Proinflammatory Molecule Follistatin-Like Protein-1 Is a Novel

Follistatin-Like Protein-1 Is a Novel Proinflammatory Molecule Takako Miyamae, Anthony D. Marinov, Dawn Sowders, David C. Wilson, Jason Devlin, Robert Boudreau, Paul This information is current as Robbins and Raphael Hirsch of September 28, 2021. J Immunol 2006; 177:4758-4762; ; doi: 10.4049/jimmunol.177.7.4758 http://www.jimmunol.org/content/177/7/4758 Downloaded from

References This article cites 18 articles, 5 of which you can access for free at: http://www.jimmunol.org/content/177/7/4758.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 28, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Follistatin-Like Protein-1 Is a Novel Proinﬂammatory Molecule1

Takako Miyamae,* Anthony D. Marinov,* Dawn Sowders,† David C. Wilson,* Jason Devlin,‡ Robert Boudreau,§ Paul Robbins,¶ and Raphael Hirsch2*

While analyzing gene expression in collagen-induced arthritis, we discovered that a poorly characterized gene, follistatin-like protein 1 (FSTL-1), is highly overexpressed in mouse paws during early arthritis, especially at the interface of synovial pannus and eroding bone. In this study, we show that FSTL-1 is a novel proinflammatory molecule with a previously unrecognized role in inflammation. Transfection of FSTL-1 into macrophages and fibroblasts leads to up-regulation of proinflammatory cytokines, including IL-1␤, TNF-␣, and IL-6. Overexpression of FSTL-1 in mouse paws by gene transfer results in severe paw swelling and arthritis. The Journal of Immunology, 2006, 177: 4758–4762. Downloaded from ollistatin-like protein 1 (FSTL-1)3 is a poorly character- (RA) synovial tissue, demonstrated anti-FSTL-1 Abs in the serum ized protein that has never previously been associated with and synovial fluid of RA patients, and suggested that FSTL-1 was F inflammation. FSTL-1, also known as FRP and TSC-36, is an autoantigen. an extracellular glycoprotein belonging to the BM-40/SPARC/os- teonectin family of proteins containing both extracellular calcium- Materials and Methods

