Oncostatin M Regulation of Inflammatory Responses By

Regulation of Inﬂammatory Responses by Oncostatin M

Philip M. Wallace,1,2* John F. MacMaster,† Katherine A. Rouleau,† T. Joseph Brown,* James K. Loy,† Karen L. Donaldson,3* and Alan F. Wahl3*

Oncostatin M (OM) is a pleiotropic cytokine produced late in the activation cycle of T cells and macrophages. In vitro it shares properties with related proteins of the IL-6 family of cytokines; however, its in vivo properties and physiological function are as yet ill defined. We show that administration of OM inhibited bacterial LPS-induced production of TNF-␣ and lethality in a dose-dependent manner. Consistent with these findings, OM potently suppressed inflammation and tissue destruction in murine models of rheumatoid arthritis and multiple sclerosis. T cell function and Ab production were not impaired by OM treatment. Taken together these data indicate the activities of this cytokine in vivo are antiinflammatory without concordant immunosuppression. The Journal of Immunology, 1999, 162: 5547–5555.

he normal development of an inflammatory response must from the inflammatory effector phase back to homeostasis also are be rapidly followed by the engagement of a feedback sys- being evaluated for their clinical potential as drugs. The cytokines T tem to minimize adventitious tissue damage and regulate IL-10 and IL-11 both appear to accelerate this process and their the eventual return to homeostasis. This system involves a multi- administration have proven effective in resolving several animal tude of regulators including cytokines, adhesion molecules, pro- models of chronic inflammatory disease (10). teases, corticosteroids, and subsequent regulators of each of these Oncostatin M (OM)4 is a pleiotropic cytokine that is produced agents. A normal response to infection or other insult is a self- by activated T cells and macrophages and has shown in vitro prop- limiting process that by way of temporal expression of both reg- erties that would be expected to influence the course of inflamma- ulators, and effector molecules, causes the resolution of the initi- tory responses (11, 12). The protein is structurally and functionally ating event. The failure to resolve the causative insult or to redress related to IL-6, leukemia inhibitory factor (LIF), and IL-11, pro- the balance of pro- and antiinflammatory agents results in tissue teins that also influence immune and inflammatory function (13). injury and destruction that characterize the pathology of various Despite each protein signaling via a family of related receptors and chronic inflammatory diseases (1, 2). sharing various common properties, each is endowed with a unique The exact participation of each cytokine in the inflammatory disease process is poorly understood in part due to their complex array of biological functions (13). Numerous activities have been interplay. However, the ability of a variety of cytokine and cyto- ascribed to OM in vitro, including the differentiation of kine agonists to alter the severity or course of various inflamma- megakaryocytes, inhibition of tumor cell growth, induction of neu- tory diseases is an impressive testament to the clinical value of rotrophic peptides, regulation of cholesterol metabolism, and ef- cytokines as a target for therapeutic intervention (2, 3). Such data fects on bone-derived cells (7, 14, 15). Recently a collective pic- has been accrued using animal models of disease, transgenic ani- ture of OM has emerged that strongly suggests a natural role of the mals and, more recently, clinical trials of cytokine inhibitors (4, 5). cytokine in the wound healing process and attenuation of the in- A variety of approaches are currently being studied to alter cyto- flammatory response. We have previously found that OM can kine function to bring about the regulation of aberrant inflamma- modulate the expression of IL-6, an important regulator of various tory responses (6). Inhibitors of proinflammatory cytokines, most aspects of the host defense system (16). OM has been shown to notably TNF-␣ inhibitors, have been successful in moderating un- regulate the expression by human cells of acute phase proteins and toward inflammatory responses (2). Abs to TNF-␣ and soluble protease inhibitors that have been implicated in modulating cyto- receptors are currently in clinical trials against a variety of diseases kine function and limiting tissue damage at sites of inflammation. including rheumatoid arthritis, multiple sclerosis, and Crohn’s dis- Recently many of these in vitro effects have been found to occur ease (7–9). Their efficacy has helped establish a set of common in rodents and nonhuman primates following OM administration effectors in these apparently disparate diseases. Alternatively, the (17). Here we have extended these in vivo findings to further un- cytokines whose normal physiological role is to usher a response derstand the role of OM in regulating cytokine networks following inflammatory stimuli. We have also examined the effects of OM *Bristol-Myers Squibb Pharmaceutical Research Institute, Seattle, WA 98121; and treatment in two murine models of disease in which common † Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ 08543 proinflammatory cytokines have been previously shown to play Received for publication November 19, 1998. Accepted for publication February key roles. 16, 1999. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Address correspondence and reprint requests to Dr. Philip M. Wallace, Xcyte Ther- apies, 2203 Airport Way South, Suite 300, Seattle, WA 98134. E-mail address: 4 Abbreviations used in this paper: OM, oncostatin M; LIF, leukemia inhibitory fac- [email protected] tor; ␣1-Pi, ␣1-proteinase inhibitor; EAE, experimental autoimmune encephalomyelitis; KLH, keyhole limpet hemocyanin; MMP, matrix metalloproteinase; PLP, prote- 2 Current address: Xcyte Therapies, 2203 Airport Way South, Suite 300, Seattle, WA olipid protein; SAA, serum amyloid A; TIMP-1, tissue inhibitor of 98134. metalloproteinase-1; BMS-PRI, Bristol-Myers Squibb Pharmaceutical Research 3 Current address: Seattle Genetics, 22215, 26th Avenue SE, Bothell, WA 98021. Institute.

