By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Rheumatoid arthritis (RA) is an autoimmune disease characterized by a heavy lymphocytic infiltration into the synovial cavity, resulting in the secretion of a variety of cytokines which ultimately leads to destruction of joint tissue. Among the infiltrating cells are activated T cells which produce specific cytokines capable of osteoclast progenitor cell expansion, fusion, and activation. Cultures of activated human T cells and human osteoblasts (hOBs) were used to study the possibility that lymphokines may act on osteoblasts to produce the osteoclastogenic factor interleukin-6 (IL-6). Purified T cells were activated with a combination of anti-CD3 and anti-CD28 antibodies, cocultured with hOBs in direct physical contact or separated by a transwell system, and conditioned media (CM) were assayed for IL-6 production. After a 72 h incubation period, activated T cell–hOB interaction resulted in a 100-fold increase of IL-6 production over basal levels. The immunosuppressant cyclosporine A (CsA) inhibited T cell tumor necrosis factor alpha and IL-6 production but did not inhibit the T cell induction of IL-6 from hOB. Assay of activated T-cell CM on hOB revealed that a soluble factor, not cell-cell contact, was the major inducer of IL-6. The induction of IL-6 mRNA by both activated T cell CM and CsA-treated activated T cell CM was confirmed by Northern blot analysis. Neutralizing antibodies to IL-13 and IL-17 did not affect IL-6 production. These findings suggest that activated T cells produce a novel, potent, IL-6 inducing factor that may be responsible for the bone loss observed in RA patients.
The earliest observation that osteoporosis is involved in rheumatoid arthritis (RA) was made by Barwell in 1865.(1) Osteoporosis is frequently recognized in patients with RA and is observed in two characteristic patterns: juxta-articular bone loss, occurring around inflamed joints, and generalized bone loss with reduced bone mass (reviewed elsewhere(2-6)). Juxta-articular bone loss occurs early after onset of the disease and may be mediated by the inflammatory process driven by T cell infiltration and activation,(7,8) increased vascularity of the synovial space,(4) and localized production of cytokines.(9,10) Although presently controversial, the generalized osteopenia, often seen in patients with a longstanding history of RA,(5,11,12) may derive from either long-term corticosteroid treatment(2,4,6) or treatment with cyclosporine A (CsA).(13–15) In untreated patients, circulating cytokines and reduced mobility due to functional impairment may be critical factors. However, the pathogenesis of both RA and the associated bone loss is still not well established, but a growing body of evidence has implicated T cells as pivotal in this process.(16,17).
The role of T cells in regulating bone turnover has not been well explored at this time. T cells are classified into three subsets, Th1, Th2, and Th0, based on their pattern of cytokine production. Th1 cells produce interleukin-2 (IL-2) and interferon-γ (IFN-γ); Th2 cells produce IL-4, IL-5, IL-6, IL-8, IL-10, and IL-13; and Th0 cells produce a mixture of cytokines produced by both Th1 and Th2 cells (reviewed elsewhere(18,19)). CD4+ T helper cells secrete the recently discovered cytokine IL-17.(20) T cells also produce transforming growth factor-beta (TGF-β) and tumor necrosis factor-alpha (TNF-α). Of the cytokines listed, TGF-β, IFN-γ, and IL-4 have all been shown to inhibit osteoclastogenesis (reviewed elsewhere(21)), while TNF-α(22) and IL-6(23) have been shown to induce osteoclastogenesis. CsA is an immunosuppressant that selectively suppresses T cell activation by blocking the gene transcription of various cytokines such as IL-2, IL-3, IL-4, IL-5, IL-6, IL-8, IFN-γ, granulocyte macrophage colony stimulating factor,(19,(24-26) and TNF-α.(27) CsA is one of several therapeutic agents used to treat RA.(28-30) However, since bone resorption can ensue in RA patients during CsA therapy, we have examined activated T cells for cytokine synthesis which can induce IL-6 in osteoblasts and thus may account for the juxta-articular bone loss seen in RA. The results of this study demonstrate that a novel T cell cytokine activity, resistant to CsA, is a potent inducer of IL-6 in human osteoblasts (hOBs).