binding and follistatin-like domains. FSTL-1 was originally cloned Adenoviral vectors http://www.jimmunol.org/ ␤ from an osteoblastic cell line as a TGF- -inducible gene (1). The A recombinant, E1a-E3-deleted replication defective adenovirus type 5 protein occurs in two isoforms resulting from differential sialyla- vector encoding the mouse FSTL-1 gene (National Center for Biotechnol- tion. FSTL-1 was detected in the medium of all osteosarcoma and ogy Information Nucleotide Database accession number BC028921) was chondrosarcoma cell lines tested and in medium of some cells of generated through Cre-lox recombination as described by Hardy et al. (10). The control vector, Ad-BglII, is an E1a/E3-deleted replication-defective the fibroblast lineage. In mice, the highest expression of FSTL-1 adenovirus type 5 lacking an insert. The vectors were grown in 293 cells was observed in the lung (2). and purified by CsCl gradient ultracentrifugation, dialyzed at 4°C against The action of FSTL-1 is unclear, although a number of putative sterile virus buffer, aliquoted, and stored at Ϫ80°C. FSTL-1 expression was functions have been suggested. Both proliferative and antiprolif- verified by both RT-PCR and Western blot from infected COS-7 cells. erative effects have been reported. FRP may play a role in neural- Cell transfection by guest on September 28, 2021 ization during embryogenesis (3), and its expression is up-regulated by estrogen (4). In contrast to other BM-40 family members, U937 and COS-7 cells were electroporated with pRcCMV-hFSTL-1 using a Bio-Rad Gene Pulser (Bio-Rad) according to the manufacturer’s proto- the extracellular calcium-binding domain of FSTL-1 is nonfunc- col. Cells were maintained in G418 selection medium for a period of 3–4 tional (5), suggesting that, despite its sequence homology to BM- wk, and clones were isolated. 40, it has evolved clearly distinct properties. Analysis of prostate Histologic analysis cancers revealed that overexpression of FSTL-1 was associated with higher metastatic potential (6). In contrast, FSTL-1 expres- Mouse knee joints were fixed in 10% neutral buffered formalin, decalcified, sion was extinguished in v-ras-transformed rat fibroblasts, and dehydrated in a gradient of alcohols, paraffin embedded, sectioned, mounted on glass slides, and stained with H&E. For immunohistochemis- transfection of FSTL-1 into these cells inhibited in vitro invasion try, knees were fixed in 2% paraformaldehyde, cut longitudinally with a (7) and led to growth inhibition in human lung cancer cells (8). In scalpel blade, mounted onto a cork disk with a small drop of Cryogel 1998, Tanaka et al. (9) cloned FSTL-1 from rheumatoid arthritis (Cancer Diagnostics), and placed in liquid nitrogen-cooled isopentane for 45 s. Sections of 7 ␮m were obtained using a cryostat. Slides containing knee sections were fixed with 2% paraformaldehyde for 20 min and washed with PBS followed by BSA buffer (0.5% BSA and 0.15% glycine *Division of Rheumatology, Children’s Hospital of Pittsburgh, University of Pitts- † in PBS). Slides were blocked for 45 min with a 1/20 dilution of normal burgh, School of Medicine, Pittsburgh, PA 15213; Cincinnati Children’s Hospital donkey serum (Sigma-Aldrich) in BSA buffer and washed three times with Medical Center, Cincinnati, OH 45229; ‡Department of Cell Biology and Physiology, § BSA buffer. The primary Ab goat anti-mouse FSTL-1 (R&D Systems) University of Pittsburgh School of Medicine, Pittsburgh, PA 15261; Department of ␮ Epidemiology, Graduate School of Public Health, Pittsburgh, PA 15261; and ¶De- diluted to 10 g/ml in BSA buffer was added, incubated for 2 h, and partment of Molecular Genetics and Biochemistry, University of Pittsburgh School of washed three times with BSA buffer. The secondary Ab Alexa Fluor 488 Medicine, Pittsburgh, PA 15261 donkey anti-goat IgG (Molecular Probes) diluted 1/500 in BSA buffer was Received for publication April 25, 2006. Accepted for publication July 11, 2006. added, incubated for 1 h, and washed three times with BSA buffer and three times with PBS. The nucleus stain DRAQ5 (Biostatus) diluted 1/1000 in The costs of publication of this article were defrayed in part by the payment of page PBS was added, incubated for 30 min, and washed three times with PBS. charges. This article must therefore be hereby marked advertisement in accordance A coverslip containing a drop of gelvatol was added, and slides were stored with 18 U.S.C. Section 1734 solely to indicate this fact. at 4°C until observation with confocal microscope. For the double staining 1 This work was supported in part by National Institutes of Health Grants AI44566 (to of fibroblast-like synoviocytes and FSTL-1, the primary Ab rat anti-mouse R.H) and AI 56374 (to P.D.R). Thy-1.2 (BD Pharmingen) was included at the same time as the goat anti- 2 Address correspondence and reprint requests to Dr. Raphael Hirsch, Division of mouse FSTL-1 at a concentration of 1.5 ␮g/ml. The secondary Ab Alexa Rheumatology, Children’s Hospital of Pittsburgh, 3705 Fifth Avenue, Pittsburgh, PA Fluor 594 donkey anti-rat IgG was used at 1/1000 dilution. Appropriate 15213. E-mail address: [email protected] isotype control primary Abs were used each time immunohistochemistry 3 Abbreviations used in this paper: FSTL-1, follistatin-like protein 1; RA, rheumatoid was performed. Immunohistochemistry slides were observed using a arthritis; CII, type II collagen; CIA, collagen-induced arthritis. Olympus Fluroview 500 (Olympus) confocal microscope.

FIGURE 1. FSTL-1 is overexpressed in fibroblast-like synoviocytes in early CIA. Mouse knee joints were harvested from mice before immunization with CII (day 0), during acute arthritis (day 28), or during late arthritis (day 49). Tissues were sectioned and stained for FSTL-1 and observed under a confocal microscope at ϫ40 (A–C). Higher magnification (ϫ100) views of day 28 synovium (D–F) show CD90ϩ fibroblasts in red (D), FSTL-1 in green (F), and double staining of CD90 and FSTL-1 (E). Colocalization of FSTL-1 with CD90ϩ fibroblasts is indicated by the arrows in E.

