128 f 14.4 46 + 7.34
+
*Doses of cytokines a.re 3 ng/ml for interleukin-la (IL-la), 1 ng/ml for interleukin-l/l (IL-l/l), 1 nM for tumour necrosis factor-a (‘TNF-a), and 100 U/ml for interferon-y (IFN-y). tData are expressed as mean + SD @g PGE/104 cells, n = 4) after 2 h incubation. $Significantly different from the corresponding cytokines or vehicle in the abslence of indomethacin (p c: 0.01).
The effect of indomethacin is shown in Table 1; cytokine-induced PGE synthesis was inhibited by the addition of this compound (10m6M), and the level of PGE production was even lower than that of control cells. The localization of PGE in cells incubated with interleukin-l/l is shown in Plate Fig. 5. In control cells, the 3,3’-diaminobenzidine reaction products were distributed uniformly and the staining was light [Plate Fig. 51(A)]. When cells were incubated with interleukin-l,Il for 1 h, there was an increase in the cytoplasmic staining intensity for PGE [Plate Fig. 5(B)]. Further incubation (4 h) caused intense staining of the cytoplasm, nuclear membrane and the nucleus [Plate Fig. 5(C)]. Incubation with indomethacin ( 10e6 M) reversed the staining characteristics; cells incubated with 1.Ong/ml of interleukin- 1/I in the presence of indomethacin for 1 h showed a significant decrease in cytoplasmic staining [Plate Fig. 5(D) and Table 21. Table 2. Effect of interleukin-la (1.0 ng/ml) and indomethacin (10e6M) on the levels of PGE in human periodontal ligament fibroblasts using immunohistochemical/ microphotometric method
Control
Interleukin-l/3
Incubation time (h) 1 4 24 1 4 24
PGE production 83.6 + 1.85* 80.0 k 2.97 76.9 + 3.17 64.5 + 2.18t 59.2 + 2.6Ot 67.5 + 2.467
0
1 4 24
76.1 + 3.18 76.5 & 2.86 84.1 k 3.19t
*Data are expressed as mean f SD (% light transmitted, ?I =4). tsignificantly diffemnt from control at the same time period (p < 0.05).
CAMP production 0.45 * 0.04*
IL-la IL-la IL-la IL-la
0.3 1.0 3.0 (ng/ml) 10.0
6.58 + 12.00 k 7.92 k 3.96 k
IL-l/I IL-l/I IL-l/I IL-lb
0.1 0.3 1.O (ng/ml) 3.0
4.68 + 0.35t 14.56 + 2.04t 10.28 + 1.78t 6.32 & 0.88t
0.48t 1.48t 0.95t 0.56t
TNF-a TNF-a TNF-a TNF-a
0.3 1.0 3.0 (nM) 10.0
0.98 f 2.21 + 3.98 + 1.76 +
IFN-y IFN-y IFN-y IFN-Y
30 100 300 (U/ml) 1000
0.56 + 0.08 0.78 _+0.09t 1.78 + 0.30t 1.09+0.18t
0.12t 0.347 0.48t 0.297
*Data are expressed as mean + SD (pmobl.5 x lo4 cells, n = 4) after 1 h incubation. tSignificantly different from control (p < 0.01).
The dose effect of cytokines on CAMP production is shown in Table 3. The fibroblasts responded to the administration of each cytokine in a dose-dependent fashion. The doses of interleukin-la, -l/I, tumour necrosis factor-a and interferon-y that stimulated CAMP production were 1.Ong/ml, maximum 0.3 ng/ml, 3.0 nM and 300 U/ml, respectively. The time course for the cytokine-induced production of CAMP is shown in Text Fig. 6. Small but significant elevations in CAMP were seen as early as after 5 min incubation with interleukin- 1fi, tumour necrosis factor-a and interferon-y. At 30 min, all cytokines caused significant elevations in the production of CAMP. A further significant increase in the accumulation of CAMP was seen at 60min of incubation. In addition, differences in the level of CAMP production were found among the 4 cytokines tested; these differences were more apparent at 60 min of incubation. The effect of indomethacin on the production of CAMP by the 4 cytokines is shown in Table 4. In each .
10 f uu
9
‘0
3
5.
4
c 5 E 4
4 0
a q q S
IL-le (0.3 nglml) IL-la (1.0 nglml) TNF-a (3.0 nM) IFN-y (300 U/ml)
3 2 1 0
Interleukiwlj3 + indomethacin
Dose
-
6
15
30
50
Time (min) Fig. 6. Time course effect of cytokines on the production of CAMP in human periodontal ligament fibroblasts. Values are mean + SE for 4 cultures. *Significantly different from control (p < 0.01). *Significantly different from control (p < 0.05).
S.
392
SAITO et al.
Table 4. Effect of indomethacin on the cytokine-stimulated CAMP production by human periodontal ligament fibroblasts and the effect of addition of exogenous PGE (abbreviations as in Table 1) Cytokine* -
Indomethacin (10-6M) -
+ IL-la IL-la IL-la IL-la IL-l/? IL-ID IL-lb IL-ID TNF-a TNF-a TNF-a TNF-a IFN-y IFN-y IFN-y INF-y
+ + + + + -
PGE ( 1O-6 M) CAMP production 0.78 k 0.13t tumour necrosis factor-cc (1 .O nM = lO-9 M, 6.36 ng PGE/104 cells) > interferon-y (100 U/ml = 2.3 x lo-” M, 3.00 ng PGE/lO+’ cells). The time-course experiment for CAMP production showed that fibroblasts also responded to the 4 cytokines in a time-related fashion. Note that a shorter time course was used for the CAMP experiment because: (1) unlike PGE in the medium, CAMP does not accumulate in cells; and (2) the level of CAMP is regulated by the enzyme phospho-
diesterase and, in our experiment, cells were preincubated with IBMX to inhibit this enzyme. The potency of cytokine effects on CAMP production can be ranked as follows (Table 3): interleukin-l/l (0.3 ng/ml, 14.56 pmol/l.5 x lo4 cells) > interleukinICC (1.0 ng/ml, 12.00 pmol/l.5 x lo4 cells) > tumour necrosis factor-cc (3.0 nM, 3.98 pmol/l.5 x lo4 cells) > interferon-y (300 U/ml, 1.78 pmol/l.5 x lo4 cells). The binding of cytokines to cell surface receptors is the first step in the initiation of cytokine action on target cells, so our findings, together with those from receptor studies in other cell systems (Vilcek et al., 1987; Dinarello and Savage, 1989) suggest that specific receptors for these ligands are present on periodontal ligament cells. The mechanism of action of cytokines on fibroblast PGE production is not completely clear. Interleukin1 may stimulate PGE synthesis by hydrolysis of membrane phospholipid by specific phospholipases with release of arachidonic acid substrate (Burch and Axelrod, 1987; Parker et al., 1987), or by conversion of release of arachidonate into PGs by cyclooxygenase activity (Raz et al., 1988). Interleukin-1 also increases the synthetic rate of newly formed cycle-oxygenase, suggesting that its effect on PGE synthesis is mediated mainly, if not solely, via induction of cycle-oxygenase synthesis (Korn et al., 1989). We also found a concomitant increase in the