Materials and Methods decalcified in HCl, processed by routine methods, and embedded into par- ␮ Animals affin. The specimens were sectioned at 4–6 m, stained with hematoxylin and eosin, and examined by light microscopy. Sections were graded with- Studies used female mice (ϳ8 wk old) that were held in quarantine for out prior knowledge of the treatment group. Tibiotarsal (hock) joints were 2 wk before admission to any study, during which time serological graded as to the severity of inflammation, pannus formation, cartilage dam- examination was performed. BALB/c and C57BL/6 mice were obtained age, and osseuos changes. Each parameter was examined separately and from Taconic (Germantown, NY), and B10.S-H2(S)SgMcdJ mice were graded as follows: grade 0, unremarkable; grade 1, minimal change; grade obtained from The Jackson Laboratory (Bar Harbor, ME). Animals 2, mild; grade 3, moderate; and grade 4, severe. The inflammation score were housed according to the American Association for the Accredita- was derived from evaluation of soft tissue inflammation, synovitis, and tion of Laboratory Animal Care and institutional guidelines. Experi- angiogenesis. Pannus formation was defined as hypertrophic synovial tis- ments shown are representative of at least three independent studies. sue composed of intraarticular inflammatory exudate accompanied by sy- Statistical analyses were performed using a Wilcoxon rank test (Primer novial cell hyperplasia. Cartilage destruction and loss of matrix were eval- for Biostatistics, McGraw-Hill, NY). uated on the articular surfaces of the distal tibia, the talus, the calcaneus, and the tarsal bones to yield the cartilage damage score. The depth of Reagents erosion of the subchondral bone and the amount of periosteal exocytosis in the distal tibia, the talus, the calcaneus, and the tarsal bones were evaluated Recombinant human OM was expressed in Chinese hamster ovary cells to yield the osseous changes score. The above parameters were then eval- and purified as described (18). OM was administered via various routes of uated as to the percent of tissue involved in the disease process: 1, 0–25%; injection in PBS. Escherichia coli LPS (#L3012) and IFA were purchased ␣ 2, 26–50%; 3, 51–75%; 4, 76–100%. The severity and extent of involve- from Sigma (St. Louis, MO). TNF- and IL-6 ELISA was obtained from ment were then combined to yield the global arthritis score for each joint Endogen (Woburn, MA), and no cross reactivity was found with OM (data (maximum possible score, 32). not shown). Anti-collagen II monoclonal hybridomas were purchased from Chondrex (Redmond, WA). Abs were produced from hybridoma superna- Peptide synthesis tant and purified by protein A Sepharose chromatography. Mycobacterium tuberculosis was purchased from Difco (Detroit, MI), and pertussis toxin The peptide 139–151 from proteolipid peptide (PLP) was assembled on a was obtained from List Biological Laboratories (Campbell, CA). SRBC Gilson multiple peptide synthesizer (Middleton, WI) using F-moc amino were purchased from PML Microbiologicals (Tualatin, OR), and keyhole acids. The peptide resin was treated with trifluoroacetic acid-water-thio- limpet hemocyanin (KLH) was obtained from Pacific BioMarine Lab anisole-ethanedithiol (100/5/5/2.5) for 2 h, and the cleaved, deprotected (Venice, CA). Ab G19-4 (anti-CD3) was provided by Jeff Ledbetter (Bris- peptide was purified by reversed-phase chromatography on a Dynamax C-8 tol-Myers Squibb Pharmaceutical Research Institute (BMS-PRI), Seattle, column. The final product was shown to have the expected m.w. by mass WA), mAb 2E12 (anti-CD28) was provided by Bob Mittler (BMS-PRI), spectrometry on a Bio-Ion 20 instrument (Bio-Ion, Uppsala, Sweden). and mAb MR1 (anti-murine CD154) was provided by Dr. Tony Siadak Induction of experimental autoimmune encephalomyelitis (EAE) (BMS-PRI). EAE was induced using a protocol similar to that previously described PBMC assays (21). B10.S-H2(S)SgMcdJ mice were immunized by s.c. injection at two ␮ Human PBMC were prepared from blood obtained from healthy donors by sites in the abdominal flanks on day 0 with PLP 139–151 peptide (125 g) ␮ ␮ separation on Ficoll. T cells were isolated from this fraction by rosetting and 300 g M. tuberculosis H37RA in 200 l of a 1:1 mixture PBS and with SRBC, and the monocytes were separated from the remaining PBMC IFA. Mice were then injected i.p. with 400 ng pertussis toxin diluted in by elutriation. T cell populations were Ͼ95% CD3ϩ and monocyte pop- PBS immediately following the peptide injection. The animals received a ulations were Ͼ95% CD14ϩ as determined by immunostaining. Mono- second injection of 400 ng pertussis toxin i.p. at 48 h postimmunization. ␮ ␮ cytes were activated by treatment with 5 ng/ml bacterial LPS and T cells Animals were treated with OM (100 l, 100 g/ml, PBS, i.p.) or control were activated by costimulation with immobilized anti-CD3 Ab/soluble diluent (PBS/BSA) on days 4–7 and 12–18. All mice were examined daily anti-CD28 Ab (10 ␮g/ml). Cells were cultured using RPMI 1640 basal for neurological signs of disease. Disease was evaluated as previously de- medium supplemented with 10% FBS and penicillin/streptomycin. Culture scribed (22) using the following scale: 0, no abnormality; 1, floppy tail with supernatants were collected at various times following activation for mea- mild hind limb weakness; 2, floppy tail with moderate hind limb weakness; surement of cytokine content by cytokine-specific ELISA assay. 3, hind limb paresis with or without mild forelimb weakness; 4, hind leg paralysis with or without moderate forelimb weakness; 5, quadriplegia; 6, Cytokine and survival studies dead or moribund requiring sacrifice. Mice (C57BL/6) were coinjected i.v. with various doses of OM alone or Lymph node cell stimulation assay with 1 ␮g LPS in PBS. OM was injected i.v. at various time points (100 ␮l, 100 ␮g/ml) before 1 ␮g LPS injection, when treatment was delayed. Blood Animals were immunized with PLP peptide and treated with OM or control samples were collected via retroorbital sinus into heparinized tubes 1 h diluent as described above. At day 18, animals (5/group) treated with OM after LPS administration. Plasma was removed from the blood following or control diluent were sacrificed, and the inguinal and axillary lymph centrifugation and stored frozen at Ϫ20°C before assay by ELISA. Control nodes were removed. The nodes from each animal were pooled, a single studies showed that comparable cytokine levels were measured in freshly cell suspension was prepared, and the red cells were lysed. Cells were plated at 500,000 cells per well in media (200 ␮l, RPMI 1640, 10% FBS, isolated plasma. In survival studies, BALB/c mice were injected with OM ␮ (10 ␮g, 100 ␮l, 100 ␮g/ml) i.p. at 4, 2, and 1 h before coinjection of OM 10 mM HEPES, 50 M 2-ME) to which was added PLP peptide at various and LPS. The final injection contained OM (10 ␮g) and various doses of concentrations. Following culture at 37°C for 3 days, proliferation was 3 ␮ LPS in PBS (200 ␮l). Animals were monitored daily for survival and signs measured by addition of [ H]thymidine (1 C/well, 6.7 Ci/mmol; DuPont of shock. Moribund animals were sacrificed. NEN, Boston, MA) incorporation for 24 h. Media and cells were trans- ferred and washed onto glass fiber filters using a Tomtec cell harvester, air Induction of arthritis by anti-collagen Abs and LPS dried, aqueous scintillation fluid was added, and the filters were counted in a Betaplate scintillation counter (LKB Wallac, Gaithersburg, MD). Arthritis was induced using the method of Terato et al. (19). BALB/c mice were injected with 400 ␮l of a mixture of four Abs (D1, D8, A2, F10) in Immune response to a T cell-dependent Ag PBS (2.5 mg/ml/mAb, i.v.). At 72 h after mAb injection, mice were in- BALB/c mice were injected with 1 ϫ 108 SRBC i.v. (100 ␮l, PBS) or KLH jected with LPS (100 ␮l, 250 ␮g/ml, i.p.). Treatment with OM (100 ␮l, 100 (250 ␮g, 100 ␮l, PBS) i.p. on day 0 then treated with OM (30 ␮g, i.v., days ␮g/ml in PBS, i.v.) or control diluent (100 ␮l, PBS, i.v.) began 24 h fol- 0–10), PBS (100 ␮l, i.v., days 0–10), or anti-CD154 mAb MR1 (23) (200 lowing LPS and continued until day 10. Histopathology and scoring of ␮g i.v., days 0, 2, and 4). Mice were bled at 7-day intervals and assayed for arthritic disease were adapted from the methods of Wooley et al. (20). titers to SRBC or KLH by ELISA as previously described (24). Briefly, the extent of disease was scored in a blinded fashion by both visual observation and by measurement of limb swelling with calipers (in 1/1000 of an inch) using the following scale: 0, normal; 1, disease confined to Results single joints (Ͻ80); 2, minimal swelling, minimal redness (90–100); 3, Regulation of OM expression following leukocyte activation significant swelling, severe reddening, slight foot malformation (100–115); 4, maximal swelling, maximal redness, deformed feet (Ͼ115). Measure- OM is produced by both activated T cells and macrophages (11, ments of normal animals were in the range 65–75 1/1000th inch. At nec- 16), therefore we examined the temporal expression of OM in the ropsy, the distal one-third of the limbs were immersion-fixed in formalin, context of other cytokines expressed by these cells in response to The Journal of Immunology 5549