Quantitative RT-PCR of entire paw and digits). The arthritic index for each mouse was calculated Downloaded from by adding the score of the four individual paws. The statistical significance Total RNA was isolated from frozen liver using the ToTALLY RNA Iso- of difference was determined using the exact Wilcoxon test. p Ͻ 0.05 was lation Kit (Ambion) according to the manufacturer’s instructions. To re- considered significant. Mice were sacrificed at various times, and joints move possible genomic DNA contamination, RNA was treated with DNase were used for histopathology, whereas livers were snap frozen and stored I (Ambion). cDNA was synthesized with random hexamer oligonucleotides in a liquid nitrogen freezer. using the SuperScript II Reverse Transcriptase Kit (Invitrogen Life Tech- nologies). PCR was performed in a LightCycler (Mx3000P; Stratagene) http://www.jimmunol.org/ using the Brilliant SYBR Green QPCR Master Mix (Stratagene) according Results to the protocol (95°C hot start for 10 min followed by 40 amplification FSTL-1 is overexpressed in acute CIA by fibroblast-like cycles, denaturation at 95°C, primer annealing at 59°C, and amplicon ex- synoviocytes tension at 72°C) using oligonucleotide primer sets for IL-1␤, IL-6, and TNF-␣ (11). The copy number (number of transcripts) of amplified prod- While analyzing gene expression profiles in CIA using DNA mi- ucts was calculated from a standard curve obtained by plotting known input croarrays, we discovered that FSTL-1 mRNA is highly up-regu- concentrations of plasmid DNA and normalizing to 18S rRNA. lated in mouse paws during early arthritis, most prominently at the Cytokine analysis interface of synovial pannus and eroding bone (13), suggesting an active role in joint destruction. We also observed FSTL-1 expres- Titers of IL-1␤, TNF-␣, and IL-6 were determined using commercial by guest on September 28, 2021 ELISAs according to the manufacturer’s instructions (R&D Systems). sion in RA synovium. To further explore the role of FSTL-1 in arthritis, we examined its cellular distribution in the joints of mice Induction and assessment of collagen-induced arthritis (CIA) by immunohistochemistry at various times during the course of Male DBA/1 mice, 6–10 wk of age, were purchased from Harlan Sprague arthritis (Fig. 1). An increase in expression was observed in early Dawley and The Jackson Laboratory and housed in the animal resource CIA (day 28), whereas only minimal expression was observed in facility at the Children’s Hospital of Pittsburgh Rangos Research Center late CIA (day 49). The expression of FSTL-1 localized to fibro- (Pittsburgh, PA). The study was approved by the Children’s Hospital of blasts, as evidenced by CD90 staining. No expression was ob- Pittsburgh’s Animal Research and Care Committee. Mice were treated by i.v. injection with 1 ϫ 1010 particles of adenoviral vectors in 200 ␮lof served in macrophages, neutrophils, or T cells (data not shown). PBS. CIA was induced by intradermal immunization with bovine type II collagen (CII; Elastin Products), and a booster was given 21 days later as FSTL-1 induces secretion of proinflammatory cytokines from previously described (12). Mice were evaluated for arthritis several times fibroblasts and macrophages weekly using a macroscopic scoring system ranging from 0 to 4 (0 ϭ no detectable arthritis; 1 ϭ swelling and/or redness of paw or 1 digit; 2 ϭ two Arthritic synovium contains large numbers of fibroblasts and mac- joints involved; 3 ϭ three to four joints involved; and 4 ϭ severe arthritis rophages that are centrally involved in joint destruction. Having

FIGURE 2. FSTL-1 induces secretion of proinflammatory cytokines from fibroblasts and macrophages. A, Monkey COS-7 fibroblasts were stably transfected by electroporation with a plasmid (pRcCMV) encoding the neomycin resistance gene with or without human FSTL-1. After selection for 4 wk, cells were plated for 3 days, and supernatants were assayed for IL-6. B—D, Human U937 monocyte cells were similarly transfected. After selection for 4 wk, cells were stimulated with 10 ng/ml PMA in the presence or absence of 100 ng/ml LPS and assayed for IL-1␤, TNF-␣, and IL-6. Bars, mean and .p Ͻ 0.05 ,ء .SEM of triplicate samples 4760 FSTL-1 IS A NOVEL PROINFLAMMATORY MOLECULE