production of CAMP with all 4 cytokines; prominent increases in CAMP levels were seen at 60min of incubation. This elevation of CAMP is seemingly dependent on the synthesis of PGE because it was inhibited by the addition of indomethacin, and because this inhibition could be reversed by the addition of exogenous PGE (Table 4). Cyclic AMP levels increase in response to PGE administration via the adenylate cyclase system (Sutherland and Rall, 1960). On the other hand, the effects of tumour necrosis factor and interleukin-I on CAMP accumulation occur through receptor-G protein interactions (Zhang et al., 1988). We attempted to localize and semi-quantitate the effect of interleukin-1 jj’ on the level of PGE and CAMP in periodontal ligament by using an immunohistochemical procedure. We found an increase in staining for PGE in the cytoplasm and to a lesser extent in the nucleus after 60 min of incubation with the interleukin-l/I. Normally, PGE is not stored in cells but is released into the extracellular space (Piper and Vane, 1971). The staining for PGE found in these cells suggests that it is either bound to surface receptors and/or being taken up from the
extracellular medium by a process of internalization (Rao and Cheginin, 1983). PGE receptors have been described in the cell plasma membrane in a number of tissues, including alveolar bone (Moore and Wolfe, 1973; Rao, 1974; Yoshizawa, Abiko and Takiguchi, 1983). PGE binding sites have also been found in subcellular organelles such as endoplasmic reticulum, Golgi apparatus (Mitra and Rao, 1978a), lysosomes (Mitra and Rao, 1978b) and in nuclei (Rao and Cheginin, 1983). Interleukin-l/l enhanced both the cytoplasmic and nuclear staining for CAMP in the periodontal ligament cells. This increase in staining intensity indicates a true elevation in the synthesis of CAMP, as confirmed by the protein binding assay. The nuclear CAMP staining merits further explanation, because AMP binds to the regulatory subunits of protein kinase and releases the catalytic subunits. In one study, Gancedo, Mazon and Eraso (1985) have shown that the binding of CAMP to the regulatory subunits in prokaryotes might function directly as intranuclear regulatory proteins. The increases in PGE and CAMP production in response to all of the cytokines tested suggest that the function of periodontal ligament fibroblasts may be regulated, at least in part, by these cytokines, either alone or in concert with systemic hormones. There are indications that the effect of these cytokines on fibroblast function can be synergistic (Elias et al., 1988) additional or overlapping (Beresini, Lempert and Epstein, 1988). REFERENCES
Beresini M., Lempert M. J. and Epstein L. B. (1988) Overlapping polypeptide induction in human fibroblasts in response to treatment with interferon-a, interferon-y, interleukin la, interleukin Ip. and tumor necrosis factor. J. Immunol. 140, 485-493. Burch R. M. and Axelrod J. (1987) Dissociation of bradykinin-induced prostaglandin formation from phosphatidylinositol regulation of phospholipase A,. Proc. natn. Acad. Sci. U.S.A. 84, 6374-6378. Davidovitch Z., Shanfeld J. and Lally E. (1982) Prostaglandin Fz. is associated with alveolar bone cells: immunohistochemical evidence utilizing monoclonal antibodies. In: Factors and Mechanisms InJuencing Bone Growrh (Edited by Dixon A. D. and Sarnat B. G.) pp. 125-134. Liss, New York. Davidovitch Z., Nicolay O., Ngan P. W. and Shanfeld J. L. (1988) Neurotransmitters, cytokines and the control of alveolar bone remodeling in orthodontics. Denr. Clin. North Am. 32, 41 l-435. Dinarello C. A. (1988) Biology of interleukin-1. FASEB J. 2, 108-l 15. Dinarello C. A. and Savage N. (1989) Interleukin-I and its receptor. CRC Crit. R&J. Immunol. 9, l-20. Elias J. A.. Gustilo K.. Baeder W. and Freundlich B. (1988) Synergistic stimulation of fibroblast prostaglandin pro: duction by recombinant interleukin 1 and tumor necrosis factor. J. Immunol. 138, 3812-3816. Gancedo J. M., Mazon M. J. and Eraso P. (1985) Biological roles of CAMP: similarities and differences between organisms. Trends Biochem. Sci. 10, 210-212. Geisser S. and Greenhouse S. W. (1958) An extension of Box’s results on the use of the F distribution in multivariate analysis. Ann. Math. Stat. 29, 885-891. Gilman A. G. (1970) A protein binding assay for adenosine 3’,5’-monophosphate. Proc. natn. Acad. Sci. U.S.A. 67, 305-312.
Cytokines and fibroblast PGE and CAMP Hughes T. K. and Baron S. (1987) Do the interferons act singly or in combination? J. Interferon Res. 7, 603614. Kapschmidt R. F. and Upchurch H. C. (1982) Neutrophil release after injecti’sns of endotoxins of leukocytes endogeneous mediators into rats. J. Reric. Sot. 28, 191-201. Korn J. H., Downie El., Roth G. J. and Ho Shih-Yieh (1989) Synergy of interleukin 1 (IL-I) with arachidonic acid and A23187 in stimulating PGE synthesis in human fibroblasts: IL-1 stimulates fibroblast cyclooxygenase. Clin. Immunol. Immunopath. 50, 196-204.
Kuehl F. A. Jr (1974:l Prostaglandins, cyclic nucleotides and cell function. Prostaglandins 5, 325-340. Le J. and Vilcek J. (1987) Biology of disease: tumor necrosis factor and interleukin 1; cytokines with multiple overlapping biological activities. Lab. Invest. 56, 234-248. Mitra S. and Rae C. (1978a) Gonadotropin and prostaglandins binding sites in rough endoplasmic reticulum and golgi fractions of bovine corpore luteum. Archs Biochem. Biophys. 191, 332-340. Mitra S. and Rao C. (1978b) Receptors for gonadotropins and prostaglandins in lysosomes of bovine corporan luteum. Archs Biochem. Biophys. 185, 126-133. Miyamoto E., Kuo J. F. and Greengard P. (1969) Cyclic nucleotide dependent protein kinases. J. biol. Chem. 244, 6395-6409.
Mizel S. B., Kilian I’. L., Lewis J. C., Paganelli K. A. and Chizzonite R. A. (1987) The interleukin 1 receptor. Dynamics of interleukin 1 binding and internalization in T cells and fibroblasts. .I. Immunol. 138, 29062912. Mochan E., Uhl J. and Newton R. (1986) Evidence that interleukin inductions of synovial cell plasminogen activator is mediated via prostaglandin E and CAMP. Arfhritis & Rheum. 29, 1078-1083.