MATERIALS AND METHODS
All reagents were obtained from Sigma Chemical Co. (St. Louis, MO, U.S.A.) unless otherwise indicated. CsA was prepared as a stock solution at 1 mg/ml in ethanol then diluted to the desired concentration in medium.
Antibodies and recombinant cytokines
Mouse anti-human CD3 monoclonal antibody (MAb) (clone HIT3a) and mouse anti-human CD28 MAb (clone 28.2) were obtained from Pharmingen (San Diego, CA, U.S.A.), as sterile, azide free, and low endotoxin preparations. Neutralizing antibodies against human IL-13 and IL-17 were obtained from R&D Systems (Minneapolis, MN, U.S.A.). Antibodies for fluorescence-activated cell sorting analysis of isolated T cells, mouse anti-human CD3/CD16+CD56 (Simultest conjugates) were obtained from Becton Dickinson (Franklin Lakes, NJ, U.S.A.).
T cell isolation and culture
Peripheral blood mononuclear cells (PBMCs) were obtained in the form of buffy coats from the American Red Cross and further purified by separation on Histopaque (1.077 g/ml) lymphocyte separation medium. Briefly, the buffy coat preparations were diluted 1:1 in phosphate-buffered saline. Twenty milliliters of diluted buffy coat was over layered onto 15 ml of Histopaque and centrifuged at 400g for 30 minutes at room temperature. Mononuclear cells were recovered from the interface, washed twice with phosphate-buffered saline (Ca2+-, Mg2+-free; Sigma) by centrifugation at 300g, 5 minutes, then T cells were isolated from the PBMC using CD3+ T cell enrichment columns (R&D Systems) according to the manufacturer's instructions. Briefly, PBMCs were suspended in column wash buffer at 1 × 108 MNCs/ml. Two milliliters were applied to the column and incubated at RT for 10 minutes to allow non–T cells (B cells and monocytes) to bind. B cells bound, via F(ab)-surface immunoglobulin interactions, to glass beads coated with anti-immunoglobulin while monocytes bound, via Fc interactions, to the glass beads coated with immunoglobulin. T cells were eluted by sequentially applying 4 × 2 ml column wash buffer. The cells were diluted in medium, counted, pelleted by centrifugation at 300g, then resuspended in complete medium at 1–4 × 106 cells/ml. This process results in a population of T cells enriched ∼90% as determined by fluorescence-activated cell sorting analysis using antibodies against CD3 (T cells) and CD15 and CD56 (natural killer cells).
T cell cultures
T cells were cultured at 1 × 106 cells/ml in alpha modified minimum essential medium (α-MEM), supplemented with 1% L-glutamine, 10% heat-inactivated fetal bovine serum (HIFBS), 100 U/ml penicillin, and 100 μg/ml streptomycin. T cells were activated by the addition of mouse anti-human CD3 MAb (1 μg/ml) and mouse anti-human CD28 MAb (5 μg/ml) in the absence or presence of CsA. The cells were incubated for 72 h, then the conditioned media (CM) were harvested and frozen at −80°C until assayed for cytokine production or induction of osteoblast IL-6.