FIGURE 3. Overexpression of FSTL-1 in liver induces expression of proinflammatory cytokines. Mice were injected i.v. with 1 ϫ 1010 particles of adenovirus encoding either mouse FSTL-1 (Ad-mFSTL-1) or a control virus lacking a transgene (Ad-BglII). Total RNA was isolated from liver 7 days later and amplified by real time PCR. The copy number of amplified product was calculated from a standard curve obtained by plotting known input concentrations of plasmid DNA. The copy numbers for TNF-␣, IL-1␤, and IL-6 were normalized to the housekeeping gene 18s. Bars, mean and SEM of 10 mice. .p Ͻ 0.05 ,ء Downloaded from

demonstrated that fibroblast-like synoviocytes overexpressed FSTL-1 induces spontaneous inflammation and synovitis FSTL-1 during acute arthritis, we wished to examine the down- To determine whether the observed proinflammatory effects of stream effects of such overexpression on this cell phenotype. We FSTL-1 required an inflammatory costimulus, such as immuniza- were unsuccessful at transfecting primary mouse or human syno- tion with CII, Ad-mFSTL-1 was injected intradermally into paws http://www.jimmunol.org/ vial fibroblasts; however, we were able to stably transfect the mon- of unimmunized DBA/1 mice. To our surprise, severe paw swell- key kidney fibroblast cell line, COS-7, by electroporation. After ing and erythema was observed, beginning 8 days after injection transfection with human FSTL-1, a spontaneous increase in IL-6 (Fig. 6, A and B) and persisting for 1 week before subsiding. His- expression was observed (Fig. 2A). We also stably transfected the tologic analysis revealed synovitis, with infiltration of inflamma- human monocyte cell line, U937, with FSTL-1. Although these tory cells into the synovium and surrounding tissue (Fig. 6, C and cells did not spontaneously produce inflammatory cytokines, addition of PMA to induce differentiation into macrophages led to a significant increase in secretion of IL-1␤,TNF-␣, and IL-6 (Fig. 2, B–D). A substantial synergistic effect was observed on addition of by guest on September 28, 2021 LPS, suggesting that FSTL-1 utilizes a distinct signaling pathway different from the LPS receptor.

FSTL-1 induces production of proinﬂammatory cytokines in vivo To determine the in vivo effects of overexpression of FSTL-1, DBA/1 mice were given an adenovirus encoding either mouse FSTL-1 (Ad-mFSTL-1) or a backbone virus lacking a transgene (Ad-BglII) by i.v. injection. Administration of adenovirus i.v. pre- dominantly targets the liver. Seven days after adenovirus administration, livers were removed and analyzed for expression of TNF-␣, IL-1␤, and IL-6 by real time PCR. A pronounced increase was observed in mRNA for all 3 cytokines (Fig. 3), demonstrating that FSTL-1 up-regulates their expression in vivo.

FSTL-1 exacerbates CIA To demonstrate that FSTL-1 could function as a proinflammatory mediator in arthritis, CIA was induced in DBA/1 mice by immunization with CII. Mouse FSTL-1 was overexpressed before onset of arthritis by i.v. administration of Ad-mFSTL-1. Control mice received Ad-BglII. Severity of arthritis, as measured by the arthritic index (12), was significantly greater in mice receiving Ad- mFSTL-1, compared with controls (Fig. 4A). There was also a significant increase in the percent of mice developing severe ar- FIGURE 4. FSTL-1 exacerbates CIA. DBA/1 male mice were immu- thritis in the Ad-mFSTL-1 group (Fig. 4B). Histologic analysis nized with CII on days 0 and 21. Mice were injected i.v. with 1 ϫ 1010 correlated with the arthritic index scores (Fig. 5). Substantially particles of adenovirus encoding either mouse FSTL-1 (Ad-mFSTL-1) or a more synovial inflammation, with infiltration of neutrophils and control virus lacking a transgene (Ad-BglII) (20 mice per group) before the mononuclear cells, and cartilage destruction was observed in mice onset of arthritis (day 17). The arthritic index (A) and the percent of mice .p Ͻ 0.05 ,ء .treated with FSTL-1. with a maximum arthritis index (B) is shown The Journal of Immunology 4761

FIGURE 5. FSTL-1 exacerbates synovitis and joint destruction in CIA. Downloaded from Arthritic mice treated with either Ad-mFSTL-1 (A) or with a control virus, Ad-BglII (B) were sacrificed on day 28, and knee joints were sectioned and stained with H&E. C, An age-matched healthy control. A, Severe inflammation of the synovium with invasion of cartilage and bone (arrow). F, femur; S, synovium (magnification, ϫ200). http://www.jimmunol.org/ D). Protein homogenates from these paws had a significant increase in IL-6 and IL-1␤, compared with control paws (Fig. 6, E and F).