a peak level of 8000 pg/ml at 24 h postactivation, which then declined to near baseline levels at 72 h. OM was produced significantly later than IL-2 following T cell activation. OM was first detectable between 24 and 48 h following cell activation, rising over the next 48 h reaching peak level of 3000 pg/ml. OM appears coincident with the decline of IL-2 levels and was further delayed from the induction of IFN-␥ (Fig. 1B). Therefore, expression of OM by the two predominant cell types present at sites of inflammation occurred significantly later than expression of cytokines most closely associated with the initiation of inflammation. Regulation of inflammatory cytokines in vivo by OM Levels of TNF-␣ correlate with the severity of a variety of chronic inflammatory diseases including rheumatoid arthritis and multiple sclerosis, and its action is considered to be central in the patho- genesis of many of these diseases (5, 28). We initially asked if TNF-␣ production following challenge of mice with LPS was af- fected by OM. Following i.v. injection of LPS (1 ␮g), control animals had the expected rapid increase of TNF-␣ with maximal levels of 50 ng/ml being measured 1 h postinjection. OM inhibited the induction of TNF-␣ expression in a dose-dependent manner when administered concurrently with LPS (Fig. 2A). Maximal inhibition of TNF-␣ (ϳ75%) occurred at a dose of 1 ␮g OM. OM treatment did not need to be concurrent with LPS, as administration up to 24 h before LPS was still effective in inhibiting TNF-␣ induction (Fig. 2B). Because IL-6 is regulated in vivo by both OM and TNF-␣ (17, 25), levels of IL-6 in mice receiving a combination of LPS and OM were also measured. The combination of LPS and OM produced levels of IL-6 that were significantly greater than the levels produced by LPS alone (threefold) or OM alone (100-fold) (Fig. 2A). This result was not expected, as LPS-induced IL-6 expression is a result of TNF-␣ production (29) and OM blocked the production of TNF-␣. Increased levels of IL-6 dimin- ished rapidly with increased time between OM and LPS injections. Enhanced levels of IL-6 were only seen when OM was administered with a delay of less than 2 h (Fig. 2B). If OM preceded LPS by 24 h, the IL-6 levels were then consistent with the reduced FIGURE 1. Expression of OM after leukocyte activation. A, Time TNF-␣ levels. IL-1 levels were also measured in these experiments course of OM production from human monocytes following activation. and were not significantly altered by OM administration (data not Peripheral blood monocytes were activated by treatment with E. coli LPS. shown). We next asked whether the increases in IL-6 expression Media was sampled at the indicated times following activation, and the levels of cytokines were quantified by ELISA. Shown are the levels of seen with OM in combination with LPS were the result of a syn- ␣ TNF-␣ and OM Ϯ SD at the indicated times after activation. B, A time ergistic or additive effect between OM and TNF- produced by course of OM production from human T lymphocytes following activation. LPS. Combinations of OM (10 ␮g) with TNF-␣ (1 ␮gor10␮g) Peripheral blood T lymphocytes were activated by treatment with anti-CD3 produced significantly higher levels of IL-6 (more than sevenfold) and anti-CD28 Abs. Following activation, media were sampled at the in- than each cytokine alone and at greater levels than would be ex- dicated times, and the levels of secreted lymphokines were quantified by pected from just an additive effect (Fig. 2C). In vitro studies using ELISA. Shown are the concentrations of IFN-␥, IL-2, and OM (ϮSD) in human or murine macrophages isolated from either peripheral the media at the indicated times after activation. blood or peritoneal exudates were unable to demonstrate a direct effect of OM on the production of TNF-␣ or IL-6. Also, macrophages isolated from mice treated with OM produced similar lev- proinflammatory stimuli. Treatment of peripheral blood mono- els of soluble TNF-␣ following LPS stimulation in vitro to normal cytes with E. coli-derived LPS resulted in a rapid induction of controls (data not shown). These findings suggest that the effects of TNF-␣ (Fig. 1A) (25, 26). Secreted TNF-␣ levels in the media OM on TNF-␣ production in vivo are indirect. The levels of IL-10, peaked at 2 h postactivation and then declined over the next 46 h. another antiinflammatory cytokine that inhibits TNF-␣ production Analysis of the same supernatants demonstrated that OM was pro- (10), and the proinflammatory cytokine IL-1␤ were not signifi- duced significantly later than TNF-␣. Increased OM levels were cantly changed in mice treated with OM and LPS compared with first detectable at 24 h post-LPS (Ͼ10 pg/ml), then continued to mice receiving LPS alone (data not shown). In conclusion, OM is rise over the next 24 h, reaching a maximum of 1000 pg/ml at 48 h. able to inhibit TNF-␣ production while augmenting the normal These results demonstrate that OM secretion from monocytes is feedback loop of IL-6 expression. significantly delayed following cell activation with a pro-inflam- The alteration in cytokine production was accompanied by a matory stimulus. In T cells, induction of IL-2 and other activation reduced lethality of OM-treated mice following exposure to LPS. markers rapidly follows receptor-mediated signaling (27). As Animals were injected with increasing doses of LPS (25–200 ␮g) shown in Fig. 1B, activation of human peripheral blood T cells by either alone or with OM. OM (10 ␮g) was administered i.p. 4, 2, the cross-linking of Abs to CD28 and CD3 rapidly induced IL-2 to and 1 h prior and at the time of LPS injection. As shown in Table 5550 OM ATTENUATES INFLAMMATION

FIGURE 2. LPS-induced TNF-␣ regulation by OM. A, Dose-response of TNF-␣ inhibition by OM. Groups of three C57BL/6 mice were injected with various doses of OM coincident with administration of LPS. Blood was sampled 1 h after LPS administration and the plasma removed and assayed for TNF-␣ and IL-6. Shown are the mean concentrations of cytokine (ϮSD). B, Effect of timing of OM administration on TNF-␣ inhibition. Animals were treated with a ﬁxed dose of OM (1 ␮g, i.v.) at the designated time points before LPS administration and cytokines measured as in A. Shown are the mean concentrations of cytokine (ϮSD) 1 h after LPS injection. C, Combined effects of OM and TNF-␣ on IL-6 production. Mice were injected with OM only, TNF-␣ only, or coinjected with OM and TNF-␣ combined at the indicated doses. IL-6 levels (ϮSD) in the blood 1 h after administration are shown.

