AUTOCRINE REGULATION OF RHEUMATOID ARTHRITIS SYNOVIAL CELL GROWTH IN VITRO David H. Goddard,
Scott L. Grossman,
Mary E. Moore
Rheumatoid arthritis (RA), and not osteoarthritis (OA) synovial cells proliferate in serum-free medium, a finding that suggests that, in vitro, RA synovial cells may be stimulated to grow by the continuous autocrine production of at least one polypeptide growth factor. Adding monoclonal antibody lD11.16, or rabbit polyclonal anti-tumor growth factor p (anti-TGF-/3) antibodies (both neutralizing antibodies to TGF-PI and TGF-&) to RA synovial cells, in culture, caused a significant reduction in cell growth, an effect not seen when other growth factor antibodies (platelet-derived growth factor [PDGF], epidermal growth factor [EGF], or EGF receptor) were added to the culture medium. Taken together, these data are consistent with the concept that RA synovial cell growth in vitro is driven endogenous TGF-/3. Moreover, when EGF was added to the culture medium, this caused the numbers of RA, and not OA, synovial cells to increase significantly. This finding suggests that RA synovial cells are in G, phase of the cell cycle; an effect that could be mediated by endogenous TGF-& o 1990 by W.B. Saunders Company.
Excessive cellular proliferation and turnover of extracellular matrix are common features in several different diseases, including rheumatoid arthritis, (RA), atherosclerosis, and several cancers.’ Recent cancer studies, as well as studies in fetal tissues and placenta, have established that polypeptide growth factors play an important role in regulating cell growth.2-5 Furthermore, the discovery that the protein products of several viral oncogenes show close homology with polypeptide growth factors, coded for by their human homologs, suggests a possible link between certain viruses and excessive cell growth.6-g RA is a disease of unknown etiology, which is characterized by excessive growth of the synovial mem-
From the Division of Rheumatology, Department of Medicine, Albert Einstein Medical Center, and Temple University School of Medicine, Philadelphia, PA. Presented in part at the 53rd Annual Scientific Meeting of the American College of Rheumatology, Cincinnati, Ohio, June 1989, and the Second International Workshop on Cytokines, Hilton Head, South Carolina, December 1989. Supported in part by an Albert Einstein Society grant #2-0731-601000. Address reprint requests to David H. Goddard MD, MRCP Rheumatic Diseases Research Laboratories. Room 107. Korman Research Pavilion, 12th Street and Tabor Roads, Philadelphia, PA 19141. o 1990 by W.B. Saunders Company. 1043/90/0202-0007$05.00/0 KEY WORDS: Rheumatoid arthritis/Osteoarthritis/Synovial Transforming growth factor-@/Autocrine growth
CYTOKINE,
Vol. 2, No. 2 (March),
1990: pp 149-155
cells/
brane, causing cartilage and bone destruction. In this disease, the synovium acts like a localized malignancy. In vitro, cultured RA synovial cells display several of the growth characteristics of neoplastic and virally transformed cells. These characteristics include the tendency to grow as disorganized monolayers with loss of normal contact-inhibited growth and piling-up of cells to form foci, the presence of many multinucleated cells, colony formation in soft agar, and tumor formation in nude mice.‘0-‘3 Similarities between rheumatoid synovitis and neoplasia have led to speculation that polypeptide growth factors play a role in causing synovial proliferation. The recently reported identification of several polypeptide growth factors and cytokines in synovial fluids, including transforming growth factor-p (TGF-P), plateletderived growth factor (PDGF), epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), the cytokines interleukin 1 (IL l), interleukin 6, and tumor necrosis factor-a further strengthens such speculation.‘4’15 We present evidence showing that RA synovial cell growth, in vitro, is driven by the continuous autocrine secretion of at least one polypeptide growth factor. Based on blocking studies using growth factor antibodies, we speculate that this growth factor may be TGF-0. An autocrine-secreted growth factor of this type, acting synergistically with other growth factors, could cause enhanced RA synovial cell growth.
149
150 / Goddard,
Grossman,
Moore
CYTOKINE,
RESULTS
Phenotypic Analysis of Synovial Cell Lines Phenotypic analysis was performed on cells taken from three RA and three OA synovial cell cultures. Paired cell cultures (RA and OA) were selected from representative passages (passage numbers 3, 10, and 23). Positive cytoplasmic staining of varying intensity was observed in all cells when reacted with monoclonal antibodies (MAbs) to vimentin but not cytokeratin. In contrast, cell-surface staining was negative when cells were reacted with antibodies, recognizing lymphocyte and monocyte/macrophage markers, or factor VIII antigen. Synovial cells grown in medium alone did not react with MAbs, recognizing human leukocyte antigensD-related (HLA-DR) antigens, Adding recombinant y-interferon to the culture medium (48 hr at 37OC; Du Pont, Wilmington, DE) caused these cells to express HLA-DR on their surface, as evidenced by positive surface staining.
Synovial Cell Growth in Serum-Free Medium Figure 1 shows the significant increase in mean (* SEM) cell number in three RA, when compared with three OA synovial cell cultures, whose phenotypes had been determined in the preceding experiments, and which were grown for 14 d in RPM1 plus 2% bovine serum albumin (BSA) (RA d 14 = 3.71 k 0.62, d 1 = 1.76 * 0.21 x 104, P < 0.05; t-test, two-tailed; OA d 14 = 1.55 * 0.3 x 104;d 1 = 1.43 * 0.18 x 104). Since RA synovial cell growth could have resulted from the effects of contaminating mitogens in BSA, five RA synovial cell lines were then grown in RPMI, supplemented with only nonessential amino acids, and
Vol. 2, No. 2 (March
1990:
149-155)
antibiotics. Figure 2 shows that by d 14, the number of RA synovial cells had increased significantly (d 14 - day 1 = 2.16 + 0.68 x lo4 cells; P < 0.04; paired t-test, two-tailed). By d 35, this had increased further to 8.1 f 1.56 x 1O4cells, which was significantly greater than d 14 (P < 0.002; paired t-test, two-tailed). To exclude the possibility that synovial cell growth in serum-free RPM1 occurred because of the persisting effects of fetal bovine serum (FBS), RA synovial cell growth in serum-free RPM1 (RPM1 plus nonessential amino acids and antibiotics) was compared with that in serum-free Dulbecco’s modified essential medium (DMEM) (DMEM plus nonessential amino acids and antibiotics). Because serum-free DMEM rapidly induces G, arrest in fibroblasts in culture, we hypothesized that significant growth in both media would indicate that previous exposure to FBS continued to provide the stimulus for synovial cell growth. Figure 3 shows that by d 14 there was a significant increase in cell number (mean * SEM) in four RA cell lines grown in RPMI, but not DMEM (RPM1 d 14 = 16.2 f 3.8 x 104, d 1 = 3.2 ? 0.6 x 104, P K 0.05; t-test, two-tailed. DMEMd 14 = 3.0 f 0.7 x 104,d 1 = 3.2 k 0.6 x 104).
Influence of Growth Factor Antibodies on RA Synovial Cell Growth We then began experiments with the aim of blocking RA synovial cell growth by adding specific growth factor antibodies to the culture medium. Figure 4 shows that a significant reduction in cell number occurred when four RA synovial cell cultures were grown for 14 d in serum-free RPM1 plus rabbit antibodies to TGF-P. This effect was observed at an antibody concentration of 5 pg/ml (P < 0.04; paired t-test, one-tailed) and was
Figure 1. Growth of RA and OA synovial cells in serum-free medium.
1
8
Days
in Culture
14
Values shown are the mean + SEM number of RA (W), and OA cells (O-O) taken from paired cell cultures, which had undergone phenotypic analysis and which were grown in RPM1 plus 2% BSA. By d 14, the number of RA, but not OA synovial cells had increased significantly compared with d 1 (*RA cell number compared with day 1 P < 0.05; 2-tailed; **RA vs OA P < 0.05; 2-tailed).
Autocrine
growth
of synovial
cells
/ 151
Influence of ExogenousGrowth Factors on Synovial Cell Growth
L
= u” 6.E f b -r
I
/
1
4-
I/
./. A’
f6 2-
.'
0’
7
14
21 Days
Figure 2. non-essential
Growth of RA synovial amino-acids.