Preparation of human osteoblast cultures
hOB cultures were prepared as previously described from rib specimens.(31,32) Briefly, the trabecular surfaces of human ribs were exposed by using bone cutters and cleaned of loosely adherent marrow by washing repeatedly with culture medium. Trabecular bone was removed by scooping the surface with a size 0 bone curette and minced further with straight microdissecting scissors in a conical bottom centrifuge tube. The minced chips were washed extensively then subjected to collagenase treatment at 37°C on an orbital shaker, in Dulbecco's modified Eagle's medium/F-12 medium containing 240 U/ml of highly purified (low clostripain) collagenase (Boehringer Mannheim, Indianapolis, IN, U.S.A.), 2% fetal bovine serum, and 1 μg/ml DNAse in 50 ml centrifuge tubes. After 2 h, the bone chips were washed with serum-containing medium, the released cells discarded, and the resultant bone chips incubated in low calcium Dulbecco's modified Eagle's medium/F-12 (50:50, v:v), supplemented with 10% HIFBS, 50 μg/ml ascorbic acid, 100 U/ml penicillin, and 100 μg/ml streptomycin, and incubated in a humidified atmosphere of 4% CO2/95% air. The culture medium was changed every 3–4 days. After 1 week, the antibiotics were removed from the medium. Cells migrated from the chips within 2 weeks and were confluent in cultures within 4 weeks. At that time cells were subcultured for experiments by first washing the cells with serum-free α-MEM, then incubating the cell layer with collagenase (1 mg/ml Collagenase A in α-MEM; Boehringer Mannheim) for 10 minutes followed by an incubation with trypsin-EDTA (0.05–0.02%) for an additional 15 minutes. The cell suspensions were collected, added to an equal amount of α-MEM, 10% HIFBS, then pelleted by centrifugation at 300g for 10 minutes with reverse transcriptase. The cells were washed twice with serum-containing medium then seeded into culture plates in α-MEM, 10% HIFBS. Cultures were examined for the absence of mycoplasma using a Gen-Probe hybridization kit (Fisher Scientific Co., St. Louis, MO, U.S.A.). Cells were subcultured at confluence by treatment with trypsin-EDTA (0.05–0.02%, respectively), washed and suspended in α-MEM, supplemented with 10% HIFBS. The cells have the characteristics of osteoblasts as previously reported.(31–37)
Transwell cocultures of T cells and hOB
hOBs were plated in 24-well multiwell dishes at 5 × 104 cells/well in order to achieve confluence upon seeding, and incubated for 48 h. The medium was changed, and 0.8 ml of medium (α-MEM, 10% HIFBS) was added to the wells. Transwells (Corning Costar Corp., Cambridge, MA, U.S.A.) were placed in half of the wells, and either no T cells or 4 × 105 T cells, in 0.2 ml of medium, were added either to the transwell or to hOB cultures without transwells. Activation of T cells was performed by the addition of mouse anti-human CD3 MAb (1 μg/ml) and mouse anti-human CD28 MAb (5 μg/ml) in the absence or presence of CsA (100 ng/ml). Unactivated T cells received vehicle as control. Cultures were incubated for 72 h and the CM collected and frozen at −80°C until assayed for cytokine production.
Analysis of IL-6 induction in osteoblasts by T cell supernatants
T cells (1 × 106 cells/ml) were incubated in α-MEM, 10% HIFBS in the absence or presence of activating antibodies and in the absence or presence of CsA (100 ng/ml) for a 72 h period. The medium was collected and frozen at −80°C. For analysis of IL-6 inducing activity, hOBs were seeded into 96-well plates at 1 × 104 cells/well in 100 μl of α-MEM, 10% HIFBS. T cell supernatants (100 μl) were added to the wells, and the cells were incubated for 72 h. The resultant CM were collected and assayed for IL-6 production by enzyme-linked immunosorbent assay (ELISA). T cell supernatants were assayed for IL-6 by ELISA as well.
Dose response curves of T cell CM were performed as above except that CM were diluted in α-MEM, 10% HIFBS to achieve a final dilution of 1:2 to 1:16 in a final volume of 200 μl in the appropriate wells. Osteoblasts were incubated for 72 h, then the resultant CM were assayed for IL-6 by ELISA.
ELISA analysis of cytokine production
TNF-α, IL-1, and IL-6 were assayed in CM using kits obtained from Amersham (Arlington Heights, IL, U.S.A.) according to the manufacturer's instructions.