I, OM treatment increased the maximum tolerated dose of LPS in 3.8 Ϯ 1.35. In contrast, only one OM-treated animal had macro- mice twofold when compared with animals treated with control scopic evidence of disease with a score Ͼ1 and the median ar- diluent. thritic score was 0.4 Ϯ 1. At day 11 postinduction of disease, the animals were sacrificed, and the rear limbs were subjected to mi- Suppression of inflammation and joint destruction in a mouse croscopic examination of disease. Shown in Fig. 3B are represen- model of rheumatoid arthritis tative histological sections of joints from OM-treated mice and To further investigate the antiinflammatory properties of OM fol- those treated with control diluent. Histological examination lowing inflammatory stimuli, its effects were studied in an Ab- showed that treatment with OM completely inhibited the influx of induced model of rheumatoid arthritis (19). In this model, inflam- inflammatory cells seen in control animals and prevented the tissue mation occurs in the absence of a primary immune response, damage associated with a severe inflammatory reaction. The in- allowing one to distinguish between two immunoregulatory path- flammation and tissue injury was quantitated. Nine of 10 control ways, immune response and inflammation, which are often inter- animals had severe inflammation and tissue injury including pan- dependent and therefore difficult to separate experimentally. To nus formation, connective tissue destruction and erosion of cart- induce joint inflammation, animals received a mixture of four lidge and bone, with an average score of 26.9 Ϯ 16.2. The OM- mAbs to collagen type II (1 mg each) followed 72 h later by LPS treated mice had histological measures of inflammation and injury (25 ␮g) (19). This protocol induces a severe arthritis ϳ24 h fol- consistent with macroscopic evidence, and this group had signif- lowing LPS injection. OM treatment was initiated on day 4, after icantly better score of 2.4 Ϯ 7.6 ( p Ͻ 0.001). In two additional joint inflammation was clearly established, and continued for 7 independent studies, similar efficacy was seen and the cessation of days thereafter (10 ␮g, i.v.). As shown in Fig. 3A, the severity of OM treatment at day 7 was not followed by a delayed onset of joint inflammation was significantly reduced in OM-treated mice inflammation in animals monitored for an additional 14 days (data compared with control animals when assessed for the incidence not shown). and severity of arthritis (20). Nine of 10 control animals were OM inhibits EAE in the absence of immunosuppression afflicted with arthritic injury by day 6 and had a median score of To further evaluate the ability of OM to suppress the inflammatory process, its effects were studied in a murine model of multiple

a sclerosis, EAE. This model shares many inflammatory components Table I. Effect of OM on the survival of mice following LPS treatment that are key to the destruction of the neural sheath and disease progression in multiple sclerosis (30) and to joint destruction in Treatment Control Diluent OM Treatment (␮g LPS) (no. alive/group total) (no. alive/group total) rheumatoid arthritis. In contrast to the arthritis model, T cells are responsible for the inflammatory stimulus in EAE (31). Further, 200 1/15 3/15 the model also provided a means to independently measure the 100 1/30 19/20 effects of OM on the immune and inflammatory components of the 50 12/20 10/10 25 10/10 10/10 disease. Groups of susceptible mice (B10.S-H2 strain) were immunized a BALB/c mice were injected with OM (10 ␮g, i.p., 4, 2, 1, and 0 (coinjected) h) before LPS challenge. Shown is the number of mice surviving to day 7 after LPS with a peptide from a myelin sheath protein, PLP. The peptide treatment. contained amino acids 139–151 of PLP and has been previously The Journal of Immunology 5551

of damage in the histological examination correlated with the symptoms of disease when quantitated (OM 0.10 ϩ 0.10, Control 1.60 ϩ 0.31; p Յ 0.001). In two additional, independent studies using other strains of mice with either myelin basic protein or PLP as the immunogen, OM was comparably effective in blocking the manifestation of disease (data not shown). To further examine the mechanism underlying the inhibition of encephalomyelitis, animals were sacrificed on day 18 and the immune response to PLP was measured. T cell response was assessed by measuring proliferation of the isolated cells from the draining lymph nodes in response to the immunizing peptide in vitro. Treat- ment of the cells with the immunogen induced a concentration- dependent increase in proliferation with no significant differences seen between the control and OM-treated animals (Fig. 5A). Con- sistent with this, there was no significant difference in levels of circulating Abs to the PLP peptide found in control and OM- treated animals (data not shown). In parallel studies, we asked if OM treatment comparable to that which inhibited inflammatory responses modified the immune response to either of two T cell- dependent Ags, KLH and SRBC (Fig. 5B). Injection of SRBC elicited a strong immune response with an IgG1 titer on day 21 (1/3300) consistent with previous findings (24). An Ab to murine CD154 was included in the study at doses previously shown to be suppressive (P. M. Wallace, unpublished observations) as a posi- tive control. No significant difference was seen in animals that received OM (10 ␮g) daily for 10 days following immunization. Similarly, a comparable treatment with OM had no effect on the immune response to injected KLH. These collective results suggest that treatment of animals with OM at levels sufficient to inhibit the inflammatory/degradative aspects of the disease does not suppress normal Ab production or T cell responsiveness.