28
35
in Culture
cells in RPM1
supplemented
with
Values shown are the mean + SEM increase in RA synovial cell numbers over baseline, when grown in RPM1 supplemented with non-essential amino-acids. By d 14, the number of cells had increased significantly compared with day 1 (P -C 0.04; paired t-test, two-tailed). By d 35, cell number had increased further and was significantly greater than d 14 (P < 0.002; paired t-test, two-tailed).
maximal at an antibody concentration of 100 pg/ml. In contrast, no reduction in growth occurred when synovial cells were grown in serum-free RPM1 containing antibodies to PDGF, or EGF (0.1 to 50 /*g/ml IgG), or MAb to EGF receptor (0.1 to 50 pg/ml IgG; data not shown). Next, three of the four RA synovial cell cultures used in the preceding experiments were grown in serum-free RPM1 plus MAb lDll.16. Figure 4 shows that, when added to the culture medium in concentrations as low as 0.1 pg/ml, this antibody caused a significant reduction in cell growth (P < 0.02; paired t-test, one-tailed), an effect that was maximal at an antibody concentration of 5 pg/ml (P < 0.03; paired t-test, one-tailed). To confirm that neither MAb lD11.16 nor polyclonal antibody to TGF-0 were cytotoxic for RA synovial cells, cell lines used in the preceding experiments were grown for 15 d under the following conditions: serum-free medium alone; serum-free RPM1 plus TGFfi (50 rig/ml); or serum-free RPM1 plus MAb lD11.16 (1 pg/ml), or rabbit TGF-/3 antibody (50 pg/ml); or serum-free RPM1 plus MAb 1D 11.16 (1 pg/ml) plus TGF-P (11 rig/ml), or rabbit antibody to TGF-P (50 pg/ml) and TGF-fi (50 rig/ml). As illustrated in Fig. 5, the significant reduction in synovial cell growth that occurred in serum-free RPM1 plus either MAb 1D 11.16 (serum-free RPM1 plus MAb 1Dll. 16 vs serum-free RPM1 with and without TGF-0, P < 0.02; t-test, onetailed) or rabbit TGF-fl antibody (serum-free RPM1 plus rabbit antibody vs serum-free medium with and without TGF-fi P < 0.03; one-tailed) was totally abolished by adding an excess of the growth factor.
The preceding experiments suggest that RA synovial cells in culture secrete TGF-0. Because TGF-P and PDGF stimulate cells to enter G, phase, while EGF stimulates cells in G, to enter S phase, the next experiments compared the growth of three RA and three OA synovial cell lines in RPM1 plus 2% BSA (serum-free medium), with growth in serum-free medium plus EGF (6.5 rig/ml), or PDGF (5 rig/ml). Figure 6 shows that by d 21, the number of RA synovial cells grown in serum-free medium plus EGF was significantly greater than cells grown in serum-free alone (P < 0.04; t-test, one-tailed). The capacity of EGF to induce RA cell growth was similar to that seen with PDGF. This finding contrasted with the modest increase in the number of OA synovial cells grown in serum-free medium, containing either of these growth factors.
DISCUSSION It is now clear that polypeptide growth factors control the growth of normal as well as neoplastic cells.16-22The recent identification of several of these in rheumatoid synovial effusions suggests that they may cause excessive synovial cell growth.14 RA and OA synovial cells grow differently in culture,” with RA synovial cells displaying several characteristics, which suggest transformation.“-I3 Since growth-factor-regulated genes are believed to play a critical role in cellular transformation,21’22 and because the products of some
25
Day 1 Day 14 Days in Culture Figure growth
3. Growth in RPMI.
of RA
synovial
cells
in DMEM
compared
with
Values shown are the mean + SEM number of RA synovial cells grown for 14 d in DMEM (O), and RPM1 (W), supplemented with non-essential amino acids and antibiotics. By d 14, the number of cells grown in RPM1 had increased significantly (d 14 vs d 1; P < 0.05, paired t-test). This contrasted with the lack of cell growth in DMEM.
152 /
Goddard,
Grossman,
CYTOKINE,
Moore
Figure 4. Influence of TGF-j3 antibodies on RA synovial cell growth.
80 60
0.01
0.1
1.0
10.0
100
ANTIBODY CONCENTRATION (micrograms
/ ml of IgG)
factor properties, we have investigated the possibility that RA synovial cell growth could be regulated by these factors. Our finding that RA, but not OA, synovial cells in long-term culture, grow in RPM1 lacking added growth factors, suggests that in vitro RA synovial cells have the capacity to synthesize those factors essential for growth. We believe that it is unlikely that RA synovial cell growth was mediated by a growth factor-like contaminant present in the medium, because the experiments used RPM1 with different batch numbers, and because OA synovial cells failed to grow in this medium. Phenotypic analysis confirmed that, our synovial cell cultures are not contaminated with either macrophages or longlived lymphocytes, whose secretory products could modactivated
ulate
oncogenes
synovial
Vol. 2, No. 2 (March 1990: 149- 155)
cell
have growth
function
in vitro.
Moreover,
Values shown are mean + SEM number of RA synovial cells grown for 15 d in serum-free medium plus MAb lD11.16 (0.1 - 10 pglml IeG). or rabbit oolvclonal TGF-B antibody (1 to -106 pg/ml Igd) expressed as percent of control (control = increase in synovial cell number at d 15 over d 1 in RPM1 supplemented with 2% BSA). A significant reduction in growth occurred at a MAb antibody concentration 0.1 rg/ml IgG (P < 0.02; paired t-test, one-tailed) which was maximal at an antibody concentration of 10 pg/ml IgG (P < 0.03; paired t-test, one-tailed), and at a rabbit polyclonal antibody concentration of 5 rg/ml IgG (P < 0.04; paired t-test, one-tailed), which was maximal at an antibody concentration of 100 pg/ml IgG.
pyridoxine hydrochloride, vitamin B- 12, aminopterin, reduced glutathione, hypoxanthine, and thymidine, the absence of which may affect synovial cell growth. We found that when added to serum-free RPMI, both MAb 1Dl 1.16 and polyclonal antibody to TGF-@, but not PDGF, or EGF, or EGF receptor antibodies, caused a significant reduction in RA synovial cell
0 El+ m+mAb E!J + eZa +
RPM TGF-Beta Rabbit Ab TGF-beta +Ab
the
finding that cells from each of the cultures showed positive staining with MAbs to vimentin, but not cytokeratin,
indicates
that they are of mesenchymal
origin.
Our observations that RA synovial cell numbers significantly increased in RPMI, but not DMEM, when supplemented only with nonessential amino acids, shows first, that RA cell growth occurs independently of mitogens, which may be present as contaminants in BSA, and second, that RA synovial cell growth is not likely to have occurred through a persistent effect of previous exposure to FBS. We chose serum-free DMEM because it rapidly induces G, arrest in mesenchymal cells in culture.23 At present, we are unsure why RA synovial cells grow in serum-free RPM1 and not serumfree DMEM, but we speculate that this may be related to differences in the composition of the two types of media. Specifically, DMEM lacks L-Asn, ~-Asp, L-Glu, hydroxy-L-Pro,
L-Pro,
D-Biotin,
p-amino
benzoic
acid,
2
Figure synovial
+ p(O.02 **p(o.o3
5. Lack of cytotoxic cell growth in serum-free
effect of TGF-8 medium.
antibodies
on RA
Values shown are mean + SEM increases in the numbers of cells grown for 15 d in serum-free medium alone, or serum-free medium plus TGF-8 (50 rig/ml), or mab lD11.16 (1 rig/ml), or TGF-P antibody (50 pg/ml), or an excess of TGF-@ plus antibody. The antiproliferative effects of both antibodies could be overcome by adding TGF-8 in excess (serum-free RPM1 plus mab lD11.16 vs serum-free RPM1 with and without TGF-fl P < 0.01; serum-free RPM1 plus rabbit TGF-fi antibody vs serum-free RPM1 with and without TGF-& P < 0.03; both t-test, two-tailed).
Autocrine growth of synovial cells / 153 4.0,
*
OA Figure growth.
6.
Influence
of exogenous
RA growth
factors
on synovial
cell
Values shown are mean * SEM increase in cell number at d 21 over baseline. Significant growth occurred in RA synovial cells in serumfree medium plus EGF (Es9)compared with serum-free medium alone @Zl$1):*P < 0.041. A similar increase occurred in the number of RA dynivial cells &own in serum-free medium plus PDGF (0). These findings contrasted with the lack of significant growth of OA cells in serum-free medium alone, or serum-free medium plus either EGF, or PDGF.
