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-a 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.

T

he normal development of an inflammatory response must be rapidly followed by the engagement of a feedback system to minimize adventitious tissue damage and regulate the eventual return to homeostasis. This system involves a multitude of regulators including cytokines, adhesion molecules, proteases, corticosteroids, and subsequent regulators of each of these agents. A normal response to infection or other insult is a selflimiting process that by way of temporal expression of both regulators, and effector molecules, causes the resolution of the initiating event. The failure to resolve the causative insult or to redress the balance of pro- and antiinflammatory agents results in tissue injury and destruction that characterize the pathology of various chronic inflammatory diseases (1, 2). The exact participation of each cytokine in the inflammatory disease process is poorly understood in part due to their complex interplay. However, the ability of a variety of cytokine and cytokine agonists to alter the severity or course of various inflammatory diseases is an impressive testament to the clinical value of cytokines as a target for therapeutic intervention (2, 3). Such data has been accrued using animal models of disease, transgenic animals and, more recently, clinical trials of cytokine inhibitors (4, 5). A variety of approaches are currently being studied to alter cytokine function to bring about the regulation of aberrant inflammatory responses (6). Inhibitors of proinflammatory cytokines, most notably TNF-a inhibitors, have been successful in moderating untoward inflammatory responses (2). Abs to TNF-a and soluble receptors are currently in clinical trials against a variety of diseases including rheumatoid arthritis, multiple sclerosis, and Crohn’s disease (7–9). Their efficacy has helped establish a set of common effectors in these apparently disparate diseases. Alternatively, the cytokines whose normal physiological role is to usher a response *Bristol-Myers Squibb Pharmaceutical Research Institute, Seattle, WA 98121; and † Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ 08543 Received for publication November 19, 1998. Accepted for publication February 16, 1999.

from the inflammatory effector phase back to homeostasis also are being evaluated for their clinical potential as drugs. The cytokines IL-10 and IL-11 both appear to accelerate this process and their administration have proven effective in resolving several animal models of chronic inflammatory disease (10). Oncostatin M (OM)4 is a pleiotropic cytokine that is produced by activated T cells and macrophages and has shown in vitro properties that would be expected to influence the course of inflammatory responses (11, 12). The protein is structurally and functionally related to IL-6, leukemia inhibitory factor (LIF), and IL-11, proteins that also influence immune and inflammatory function (13). Despite each protein signaling via a family of related receptors and sharing various common properties, each is endowed with a unique array of biological functions (13). Numerous activities have been ascribed to OM in vitro, including the differentiation of megakaryocytes, inhibition of tumor cell growth, induction of neurotrophic peptides, regulation of cholesterol metabolism, and effects on bone-derived cells (7, 14, 15). Recently a collective picture of OM has emerged that strongly suggests a natural role of the cytokine in the wound healing process and attenuation of the inflammatory response. We have previously found that OM can modulate the expression of IL-6, an important regulator of various aspects of the host defense system (16). OM has been shown to regulate the expression by human cells of acute phase proteins and protease inhibitors that have been implicated in modulating cytokine function and limiting tissue damage at sites of inflammation. Recently many of these in vitro effects have been found to occur in rodents and nonhuman primates following OM administration (17). Here we have extended these in vivo findings to further understand the role of OM in regulating cytokine networks following inflammatory stimuli. We have also examined the effects of OM treatment in two murine models of disease in which common proinflammatory cytokines have been previously shown to play key roles.

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 Therapies, 2203 Airport Way South, Suite 300, Seattle, WA 98134. E-mail address: [email protected] 2 Current address: Xcyte Therapies, 2203 Airport Way South, Suite 300, Seattle, WA 98134. 3

Current address: Seattle Genetics, 22215, 26th Avenue SE, Bothell, WA 98021.

Copyright © 1999 by The American Association of Immunologists

4 Abbreviations used in this paper: OM, oncostatin M; LIF, leukemia inhibitory factor; a1-Pi, a1-proteinase inhibitor; EAE, experimental autoimmune encephalomyelitis; KLH, keyhole limpet hemocyanin; MMP, matrix metalloproteinase; PLP, proteolipid protein; SAA, serum amyloid A; TIMP-1, tissue inhibitor of metalloproteinase-1; BMS-PRI, Bristol-Myers Squibb Pharmaceutical Research Institute.

0022-1767/99/$02.00

5548

Materials and Methods Animals Studies used female mice (;8 wk old) that were held in quarantine for 2 wk before admission to any study, during which time serological examination was performed. BALB/c and C57BL/6 mice were obtained from Taconic (Germantown, NY), and B10.S-H2(S)SgMcdJ mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals were housed according to the American Association for the Accreditation of Laboratory Animal Care and institutional guidelines. Experiments shown are representative of at least three independent studies. Statistical analyses were performed using a Wilcoxon rank test (Primer for Biostatistics, McGraw-Hill, NY).