FIGURE 3. Inhibition of joint inflammation by OM. A, Groups of 10 BALB/c mice were injected i.v. with 1 mg each of four different anti- Discussion collagen mAbs. At 72 h after mAbs injection, an i.v. boost of 25 ␮g LPS was given to accelerate the progression of disease. Joint and limb inflam- The temporal expression of OM by activated T cells and macro- mation was apparent within 24 h. Treatment with OM (10 ␮g/day, i.v.) or phages follows the production of cytokines associated with the control diluent of animals began 24 h following LPS and continued until initiation phase of host defense IL-2 and TNF-␣, respectively. The day 10. Arthritic disease was assessed as described in the Materials and role IL-2 may play in the subsequent induction of OM by T cells Methods. (Representative limbs for each score are shown in the insert.) was not addressed in these studies. However the kinetics of cyto- Shown are the median arthritic scores of control (F) and OM-treated (E) kine expression are consistent with the recent findings that the p Յ .001). B, Representative histology ,ء animals Ϯ SD (ϩ, p Յ .005; and of a hind articular joints at day 11 following initiation of Ab-induced ar- murine OM gene is inducible by IL-2 (33), and the idea that de- thritis from animals treated with OM or control diluent (Pos. Control). layed OM production may represent a regulatory function involved in a feedback mechanism following an initial response. The inhibition of endotoxin-induced TNF-␣ production by OM further sup- demonstrated to be encephalogenic (32). Immunization with the ports a role for OM in an attenuation phase following an inflam- peptide (in IFA and 3 mg/ml M. tuberculosis) was followed by two matory stimulus rather than in the initiation or effector phases. The injections with pertussis toxin as previously described (32). Ani- significance of this attenuation was demonstrated as OM increased mals were treated with OM (10 ␮g, i.p.) or control diluent i.p. on the survival of endotoxin-treated animals. Interestingly, patients days 4–7 and days 12–18. At day 11 postimmunization, control with septic shock have elevated OM levels (34). The efficacy of animals began to exhibit neurological symptoms of the disease, OM treatment is also demonstrated in murine disease models of particularly paralysis. By day 15, 9 of 10 animals in the control rheumatoid arthritis and multiple sclerosis. Studies of both rheu- group had succumbed to the disease, and the median score was 4 matoid arthritis and multiple sclerosis have clearly demonstrated (Fig. 4A). In contrast, no animal that received treatment with OM the importance of TNF-␣ in disease progression, and in models of showed overt signs of the disease in the 18 days following initi- each disease a variety of TNF-␣ antagonists have proven effica- ation of the disease-inducing immunization. Inhibition of the in- cious (35). Therefore, it is reasonable that inhibition of TNF-␣ by flammation associated with this disease was confirmed by histo- OM would protect from other inflammatory stimuli, such as in logical examination of the brain and spinal cord. Control animals autoimmune disease, in addition to LPS. Others have reported that treated with diluent had lesions typical of EAE (Fig. 4B). The commercially available OM, expressed in bacteria, is proinflam- majority of the infiltrate were mononuclear cells (mainly T cells matory and induces the expression of adhesion molecules on en- and smaller numbers of macrophages) and a few granulocytes. dothelial cells in vitro, and an inflammatory infiltrate in vivo (36). Infiltration involved the meninges with extension in a perivascular, Using highly purified mammalian protein or yeast derived protein, white matter orientation. In contrast, inflammatory infiltrate was we have previously studied the effects of OM on endothelial cells not detectable in the histology of OM-treated animals. The extent in vitro (16, 37) and injected protein in vivo in five species (17) 5552 OM ATTENUATES INFLAMMATION

FIGURE 4. Inhibition of EAE by OM. A, Animals (10/group) were immunized with PLP peptide of myelin basic protein in adjuvant. Animals were treated with OM (10 ␮g/day, i.p) or control diluent (PBS) on days 4–7 and 12–18. The extent of disease was assessed in a blinded fashion as described in the Mate- rials and Methods. Shown are the median scores of control (F) and OM-treated (E) animals Ϯ SD. B, EAE-related histopathologic changes. OM, Represen- tative histopathology of meninges between the mes- encephalon (brain stem, bottom portion) and dentate gyrus (cortex, top portion) from EAE animal treated with OM is shown. Control, Histopathology of the similar section from a control animal with EAE, showing prominent inﬂammation and mononuclear inﬁltration is shown.

(P. M. Wallace, unpublished observations), with no evidence of most closely associated with TNF-␣ in mediating tissue damage these findings. These differences may reflect differences in the at sites of inflammation (38). In vivo, inflammatory stimuli result- sources and purity of the proteins used. ing in production of IL-1 initiate a cascade of effectors including The mechanisms by which OM inhibits TNF-␣ remain to be IL-8 and GM-CSF that amplify the inflammatory response by re- elucidated. No direct effects of OM on TNF-␣ production could be cruiting, expanding, and activating inflammatory cells (39). While demonstrated on macrophages either treated ex vivo or isolated OM has no apparent effect on IL-1 production, Richards et al. have from OM-treated animals. IL-6 has been found to inhibit TNF-␣ demonstrated that gene expression of GM-CSF and IL-8 induced production in vivo (34) and is induced by OM (17). However, both by treatment of synovial fibroblasts with IL-1 are suppressed by the inhibition of TNF-␣ by OM in IL-6-deficient mice (P. Wallace, cotreatment with OM in a dose-dependent manner (39). One of the unpublished results) and the kinetics of IL-6 induction in the primary chemoattractants of neutrophils, IL-8 also stimulates the present study make it unlikely that IL-6 is necessary for TNF-␣ production of neutrophil peroxide and the exocytosis of tissue deg- inhibition when LPS and OM are coadministered. However, IL-6 radative granules at sites of inflammation (40). Coincident with may participate in the sustained effects of OM when LPS admin- suppression of GM-CSF and IL-8, OM acted synergistically with istration was delayed. IL-1 to induce expression of IL-6 and tissue inhibitor of metallo- In addition to the inhibition of TNF-␣, collateral antiinflam- proteinase-1 (TIMP-1) in these same cells. Interestingly, the ability matory properties of OM are likely to contribute to the effects in of OM to synergize with IL-1 to suppress inflammatory cytokine these disease models. Cell culture studies have shown that OM expression and induce expression of IL-6 was not paralleled by can block and modify the response to IL-1, the agent provocateur other members of this cytokine family (39). The Journal of Immunology 5553

FIGURE 5. Effects of OM on immune responses. A, Lymph node cell stimulation assay. Lymph nodes were removed from control or OM-treated mice 18 days following immunization with PLP peptide (as in Fig. 4). Lymph node cells were plated at 500,000 cells per well, and PLP peptide was added at the indicated concentrations. Ag-dependent proliferation was measured by [3H]thymidine incorporation during the ﬁnal 24 h of a 4-day culture. Values given are mean cpm (ϮSD) over control (no peptide stimulus) for control and OM-treated animals (5/group). B, Immune response to a T cell-dependent Ag. BALB/c mice were injected with 1 ϫ 108 SRBC i.v. or 250 ␮g KLH i.p. on day 0 then treated with OM (30 ␮g, i.v., days 0–10), PBS (100 ␮l i.v., days 0–10), or anti-CD154 mAb MR1 (200 ␮g i.v., days 0, 2, and 4). Mice were bled at 7-day intervals and assayed for titers to SRBC or KLH by ELISA. Shown are the mean Ab titers (ϮSD) of groups of ﬁve mice.