growth. Together with the observation that the effects of both of these antibodies could be overcome by adding an excess of exogenous TGF-/3, it is suggested that endogenous TGF-P may provide this mitogenic stimulus. Moreover, they show that neither the monoclonal, nor rabbit polyclonal, antibodies to TGF-P antibody were cytotoxic for RA synovial cells, in culture. These data are strongest for MAb 1Dll. 16, which recognizes TGF-/3, in Western analysis, and completely neutralizes the bioactivity of TGF-/3, and TGF-&.24 These findings contrast with the much weaker effect of the rabbit polyclonal antibodies on RA synovial cell growth. While the explanation for these differences is not known, we speculate that they may be related to difficulties in raising antibodies against TGF-/3. Because of the highly conserved structure of this peptide, its ubiquity, and its immunosuppressive activity, TGF-@ is a weak immunogen. 25These properties could lead to the generation of polyclonal antibodies in low-titer, which only weakly neutralize the biologic activity of TGF-P. Polypeptide growth factors can be broadly divided into two groups, depending on the point at which they act in the cell-cycle.6 TGF-P, and PDGF, are examples of growth factors, which stimulate quiescent cells (G, phase) to enter G, phase, and have been called competence factors. EGF, somatomedins, and IL 1, which stimulate cells in G, phase to enter S phase and divide, have been called progression factors.26*27 While competence and progression factors act synergically to stimulate cell division, certain differences in their individual properties are apparent. For example, only brief exposure to PDGF is required for Go-arrested cells to enter
G, phase. Longer exposure to TGF-P is necessary for quiescent cells to enter G, phase. Leof et a1.28 have proposed that TGF-P may act as an indirect mitogen by inducing c-sis mRNA transcription in quiescent fibroblasts. In these studies, they have shown that c-sis induction is accompanied by the release of a peptide with PDGF-like properties, which they suggest may provide the mitogenic stimulus for cell growth. Moreover, they have shown that TGF-fi induces at least two other PDGF-inducible genes, c-myc and c-jiis. Our finding, of a lack of effect of PDGF antibody on RA synovial cell growth, is not in conflict with this proposal, because induced PDGF, responsible for stimulating these cells, could remain in a cellular compartment inaccessible to the antibody.28 TGF-fl and PDGF both enhance the mitogenic effects of EGF in some cells by increasing the affinity and number of receptors for this growth factor.2g330Our observation that RA, but not OA synovial cells grown in serum-free medium with EGF, increased significantly in number, when compared with those grown in serum-free medium alone, indicates that these cells are in G, phase, an effect which could be mediated by endogenous TGF-P. While TGF-/3 has been reported to have variable biologic effects, depending on the cell type and conditions of culture, this growth factor regulates cell replication and differentiation, angiogenesis, extracellular matrix turnover, as well as stimulating mononuclear cell release of PDGF and IL 1.11,30-34Based on our results, together with the known biologic properties of TGF-/3, we speculated that autocrine-secreted TGF-P acts synergically with other growth factors and cytokines to stimulate RA synovial cell growth. The recent report by Lafyatis et a1.,35identifying TGF-/3 in the intact rheumatoid synovium using in situ immunohistochemical techniques, further strengthens such speculation. We both believe that many of the features of rheumatoid synovitis can be explained on the basis of local production and release of this growth factor, which could play a key role in synovial cell proliferation and subsequent fibrosis.
MATERIALS
AND
METHODS
Matevials Monoclonal antibody (MAb) 1Dll. 16, which neutralizes TGF-/3, and &, was a gift from Dr. James Dasch, Celtrix Laboratories, Palo Alto, CA. All other antibodies were purchased as follows: Polyclonal antibody to TGF-P (R & D Systems, Minneapolis, MN); antibodies to PDGF and EGF (Genzyme, Boston, MA); EGF receptor antibody (ICN Immunochemicals, Lisle, IL); anti-Leu 4 (CD3, T-cells), anti-Leu 14 (CD22, B-cells), anti-Leu M3 (mature monocyte), anti-leu1 la (large granular lymphocyte/NK cells) and peroxidase goat-anti mouse immunoglobulin (Becton Dickinson, Mountain View, CA); MAbs to vimentin and cytokeratin, sheep polyclonal antibody to Factor VIII antigen; and peroxidase goat anti-sheep immunoglobulin (ICN). Recombinant y inter-
154 / Goddard,
Grossman,
CYTOKINE, Vol. 2, No. 2 (March 1990: 149-155)
Moore
feron was a gift from Dr. Robert Newton (E. I. DuPont de Nemours & Co, Glenolden, PA). PDGF was purchased from R & D Systems, and EGF from Sigma Chemical Co., St. Louis, MO. Monoclonal antibody lD11.16 neutralizes the biologic activity of TGF-0, and & in vitro in a dose-dependent manner. 24TGF-@ antibody is a polyclonal, neutralizing antibody raised in rabbits that recognizesTGF-/3, and & and does not crossreact with acidic or basic FGF, PDGF, or EGF. Rabbit anti-PDGF is an IgG polyvalent antibody generated against purified, recombinant PDGF-BB homodimer. This antibody recognizes both human PDGF-BB BB homodimer and PDGF-AB heterodimer, and does not crossreact with human PDGF-AA. Monoclonal EGF receptor antibody is a mouse IgG,, obtained from C57BL/6 mice immunized with the paraformaldehyde-fixed A43 1 carcinoma cell line containing approximately 2 x lo6 EGF receptors per cell. This MAb reacts with human and mouse EGF receptors. Rabbit antiEGF is a polyclonal IgG neutralizing antibody raised in rabbits against natural murine EGF (receptor grade), and in ‘H uptake assaysusing rat primary hepatocytes, 2 to 4 kg of this antibody neutralizes 50% of the biologic activity of 10 ng of EGF. PDGF isolated from human platelets is >95% pure when analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). EGF is extracted from the submaxillary glands of mice and lyophilized from a solution of 5 mM ammonium acetate, pH 5.6. RPM1 1640 (RPMI), DMEM, Hank’s balanced salt solution (HBSS), nonessential amino acids, streptomycin, penicillin, and fungazone, were all purchased from Flow Labs (McLean, VA). FBS was purchased from Hyclone (Logan, UT), crude collagenasefrom Sigma Chemical Co., BSA from Miles Scientific Co. (Naperville, IL), and trypsin/EDTA from Gibco Laboratories (Grand Island, NY). All tissue culture plastics were purchased from Falcon (Oxnard, CA). Synovial samples were obtained with informed consent at the time of joint replacement surgery from six patients with a diagnosis of definite RA,36and three patients with OA.
Synovial
Cell Culture
Synovial cell cultures were established using a modification of the method described by Dayer et a1.37Synovial samples cut into small pieces (~5 mm’), were incubated overnight in RPMI, and supplemented with nonessential aminoacids (100 pg/ml), streptomycin (100 pg/ml), penicillin (100 h/ml), fungazone (25 pg/ml), 10% FBS, and crude collagenase (200 p/ml), at 37OCin 10% CO, in air. Dispersed cells were then collected by low speed centrifugation, washed twice in HBSS, once in RPM1 plus 10% FBS, resuspended in 8 ml of RPM1 plus 10% FBS, and seeded into two tissue culture flasks (25 cm’). Once primary cultures had reached confluence they were split weekly, and used in passages3 through 23.
Phenotypic
Analysis
of Synovial
Cells
Synovial cells (RA and OA) grown in RPM1 plus 10% FBS were detached from the culture flasks by adding trypsin/ EDTA, washed with RPM1 plus 10% FBS, and seededat low density on to tissue culture slides. Synovial cells were allowed to grow to the point where visual inspection confirmed adequate numbers of cells for staining, but before the cultures had
reached confluence. After washing in several changes of DMEM to remove traces of biotin present in RPMI, cells were fixed in acetone at 4OC for 30 min and then air-dried. Cells were then reacted with the following antibodies: anti-Leu 4, anti-Leu 14, anti-Leu M3, anti-Leu 1la, anti-vimentin, anticytokeratin, and anti-Factor VIII antigen. These were all used in accordance with the manufacturers’ recommendations. After washing in phosphate buffered saline (PBS), the cells were then incubated with either peroxidase goat anti-mouse immunoglobulin or peroxidase goat anti-sheep immunoglobulin as appropriate, washed in PBS, reacted with amino-ethylcarbazole (AEC) solution, washed in distilled water, counterstained with progressive hematoxylin, cleared in tap water, and then mounted in glycerin jelly. The slides were then examined in an Olympus BHS light microscope (Olympus Corporation of America, Hyde Park, NY).
Synovial
Cell Growth Studies
In the period preceding the start of each study, synovial cells (RA and OA) were grown to confluence in RPM1 containing 5% FBS. Cells were then detached by adding trypsin/EDTA, and plated in 24-well tissue culture plates (1.5 x lo4 cells/well). The next day, the medium was changed to serum-free medium (RPM1 plus 2% BSA) alone, or serum-free medium plus either growth factors (PDGF or EGF), or growth factor antibodies (TGF-fi, PDGF, EGF, or EGF receptor). Cultures were fed twice weekly by removing half of the medium and replacing it with a similar volume of fresh medium. Cell growth was measured by determining the increasein cell number over baseline at specifictime intervals during each experiment. Cells were detached from the plates by adding trypsin/EDTA, transferred immediately into isotonic solution (Isoton, Coulter Diagnostics, Hialeah, FL), and counted using a Coulter counter (model ZBI, Coulter Diagnostics). Triplicate counts were made on duplicate samples, and the mean value of these counts was used in further data analysis, using paired t-tests. Significance was set at the 5% level.