RNA isolation and Northern blot analysis
Primary hOBs were passed into 100 mm diameter tissue culture plates at 4 × 106 cells in 10 ml of α-MEM, 10% HIFBS, and allowed to settle and attach overnight. The next day, the medium was changed and the cells incubated with either medium alone, 25% activated T cell CM, or 25% CM from T cells activated in the presence of 100 ng/ml CsA. After a 72 h incubation period, the medium was removed and frozen at −80°C until assayed by ELISA and total RNA was extracted from the cell monolayers with an RNeasy kit (Qiagen, Chatsworth, CA, U.S.A.) according to the manufacturer's instructions. The quantity and quality of RNA were routinely tested spectrophotometrically using A260/A280. Twenty micrograms of total RNA from each preparation were electrophoresed on formaldehyde-containing 1% agarose gels containing 0.5 μg/ml ethidium bromide in order to visualize ribosomal RNA bands. Ribosomal RNA bands were photographed and RNA was then transferred to Nytran nylon membranes using the Turboblotter system (Schleicher & Schuell, Keene, NH, U.S.A.). The membranes were prehybridized with Hybrisol I (Oncor, Gaithersburg, MD, U.S.A.) at 42°C overnight followed by hybridization in the same solution but with cDNA probe overnight at 42°C. The membranes were probed first with [32P]deoxy-CTP-labeled cDNA for human IL-6,(38) a gift of Dr. Pravin B. Sehgal (New York Medical College, Valhalla, NY, U.S.A.), and reprobed with [32P]deoxy-CTP-labeled cDNA for human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a gift of Drs. Theresa Sunyer and Phil Osdoby (Washington University, St. Louis, MO, U.S.A.). The labeled probes were generated using the multiprime random primer labeling kit obtained from Amersham according to the procedures provided. The membranes were washed twice with 2× SSC and 0.1% SDS for 20 minutes at 42°C and a single wash with 0.2× SSC and 0.1% SDS at 52°C. Visualization of the extent of hybridization was performed by autoradiography at −80°C using Hyperfilm-MP (Amersham, Arlington Heights, IL, U.S.A.). Relative loading efficiency was determined by GAPDH levels.
Group mean values were compared by two-tailed Student's t-test or multifactor analysis of variance (ANOVA) as appropriate. Subsequent multiple comparison tests were performed by using the Fisher protected least significant difference test.
Activated T cells produce a soluble cytokine that induces IL-6 in hOB
We have examined whether activated T cells induce IL-6 in hOB by coculturing the two cell types either together or separated by a transwell system. The transwell system allows for the interaction of the two cell types without physical contact. Coculture of hOB with unactivated T cells did stimulate IL-6 secretion after a 72-h period of incubation, indicating that T cells may interact with osteoblasts resulting in a small (2-fold) increase in IL-6 induction (Fig. 1). When activated T cells were cocultured with hOB, IL-6 secretion was strikingly high (109-fold vs. control, p < 0.001), demonstrating that activated T cells induced hOB to produce IL-6 either by cell–cell contact or a soluble factor or both. Furthermore, when T cells were cocultured with hOB separated by a transwell system, the amount of IL-6 produced was only 30% lower (77-fold vs. control, p < 0.001) as compared with coculture of the two cell types in physical contact. CsA had no effects on basal IL-6 secretion from hOB. When T cells were cultured either physically with hOB or separated by transwells, IL-6 production was only partially and equally inhibited by CsA (∼25%). IL-6 secretion was unaffected by the presence of an anti–IL-13 neutralizing antibody in the culture (data not shown).
To determine whether PBMC cytokines known to induce IL-6 were responsible for the observed induction, we examined T cell supernatants for IL-1 and TNF-α activity. Neither activated T cells nor T cells activated in the presence of CsA produced any detectable IL-1 (data not shown). Furthermore, hOB did not constitutively produce IL-1, and incubation with CsA did not result in IL-1 production in the hOB cells (data not shown). Since TNF-α can induce IL-6 in hOB, we also determined whether TNF-α production by activated T cells was inhibited by the CsA treatment. The results showed that CsA inhibited TNF-α production by ∼90% (Fig. 2). Furthermore, the source of IL-6 was not activated T cells since these cells produce very low levels of IL-6 which were significantly inhibited by CsA treatment (Fig. 2).