Moore W. and WolFe I. (1973) Binding of prostaglandin E, to beef thyroid membranes. J. biol. Chem. 248, 570555711. Newton R. C. and Covington M. (1987) The activation of human fibroblast prostaglandin E production by interleukin 1. Cell. Imm&ol. li0, 338-349. Nean P. W.. Zadeh Y. Z.. Shanfeld J. and Davidovitch 2. 71988) The effect of interleukin-1 fi and parathyroid hormone on cyclic nucleotide and prostaglandin levels in human periodontal ligament fibroblasts in vitro. In: The Biological Mechanisms of Tooth Eruption and Root Resorption (Edited by Davidovitch Z.) pp. 261-267.
ESCO Media, Birmingham, Ala. Owens R. B. (197411Glandular epithelial cells from mice: a method for selective cultivation. J. natn. Cancer Inst. 52, 1375-1378.
395
Parker J., Daniel L. W. and Waite M. (1987) Evidence of proteinkinase C involvement in phorbol diester-stimulated arachidonic acid release and prostaglandin synthesis. J. biol. Chem. 262, 5385-5390. Piper P. and Vane J. (197 1) The release of prostaglandins from lung and other tissues. Ann. N. Y. Acad. Sci. H&363-385. Ragnarsson B., Carr G. and Daniel J. C. (1985) Isolation and growth of human periodontal ligament cells in vitro. J. dent. Res. 64, 1026-1030. Rao C (1974) Characterization of prostaglandin receptors in the bovine corpus luteum cell membrane. J. biol. Chem. 249, 7203-7209.
Rao C. and Cheginin N. (1983) Gonadotropins and prostaglandins binding sites in nuclei of bovine corpora luteum. Biochim. biophys. Acta 584, 454466. Raz A., Wyche A., Siegel N. and Needleman P. (1988) Regulation of fibroblast cyclooxygenasc synthesis by interleukin-1. J. biol. Chem. 263, 3022-3028. Rippin J. W. (1976) Collagen turnover in the periodontal ligament under normal and altered functional forces. J. periodont. Res. 11, 101-107. Roberts W. E. and Chase D. C. (1981) Kinetics of cell proliferation and migration associated with orthodontically-induced osteogenesis. J. dent. Res. 60, 174-18 I. Sutherland E. W. and Rail T. W. (1960) The relation of adenosine 3’,5’-phosphate and phosphorylase to the actions of catecholamines and other hormones. Pharmac. Rev. 12, 265-299.
Ten Cate A. R., Deporter D. A. and Freeman E. (1976) The role of fibroblasts in the remodeling of periodontal ligament during physiological tooth movement. Am. J. Orthod. 69, 155-167.
Vilcek J., Tsujumoto M., Palombella V. L., Kohase M. and Le J. (1987) Tumor necrosis factor: Receptor binding and mitogenic action in fibroblasts. J. Cell. Physiol. Suppl. 5, 57-61. Weissmann G., Smolen J. E. and Korchak H. M. (1980) Release of inflammatory mediators from stimulated neutrophils. New Engl. J. Med 303, 27-34. Whiteley P. J. and Needleman P. (1984) Mechanism of enhanced fibroblast arachidonic acid metabolism by mononuclear cell factor. J. clin. Invest. 74, 2249-2253. Yoshizawa Y., Abiko Y. and Takiguchi H. (1983) Some properties of prostaglandin E, receptors in rabbit alveolar bone cell membrane. Inr. J. Biochem. 15. 991-995. Zhang Y., Lin J. X., Yip Y. K. and Vilcek J. (1988) Enhancement of CAMP levels and of protein kinase activity by tumor necrosis factor and interleukin 1 in human fibroblasts: role in the induction of interleukin 6. Proc. natn. Acad. Sci. U.S.A. 85, 6802-6805.
Vol. 35,lVo.5,pp. 387-395,1990
0003-9969/90 53.00+ 0.00
Printed in GreatBritain. All rightsreserved
Copyright 0 1990Pergamon Pressplc
EFFECTS OF CYTOKINES ON PROSTAGLANDIN E AND CAMP LEVELS IN HUMAN PERIODONTAL LIGAMENT FIBROBLASTS IN VITRO S. SAITO,’ P. NGAN,~* M. SAITO,’ K. KIM,] R. LANESE,~J. SHANFELD’and
Z. DAVIDOVITCH’
‘Department of Orthodontics, School of Dentistry, Showa University, Tokyo, Japan, Departments of *Orthodontics, 4Preventive Medicine and ‘Oral Biology, The Ohio State University, Columbus, OH 43210, U.S.A., I’Department of Orthodontics, College of Dentistry, Yon Sei University, Seoul, Korea (Accepted 31 October 1989)
Summary-The stimulation of PGE synthesis and CAMP production by cytokines have important physiological effects in many target tissues. The effects of interleukin-la and -11. tumour necrosis factor-a and interferon-y on PGE and CAMP production by periodontal ligament fibroblasts were studied. Fibroblasts in the 4th-dth passage, grown and maintained in DMEM supplemented with 10% equine serum, were incubated with graded doses of the various cytokines for 0.25, 0.5, 1, 2, 4, 24, 48 or 72 h. At the end of each incubation, PGE in the medium and the cellular content of CAMP were evaluated by a combined immunohistochemical microphotometric procedure, and conventional radiometric assays. The fibroblasts responded to all the cytokines with a dose- and time-related increase in the levels of PGE and CAMP. Such increases were inhibited by the inclusion of indomethacin in the medium. The addition of exogenous PiGE reversed that inhibition in respect of CAMP production. Immunohistochemical localization showed PGE predominantly in the cytoplasm and CAMP in the nucleus. These findings indicate that: (1) human periodontal ligament fibroblasts respond to these cytokines by increased synthesis of PGE and the production of CAMP; and (2) the CAMP production is secondary to the PGE synthesis. They suggest that these cytokines may regulate the function of these fibroblasts in physiological remodelling of the periodontium, as well as in inflammatory reactions. Key words: ‘cytokines, inflammation, growth, periodontium, fibroblasts.
INTRODUCTION
Periodontal ligament fibroblasts are a unique population of cells whose function is the maintenance and remodelling of the ligament itself, and also to some extent that of the associated alveolar bone and
cementurn. Ten Gate, Deporter and Freeman (1976) observed morphological features which suggested that these fibroblasts are capable of synthesizing as well as degrading collagen during physiological tooth movement. Bippin (1976) examined the rate of protein turnover in young rat molars subjected to altered functional forces and found it increased throughout the whole width of the ligament, which would facilitate the tissue remodelling associated with tooth movement. Boberts and Chase (1981) reported that the osteogenic component of the periodontal ligament cells responds to orthodontic force by increased proliferation as well as by differentiation of such cells into osteoblasts. *Address correspondence to: Dr P. Ngan, Department of Orthodontics, The Ohio State University, College of Dentistry, 305 W. 12th Avenue, Columbus, OH 43210, U.S.A. Abbreoiarions: ANOVA, analysis of variance; BSA, bovine serum albumin; DMEM, Dulbecco’s modified Eagle medium; HBSS, Hank’s balanced salt solution; IBMX, 3-isobutyl-l-meth yl xanthine; PG(E,, E2), prostaglandin (E,, E2).