OM treatment in vivo may also act indirectly on the function of sin, an inhibitor of cathepsin G and other chymotrypsin-like en- IL-1 and TNF-␣ by inducing a constellation of protein antagonists. zymes secreted during inflammation, also inhibits superoxide gen- The acute phase proteins serum amyloid A (SAA) and ␣-1 glyco- eration by activated neutrophils (47). Anti-chymotrypsin is protein are produced locally following tissue injury to minimize produced following OM treatment of hepatic and numerous non- damage proximal to the site of injury. In addition, systemic release hepatic human cells (48). Its expression by epithelial cells in re- of cytokines results in an acute phase response by the liver to sponse to treatment with OM is synergistically increased by co- down-regulate the inflammatory response and reestablish ho- treatment with OM and corticosteroid (48). meostasis. These proteins are normally produced by adult mam- In vivo, the action of collagenase and gelatinase are regulated by mals in response to tissue injury and/or infection (41) and in a the relative level of their cognate inhibitor, TIMP-1 (49). OM in- ␣ normal, self-limiting process are induced by IL-1 and TNF- creased the expression of TIMP-1 by synovial fibroblast and did so ␣ themselves. Administration of -1 glycoprotein protects animals more effectively than other members of the IL-6 cytokine family ␣ ␣ from TNF- -induced lethality (42). SAA and -1 glycoprotein (50). Similarly, OM stimulated the production of TIMP-1 from produced during the acute phase are thought to decrease inflam- human articular chondrocytes and cartilage explant culture more ␣ mation by sequestering circulating IL-1 and decreasing TNF- ex- effectively than IL-11, LIF, or IL-6 (51). Expression of TIMP-1 pression, respectively (22, 43, 44). We have previously demon- from fibroblasts in response to OM occurs with no effect on matrix strated that administration of OM can up-regulate the expression of metalloproteinase (MMP) levels, resulting in a net decrease in SAA and ␣-1 glycoprotein in vivo in both mice and in nonhuman MMP activity (50). The effects of OM on ␣1-Pi, anti- primates (17). Corticosteroids are also potent inhibitors of proin- chymotrypsin, and TIMP-1 expression are greatly enhanced in the flammatory cytokines, including IL-1, IL-8, and TNF-␣.OMin presence of IL-1 (52), again suggesting that OM works in consort combination with IL-1 stimulates the hypothalamus-pituitary-ad- with proinflammatory molecules as part of an antiinflammatory renal axis to secrete corticosterone (45), providing an additional feedback loop. Based on the numerous cell types responsive to mechanism whereby it can feedback to attenuate inflammation. OM, it is reasonable to expect this synergistic feedback could oc- Infiltration of inflammatory cells into the articular synovium or the CNS during acute and chronic inflammation results in tissue cur at sites of inflammation. damage. The secretion of reactive oxygen intermediates, and de- The synergy of OM with proinflammatory mediators is also seen structive proteases including neutrophil elastase, cathepsins, and for the induction of IL-6. Although the role of IL-6 in inflamma- matrix metalloproteinases by these activated cells degrade connec- tion remains controversial (53), in vivo IL-6 induces IL-1 receptor tive tissue and cartilage in rheumatoid arthritis and the neural antagonist and soluble TNF receptor p55 that attenuate inflamma- sheath in multiple sclerosis (46). In addition to attenuating the tion (54). IL-6 is a key inducer of protease- and cytokine-inhibitors cytokines that stimulate secretion of these proteases, in vitro stud- that reduce inflammation and initiate healing. Its protective effects ies have demonstrated that OM is capable of inducing a spectrum are also inferred from the failure of IL-6-deficient mice to repair of protease inhibitors. The acute phase proteins induced by OM tissue and recover from inflammation, to attenuate TNF-␣, or pro- also include two major serine proteinase inhibitors: ␣1-proteinase duce proteins that limit damage at sites of injury (55). In this light, inhibitor (␣1-Pi) and anti-chymotrypsin. ␣1-Pi, the primary inhib- the ability of OM to enhance IL-6 production is consistent with a itor of neutrophil elastase, is secreted from lung epithelial cells and role for the protein in tissue repair. The synergistic interaction of synovial fibroblasts stimulated by OM (C. Richards, unpublished OM and TNF-␣ on increased IL-6 expression supports the concept observations). In comparison, LIF and IL-6 have little or no effect that OM activity is enhanced in the presence of this proinflamma- on expression of ␣1-Pi. Interestingly, the stimulation of ␣1-Pi by tory cytokine at sites of injury and inflammation. This synergy OM is greatly enhanced in the presence of IL-1. Anti-chymotryp- between pro- and antiinflammatory cytokines is not unique to OM 5554 OM ATTENUATES INFLAMMATION and TNF-␣, as OM combined with IL-1 also yields a similar en- necrosis factor (TNF) activity within the central nervous system using monoclo- hancement of IL-6 (45). As described above, the production of nal antibodies and TNF receptor-immunoglobulin fusion proteins. Eur. J. Immu- nol. 24:2040. acute phase proteins is also maximized by the combined presence 6. Burger, D., and J.-M. Dayer. 1995. Inhibitory cytokines and cytokine inhibitors. of OM and inflammatory mediators. Neurology 45:39. We have demonstrated that expression of OM from activated T 7. Lorenz, H.-M., C. Antoni, T. Valerius, R. Repp, M. Gru¨nke, N. Schwerdtner, H. Nu¨␤lein, J. Woody, J. R. Kalden, and B. Manager. 1996. In vivo blockade of cells and macrophages is temporally delayed and increases coin- TNF-␣ by intravenous infusion of a chimeric monoclonal TNF-␣ antibody in cident with a decline in expression of TNF-␣ and IL-2. We, and patients with rheumatoid arthritis. J. Immunol. 156:1646. 8. Davidsen, B., O. H. Nielsen, B. Vainer, and I. Kirman. 1996. The immunological others, have demonstrated synergy between OM and inflammatory network: novel approaches to the treatment of Crohn’s disease. Exp. Opin. Invest. cytokines in suppression of inflammatory mediators, and we have Drugs 5:555. shown herein that, administered systemically, the molecule is ef- 9. Raine, C. S. 1995. Multiple sclerosis: TNF revisited, with promise. Nat. Med. 1:211. ficacious in three different models of acute disease with common 10. de Vries, J. E. 1995. Immunosuppressive and anti-inflammatory properties of proinflammatory mediators. The recent cloning of the murine OM- interleukin 10. Ann. Med. 27:537. specific receptor has called into question the interaction of the 11. Zarling, J. M., M. Shoyab, H. Marquardt, M. B. Hanson, M. N. Lioubin, and G. J. Todaro. 1986. Oncostatin M: a growth regulator produced by differentiated human protein in murine studies (56). However, cytokine produc- histiocytic lymphoma cells. Proc. Natl. Acad. Sci. USA 83:9739. tion and the generation of acute phase proteins also occur from 12. Richards, C. D., and M. Shoyab. 