Effect of antibodies on IL-6 production in hOB
To rule out the effect of the T cell activating antibodies on IL-6 production in hOB, osteoblasts were incubated in the absence or presence of mouse anti-human CD3 MAb (1 μg/ml) and mouse anti-human CD28 MAb (5 μg/ml) in the absence or presence of CsA (100 ng/ml). Neither antibody alone nor in combination in the absence or presence of CsA had a positive effect on basal hOB IL-6 secretion (Fig. 3).
To determine whether an interaction between activated T cells and hOB was necessary to induce hOB IL-6, T cells were cultured separately and activated for 72 h. The CM were then assayed on hOB for IL-6–inducing activity. CM from T cells activated in the absence of CsA induced a 118-fold increase in hOB IL-6 compared with control (p < 0.001), while CM from T cells treated with CsA induced a 77-fold increase in hOB IL-6 compared with control (p > 0.007) (Fig. 4). A dose response curve of the T cell CM demonstrated that as little as 6% CM induced a 12-fold increase in IL-6 (Fig. 5).
To examine the difference in hOB IL-6 induction by activated T cell–CM and CM from CsA-treated activated T cells, total RNA from hOB incubated with the two activated T cell–CM were subjected to Northern blot analysis and the resultant CM assayed for IL-6 by ELISA. The steady-state level of IL-6 mRNA was not detectable in unstimulated hOB. After stimulation with activated T cell–CM, a high induction of steady-state IL-6 mRNA was found. hOB treated with CsA-treated activated T cell CM exhibited a strong steady-state mRNA level which was about half that after stimulation with activated T cell CM (Fig. 6). Analysis of the CM for IL-6 by ELISA from the same cultures confirmed the Northern blot results.
We have also examined whether IL-17, a cytokine produced by activated CD4+ memory T cells which induces IL-6 in synovial fibroblasts,(39) may be responsible for the T cell induction of IL-6 in hOB. Incubation of CM from T cells, activated in the absence or presence of CsA, with an excess of neutralizing antibody against human IL-17, did not affect T cell CM induction of IL-6 in hOB (Fig. 7), suggesting that the T cell population consisted primarily of naive T cells.
We have examined several aspects of activated T cell regulation of IL-6 production in hOB cultures. Our data demonstrate that T cells are potent regulators of IL-6 in hOBs and that they secrete a variety of cytokines which together result in maximum IL-6 stimulation. We have determined that a soluble factor(s) is critical in the T cell induction of IL-6 production by hOB since the separation of activated T cells from osteoblasts via a transwell system had only a marginal effect on IL-6 induction. Differences in fold stimulation of IL-6 in different culture preparations reflects variance in basal IL-6 secretion by the osteoblasts obtained from 10 different donors. Nonetheless, the soluble factor(s) secreted by T cells was consistently found to be a potent inducer of IL-6.
Since activated T cells produce many factors that may induce IL-6 in osteoblasts and are sensitive to inhibition by the calcineurin inhibitor CsA,(19, 24–27) we have used CsA as a tool to unmask the T cell activity. Thus, this factor(s) is not likely to be one of the major known regulators of IL-6, such as IL-1 or TNF-α, since TNF-α is significantly inhibited by CsA and the T cells do not produce IL-1. Nonetheless, Northern blot analysis has confirmed that despite the inhibition of known IL-6–inducing cytokines, T cells produce a factor that is a potent inducer of IL-6 mRNA. The difference noted between the CsA-treated activated T cell CM and the nontreated activated T cell CM most probably reflects the additive effects of TNF-α and perhaps other CsA-sensitive IL-6–inducing cytokines produced in the absence of CsA. Furthermore, the factor is neither IL-13 nor IL-17 since neutralizing antibodies had no effect on activity.