Interleukin- 1, tumour necrosis factor-a and interferon-y are cytokines known to mediate a variety of actions important in host defence, inflammation and autoimmune responses. The interleukin and the necrosis factor are largely products of monocytes/ macrophages, whereas the interferon is produced primarily by lymphocytes. The pleiotropic biological effects of these cytokines on a variety of target tissues have been reviewed in detail (Dinarello, 1988; Le and Vilcek, 1987; Hughes and Baron, 1987). Augmented production of PGE, by fibroblastic cells in response to stimulation by cytokines is a prominent feature of inflammatory reactions (Whiteley and Needleman, 1984). The sequelae of PGE, release include vasodilatation, increased vascular permeability and altered immune cell function. In addition, PGE, acts in an autocrine manner to regulate fibroblast proliferation, collagen synthesis and generation of tissue factor such as transforming growth factor. Interleukin-1 is involved in the process of alveolar bone remodelling during orthodontic tooth movement, acting via PGE synthesis by periodontal ligament fibroblasts (Davidovitch er al., 1988). Human periodontal ligament fibroblasts respond to the administration of parathyroid hormone and interleukin-l/l with an increase in the synthesis of PGE and CAMP, indicating an interaction between hormone and paracrine/autocrine factors in their regulation (Ngan et al., 1988).
S. S~rro et al.
388
The mechanism of action of cytokines on PGE production by fibroblasts is not completely clear. Receptors have been identified for interleukin-1 and tumour necrosis factor (Vilcek et al., 1987; Dinarello and Savage, 1989) in other cell types. Interleukin-1 apparently activates arachidonic acid metabolism in fibroblasts by release of arachidonic acid from cellular phospholipids by specific phospholipases (Burch and Axelrod, 1987; Parker, Daniel and Waite, 1987), and through conversion by cyclooxygenase of released arachidonate into PGs (Raz et al., 1988). Cyclic AMP mediates the effects of various hormonal stimuli on the activity of target cells as second messengers (Kuehl, 1974). Interleukin-1 can induce the release of neutrophils from bone marrow in rats (Kapschmidt and Upchurch, 1982); the induction of synovial-cell plasminogen activator (Mochan, Uhl and Newton, 1986) is mediated by CAMP. In fibroblasts, the induction of interleukin-6 is mediated by CAMP-dependent protein kinase upon stimulation by interleukin-1 and tumour necrosis factor-cc (Zhang et al., 1988). Our objective was to determine whether human periodontal ligament fibroblasts respond to cytokines by altered production of PGE and CAMP.
MATERIALS AND METHODS
Collection of human periodontal ligament jibroblasts
Fibroblasts were prepared according to the method of Ragnarsson, Carr and Daniel (1985), with slight modifications, Premolars extracted in the course of orthodontic treatment were obtained fresh from the Oral Surgery department. The teeth were washed in HBSS, and the gingival attachments were carefully removed with a sharp scalpel. The crowns were dipped in a 5.25% sodium hypochlorite solution for 2 min to kill bacteria and any remaining gingival cells. The teeth were then rinsed in three changes of 0.1% collagenase (Cooper Biomedical, West Chester, Pa, U.S.A.) and incubated at 37°C for 6 h. After incubation, the tube was centrifuged at IOOOg for lOmin, the tooth was removed and the cell pellet collected. The pellet was washed in complete medium consisting of DMEM (GIBCO, Grand Island, N.Y., U.S.A.) containing 10% decomplemented equine serum (Hyclone Laboratories, Logan, Utah, U.S.A.; heat inactivated at 56°C for 30min), 50pg/ml gentamycin (GIBCO) and 50 pg/ml nystatin (GIBCO). It was then plated onto 60 mm culture dishes, which were incubated in a humid environment of 95% air, 5% CO,, at 37°C for 48 h, to permit attachment to occur, then washed free of debris and refilled daily with complete medium. To separate fibroblasts from epithelial cells, the cell masses were subcultured after differential trypsinization as follows. The dishes were washed with HBSS and the cells removed by incubation in 0.15% trypsin (GIBCO) for 7 min at 37°C; cell removal was closely monitored by phase contrast microscopy, which revealed that the trypsin removed the fibroblasts more rapidly than the epithelial cells (Owens, 1974). Cells in the 4th-6th passage were used for the experiments.
Incubation and assay for PGE
Cells (1 x 105; counted by haemocytometer) were seeded into 12 x 75 mm plastic tissue culture tubes. Satisfactory attachment of fibroblasts was usually obtained in 6-12 h. The doses tested for recombinant human interleukin-la (a gift from Dr S. Gillis; Immunex, Seattle, Wash., U.S.A.) were 0.3, 1.0, 3.0 and 10.0 ng/ml; those for recombinant human interleukin-l/I (a gift from Dr S. Gillis, Immunex) were 0.1, 0.3, 1.0 and 3.0 ng/ml; those for recombinant human tumour necrosis factor-u (Genentech, San Francisco, Calif., U.S.A.) were 0.1, 0.3, 1.0 and 3.0 nM; and those for recombinant human interferon-y (Genzyme, Boston, Mass., U.S.A.) were 10, 30, 100 and 300 U/ml. Indomethacin (a gift from Dr C. A. Stone, Merck Sharp & Dohme Research Laboratories, West Point, Pa., U.S.A.), when used, was added to a final concentration of 1 x 10m6M. These agents were added to the test tubes with 0.5% BSA (RIA grade, Sigma, St Louis, MO., U.S.A.) for 0.25, 0.5, 1, 2, 4, 24, 48 or 72 h. The PGE content of sample media was measured in duplicate by radioimmunoassay. Subsamples of media were incubated with diluted rabbit anti-PGE serum (ICN Immuno Biologicals, Lisle, Ill., U.S.A.tthe term PGE instead of PGE2 is used here because the antibodies cross-react with PGE,-and [3H]-PGE, (10,000 dis/min, New England Nuclear, Boston, Mass., U.S.A.) for 90min at 4°C. After incubation, cold dextran-coated charcoal suspension (0.5% Norit A, Fisher Scientific, Fair Lawn, N.J., U.S.A.; 0.05% dextran T 70, Pharmacia, Uppsala, Sweden; in 0.1 M potassium phosphate buffer, pH 7.4, plus 0.1% BSA) was added to all assay tubes simultaneously, 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 2, Fisher Scientific) were added, then the mixture was vortexed and measured in a scintillation counter (Packard, Model B4430, Downers Grove, Ill., U.S.A.). The sensitivity of the assay was lOpg, and the average absolute binding (zero standard) was 3540%. In order to determine if cytokines in the media have any effect on the PGE assay, 10.0 ng/ml of interleukin-1 ct, 3.0 ng/ml of interleukin-lp, 3.0 nM of tumour necrosis factor-cc or 300 U/ml of interferon-y were added; these additions had no effect on the quantification of PGE. Incubation and assay for CAMP
For the assay of CAMP, cells (1.5 x 105) were pre-incubated for 20 min in 1.Oml of DMEM containing 2.0mM IBMX (Sigma) and 0.5% BSA, pH 7.4, at 37°C. Interleukin-lcr at doses of 0.3, 1.0, 3.0 and 10.0 ng/ml; interleukin-l/3 at doses of 0.1,0.3, 1.O and 3.0 ng/ml; tumour necrosis factor-cc at doses of 0.3, 1.0, 3.0 and 10.0 nM and interferon-y at doses of 30, 100, 300 and 1000 U/ml were added to the test tubes for either 5, 15, 30 or 60 min. The reaction was terminated with 1.0 ml of cold buffer (50 mM tris-HCl, 5 mM EDTA and 0.2% BSA; pH 7.6). Cells were sonicated for 30 s and centrifuged 3000g for 10min. Subsamples of the supernatants were used to measure CAMP content by a modification of the protein binding method of Gilman (1970).