Acknowledgement We thank J. Dasch PhD (Celtrix Laboratories, Palo Alto, CA), for the gift of MAb 1Dll. 16 and R. Newton PhD (E. I. DuPont de Nemours & Co), for the kind gift of recombinant y-interferon. We also thank B. Wilson PhD for her helpful comments and criticisms.
REFERENCES 1. SpornMB, Harris ED Jr (1981):Proliferativediseases. Am J Med 170:1231-1236 2. SpornMB, RobertsAB (1985):Autocrinegrowth factorsand cancer.Nature 313:745-747 3. JamesR, BradshawRA (1984):Polypeptidegrowth factors. Ann RevBiochem53~259-292 4. RossR, GlomsetJ, Kariya B, Harker LA (1984):A plateletdependentserum factor that stimulatesthe proliferation of arterial smoothmusclecellsin vitro. ProcNat1 AcadSci USA 71:1207-1210 5. GajdusekC, DiCorleto P, RossR, SchwartzSM (1980): An endothelialcell-derivedgrowth factor.J Cell Bio185:467-472
Autocrine
6. Armelin HA, Armelin MCS peptide growth factors and oncogenes. genes, and growth factors. G. Wiley 361-363
(1987): The interactions of In Guroff G (ed). Oncogenes, Interscience, New York, pp
growth
of synovial
cells
/ 155
22. Anastassiades TP, Ley L, Wood A, Irwin D (1978): growth kinetics of synovial fibroblastic cells from inflammatory noninflammatory arthropathies. Arthritis Rheum 21:461-466 23. Greenberg transcription
ME, Ziff EB (1984): Stimulation of the c-j& protooncogene. Nature
The and
of 3T3 cells 3 11:433-438
7. Waterfield MD, Scrace GT, Whittle N, Stroobant P, Johnsson A, Wasteson A, Westermark B, Heldin CH, Huang JS, Deuel TF (1983): Platelet-derived growth factor is structurally related to the putative transforming protein ~28”‘” of simian sarcoma virus. Nature 304:35-39
induces
8. Downward J, Yarden E, Mayes E, Scrace G, Totty N, Stockwell P, Ulrich A, Schlessinger J, Waterfield MD (1984): Close similarity of epidermal growth factor and v-erb-0 oncogene protein sequences. Nature 307:521-527
25. Leung BS, Zhou L (1988): A fluorimetric enzyme-linked immunosorbent assav (F ELISA) for TGF-B. In Powanda MC. Oppenheim JJ, Kluger‘ MJ, Dinarello CA (Eds.) Monokines and Other Non-Lymphocytic Cytokines. Alan R Liss, New York, pp 397-400
9. Coussens L, Yang-Feng TL, Chen Y-CLE, Gray A, J, Seeburg PH, Liberman TW, Schlessinger J, Francke U, A, Ulrich A (1985): Tyrosine kinase receptor with extensive to EGF receptor shares chromosomal location with neu Science 230: 1132- 1139 sponses require Arthritis
10. Brinckerhoff CE (1983): of rabbit synovial fibroblasts transforming growth factor Rheum 26:1370-1379
McGrath Levinson homology oncogene.
Morphologic and mitogenic reto transforming growth factor p LY or epidermal growth factor.
11. Brinckerhoff CE, Harris ED Jr (1978): Collagenase tion by cultures containing multinucleated cells derived from tibroblasts. Arthritis Rheum 21:745-753
producsynovial
12. Brinckerhoff CE, Harris ED Jr (1981): Survival of rheumatoid synovium implanted into nude mice. Am J Path01 103:41 l-419 13. Lafyatis R, Remmers EF, Roberts A, Yocum D, Sporn MB, Wilder RL (1989): Anchorage-independent growth of arthritic synoviocytes: Dependence on platelet-derived growth factor (PDGF) and inhibition by transforming growth factor-p (TGF-6) and retinoids. J Clin Invest 83:1267-1276 14. Lafyatis R, Remmers EF, Roberts A, Sporn MB, Wilder (1988): Regulation of rheumatoid synovial cell proliferation cytokines. Arthritis Rheum 31 (suppl l):A62
RL by
15. Hamerman D, Taylor S, Kirschenbaum I, Kalgsbrun M, Raines EW, Ross R, Thomas KA (1987): Growth factors with heparin binding affinity in human synovial fluid. Proc Exp Biol Med 186:384389 16. Stiles CD, Capone GT, Scher Wyk JJ, Pledger JJ (1979): Dual control dins and nlatelet-derived growth factor. 76:1279-l&3
CD, Antoniades HN, Vanof cell growth by somatomeProc Nat1 Acad Sci USA
17. Assoian RK (1985): Biphasic effects of type fl transforming growth factor on epidermal growth factor receptors in NRK fibroblasts. Functional consequences for epidermal growth factorstimulated mitosis. J Biol Chem 260:9613-9617 18. Beach D, Basilic0 C, Newport J (1988): In D Beach, C Basilica, and J Newport (eds.) Cell Cycle Control in EukayotesCurrent Communications in Molecular Biology. Cold Spring Harbor Laboratory, New York, pp 2-8 19. Heine U, Munoz EF, Flanders KC, Ellingsworth LR, Lam HY, Thompson NL, Roberts AB, Sporn MB (1987): Role of transforming growth factor-p in the development of the mouse embryo. J Cell Biol 105:2861-2876 20. Frolik CA, Dart LL, Myers CA, Smith DM, Sporn MB (1983): Purification and initial characterization of a type fl transforming growth factor from human placenta. Proc Nat1 Acad Sci USA 80:3676-3680 21. Derynck R, Jarrett JA, Chen EY, Eaton DH, Bell JR, Assoian RK, Roberts AB, Sporn MB, Goeddel DV (1985): Human transforming growth factor-p complementary DNA sequence and expression in normal and transformed cells. Nature 316:701-705
24. Dasch JR, Pace DR, Waegell W, Inenaga D, Ellingsworth L (1989): Monoclonal antibodies recognizing transforming growth factor@. Bioactivity, neutralization, and transforming growth factor p2 affinity purification. J Immunol 142:1536-1541
26. Bravo R (1988): Complexity of the early to growth factors in mouse fibroblasts. In D Beach, Newport (eds) Cell Cycle Control in Eukaryotes-Current cations in Molecular Biology. Cold Spring Harbor York, pp 15-21
genomic response C Basilica, and J CommuniLaboratory, New
27. Singh JP, Adams LD, Bonin PD (1988): Mode of fibroblast enhancement by human interleukin 1. J Cell Biol 106:8 13-8 19 28. Leof EB, Proper JA, Goustin AS, Shipley GD, DiCorleto PE, Moses HL (1986): Induction of c-sis mRNA and activity similar to platelet-derived growth factor by transforming growth factor p: A proposed model for indirect mitogenesis involving autocrine activity. Proc Nat1 Acad Sci USA 83:2453-2457 29. Davis RJ, Czech MP (1987): Stimulation of epidermal growth factor receptor threonine 654 phosphorylation by plateletderived growth factor in protein kinase C-deficient human fibroblasts. J Biol Chem 262:6832-6841 30. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS, Wakefield LM, Heine U, Liotta LA, Falanga V, Kehrl JH, Fauci AS (1986): Transforming growth factor type B: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Nat1 Acad Sci USA 83:4167-4171 31. Sporn MB, Roberts MB, Wakefield LM, de Crombrugghe (1987): Some recent advances in the chemistry and biology transforming growth factor-@. J Cell Biol 105:1039-1045
B of
32. Wahl SM, Hunt DA, Wakefield LM, McCartney FN, Wahl LM, Roberts AB, Sporn MB (1987): Transforming growth factor type 0 induces monocyte chemotaxis and growth factor production. Proc Nat1 Acad Sci USA 84:5788-5792 33. Martinet Y, Bitterman PB, Mornex JF, Grotendorst GR, Martin GR, Crystal RG (1986): Activated human monocytes express the c-k proto-oncogene and release a mediator showing PDGF-like activity. Nature. 319:158-160 34. Baird A, Mormede P, Bohlen P (1985): Immunoreactive fibroblast growth factor in cells of peritoneal exudate suggests its identity with macrophage-derived growth factor. Biochem Biophys Res Comm 126:358-364 35. Lafyatis R, Thompson NL, Remmers EF, Flanders KC, Roche NS, Kim S-J, Case JP, Sporn MB, Roberts AB, Wilder R (1989): Transforming growth factor-B production by synovial tissues from rheumatoid patients and streptococcal cell wall arthritic rats: Studies on secretion by synovial fibroblast-like cells and immunhistologic localization. J Immunol 143:1142-l 148 36. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, et al. (1988): The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 3 I:3 15324 37. Dayer J-M, Krane SM, Russell RGG, Robinson DR (1976): Production of collagenase and prostaglandins by isolated adherent rheumatoid synovial cells. Proc Nat1 Acad Sci USA 73:945-949
Scott L. Grossman,
Mary E. Moore
Rheumatoid arthritis (RA), and not osteoarthritis (OA) synovial cells proliferate in serum-free medium, a finding that suggests that, in vitro, RA synovial cells may be stimulated to grow by the continuous autocrine production of at least one polypeptide growth factor. Adding monoclonal antibody lD11.16, or rabbit polyclonal anti-tumor growth factor p (anti-TGF-/3) antibodies (both neutralizing antibodies to TGF-PI and TGF-&) to RA synovial cells, in culture, caused a significant reduction in cell growth, an effect not seen when other growth factor antibodies (platelet-derived growth factor [PDGF], epidermal growth factor [EGF], or EGF receptor) were added to the culture medium. Taken together, these data are consistent with the concept that RA synovial cell growth in vitro is driven endogenous TGF-/3. Moreover, when EGF was added to the culture medium, this caused the numbers of RA, and not OA, synovial cells to increase significantly. This finding suggests that RA synovial cells are in G, phase of the cell cycle; an effect that could be mediated by endogenous TGF-& o 1990 by W.B. Saunders Company.