Discussion These findings demonstrate that FSTL-1 is a previously unrecognized proinflammatory mediator capable of inducing spontaneous inflammation. In the context or arthritis, FSTL-1 overexpression in FIGURE 6. FSTL-1 induces spontaneous inflammation and synovitis. by guest on September 28, 2021 the joints might lead to secretion of IL-6 by fibroblast-like syno- Paws of healthy male DBA/1 mice were injected with Ad-mFSTL-1 or viocytes which serves as a proinflammatory signal. In addition, with a control virus, Ad-BglII, at viral titers ranging from 1 ϫ 109 to 1 ϫ FSTL-1 secreted by these cells might stimulate resident macro- 108 particles. Severe inflammation is observed in the Ad-mFSTL-1 group phages to secrete proinflammatory cytokines. Efforts are currently on day 8. A, C, and D, Representative paws (at a titer of 1 ϫ 108 particles). under way to identify an FSTL-1 receptor on these cells. C and D are at a magnification of ϫ100; B, bone; S, synovium. B, Mean Tanaka et al. recently reported that administration of human paw swelling and SEM (n ϭ 8 paws) at each titer. Paw homogenates were FSTL-1 to BALB/c mice with Ab-induced arthritis ameliorated assayed by ELISA for IL-6 (E) and IL-1␤ (F); data represents the mean and Ͻ ء ϭ disease (14), possibly by reducing synovial production of matrix SEM of six paws, except for untreated paws (n 3). , p 0.05. metalloproteinases (15). Although seemingly contradictory to our findings, the effect was mild and may be a consequence of using the xenogeneic human, rather than the endogenous mouse protein References or using a mouse with a predominant Th2 phenotype (BALB/c). 1. Shibanuma, M., J. Mashimo, A. Mita, T. Kuroki, and K. Nose. 1993. Cloning We have not been able to reproduce this finding in the CIA model, from a mouse osteoblastic cell line of a set of transforming-growth-factor-␤1- and polymorphisms in FSTL-1 were not found to be associated regulated genes, one of which seems to encode a follistatin-related polypeptide. Eur. J. Biochem. 217: 13–19. with genetic susceptibility to RA (16). However, we cannot rule 2. Mashimo, J., R. Maniwa, H. Sugino, and K. Nose. 1997. Decrease in the expres- out the possibility that FSTL-1 might have additional effects, in- sion of a novel TGF␤1-inducible and ras-recision gene, TSC-36, in human cancer cells. Cancer Lett. 113: 213–219. cluding immunosuppressive effects in certain circumstances. 3. Okabayashi, K., H. Shoji, Y. Onuma, T. Nakamura, K. Nose, H. Sugino, and Based on their ability to bind various cell growth factors and M. Asashima. 1999. cDNA cloning and distribution of the Xenopus follistatin- extracellular matrices, it has been proposed that follistatin modules related protein. Biochem. Biophys. Res. Commun. 254: 42–48. 4. Ohashi, T., S. Sato, A. Yoshiki, and M. Kusakabe. 1997. TSC-36 (follistatin- may function as regulators of growth factors and/or cytokines (8). related polypeptide) gene expression in estrogen receptor positive osteoblastic Our observations therefore suggest a model for FSTL-1 activity in cell line, CDO7F. Calcif. Tissue Int. 61: 400–403. the context of inflammation. FSTL-1 secreted by fibroblasts acts 5. Hambrock, H. O., B. Kaufmann, S. Muller, F. G. Hanisch, K. Nose, M. Paulsson, P. Maurer, and U. Hartmann. 2004. Structural characterization of TSC-36/Flik: on macrophages to increase production of TNF-␣ and IL-1␤, analysis of two charge isoforms. J. Biol. Chem. 279: 11727–11735. which in turn increase activin production (17). Activin normally 6. Trojan, L., A. Schaaf, A. Steidler, M. Haak, G. Thalmann, T. Knoll, N. Gretz, P. Alken, and M. S. Michel. 2005. Identification of metastasis-associated genes inhibits the activity of IL-6 (18), but FSTL-1, by binding activin, in prostate cancer by genetic profiling of human prostate cancer cell lines. Anti- leads to increased IL-6 activity. Neutralizing FSTL-1 may repre- cancer Res. 25: 183–191. sent a novel approach to treating arthritis and other inflammatory 7. Johnston, I. M., H. J. Spence, J. N. Winnie, L. McGarry, J. K. Vass, L. Meagher, G. Stapleton, and B. W. Ozanne. 2000. Regulation of a multigenic invasion conditions. programme by the transcription factor, AP-1: re-expression of a down-regulated gene, TSC-36, inhibits invasion. Oncogene 19: 5348–5358. 8. Sumitomo, K., A. Kurisaki, N. Yamakawa, K. Tsuchida, E. Shimizu, S. Sone, and Disclosures H. Sugino. 2000. Expression of a TGF-␤1 inducible gene, TSC-36, causes growth The authors have no financial conflict of interest. inhibition in human lung cancer cell lines. Cancer Lett. 155: 37–46. 4762 FSTL-1 IS A NOVEL PROINFLAMMATORY MOLECULE