Cytokines and fibroblast PGE and CAMP Each subsample was incubated at 4°C for 120 min with CAMP-binding protein and [)H]-CAMP (10,000 dis/min, sp. act. 31.2 Ci/mmol, New England Nuclear). The binding protein was prepared from rabbit skeletal muscle in 5 mM EDTA (Sigma) according to the method of Miyamoto, Kuo and Greengard (1969), without purification, and diluted with 50 mM potassium phosphate containing 0.1% BSA, pH 7.4, to p.rovide 40-50% binding. After incubation, dextran-coated charcoal suspension was added to all assay tubes at the same time, and the assay completed in manner described above for PGE. The assay sensitivity was 0.1 pmol. Immunohistochemical staining and microphotometric measurement of cellular staining intensity Cells (2 x 104) were seeded on tissue-culture chamber/slides (Miles Scientific, Naperville, Ill., U.S.A.) and incubated with cytokines as described above. After freeze-drying, cells were incubated with normal goat serum (Cappel Laboratories, Malvern, Pa, U.S.A.), 2 mg total protein/ml, for 20 min, followed by overnight incubation with either anti-PGE (diluted 1:20) or anti-CAMP (diluted 1:20) mouse monoclonal antibod.ies at 4°C. The preparation of these antibodies in ascites fluid has been described by Davidovitch, Shanfeld and Lally (1982). After a 10 min rinse in tris buffer, the cells were incubated for 20 min at 22°C with peroxidase-conjugated goat antimouse IgG (Cappel Laboratories), 26.1 mg/ml total protein, diluted 1: 60. The presence of peroxidase was determined by the substrate 3,3’-diaminobenzidine (97-99% pure, Sigma). Stained cells were examined with a Zeiss Universal microscope fitted with a model 03 microphotometer, interfaced with a Zonax microcomputer (Carl Zeiss, Oberkochen, F.R.G.). Part of the light that reached the specimen was absorbed by the stained cells; the extent of absorption was dependent on the density of the 3,3’-diaminobenzidine reaction products in the field. A fixed aperture (5 pm), positioned in front of the sensor, controlled the diameter of the field to be screened. The instrument was adjusted to 100% transmission by focusing on the area of the slide devoid of cells. All measurements were performed with a 40x objective. Fifteen cells were examined microphotometrical1.y at 600 nm wavelength in the visible light range, which is optima1 for 3,3’diaminobenzidine. SpeciJicity tests for PGE and CAMP staining Cultured cells were incubated with 3,3’diaminobenzidine only, or with goat anti-mouse IgG conjugated to horseradish peroxidase, without prior incubation with the monoclonal antibodies to PGE or CAMP. Cellular staining was absent in both cases. In other control experiments, the monoclonal antibodies to PGE were incubated at 4°C with each of the following PGs: PGB, , PGEz, PGF,,, PGF,,, PGAz, PGB,-all at a concentration of 1 x 10m4M. After 12 h of incubation, the antibody-antigen mixture was then centrifuged (13,OOOg, 30 min) and the supernatant was placed over the cells as the first step of the immunohistochemical procedure. Incubation of the anti-PGE antibodies with PGs other than PGE, and PGE, did not prevent cellular staining. Similarly, the
389
monoclonal antibody to CAMP was incubated at 4°C with the following compounds: CAMP, cGMP, ATP, GTP, 5’-GMP, adenosine, guanosine, DNA and RNA at concentrations of 1 x 10m4, 10e5, and 10e6M. Such incubation of the anti-CAMP antibodies with compounds other than CAMP did not prevent cellular staining. Statistical analysis of data Data were analysed using a two-way ANOVA, where both factors, dose and time, were repeated measures. In all of the analysis, there were 4 samples in each cell of the experimental design. The variability of PGE and CAMP responses required transformations to natural logarithms. Degrees of freedom for the main effects of dose, time and their combination in the ANOVA were adjusted by the method of Geisser and Greenhouse (1958). All dose comparisons were made with reference to control values using Dunnett’s t-test and the appropriate error term from the ANOVA. RESULTS
The production of PGE by human periodontal ligament fibroblasts in response to the administration of each of the cytokines is shown in Text Figs l-4. In each illustration, the top figure (A) depicts the results of short incubation times (within 4 h) and the bottom figure (B) the results of long incubation times (24-72 h). The effect of interleukin- ICCis shown in Text Fig. 1. A significant increase in PGE production was observed after 1 h of incubation. The dose of 3.0 ng/ml produced peak PGE stimulation between 1 and 4 h of incubation. Long-term incubation showed a timerelated increase in PGE production with 0.3 ng/ml of interleukin-or. However, at higher concentrations (1 .O, 3.0 ng/ml), PGE production reached a maximum and started to form a plateau at 72 h of incubation. The effect of interleukin-l/l is shown in Text Fig. 2. At the concentrations tested (0.1-3.0 ng/ml), the cells responded in a dose- and time-dependent fashion; this was particularly prominent between 1 and 4 h of incubation. The dose of 1.0 ng/ml of interleukin-lg produced peak PGE stimulation; the dose of 3.0 ng/ml caused less response. In contrast to the time course with interleukin-lee, there was a small but significant increase in PGE production for the first 15 and 30 min, and then production increased rapidly for the next 4 h of incubation. Long-term administration of interleukin-lfl was associated with PGE production that reached a maximum after 24 h of incubation for concentrations of 1.0 ng/ml, 48 h for 0.3 ng/ml, and at 72 h for 0.1 ng/ml. The dose- and time-dependent experiments with tumour necrosis factor-cc (Text Fig. 3) showed a pattern similar to that of interleukin-1. This effect was dose-dependent, was most prominent between 1 and 4 h of incubation, and maximal stimulation was noted at 1.0 nM. PGE production in response to the administration of this factor was increased significantly (P < 0.01) as compared to control after 30 min of incubation, then reached a maximum after 48 h. The production of PGE by fibroblasts in response to interferon-y (Text Fig. 4) at the concentrations
390
S. 1.0.
A-O c---4
0.8.
e........ e...-..