1993. The role of oncostatin M in the acute human cells treated with the human protein. Similarly, we have phase response. In Acute Phase Proteins: Molecular Biology, Biochemistry and Clinical Applications. A. Mackiewicz, I. Kushner, and H. Baumann, eds. CRC also established that OM functions in nonhuman primates to in- Press, Cleveland, Ohio, p. 321. hibit LPS-induced TNF-␣ production and up-regulate both IL-6 13. Kishimoto, T., S. Akira, M. Narazaki, and T. Taga. 1995. Interleukin-6 family of (P. M. Wallace and A. F. Wahl, unpublished observations) and cytokines and gp130. Blood 86:1243. 14. Bruce, A. G., T. M. Rose, P. S. Linsley, and P. M. Wallace. 1993. Oncostatin M. acute phase proteins (17). In the context of an inflammatory cycle, In Human Cytokines: Handbook for Basic and Clinical Research. initiators such as TNF-␣ and IL-1, which promote inflammatory B. B. Aggarwal and J. U. Gutterman, eds. Blackwell Science, Oxford, U.K., p. cell activation and secretion of chemoattractants and proteinases, 361. 15. Horowitz, M. C., and J. A. Lorenzo. 1996. Local regulators of bone: IL-1, TNF, would remain maximal at the peak of inflammatory response. Lo- lymphotoxin, interferon-␥, IL-8, IL-10, IL-4, the LIF/IL6 family, and additional cal expression of OM in the presence of these activators would cytokines. In Principles of Bone Biology. J. P. Bilezikian, L. A. Raisz, and G. A. then potentiate a return to homeostasis as the proinflammatory me- Rodan, eds. Academic Press, San Diego, CA, p. 687. 16. Brown, T. J., J. M. Rowe, J. W. Liu, and M. Shoyab. 1991. Regulation of IL-6 diators are suppressed. expression by oncostatin M. J. Immunol. 147:2175. Taken together, these in vivo findings suggest that OM partic- 17. Wallace, P. M., J. F. MacMaster, J. R. Rillema, K. A. Rouleau, M. B. Hanson, ipates in attenuating the inflammatory responses and restoring nor- S. A. Burstein, and M. Shoyab. 1995. In vivo properties of Oncostatin M. Ann. NY Acad. Sci. 762:42. mal homeostasis following tissue injury and/or infection. The abil- 18. Malik, N., D. Graves, M. Shoyab, and A. F. Purchio. 1992. Amplification and ity of OM to enhance the negative feedback of proinflammatory expression of heterologous oncostatin M in Chinese hamster ovary cells. DNA cytokine production, in addition to inhibiting their biological ef- Cell Biol. 11:453. 19. Terato, K., D. S. Harper, M. M. Griffiths, D. L. Hasty, X. J. Ye, M. A. Cremer, fects, distinguishes the therapeutic potential of this molecule from and J. M. Seyer. 1995. Collagen-induced arthritis in mice: synergistic effect of E. those of individual cytokine antagonists such as IL-1 receptor an- coli lipopolysaccharide bypasses epitope specificity in the induction of arthritis ␣ with monoclonal antibodies to type II collagen. Autoimmunity 22:137. tagonist or anti-TNF- soluble receptor. Moreover, some key an- 20. Wooley, P. H., D. L. Whalen, A. E. Berger, K. A. Richard, D. G. Aspar, and tiinflammatory activities of OM such as its direct effect on fibro- N. D. Staite. 1993. The effect of an interleukin-1 receptor antagonist protein on blasts and epithelial cells and its induction of protease inhibitors type II-induced arthritis and antigen-induced arthritis in mice. Arthritis Rheum. 36:1305. are not seen with other antiinflammatory cytokines such as IL-6, 21. Yu, M., J. M. Johnson, and V. K. Tuohy. 1996. A predictable sequential deter- IL-10, or IL-11. Such OM-specific responses may reflect the tissue minant spreading cascade invariably accompanies progression of experimental distribution of the recently characterized OM-specific receptor autoimmune encephalomyelitis: a basis for peptide-specific therapy after onset of clinical disease. J. Exp. Med. 183:1777. complex being coupled to a specific set of OM-inducible genes 22. Engler, R. 1995. Acute-phase proteins in inflammation. C. R. Seances Soc. Biol. (57). These data indicate OM could act therapeutically in the treat- Fil. 189:563. ment of inflammatory diseases by regulating the spectrum of 23. Foy, T. M., D. M. Shepherd, F. H. Durie, A. Aruffo, J. A. Ledbetter, and R. J. Noelle. 1993. In vivo CD40-gp39 interactions are essential for thymus- events that comprise a natural feedback loop to return active independent humoral immunity. II. Prolonged suppression of the humoral immune flammation to homeostasis. response by an antibody to the ligand for CD40, gp39. J. Exp. Med. 178:1567. 24. Wallace, P. M., J. N. Rodgers, G. M. Leytze, J. S. Johnston, and P. S. Linsley. 1995. Induction and reversal of long-lived specific unresponsiveness to a T-de- Acknowledgments pendent antigen following CTLA4Ig treatment. J. Immunol. 154:5885. 25. Michie, H. R., K. R. Manogue, D. R. Spriggs, A. Revhaug, S. O’Dwyer, We thank Dr. Mohammed Shoyab, Dr. Najma Malik, and Dr. Stephen C. A. Dinarello, A. Cerami, S. M. Wolff, and D. W. Wilmore. 1988. Detection of McAndrew for their endeavors in protein production and helpful discus- circulating tumor necrosis factor after endotoxin administration. N. Engl. J Med. sions. We also thank the Animal Facility of BMS-PRI (Seattle, WA), with- 318:1481. out whose diligent efforts these studies would not have been possible. Mar- 26. Brissette, W. H., D. A. Baker, E.J. Stam, J. P. Umland, and R. J. Griffiths. 1995. GM-CSF rapidly primes mice for enhanced cytokine production in response to cia Hanson made critical contributions to many aspects of the project, and LPS and TNF. Cytokine 7:291. we thank Debbie Baxter for preparation of the manuscript. 27. Kownatzki, E., A. Kapp, and S. Uhrich. 1986. Novel neutrophil chemotactic factor derived from human peripheral blood mononuclear leukocytes. Clin. Exp. Immunol. 64:214. References 28. Williams, R. O., M. Feldmann, and R. Maini. 1992. Anti-tumor necrosis factor 1. Brennan, F. M., R. N. Maini, and M. Feldmann. 1985. Cytokine expression in ameliorates joint disease in murine collagen-induced arthritis. Proc. Natl. Acad. chronic inflammatory disease. Br. Med. Bull. 51:368. Sci. USA 89:9784. 2. Maini, R. N. 1995. A perspective on anti-cytokine and anti-T cell-directed ther- 29. Chapman, P. B., T. J. Lester, E. S. Casper, J. L. Gabrilove, G. Y. Wong, apies in rheumatoid arthritis. Clin. Exp. Rheum. 13:35. S. J. Kempin, P. J. Gold, S. Welt, R. S. Warren, and H. F. Starnes. 1987. Clinical 3. Breedveld, F. C., and P. A. van der Lubbe. 1995. Monoclonal antibody therapy pharmacology of recombinant tumor necrosis factor in patients with advanced of inflammatory rheumatic diseases. Br. Med. Bull. 51:493. cancer. J. Clin. Oncol. 5:1942. 4. Garcia, I., Y. Miyazaki, K. Araki, M. Araki, R. Lucas, G. E. Grau, G. Milon, 30. Martin, R., and H. McFarland. 1996. Experimental immunotherapies for multiple Y. Belkaid, C. Montixi, W. Lesslauer, and P. Vassalli. 1995. Transgenic mice sclerosis. Semin. Immunopathol. 18:1. expressing high levels of soluble TNF-R1 fusion protein are protected from lethal 31. Zamvil, S. S., and L. Steinman. 1990. The T lymphocyte in experimental allergic septic shock and cerebral malaria, and are highly sensitive to Listeria monocy- encephalomyelitis. Annu. Rev. Immunol. 8:579. togenes and Leishmania major infections. Eur. J. Immunol. 25:2401. 32. Greer, J. M., R. A. Sobel, A. Sette, S. Southwood, M. B. Lees, and 5. Baker, D., D. Butler, B. J. Scallon, J. K. O’Neill, and M. Feldmann. 1994. Control V. K. Kuchroo. 1996. Immunogenic and encephalitogenic epitope clusters of of established experimental allergic encephalomyelitis by inhibition of tumor myelin proteolipid protein. J. Immunol. 156:371. The Journal of Immunology 5555