9. Tanaka, M., S. Ozaki, F. Osakada, K. Mori, M. Okubo, and K. Nakao. 1998. TSC-36/FSTL1 on joint inflammation in a mouse model of arthritis. Arthritis Cloning of follistatin-related protein as a novel autoantigen in systemic rheumatic Rheum. 50: 660–668. diseases. Int. Immunol. 10: 1305–1314. 15. Tanaka, M., S. Ozaki, D. Kawabata, M. Kishimura, F. Osakada, M. Okubo, 10. Hardy, S., M. Kitamura, T. Harris-Stansil, Y. Dai, and M. L. Phipps. 1997. M. Murakami, K. Nakao, and T. Mimori. 2003. Potential preventive effects of Construction of adenovirus vectors through Cre-lox recombination. J. Virol. 71: follistatin-related protein/TSC-36 on joint destruction and antagonistic modula- 1842–1849. tion of its autoantibodies in rheumatoid arthritis. Int. Immunol. 15: 71–77. 11. Overbergh, L., A. Giulietti, D. Valckx, R. Decallonne, R. Bouillon, and 16. Ehara, Y., D. Sakurai, N. Tsuchiya, K. Nakano, Y. Tanaka, A. Yamaguchi, and C. Mathieu. 2003. The use of real-time reverse transcriptase PCR for the quan- K. Tokunaga. 2004. Follistatin-related protein gene (FRP) is expressed in the tification of cytokine gene expression. J. Biomol. Tech. 14: 33–43. synovial tissues of rheumatoid arthritis, but its polymorphisms are not associated 12. Hughes, C., J. A. Wolos, E. H. Giannini, and R. Hirsch. 1994. Induction of T cell with genetic susceptibility. Clin. Exp. Rheumatol. 22: 707–712. anergy in an experimental model of autoimmunity using nonmitogenic anti-CD3 17. Mohan, A., J. Asselin, I. L. Sargent, N. P. Groome, and S. Muttukrishna. 2001. monoclonal antibody. J. Immunol. 153: 3319–3325. Effect of cytokines and growth factors on the secretion of inhibin A, activin A and 13. Thornton, S., D. Sowders, B. Aronow, D. P. Witte, H. I. Brunner, E. H. Giannini, follistatin by term placental villous trophoblasts in culture. Eur. J. Endocrinol. and R. Hirsch. 2002. DNA microarray analysis reveals novel gene expression 145: 505–511. profiles in collagen-induced arthritis. Clin. Immunol. 105: 155–168. 18. Yu, E. W., K. E. Dolter, L. E. Shao, and J. Yu. 1998. Suppression of IL-6 14. Kawabata, D., M. Tanaka, T. Fujii, H. Umehara, Y. Fujita, H. Yoshifuji, biological activities by activin A and implications for inflammatory arthropathies. T. Mimori, and S. Ozaki. 2004. Ameliorative effects of follistatin-related protein/ Clin. Exp. Immunol. 112: 126–132. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021