--+
0.3 1.0 3.0 10.0
1
u-lo
..’
sAIlI
et al.
;5
,..”
(w/ml)
,,./’ (./
0.6.
3
”
0.25 0.6
”
I
24
i
i
i
42
72
Incubation
I
Time
46
(h)
Fig. 1. Dose- and time-dependent effects of interleukin-la (IL-la) on the synthesis of PGE by human periodontal ligament fibroblasts. Cells (1 x 105) were plated into tubes and incubated for 15min to 4 h (A) or 24-72 h (B). Values are mean & SE for 4 cultures. Note the scales of X and Y axis in (A) and (B) are different. Absence of error bar indicates that SE is less than the size of the symbol. Individual data that were too close together are presented by only one symbol instead of two or more symbols. Open circle and closed circle represent significant (p i 0.01) and no significant differences from control. respectively.
Incubation
Time
(h)
Fig. 3. Dose- and time-dependent effects of tumour necrosis factor-a (TNF-a) on the synthesis of PGE by human periodontal ligament fibroblasts; all information as in legend to Fig. 1. tested (l&300 U/ml) was dose- and time-dependent. The earliest incubation time to cause a significant PGE elevation was 1 h. Maximum stimulation was noted with 100 U/ml of interferon-y. Long-term incubations showed a lower response in PGE production after 48 h of incubation. At 72 h, such production for all the doses tested was not significantly different from that of control cells.
t
n
2c 1
I-
- ,.,
24
2.
46
Incubation
Time
(h) .’
Fig. 2. Dose- and time-dependent effects of interleukin-lp (IL-1B) on the synthesis of PGE by human periodontal ligament fibroblasts. Open circle represents significant (p < 0.01) difference from control; all other information as in legend to Fig. 1.
I
24
44 Incubation
Fig. 4. Dose- and time-dependent (IFN-y) on the synthesis of PGE . hgament tibroblasts; all information
72
Time
(h)
effects of interferon-y by human periodontal as in legend to Fig. 1.
391
Cytokines and fibroblast PGE and CAMP Table 1. Effect of cytokines on the synthesis of PGE in the presence or absence of indomethacin by human periodontal ligament fibroblasts Cytokine* IL-la IL-la IL-l/I IL-l/l TNF-a TNF-a IFN-y IFN-y
Table 3. Dose-dependent effect of cytokines (abbreviations as in Table 1) on the production of CAMP by human periodontal ligament fibroblasts Cvtokine
Indomethacin (1O-6 M) + + + +
PGE production 78 + 15.4t 37 * 4.53 245 k 38.0 42 + 6.8$ 398 + 42.3 56 & 9.2f 215 + 19.8 38 5 6.51 128 f 14.4 46 + 7.34
+
*Doses of cytokines a.re 3 ng/ml for interleukin-la (IL-la), 1 ng/ml for interleukin-l/l (IL-l/l), 1 nM for tumour necrosis factor-a (‘TNF-a), and 100 U/ml for interferon-y (IFN-y). tData are expressed as mean + SD @g PGE/104 cells, n = 4) after 2 h incubation. $Significantly different from the corresponding cytokines or vehicle in the abslence of indomethacin (p c: 0.01).
The effect of indomethacin is shown in Table 1; cytokine-induced PGE synthesis was inhibited by the addition of this compound (10m6M), and the level of PGE production was even lower than that of control cells. The localization of PGE in cells incubated with interleukin-l/l is shown in Plate Fig. 5. In control cells, the 3,3’-diaminobenzidine reaction products were distributed uniformly and the staining was light [Plate Fig. 51(A)]. When cells were incubated with interleukin-l,Il for 1 h, there was an increase in the cytoplasmic staining intensity for PGE [Plate Fig. 5(B)]. Further incubation (4 h) caused intense staining of the cytoplasm, nuclear membrane and the nucleus [Plate Fig. 5(C)]. Incubation with indomethacin ( 10e6 M) reversed the staining characteristics; cells incubated with 1.Ong/ml of interleukin- 1/I in the presence of indomethacin for 1 h showed a significant decrease in cytoplasmic staining [Plate Fig. 5(D) and Table 21. Table 2. Effect of interleukin-la (1.0 ng/ml) and indomethacin (10e6M) on the levels of PGE in human periodontal ligament fibroblasts using immunohistochemical/ microphotometric method
Control
Interleukin-l/3
Incubation time (h) 1 4 24 1 4 24
PGE production 83.6 + 1.85* 80.0 k 2.97 76.9 + 3.17 64.5 + 2.18t 59.2 + 2.6Ot 67.5 + 2.467
0
1 4 24
76.1 + 3.18 76.5 & 2.86 84.1 k 3.19t
*Data are expressed as mean f SD (% light transmitted, ?I =4). tsignificantly diffemnt from control at the same time period (p < 0.05).
CAMP production 0.45 * 0.04*
IL-la IL-la IL-la IL-la
0.3 1.0 3.0 (ng/ml) 10.0
6.58 + 12.00 k 7.92 k 3.96 k
IL-l/I IL-l/I IL-l/I IL-lb
0.1 0.3 1.O (ng/ml) 3.0
4.68 + 0.35t 14.56 + 2.04t 10.28 + 1.78t 6.32 & 0.88t
0.48t 1.48t 0.95t 0.56t
TNF-a TNF-a TNF-a TNF-a
0.3 1.0 3.0 (nM) 10.0
0.98 f 2.21 + 3.98 + 1.76 +
IFN-y IFN-y IFN-y IFN-Y
30 100 300 (U/ml) 1000
0.56 + 0.08 0.78 _+0.09t 1.78 + 0.30t 1.09+0.18t
0.12t 0.347 0.48t 0.297
*Data are expressed as mean + SD (pmobl.5 x lo4 cells, n = 4) after 1 h incubation. tSignificantly different from control (p < 0.01).
The dose effect of cytokines on CAMP production is shown in Table 3. The fibroblasts responded to the administration of each cytokine in a dose-dependent fashion. The doses of interleukin-la, -l/I, tumour necrosis factor-a and interferon-y that stimulated CAMP production were 1.Ong/ml, maximum 0.3 ng/ml, 3.0 nM and 300 U/ml, respectively. The time course for the cytokine-induced production of CAMP is shown in Text Fig. 6. Small but significant elevations in CAMP were seen as early as after 5 min incubation with interleukin- 1fi, tumour necrosis factor-a and interferon-y. At 30 min, all cytokines caused significant elevations in the production of CAMP. A further significant increase in the accumulation of CAMP was seen at 60min of incubation. In addition, differences in the level of CAMP production were found among the 4 cytokines tested; these differences were more apparent at 60 min of incubation. The effect of indomethacin on the production of CAMP by the 4 cytokines is shown in Table 4. In each .