Excessive cellular proliferation and turnover of extracellular matrix are common features in several different diseases, including rheumatoid arthritis, (RA), atherosclerosis, and several cancers.’ Recent cancer studies, as well as studies in fetal tissues and placenta, have established that polypeptide growth factors play an important role in regulating cell growth.2-5 Furthermore, the discovery that the protein products of several viral oncogenes show close homology with polypeptide growth factors, coded for by their human homologs, suggests a possible link between certain viruses and excessive cell growth.6-g RA is a disease of unknown etiology, which is characterized by excessive growth of the synovial mem-
From the Division of Rheumatology, Department of Medicine, Albert Einstein Medical Center, and Temple University School of Medicine, Philadelphia, PA. Presented in part at the 53rd Annual Scientific Meeting of the American College of Rheumatology, Cincinnati, Ohio, June 1989, and the Second International Workshop on Cytokines, Hilton Head, South Carolina, December 1989. Supported in part by an Albert Einstein Society grant #2-0731-601000. Address reprint requests to David H. Goddard MD, MRCP Rheumatic Diseases Research Laboratories. Room 107. Korman Research Pavilion, 12th Street and Tabor Roads, Philadelphia, PA 19141. o 1990 by W.B. Saunders Company. 1043/90/0202-0007$05.00/0 KEY WORDS: Rheumatoid arthritis/Osteoarthritis/Synovial Transforming growth factor-@/Autocrine growth
CYTOKINE,
Vol. 2, No. 2 (March),
1990: pp 149-155
cells/
brane, causing cartilage and bone destruction. In this disease, the synovium acts like a localized malignancy. In vitro, cultured RA synovial cells display several of the growth characteristics of neoplastic and virally transformed cells. These characteristics include the tendency to grow as disorganized monolayers with loss of normal contact-inhibited growth and piling-up of cells to form foci, the presence of many multinucleated cells, colony formation in soft agar, and tumor formation in nude mice.‘0-‘3 Similarities between rheumatoid synovitis and neoplasia have led to speculation that polypeptide growth factors play a role in causing synovial proliferation. The recently reported identification of several polypeptide growth factors and cytokines in synovial fluids, including transforming growth factor-p (TGF-P), plateletderived growth factor (PDGF), epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), the cytokines interleukin 1 (IL l), interleukin 6, and tumor necrosis factor-a further strengthens such speculation.‘4’15 We present evidence showing that RA synovial cell growth, in vitro, is driven by the continuous autocrine secretion of at least one polypeptide growth factor. Based on blocking studies using growth factor antibodies, we speculate that this growth factor may be TGF-0. An autocrine-secreted growth factor of this type, acting synergistically with other growth factors, could cause enhanced RA synovial cell growth.
149
150 / Goddard,
Grossman,
Moore
CYTOKINE,
RESULTS
Phenotypic Analysis of Synovial Cell Lines Phenotypic analysis was performed on cells taken from three RA and three OA synovial cell cultures. Paired cell cultures (RA and OA) were selected from representative passages (passage numbers 3, 10, and 23). Positive cytoplasmic staining of varying intensity was observed in all cells when reacted with monoclonal antibodies (MAbs) to vimentin but not cytokeratin. In contrast, cell-surface staining was negative when cells were reacted with antibodies, recognizing lymphocyte and monocyte/macrophage markers, or factor VIII antigen. Synovial cells grown in medium alone did not react with MAbs, recognizing human leukocyte antigensD-related (HLA-DR) antigens, Adding recombinant y-interferon to the culture medium (48 hr at 37OC; Du Pont, Wilmington, DE) caused these cells to express HLA-DR on their surface, as evidenced by positive surface staining.
Synovial Cell Growth in Serum-Free Medium Figure 1 shows the significant increase in mean (* SEM) cell number in three RA, when compared with three OA synovial cell cultures, whose phenotypes had been determined in the preceding experiments, and which were grown for 14 d in RPM1 plus 2% bovine serum albumin (BSA) (RA d 14 = 3.71 k 0.62, d 1 = 1.76 * 0.21 x 104, P < 0.05; t-test, two-tailed; OA d 14 = 1.55 * 0.3 x 104;d 1 = 1.43 * 0.18 x 104). Since RA synovial cell growth could have resulted from the effects of contaminating mitogens in BSA, five RA synovial cell lines were then grown in RPMI, supplemented with only nonessential amino acids, and
Vol. 2, No. 2 (March
1990:
149-155)
antibiotics. Figure 2 shows that by d 14, the number of RA synovial cells had increased significantly (d 14 - day 1 = 2.16 + 0.68 x lo4 cells; P < 0.04; paired t-test, two-tailed). By d 35, this had increased further to 8.1 f 1.56 x 1O4cells, which was significantly greater than d 14 (P < 0.002; paired t-test, two-tailed). To exclude the possibility that synovial cell growth in serum-free RPM1 occurred because of the persisting effects of fetal bovine serum (FBS), RA synovial cell growth in serum-free RPM1 (RPM1 plus nonessential amino acids and antibiotics) was compared with that in serum-free Dulbecco’s modified essential medium (DMEM) (DMEM plus nonessential amino acids and antibiotics). Because serum-free DMEM rapidly induces G, arrest in fibroblasts in culture, we hypothesized that significant growth in both media would indicate that previous exposure to FBS continued to provide the stimulus for synovial cell growth. Figure 3 shows that by d 14 there was a significant increase in cell number (mean * SEM) in four RA cell lines grown in RPMI, but not DMEM (RPM1 d 14 = 16.2 f 3.8 x 104, d 1 = 3.2 ? 0.6 x 104, P K 0.05; t-test, two-tailed. DMEMd 14 = 3.0 f 0.7 x 104,d 1 = 3.2 k 0.6 x 104).
Influence of Growth Factor Antibodies on RA Synovial Cell Growth We then began experiments with the aim of blocking RA synovial cell growth by adding specific growth factor antibodies to the culture medium. Figure 4 shows that a significant reduction in cell number occurred when four RA synovial cell cultures were grown for 14 d in serum-free RPM1 plus rabbit antibodies to TGF-P. This effect was observed at an antibody concentration of 5 pg/ml (P < 0.04; paired t-test, one-tailed) and was
Figure 1. Growth of RA and OA synovial cells in serum-free medium.
1
8
Days
in Culture
14
Values shown are the mean + SEM number of RA (W), and OA cells (O-O) taken from paired cell cultures, which had undergone phenotypic analysis and which were grown in RPM1 plus 2% BSA. By d 14, the number of RA, but not OA synovial cells had increased significantly compared with d 1 (*RA cell number compared with day 1 P < 0.05; 2-tailed; **RA vs OA P < 0.05; 2-tailed).
Autocrine
growth
of synovial
cells
/ 151
Influence of ExogenousGrowth Factors on Synovial Cell Growth
L
= u” 6.E f b -r
I
/
1
4-
I/
./. A’
f6 2-
.'
0’
7
14
21 Days
Figure 2. non-essential
Growth of RA synovial amino-acids.