The accumulated data provide evidence that a novel soluble T cell factor(s), resistant to CsA, is secreted from activated T cells which regulates at least one osteoclastogenic factor, IL-6. The substance(s) may contribute to those factors that leads to osteopenia in RA patients and perhaps postmenopausal women.
The role of T cells in the pathogenesis of RA is presently unclear since a specific antigen has not been identified.(7–8,40) It is well established that T cells require at least two signals for activation, one delivered by receptor occupancy and the other by the interaction of costimulatory molecules with their respective ligands expressed on the T cell surface. One such T cell ligand is the T cell receptor (TCR/CD3). Furthermore, T cell activation is regulated by a homodimeric glycoprotein, CD28, expressed on the surface of T cells which is the receptor for B7–B1 or B7–B2, major costimulatory molecules expressed on the surface of antigen-presenting cells such as the monocyte or B cell (reviewed elsewhere(41)). The nature of the biochemical reactions subsequent to T cell activation via CD3 and costimulators is vast and is reviewed elsewhere.(41–44) Recent studies have deduced that binding and cross-linking of CD28 with anti-CD28 antibodies simulates the physiological signals present in allograft patients and normally delivered by the interaction of CD28 with B7. The role of CD28/B7 activation of T cells in RA is presently controversial. On the one hand, this activation scheme has been suggested to be associated with activation of both peripheral and synovial T cells in RA,(17,45–47) while on the other hand others have reported that in the periphery, but not the synovial cavity, T cells lack CD28,(48) suggesting a role for these unusual lymphocytes in the disease process. Nonetheless, we have chosen to use anti-CD3 and anti-CD28 antibodies to mimic the natural activation processes.
We have demonstrated that unactivated T cells do not induce IL-6 in hOB either via cell–cell contact or via soluble factors since these cells are basically at rest with respect to cytokine production and cell adhesion molecule expression. We have also demonstrated that activation with antibodies to the TCR–CD3 complex and CD28 ligand results not only in the release of a possible novel cytokine but that the expressed cytokine is also resistant to the down-regulating effects of CsA. Again the difference found between the effect of activated T cell CM or CsA-treated activated T cell CM induction of IL-6 protein, as assessed by ELISA, in osteoblasts most likely represents the presence of other inducing cytokines such as TNF-α and other, presently unknown cytokines, in the activated T cell CM which were inhibited in the CsA-treated T cells. Thus, we suggest that the factor(s) is most likely the primary effector of osteoclastogenesis in RA.
Our data are in contrast to that of Tanaka et al.(49) who have reported that as a result of T cell adhesion, mRNA transcription and secretion of IL-1β and IL-6 were induced in hOBs derived from osteoarthritis patients. Inhibition of adhesion with anti-CAM antibodies resulted in reduction of expression of IL-6. We have demonstrated that activated T cells do not induce IL-1 in osteoblasts and the induction of IL-6 in hOB does not require cell–cell contact, although cell–cell contact between T cells and osteoblasts did result in a small, additive increase in IL-6 production by hOB. We cannot exclude the possibility that the different methods of osteoblast preparation and culture and T cell activation used by Tanaka and coworkers (phorbol myristate acetate) may reflect differences between their results and ours. It has been established that T cell function may vary depending on the method of activation used.(50) Furthermore, we cannot exclude the possibility that osteoblasts derived from diseased tissue may be metabolically different from those of nondiseased bone. Those differences will need to be explored further.
Thus, our preliminary experiments suggest that we have found a possible new, CsA- resistant, T cell cytokine(s) which is a potent inducer of IL-6 in hOB. This apparently new cytokine(s) may be responsible for the bone loss observed in humans suffering from RA. Identification of the soluble factor may lead to the development of inhibitory agents which should prove to be of therapeutic efficacy in preventing osteoporosis in patients with RA.
The authors wish to thank Mrs. Aurora Fausto for preparation of the hOB cultures.