10 f uu
9
‘0
3
5.
4
c 5 E 4
4 0
a q q S
IL-le (0.3 nglml) IL-la (1.0 nglml) TNF-a (3.0 nM) IFN-y (300 U/ml)
3 2 1 0
Interleukiwlj3 + indomethacin
Dose
-
6
15
30
50
Time (min) Fig. 6. Time course effect of cytokines on the production of CAMP in human periodontal ligament fibroblasts. Values are mean + SE for 4 cultures. *Significantly different from control (p < 0.01). *Significantly different from control (p < 0.05).
S.
392
SAITO et al.
Table 4. Effect of indomethacin on the cytokine-stimulated CAMP production by human periodontal ligament fibroblasts and the effect of addition of exogenous PGE (abbreviations as in Table 1) Cytokine* -
Indomethacin (10-6M) -
+ IL-la IL-la IL-la IL-la IL-l/? IL-ID IL-lb IL-ID TNF-a TNF-a TNF-a TNF-a IFN-y IFN-y IFN-y INF-y
+ + + + + -
PGE ( 1O-6 M) CAMP production 0.78 k 0.13t tumour necrosis factor-cc (1 .O nM = lO-9 M, 6.36 ng PGE/104 cells) > interferon-y (100 U/ml = 2.3 x lo-” M, 3.00 ng PGE/lO+’ cells). The time-course experiment for CAMP production showed that fibroblasts also responded to the 4 cytokines in a time-related fashion. Note that a shorter time course was used for the CAMP experiment because: (1) unlike PGE in the medium, CAMP does not accumulate in cells; and (2) the level of CAMP is regulated by the enzyme phospho-
diesterase and, in our experiment, cells were preincubated with IBMX to inhibit this enzyme. The potency of cytokine effects on CAMP production can be ranked as follows (Table 3): interleukin-l/l (0.3 ng/ml, 14.56 pmol/l.5 x lo4 cells) > interleukinICC (1.0 ng/ml, 12.00 pmol/l.5 x lo4 cells) > tumour necrosis factor-cc (3.0 nM, 3.98 pmol/l.5 x lo4 cells) > interferon-y (300 U/ml, 1.78 pmol/l.5 x lo4 cells). The binding of cytokines to cell surface receptors is the first step in the initiation of cytokine action on target cells, so our findings, together with those from receptor studies in other cell systems (Vilcek et al., 1987; Dinarello and Savage, 1989) suggest that specific receptors for these ligands are present on periodontal ligament cells. The mechanism of action of cytokines on fibroblast PGE production is not completely clear. Interleukin1 may stimulate PGE synthesis by hydrolysis of membrane phospholipid by specific phospholipases with release of arachidonic acid substrate (Burch and Axelrod, 1987; Parker et al., 1987), or by conversion of release of arachidonate into PGs by cyclooxygenase activity (Raz et al., 1988). Interleukin-1 also increases the synthetic rate of newly formed cycle-oxygenase, suggesting that its effect on PGE synthesis is mediated mainly, if not solely, via induction of cycle-oxygenase synthesis (Korn et al., 1989). We also found a concomitant increase in the production of CAMP with all 4 cytokines; prominent increases in CAMP levels were seen at 60min of incubation. This elevation of CAMP is seemingly dependent on the synthesis of PGE because it was inhibited by the addition of indomethacin, and because this inhibition could be reversed by the addition of exogenous PGE (Table 4). Cyclic AMP levels increase in response to PGE administration via the adenylate cyclase system (Sutherland and Rall, 1960). On the other hand, the effects of tumour necrosis factor and interleukin-I on CAMP accumulation occur through receptor-G protein interactions (Zhang et al., 1988). We attempted to localize and semi-quantitate the effect of interleukin-1 jj’ on the level of PGE and CAMP in periodontal ligament by using an immunohistochemical procedure. We found an increase in staining for PGE in the cytoplasm and to a lesser extent in the nucleus after 60 min of incubation with the interleukin-l/I. Normally, PGE is not stored in cells but is released into the extracellular space (Piper and Vane, 1971). The staining for PGE found in these cells suggests that it is either bound to surface receptors and/or being taken up from the
extracellular medium by a process of internalization (Rao and Cheginin, 1983). PGE receptors have been described in the cell plasma membrane in a number of tissues, including alveolar bone (Moore and Wolfe, 1973; Rao, 1974; Yoshizawa, Abiko and Takiguchi, 1983). PGE binding sites have also been found in subcellular organelles such as endoplasmic reticulum, Golgi apparatus (Mitra and Rao, 1978a), lysosomes (Mitra and Rao, 1978b) and in nuclei (Rao and Cheginin, 1983). Interleukin-l/l enhanced both the cytoplasmic and nuclear staining for CAMP in the periodontal ligament cells. This increase in staining intensity indicates a true elevation in the synthesis of CAMP, as confirmed by the protein binding assay. The nuclear CAMP staining merits further explanation, because AMP binds to the regulatory subunits of protein kinase and releases the catalytic subunits. In one study, Gancedo, Mazon and Eraso (1985) have shown that the binding of CAMP to the regulatory subunits in prokaryotes might function directly as intranuclear regulatory proteins. The increases in PGE and CAMP production in response to all of the cytokines tested suggest that the function of periodontal ligament fibroblasts may be regulated, at least in part, by these cytokines, either alone or in concert with systemic hormones. There are indications that the effect of these cytokines on fibroblast function can be synergistic (Elias et al., 1988) additional or overlapping (Beresini, Lempert and Epstein, 1988). REFERENCES
Beresini M., Lempert M. J. and Epstein L. B. (1988) Overlapping polypeptide induction in human fibroblasts in response to treatment with interferon-a, interferon-y, interleukin la, interleukin Ip. and tumor necrosis factor. J. Immunol. 140, 485-493. Burch R. M. and Axelrod J. (1987) Dissociation of bradykinin-induced prostaglandin formation from phosphatidylinositol regulation of phospholipase A,. Proc. natn. Acad. Sci. U.S.A. 84, 6374-6378. Davidovitch Z., Shanfeld J. and Lally E. (1982) Prostaglandin Fz. is associated with alveolar bone cells: immunohistochemical evidence utilizing monoclonal antibodies. In: Factors and Mechanisms InJuencing Bone Growrh (Edited by Dixon A. D. and Sarnat B. G.) pp. 125-134. Liss, New York. Davidovitch Z., Nicolay O., Ngan P. W. and Shanfeld J. L. (1988) Neurotransmitters, cytokines and the control of alveolar bone remodeling in orthodontics. Denr. Clin. North Am. 32, 41 l-435. Dinarello C. A. (1988) Biology of interleukin-1. FASEB J. 2, 108-l 15. Dinarello C. A. and Savage N. (1989) Interleukin-I and its receptor. CRC Crit. R&J. Immunol. 9, l-20. Elias J. A.. Gustilo K.. Baeder W. and Freundlich B. (1988) Synergistic stimulation of fibroblast prostaglandin pro: duction by recombinant interleukin 1 and tumor necrosis factor. J. Immunol. 138, 3812-3816. Gancedo J. M., Mazon M. J. and Eraso P. (1985) Biological roles of CAMP: similarities and differences between organisms. Trends Biochem. Sci. 10, 210-212. Geisser S. and Greenhouse S. W. (1958) An extension of Box’s results on the use of the F distribution in multivariate analysis. Ann. Math. Stat. 29, 885-891. Gilman A. G. (1970) A protein binding assay for adenosine 3’,5’-monophosphate. Proc. natn. Acad. Sci. U.S.A. 67, 305-312.