28
35
in Culture
cells in RPM1
supplemented
with
Values shown are the mean + SEM increase in RA synovial cell numbers over baseline, when grown in RPM1 supplemented with non-essential amino-acids. By d 14, the number of cells had increased significantly compared with day 1 (P -C 0.04; paired t-test, two-tailed). By d 35, cell number had increased further and was significantly greater than d 14 (P < 0.002; paired t-test, two-tailed).
maximal at an antibody concentration of 100 pg/ml. In contrast, no reduction in growth occurred when synovial cells were grown in serum-free RPM1 containing antibodies to PDGF, or EGF (0.1 to 50 /*g/ml IgG), or MAb to EGF receptor (0.1 to 50 pg/ml IgG; data not shown). Next, three of the four RA synovial cell cultures used in the preceding experiments were grown in serum-free RPM1 plus MAb lDll.16. Figure 4 shows that, when added to the culture medium in concentrations as low as 0.1 pg/ml, this antibody caused a significant reduction in cell growth (P < 0.02; paired t-test, one-tailed), an effect that was maximal at an antibody concentration of 5 pg/ml (P < 0.03; paired t-test, one-tailed). To confirm that neither MAb lD11.16 nor polyclonal antibody to TGF-0 were cytotoxic for RA synovial cells, cell lines used in the preceding experiments were grown for 15 d under the following conditions: serum-free medium alone; serum-free RPM1 plus TGFfi (50 rig/ml); or serum-free RPM1 plus MAb lD11.16 (1 pg/ml), or rabbit TGF-/3 antibody (50 pg/ml); or serum-free RPM1 plus MAb 1D 11.16 (1 pg/ml) plus TGF-P (11 rig/ml), or rabbit antibody to TGF-P (50 pg/ml) and TGF-fi (50 rig/ml). As illustrated in Fig. 5, the significant reduction in synovial cell growth that occurred in serum-free RPM1 plus either MAb 1D 11.16 (serum-free RPM1 plus MAb 1Dll. 16 vs serum-free RPM1 with and without TGF-0, P < 0.02; t-test, onetailed) or rabbit TGF-fl antibody (serum-free RPM1 plus rabbit antibody vs serum-free medium with and without TGF-fi P < 0.03; one-tailed) was totally abolished by adding an excess of the growth factor.
The preceding experiments suggest that RA synovial cells in culture secrete TGF-0. Because TGF-P and PDGF stimulate cells to enter G, phase, while EGF stimulates cells in G, to enter S phase, the next experiments compared the growth of three RA and three OA synovial cell lines in RPM1 plus 2% BSA (serum-free medium), with growth in serum-free medium plus EGF (6.5 rig/ml), or PDGF (5 rig/ml). Figure 6 shows that by d 21, the number of RA synovial cells grown in serum-free medium plus EGF was significantly greater than cells grown in serum-free alone (P < 0.04; t-test, one-tailed). The capacity of EGF to induce RA cell growth was similar to that seen with PDGF. This finding contrasted with the modest increase in the number of OA synovial cells grown in serum-free medium, containing either of these growth factors.
DISCUSSION It is now clear that polypeptide growth factors control the growth of normal as well as neoplastic cells.16-22The recent identification of several of these in rheumatoid synovial effusions suggests that they may cause excessive synovial cell growth.14 RA and OA synovial cells grow differently in culture,” with RA synovial cells displaying several characteristics, which suggest transformation.“-I3 Since growth-factor-regulated genes are believed to play a critical role in cellular transformation,21’22 and because the products of some
25
Day 1 Day 14 Days in Culture Figure growth
3. Growth in RPMI.
of RA
synovial
cells
in DMEM
compared
with
Values shown are the mean + SEM number of RA synovial cells grown for 14 d in DMEM (O), and RPM1 (W), supplemented with non-essential amino acids and antibiotics. By d 14, the number of cells grown in RPM1 had increased significantly (d 14 vs d 1; P < 0.05, paired t-test). This contrasted with the lack of cell growth in DMEM.
152 /
Goddard,
Grossman,
CYTOKINE,
Moore
Figure 4. Influence of TGF-j3 antibodies on RA synovial cell growth.
80 60
0.01
0.1
1.0
10.0
100
ANTIBODY CONCENTRATION (micrograms
/ ml of IgG)
factor properties, we have investigated the possibility that RA synovial cell growth could be regulated by these factors. Our finding that RA, but not OA, synovial cells in long-term culture, grow in RPM1 lacking added growth factors, suggests that in vitro RA synovial cells have the capacity to synthesize those factors essential for growth. We believe that it is unlikely that RA synovial cell growth was mediated by a growth factor-like contaminant present in the medium, because the experiments used RPM1 with different batch numbers, and because OA synovial cells failed to grow in this medium. Phenotypic analysis confirmed that, our synovial cell cultures are not contaminated with either macrophages or longlived lymphocytes, whose secretory products could modactivated
ulate
oncogenes
synovial
Vol. 2, No. 2 (March 1990: 149- 155)
cell
have growth
function
in vitro.
Moreover,
Values shown are mean + SEM number of RA synovial cells grown for 15 d in serum-free medium plus MAb lD11.16 (0.1 - 10 pglml IeG). or rabbit oolvclonal TGF-B antibody (1 to -106 pg/ml Igd) expressed as percent of control (control = increase in synovial cell number at d 15 over d 1 in RPM1 supplemented with 2% BSA). A significant reduction in growth occurred at a MAb antibody concentration 0.1 rg/ml IgG (P < 0.02; paired t-test, one-tailed) which was maximal at an antibody concentration of 10 pg/ml IgG (P < 0.03; paired t-test, one-tailed), and at a rabbit polyclonal antibody concentration of 5 rg/ml IgG (P < 0.04; paired t-test, one-tailed), which was maximal at an antibody concentration of 100 pg/ml IgG.
pyridoxine hydrochloride, vitamin B- 12, aminopterin, reduced glutathione, hypoxanthine, and thymidine, the absence of which may affect synovial cell growth. We found that when added to serum-free RPMI, both MAb 1Dl 1.16 and polyclonal antibody to TGF-@, but not PDGF, or EGF, or EGF receptor antibodies, caused a significant reduction in RA synovial cell
0 El+ m+mAb E!J + eZa +
RPM TGF-Beta Rabbit Ab TGF-beta +Ab
the
finding that cells from each of the cultures showed positive staining with MAbs to vimentin, but not cytokeratin,
indicates
that they are of mesenchymal
origin.
Our observations that RA synovial cell numbers significantly increased in RPMI, but not DMEM, when supplemented only with nonessential amino acids, shows first, that RA cell growth occurs independently of mitogens, which may be present as contaminants in BSA, and second, that RA synovial cell growth is not likely to have occurred through a persistent effect of previous exposure to FBS. We chose serum-free DMEM because it rapidly induces G, arrest in mesenchymal cells in culture.23 At present, we are unsure why RA synovial cells grow in serum-free RPM1 and not serumfree DMEM, but we speculate that this may be related to differences in the composition of the two types of media. Specifically, DMEM lacks L-Asn, ~-Asp, L-Glu, hydroxy-L-Pro,
L-Pro,
D-Biotin,
p-amino
benzoic
acid,
2
Figure synovial
+ p(O.02 **p(o.o3
5. Lack of cytotoxic cell growth in serum-free
effect of TGF-8 medium.
antibodies
on RA
Values shown are mean + SEM increases in the numbers of cells grown for 15 d in serum-free medium alone, or serum-free medium plus TGF-8 (50 rig/ml), or mab lD11.16 (1 rig/ml), or TGF-P antibody (50 pg/ml), or an excess of TGF-@ plus antibody. The antiproliferative effects of both antibodies could be overcome by adding TGF-8 in excess (serum-free RPM1 plus mab lD11.16 vs serum-free RPM1 with and without TGF-fl P < 0.01; serum-free RPM1 plus rabbit TGF-fi antibody vs serum-free RPM1 with and without TGF-& P < 0.03; both t-test, two-tailed).
Autocrine growth of synovial cells / 153 4.0,
*
OA Figure growth.
6.
Influence
of exogenous
RA growth
factors
on synovial
cell
Values shown are mean * SEM increase in cell number at d 21 over baseline. Significant growth occurred in RA synovial cells in serumfree medium plus EGF (Es9)compared with serum-free medium alone @Zl$1):*P < 0.041. A similar increase occurred in the number of RA dynivial cells &own in serum-free medium plus PDGF (0). These findings contrasted with the lack of significant growth of OA cells in serum-free medium alone, or serum-free medium plus either EGF, or PDGF.