33. Yoshimura, A., M. Ichihara, I. Kinjyo, M. Moriyama, N. G. Copeland, induce serum amyloid A and potentiate the induction of Interleukin-6 and the D. J. Gilbert, N. A. Jenkins, T. Hara, and A. Miyajima. 1996. Mouse oncostatin activation of the hypothalamus-pituitary-adrenal axis by Interleukin-1. Blood 87: M: an immediate early gene induced by multiple cytokines through the JAK- 1851. STAT5 pathway. EMBO J. 15:1055. 46. Opdenakker, G., and J. Van Damme. 1994. Cytokine-regulated proteases in au- 34. Aderka, D., J. Le, and J. Vilcek. 1989. IL-6 inhibits lipopolysaccharide-induced toimmune diseases. Immunol. Today 15:103. tumor necrosis factor production in cultured human monocytes, U937 cells, and 47. Kilpatrick, L., J. L. Johnson, E. B. Nickbarg, Z.-M. Wang, T.F. Clifford, in mice. J. Immunol. 143:3517. M. Banach, B. S. Cooperman, S. D. Douglas, and H. Rubin. 1991. Inhibition of 35. Mori, L., S. Iselin, G. De Libero, and W. Lesslauer. 1996. Attenuation of colla- human neutrophil superoxide generation by ␣1-antichymotrypsin. J. Immunol. gen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1-treated and 146:2388. TNFR1-deficient mice. J. Immunol. 157:3178. 48. Cichy, J., J. Potempa, R. K. Chawla, and J. Travis. 1995. Stimulatory effect of 36. Modur, V., M. J. Feldhaus, A. S. Weyrich, D. L. Jicha, S. M. Prescott, inflammatory cytokines on ␣1-antichymotrypsin expression in human lung-de- G. A. Zimmerman, and T. M. McIntyre. 1997. Oncostatin M is a proinflammatory rived epithelial cells. J. Clin. Invest. 95:2729. mediator: in vivo effects correlate with endothelial cell expression of inflamma- 49. Edwards, D. R., P. P. Beaudry, T. D. Laing, V. Kowal, P. A. Leco, and M. S. Lim. tory cytokines and adhesion molecules. J. Clin. Invest. 100:158. 1996. The roles of tissue inhibitors of metalloproteinases in tissue remodelling 37. Brown, T. J., J. M. Rowe, M. Shoyab, and P. Gladstone. 1990. Oncostatin M: a and cell growth. Int. J. Obes. Relat. Metab. Disord. 20:9. novel regulator of endothelial cell properties. In Molecular Biology of Cardio- 50. Richards, C. D., M. Shoyab, T. J. Brown, and J. Gauldie. 1993. Selective regu- vascular System: UCLA Symposium on Molecular and Cellular Biology. lation of metalloproteinase inhibitor (TIMP-1) by oncostatin M in fibroblasts in R. R. Schneider, ed. Wiley-Liss, New York, p. 195. culture. J. Immunol. 150:5596. 38. Dayer, J. M., and D. Burger. 1994. Interleukin-1, tumor necrosis factor and their 51. Nemoto, O., H. Yamada, M. Mukaida, and M. Shimmei. 1996. Stimulation of specific inhibitors. Eur. Cytokine Netw. 5:563. TIMP-1 production by oncostatin M in human articular cartilage. Arthritis 39. Richards, C. D., C. Langdon, F. Botelho, T. J. Brown, and A. Agro. 1996. On- Rheum. 39:560. costatin M inhibits IL-1-induced expression of IL-8 and granulocyte-macrophage 52. Cichy, J., J. Potempa, R. K. Chawla, and J. Travis. 1995. Regulation of ␣1- colony-stimulating factor by synovial and lung fibroblasts. J. Immunol. 156:343. antichymotrypsin synthesis in cells of epithelial origin. FEBS Lett. 359:262. 40. Baggiolini, M., and I. Clark-Lewis. 1992. Interleukin-8, a chemotactic and in- 53. Scholz, W. 1996. Interleukin 6 in diseases: cause or cure? Immunopharmacology flammatory cytokine. FEBS Lett. 307:97. 31:131. 41. Steel, D. M., and A. S. Whitehead. 1994. The major acute phase reactants: C- 54. Tilg, H., E. Trehu, M. B. Atkins, C. A. Dinarello, and J. W. Mier. 1994. Inter- reactive protein, serum amyloid P component and serum amyloid A protein. leukin-6 (IL-6) as an anti-inflammatory cytokine: induction of circulating IL-1 Immunol. Today 15:81. receptor antagonist and soluble tumor necrosis factor receptor p55. Blood 83: ␣ 42. Libert, C., P. Brouckaert, and W. Fiers. 1994. Protection by 1-acid glycoprotein 1113. against tumor necrosis factor-induced lethality. J. Exp. Med. 180:1571. 55. Fattori, E., M. Cappelletti, P. Costa, C. Sellitto, L. Cantoni, M. Carelli, 43. Yonemura, Y., M. Kawakita, T. Masuda, K. Fujimoto, and K. Takatsuki. 1993. R. Faggioni, G. Fantuzzi, P. Ghezzi, and V. Poli. 1994. Defective inflammatory Effect of recombinant human interleukin-11 on rat megakaryopoiesis and throm- response in interleukin 6-deficient mice. J. Exp. Med. 180:1243. bopoiesis in vivo: comparative study with interleukin-6. Br. J. Haematol. 84:16. 56. Lindberg R. A., T. S.-C. Juan, A. A. Welcher, Y. Sun, R. Cupples, B. Guthrie, 44. Shainkin-Kestenbaum, R., G. Berlyne, S. Zimlichman, H. R. Sorin, M. Nyska, and F. A. Fletcher. 1998. Cloning and characterization of a specific receptor for and A. Danon. 1991. Acute phase protein, serum amyloid A, inhibits IL-1- and mouse oncostatin M. Mol. Cell. Biol. 18:3357. TBF-induced fever and hypothalamic PGE2 in mice. Scand. J. Immunol. 34:179. 57. Mosley, B., I. C. De, D. Friend, N. Boiani, B. Thoma, L. S. Park, and D. Cosman. 45. Benigni, F., G. Fantuzzi, S. Sacco, M. Sironi, P. Pozzi, C. A. Dinarello, 1996. Dual oncostatin M (OSM) receptors: cloning and characterization of an J. D. Sipe, V. Poli, M. Cappelletti, G. Paonessa, D. Pennica, N. Panayotatos, and alternative signaling subunit conferring OSM-specific receptor activation. J. Biol. P. Ghezzi. 1996. Six different cytokines that share GP130 as a receptor subunit, Chem. 271:32635.