Cytokines and fibroblast PGE and CAMP Hughes T. K. and Baron S. (1987) Do the interferons act singly or in combination? J. Interferon Res. 7, 603614. Kapschmidt R. F. and Upchurch H. C. (1982) Neutrophil release after injecti’sns of endotoxins of leukocytes endogeneous mediators into rats. J. Reric. Sot. 28, 191-201. Korn J. H., Downie El., Roth G. J. and Ho Shih-Yieh (1989) Synergy of interleukin 1 (IL-I) with arachidonic acid and A23187 in stimulating PGE synthesis in human fibroblasts: IL-1 stimulates fibroblast cyclooxygenase. Clin. Immunol. Immunopath. 50, 196-204.
Kuehl F. A. Jr (1974:l Prostaglandins, cyclic nucleotides and cell function. Prostaglandins 5, 325-340. Le J. and Vilcek J. (1987) Biology of disease: tumor necrosis factor and interleukin 1; cytokines with multiple overlapping biological activities. Lab. Invest. 56, 234-248. Mitra S. and Rae C. (1978a) Gonadotropin and prostaglandins binding sites in rough endoplasmic reticulum and golgi fractions of bovine corpore luteum. Archs Biochem. Biophys. 191, 332-340. Mitra S. and Rao C. (1978b) Receptors for gonadotropins and prostaglandins in lysosomes of bovine corporan luteum. Archs Biochem. Biophys. 185, 126-133. Miyamoto E., Kuo J. F. and Greengard P. (1969) Cyclic nucleotide dependent protein kinases. J. biol. Chem. 244, 6395-6409.
Mizel S. B., Kilian I’. L., Lewis J. C., Paganelli K. A. and Chizzonite R. A. (1987) The interleukin 1 receptor. Dynamics of interleukin 1 binding and internalization in T cells and fibroblasts. .I. Immunol. 138, 29062912. Mochan E., Uhl J. and Newton R. (1986) Evidence that interleukin inductions of synovial cell plasminogen activator is mediated via prostaglandin E and CAMP. Arfhritis & Rheum. 29, 1078-1083.
Moore W. and WolFe I. (1973) Binding of prostaglandin E, to beef thyroid membranes. J. biol. Chem. 248, 570555711. Newton R. C. and Covington M. (1987) The activation of human fibroblast prostaglandin E production by interleukin 1. Cell. Imm&ol. li0, 338-349. Nean P. W.. Zadeh Y. Z.. Shanfeld J. and Davidovitch 2. 71988) The effect of interleukin-1 fi and parathyroid hormone on cyclic nucleotide and prostaglandin levels in human periodontal ligament fibroblasts in vitro. In: The Biological Mechanisms of Tooth Eruption and Root Resorption (Edited by Davidovitch Z.) pp. 261-267.
ESCO Media, Birmingham, Ala. Owens R. B. (197411Glandular epithelial cells from mice: a method for selective cultivation. J. natn. Cancer Inst. 52, 1375-1378.
395
Parker J., Daniel L. W. and Waite M. (1987) Evidence of proteinkinase C involvement in phorbol diester-stimulated arachidonic acid release and prostaglandin synthesis. J. biol. Chem. 262, 5385-5390. Piper P. and Vane J. (197 1) The release of prostaglandins from lung and other tissues. Ann. N. Y. Acad. Sci. H&363-385. Ragnarsson B., Carr G. and Daniel J. C. (1985) Isolation and growth of human periodontal ligament cells in vitro. J. dent. Res. 64, 1026-1030. Rao C (1974) Characterization of prostaglandin receptors in the bovine corpus luteum cell membrane. J. biol. Chem. 249, 7203-7209.
Rao C. and Cheginin N. (1983) Gonadotropins and prostaglandins binding sites in nuclei of bovine corpora luteum. Biochim. biophys. Acta 584, 454466. Raz A., Wyche A., Siegel N. and Needleman P. (1988) Regulation of fibroblast cyclooxygenasc synthesis by interleukin-1. J. biol. Chem. 263, 3022-3028. Rippin J. W. (1976) Collagen turnover in the periodontal ligament under normal and altered functional forces. J. periodont. Res. 11, 101-107. Roberts W. E. and Chase D. C. (1981) Kinetics of cell proliferation and migration associated with orthodontically-induced osteogenesis. J. dent. Res. 60, 174-18 I. Sutherland E. W. and Rail T. W. (1960) The relation of adenosine 3’,5’-phosphate and phosphorylase to the actions of catecholamines and other hormones. Pharmac. Rev. 12, 265-299.
Ten Cate A. R., Deporter D. A. and Freeman E. (1976) The role of fibroblasts in the remodeling of periodontal ligament during physiological tooth movement. Am. J. Orthod. 69, 155-167.
Vilcek J., Tsujumoto M., Palombella V. L., Kohase M. and Le J. (1987) Tumor necrosis factor: Receptor binding and mitogenic action in fibroblasts. J. Cell. Physiol. Suppl. 5, 57-61. Weissmann G., Smolen J. E. and Korchak H. M. (1980) Release of inflammatory mediators from stimulated neutrophils. New Engl. J. Med 303, 27-34. Whiteley P. J. and Needleman P. (1984) Mechanism of enhanced fibroblast arachidonic acid metabolism by mononuclear cell factor. J. clin. Invest. 74, 2249-2253. Yoshizawa Y., Abiko Y. and Takiguchi H. (1983) Some properties of prostaglandin E, receptors in rabbit alveolar bone cell membrane. Inr. J. Biochem. 15. 991-995. Zhang Y., Lin J. X., Yip Y. K. and Vilcek J. (1988) Enhancement of CAMP levels and of protein kinase activity by tumor necrosis factor and interleukin 1 in human fibroblasts: role in the induction of interleukin 6. Proc. natn. Acad. Sci. U.S.A. 85, 6802-6805.