growth. Together with the observation that the effects of both of these antibodies could be overcome by adding an excess of exogenous TGF-/3, it is suggested that endogenous TGF-P may provide this mitogenic stimulus. Moreover, they show that neither the monoclonal, nor rabbit polyclonal, antibodies to TGF-P antibody were cytotoxic for RA synovial cells, in culture. These data are strongest for MAb 1Dll. 16, which recognizes TGF-/3, in Western analysis, and completely neutralizes the bioactivity of TGF-/3, and TGF-&.24 These findings contrast with the much weaker effect of the rabbit polyclonal antibodies on RA synovial cell growth. While the explanation for these differences is not known, we speculate that they may be related to difficulties in raising antibodies against TGF-/3. Because of the highly conserved structure of this peptide, its ubiquity, and its immunosuppressive activity, TGF-@ is a weak immunogen. 25These properties could lead to the generation of polyclonal antibodies in low-titer, which only weakly neutralize the biologic activity of TGF-P. Polypeptide growth factors can be broadly divided into two groups, depending on the point at which they act in the cell-cycle.6 TGF-P, and PDGF, are examples of growth factors, which stimulate quiescent cells (G, phase) to enter G, phase, and have been called competence factors. EGF, somatomedins, and IL 1, which stimulate cells in G, phase to enter S phase and divide, have been called progression factors.26*27 While competence and progression factors act synergically to stimulate cell division, certain differences in their individual properties are apparent. For example, only brief exposure to PDGF is required for Go-arrested cells to enter
G, phase. Longer exposure to TGF-P is necessary for quiescent cells to enter G, phase. Leof et a1.28 have proposed that TGF-P may act as an indirect mitogen by inducing c-sis mRNA transcription in quiescent fibroblasts. In these studies, they have shown that c-sis induction is accompanied by the release of a peptide with PDGF-like properties, which they suggest may provide the mitogenic stimulus for cell growth. Moreover, they have shown that TGF-fi induces at least two other PDGF-inducible genes, c-myc and c-jiis. Our finding, of a lack of effect of PDGF antibody on RA synovial cell growth, is not in conflict with this proposal, because induced PDGF, responsible for stimulating these cells, could remain in a cellular compartment inaccessible to the antibody.28 TGF-fl and PDGF both enhance the mitogenic effects of EGF in some cells by increasing the affinity and number of receptors for this growth factor.2g330Our observation that RA, but not OA synovial cells grown in serum-free medium with EGF, increased significantly in number, when compared with those grown in serum-free medium alone, indicates that these cells are in G, phase, an effect which could be mediated by endogenous TGF-P. While TGF-/3 has been reported to have variable biologic effects, depending on the cell type and conditions of culture, this growth factor regulates cell replication and differentiation, angiogenesis, extracellular matrix turnover, as well as stimulating mononuclear cell release of PDGF and IL 1.11,30-34Based on our results, together with the known biologic properties of TGF-/3, we speculated that autocrine-secreted TGF-P acts synergically with other growth factors and cytokines to stimulate RA synovial cell growth. The recent report by Lafyatis et a1.,35identifying TGF-/3 in the intact rheumatoid synovium using in situ immunohistochemical techniques, further strengthens such speculation. We both believe that many of the features of rheumatoid synovitis can be explained on the basis of local production and release of this growth factor, which could play a key role in synovial cell proliferation and subsequent fibrosis.
MATERIALS
AND
METHODS
Matevials Monoclonal antibody (MAb) 1Dll. 16, which neutralizes TGF-/3, and &, was a gift from Dr. James Dasch, Celtrix Laboratories, Palo Alto, CA. All other antibodies were purchased as follows: Polyclonal antibody to TGF-P (R & D Systems, Minneapolis, MN); antibodies to PDGF and EGF (Genzyme, Boston, MA); EGF receptor antibody (ICN Immunochemicals, Lisle, IL); anti-Leu 4 (CD3, T-cells), anti-Leu 14 (CD22, B-cells), anti-Leu M3 (mature monocyte), anti-leu1 la (large granular lymphocyte/NK cells) and peroxidase goat-anti mouse immunoglobulin (Becton Dickinson, Mountain View, CA); MAbs to vimentin and cytokeratin, sheep polyclonal antibody to Factor VIII antigen; and peroxidase goat anti-sheep immunoglobulin (ICN). Recombinant y inter-
154 / Goddard,
Grossman,
CYTOKINE, Vol. 2, No. 2 (March 1990: 149-155)
Moore
feron was a gift from Dr. Robert Newton (E. I. DuPont de Nemours & Co, Glenolden, PA). PDGF was purchased from R & D Systems, and EGF from Sigma Chemical Co., St. Louis, MO. Monoclonal antibody lD11.16 neutralizes the biologic activity of TGF-0, and & in vitro in a dose-dependent manner. 24TGF-@ antibody is a polyclonal, neutralizing antibody raised in rabbits that recognizesTGF-/3, and & and does not crossreact with acidic or basic FGF, PDGF, or EGF. Rabbit anti-PDGF is an IgG polyvalent antibody generated against purified, recombinant PDGF-BB homodimer. This antibody recognizes both human PDGF-BB BB homodimer and PDGF-AB heterodimer, and does not crossreact with human PDGF-AA. Monoclonal EGF receptor antibody is a mouse IgG,, obtained from C57BL/6 mice immunized with the paraformaldehyde-fixed A43 1 carcinoma cell line containing approximately 2 x lo6 EGF receptors per cell. This MAb reacts with human and mouse EGF receptors. Rabbit antiEGF is a polyclonal IgG neutralizing antibody raised in rabbits against natural murine EGF (receptor grade), and in ‘H uptake assaysusing rat primary hepatocytes, 2 to 4 kg of this antibody neutralizes 50% of the biologic activity of 10 ng of EGF. PDGF isolated from human platelets is >95% pure when analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). EGF is extracted from the submaxillary glands of mice and lyophilized from a solution of 5 mM ammonium acetate, pH 5.6. RPM1 1640 (RPMI), DMEM, Hank’s balanced salt solution (HBSS), nonessential amino acids, streptomycin, penicillin, and fungazone, were all purchased from Flow Labs (McLean, VA). FBS was purchased from Hyclone (Logan, UT), crude collagenasefrom Sigma Chemical Co., BSA from Miles Scientific Co. (Naperville, IL), and trypsin/EDTA from Gibco Laboratories (Grand Island, NY). All tissue culture plastics were purchased from Falcon (Oxnard, CA). Synovial samples were obtained with informed consent at the time of joint replacement surgery from six patients with a diagnosis of definite RA,36and three patients with OA.
Synovial
Cell Culture
Synovial cell cultures were established using a modification of the method described by Dayer et a1.37Synovial samples cut into small pieces (~5 mm’), were incubated overnight in RPMI, and supplemented with nonessential aminoacids (100 pg/ml), streptomycin (100 pg/ml), penicillin (100 h/ml), fungazone (25 pg/ml), 10% FBS, and crude collagenase (200 p/ml), at 37OCin 10% CO, in air. Dispersed cells were then collected by low speed centrifugation, washed twice in HBSS, once in RPM1 plus 10% FBS, resuspended in 8 ml of RPM1 plus 10% FBS, and seeded into two tissue culture flasks (25 cm’). Once primary cultures had reached confluence they were split weekly, and used in passages3 through 23.
Phenotypic
Analysis
of Synovial
Cells
Synovial cells (RA and OA) grown in RPM1 plus 10% FBS were detached from the culture flasks by adding trypsin/ EDTA, washed with RPM1 plus 10% FBS, and seededat low density on to tissue culture slides. Synovial cells were allowed to grow to the point where visual inspection confirmed adequate numbers of cells for staining, but before the cultures had
reached confluence. After washing in several changes of DMEM to remove traces of biotin present in RPMI, cells were fixed in acetone at 4OC for 30 min and then air-dried. Cells were then reacted with the following antibodies: anti-Leu 4, anti-Leu 14, anti-Leu M3, anti-Leu 1la, anti-vimentin, anticytokeratin, and anti-Factor VIII antigen. These were all used in accordance with the manufacturers’ recommendations. After washing in phosphate buffered saline (PBS), the cells were then incubated with either peroxidase goat anti-mouse immunoglobulin or peroxidase goat anti-sheep immunoglobulin as appropriate, washed in PBS, reacted with amino-ethylcarbazole (AEC) solution, washed in distilled water, counterstained with progressive hematoxylin, cleared in tap water, and then mounted in glycerin jelly. The slides were then examined in an Olympus BHS light microscope (Olympus Corporation of America, Hyde Park, NY).
Synovial
Cell Growth Studies
In the period preceding the start of each study, synovial cells (RA and OA) were grown to confluence in RPM1 containing 5% FBS. Cells were then detached by adding trypsin/EDTA, and plated in 24-well tissue culture plates (1.5 x lo4 cells/well). The next day, the medium was changed to serum-free medium (RPM1 plus 2% BSA) alone, or serum-free medium plus either growth factors (PDGF or EGF), or growth factor antibodies (TGF-fi, PDGF, EGF, or EGF receptor). Cultures were fed twice weekly by removing half of the medium and replacing it with a similar volume of fresh medium. Cell growth was measured by determining the increasein cell number over baseline at specifictime intervals during each experiment. Cells were detached from the plates by adding trypsin/EDTA, transferred immediately into isotonic solution (Isoton, Coulter Diagnostics, Hialeah, FL), and counted using a Coulter counter (model ZBI, Coulter Diagnostics). Triplicate counts were made on duplicate samples, and the mean value of these counts was used in further data analysis, using paired t-tests. Significance was set at the 5% level.
Acknowledgement We thank J. Dasch PhD (Celtrix Laboratories, Palo Alto, CA), for the gift of MAb 1Dll. 16 and R. Newton PhD (E. I. DuPont de Nemours & Co), for the kind gift of recombinant y-interferon. We also thank B. Wilson PhD for her helpful comments and criticisms.
REFERENCES 1. SpornMB, Harris ED Jr (1981):Proliferativediseases. Am J Med 170:1231-1236 2. SpornMB, RobertsAB (1985):Autocrinegrowth factorsand cancer.Nature 313:745-747 3. JamesR, BradshawRA (1984):Polypeptidegrowth factors. Ann RevBiochem53~259-292 4. RossR, GlomsetJ, Kariya B, Harker LA (1984):A plateletdependentserum factor that stimulatesthe proliferation of arterial smoothmusclecellsin vitro. ProcNat1 AcadSci USA 71:1207-1210 5. GajdusekC, DiCorleto P, RossR, SchwartzSM (1980): An endothelialcell-derivedgrowth factor.J Cell Bio185:467-472
Autocrine
6. Armelin HA, Armelin MCS peptide growth factors and oncogenes. genes, and growth factors. G. Wiley 361-363
(1987): The interactions of In Guroff G (ed). Oncogenes, Interscience, New York, pp
growth
of synovial
cells
/ 155
22. Anastassiades TP, Ley L, Wood A, Irwin D (1978): growth kinetics of synovial fibroblastic cells from inflammatory noninflammatory arthropathies. Arthritis Rheum 21:461-466 23. Greenberg transcription
ME, Ziff EB (1984): Stimulation of the c-j& protooncogene. Nature
The and
of 3T3 cells 3 11:433-438
7. Waterfield MD, Scrace GT, Whittle N, Stroobant P, Johnsson A, Wasteson A, Westermark B, Heldin CH, Huang JS, Deuel TF (1983): Platelet-derived growth factor is structurally related to the putative transforming protein ~28”‘” of simian sarcoma virus. Nature 304:35-39
induces
8. Downward J, Yarden E, Mayes E, Scrace G, Totty N, Stockwell P, Ulrich A, Schlessinger J, Waterfield MD (1984): Close similarity of epidermal growth factor and v-erb-0 oncogene protein sequences. Nature 307:521-527
25. Leung BS, Zhou L (1988): A fluorimetric enzyme-linked immunosorbent assav (F ELISA) for TGF-B. In Powanda MC. Oppenheim JJ, Kluger‘ MJ, Dinarello CA (Eds.) Monokines and Other Non-Lymphocytic Cytokines. Alan R Liss, New York, pp 397-400
9. Coussens L, Yang-Feng TL, Chen Y-CLE, Gray A, J, Seeburg PH, Liberman TW, Schlessinger J, Francke U, A, Ulrich A (1985): Tyrosine kinase receptor with extensive to EGF receptor shares chromosomal location with neu Science 230: 1132- 1139 sponses require Arthritis
10. Brinckerhoff CE (1983): of rabbit synovial fibroblasts transforming growth factor Rheum 26:1370-1379
McGrath Levinson homology oncogene.
Morphologic and mitogenic reto transforming growth factor p LY or epidermal growth factor.
11. Brinckerhoff CE, Harris ED Jr (1978): Collagenase tion by cultures containing multinucleated cells derived from tibroblasts. Arthritis Rheum 21:745-753
producsynovial
12. Brinckerhoff CE, Harris ED Jr (1981): Survival of rheumatoid synovium implanted into nude mice. Am J Path01 103:41 l-419 13. Lafyatis R, Remmers EF, Roberts A, Yocum D, Sporn MB, Wilder RL (1989): Anchorage-independent growth of arthritic synoviocytes: Dependence on platelet-derived growth factor (PDGF) and inhibition by transforming growth factor-p (TGF-6) and retinoids. J Clin Invest 83:1267-1276 14. Lafyatis R, Remmers EF, Roberts A, Sporn MB, Wilder (1988): Regulation of rheumatoid synovial cell proliferation cytokines. Arthritis Rheum 31 (suppl l):A62
RL by
15. Hamerman D, Taylor S, Kirschenbaum I, Kalgsbrun M, Raines EW, Ross R, Thomas KA (1987): Growth factors with heparin binding affinity in human synovial fluid. Proc Exp Biol Med 186:384389 16. Stiles CD, Capone GT, Scher Wyk JJ, Pledger JJ (1979): Dual control dins and nlatelet-derived growth factor. 76:1279-l&3
CD, Antoniades HN, Vanof cell growth by somatomeProc Nat1 Acad Sci USA
17. Assoian RK (1985): Biphasic effects of type fl transforming growth factor on epidermal growth factor receptors in NRK fibroblasts. Functional consequences for epidermal growth factorstimulated mitosis. J Biol Chem 260:9613-9617 18. Beach D, Basilic0 C, Newport J (1988): In D Beach, C Basilica, and J Newport (eds.) Cell Cycle Control in EukayotesCurrent Communications in Molecular Biology. Cold Spring Harbor Laboratory, New York, pp 2-8 19. Heine U, Munoz EF, Flanders KC, Ellingsworth LR, Lam HY, Thompson NL, Roberts AB, Sporn MB (1987): Role of transforming growth factor-p in the development of the mouse embryo. J Cell Biol 105:2861-2876 20. Frolik CA, Dart LL, Myers CA, Smith DM, Sporn MB (1983): Purification and initial characterization of a type fl transforming growth factor from human placenta. Proc Nat1 Acad Sci USA 80:3676-3680 21. Derynck R, Jarrett JA, Chen EY, Eaton DH, Bell JR, Assoian RK, Roberts AB, Sporn MB, Goeddel DV (1985): Human transforming growth factor-p complementary DNA sequence and expression in normal and transformed cells. Nature 316:701-705
24. Dasch JR, Pace DR, Waegell W, Inenaga D, Ellingsworth L (1989): Monoclonal antibodies recognizing transforming growth factor@. Bioactivity, neutralization, and transforming growth factor p2 affinity purification. J Immunol 142:1536-1541
26. Bravo R (1988): Complexity of the early to growth factors in mouse fibroblasts. In D Beach, Newport (eds) Cell Cycle Control in Eukaryotes-Current cations in Molecular Biology. Cold Spring Harbor York, pp 15-21
genomic response C Basilica, and J CommuniLaboratory, New
27. Singh JP, Adams LD, Bonin PD (1988): Mode of fibroblast enhancement by human interleukin 1. J Cell Biol 106:8 13-8 19 28. Leof EB, Proper JA, Goustin AS, Shipley GD, DiCorleto PE, Moses HL (1986): Induction of c-sis mRNA and activity similar to platelet-derived growth factor by transforming growth factor p: A proposed model for indirect mitogenesis involving autocrine activity. Proc Nat1 Acad Sci USA 83:2453-2457 29. Davis RJ, Czech MP (1987): Stimulation of epidermal growth factor receptor threonine 654 phosphorylation by plateletderived growth factor in protein kinase C-deficient human fibroblasts. J Biol Chem 262:6832-6841 30. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS, Wakefield LM, Heine U, Liotta LA, Falanga V, Kehrl JH, Fauci AS (1986): Transforming growth factor type B: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Nat1 Acad Sci USA 83:4167-4171 31. Sporn MB, Roberts MB, Wakefield LM, de Crombrugghe (1987): Some recent advances in the chemistry and biology transforming growth factor-@. J Cell Biol 105:1039-1045
B of
32. Wahl SM, Hunt DA, Wakefield LM, McCartney FN, Wahl LM, Roberts AB, Sporn MB (1987): Transforming growth factor type 0 induces monocyte chemotaxis and growth factor production. Proc Nat1 Acad Sci USA 84:5788-5792 33. Martinet Y, Bitterman PB, Mornex JF, Grotendorst GR, Martin GR, Crystal RG (1986): Activated human monocytes express the c-k proto-oncogene and release a mediator showing PDGF-like activity. Nature. 319:158-160 34. Baird A, Mormede P, Bohlen P (1985): Immunoreactive fibroblast growth factor in cells of peritoneal exudate suggests its identity with macrophage-derived growth factor. Biochem Biophys Res Comm 126:358-364 35. Lafyatis R, Thompson NL, Remmers EF, Flanders KC, Roche NS, Kim S-J, Case JP, Sporn MB, Roberts AB, Wilder R (1989): Transforming growth factor-B production by synovial tissues from rheumatoid patients and streptococcal cell wall arthritic rats: Studies on secretion by synovial fibroblast-like cells and immunhistologic localization. J Immunol 143:1142-l 148 36. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, et al. (1988): The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 3 I:3 15324 37. Dayer J-M, Krane SM, Russell RGG, Robinson DR (1976): Production of collagenase and prostaglandins by isolated adherent rheumatoid synovial cells. Proc Nat1 Acad Sci USA 73:945-949