Reagents Recombinant human OM was expressed in Chinese hamster ovary cells and purified as described (18). OM was administered via various routes of injection in PBS. Escherichia coli LPS (#L3012) and IFA were purchased from Sigma (St. Louis, MO). TNF-a and IL-6 ELISA was obtained from Endogen (Woburn, MA), and no cross reactivity was found with OM (data not shown). Anti-collagen II monoclonal hybridomas were purchased from Chondrex (Redmond, WA). Abs were produced from hybridoma supernatant and purified by protein A Sepharose chromatography. Mycobacterium tuberculosis was purchased from Difco (Detroit, MI), and pertussis toxin was obtained from List Biological Laboratories (Campbell, CA). SRBC were purchased from PML Microbiologicals (Tualatin, OR), and keyhole limpet hemocyanin (KLH) was obtained from Pacific BioMarine Lab (Venice, CA). Ab G19-4 (anti-CD3) was provided by Jeff Ledbetter (Bristol-Myers Squibb Pharmaceutical Research Institute (BMS-PRI), Seattle, WA), mAb 2E12 (anti-CD28) was provided by Bob Mittler (BMS-PRI), and mAb MR1 (anti-murine CD154) was provided by Dr. Tony Siadak (BMS-PRI).

PBMC assays Human PBMC were prepared from blood obtained from healthy donors by separation on Ficoll. T cells were isolated from this fraction by rosetting with SRBC, and the monocytes were separated from the remaining PBMC by elutriation. T cell populations were .95% CD31 and monocyte populations were .95% CD141 as determined by immunostaining. Monocytes were activated by treatment with 5 ng/ml bacterial LPS and T cells were activated by costimulation with immobilized anti-CD3 Ab/soluble anti-CD28 Ab (10 mg/ml). Cells were cultured using RPMI 1640 basal medium supplemented with 10% FBS and penicillin/streptomycin. Culture supernatants were collected at various times following activation for measurement of cytokine content by cytokine-specific ELISA assay.

Cytokine and survival studies Mice (C57BL/6) were coinjected i.v. with various doses of OM alone or with 1 mg LPS in PBS. OM was injected i.v. at various time points (100 ml, 100 mg/ml) before 1 mg LPS injection, when treatment was delayed. Blood samples were collected via retroorbital sinus into heparinized tubes 1 h after LPS administration. Plasma was removed from the blood following centrifugation and stored frozen at 220°C before assay by ELISA. Control studies showed that comparable cytokine levels were measured in freshly isolated plasma. In survival studies, BALB/c mice were injected with OM (10 mg, 100 ml, 100 mg/ml) i.p. at 4, 2, and 1 h before coinjection of OM and LPS. The final injection contained OM (10 mg) and various doses of LPS in PBS (200 ml). Animals were monitored daily for survival and signs of shock. Moribund animals were sacrificed.

Induction of arthritis by anti-collagen Abs and LPS Arthritis was induced using the method of Terato et al. (19). BALB/c mice were injected with 400 ml of a mixture of four Abs (D1, D8, A2, F10) in PBS (2.5 mg/ml/mAb, i.v.). At 72 h after mAb injection, mice were injected with LPS (100 ml, 250 mg/ml, i.p.). Treatment with OM (100 ml, 100 mg/ml in PBS, i.v.) or control diluent (100 ml, PBS, i.v.) began 24 h following LPS and continued until day 10. Histopathology and scoring of arthritic disease were adapted from the methods of Wooley et al. (20). 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 single joints (,80); 2, minimal swelling, minimal redness (90 –100); 3, significant swelling, severe reddening, slight foot malformation (100 –115); 4, maximal swelling, maximal redness, deformed feet (.115). Measurements of normal animals were in the range 65–75 1/1000th inch. At necropsy, the distal one-third of the limbs were immersion-fixed in formalin,

OM ATTENUATES INFLAMMATION decalcified in HCl, processed by routine methods, and embedded into paraffin. The specimens were sectioned at 4 – 6 mm, stained with hematoxylin and eosin, and examined by light microscopy. Sections were graded without prior knowledge of the treatment group. Tibiotarsal (hock) joints were graded as to the severity of inflammation, pannus formation, cartilage damage, and osseuos changes. Each parameter was examined separately and graded as follows: grade 0, unremarkable; grade 1, minimal change; grade 2, mild; grade 3, moderate; and grade 4, severe. The inflammation score was derived from evaluation of soft tissue inflammation, synovitis, and angiogenesis. Pannus formation was defined as hypertrophic synovial tissue composed of intraarticular inflammatory exudate accompanied by synovial cell hyperplasia. Cartilage destruction and loss of matrix were evaluated 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 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 to yield the osseous changes score. The above parameters were then evaluated as to the percent of tissue involved in the disease process: 1, 0 –25%; 2, 26 –50%; 3, 51–75%; 4, 76 –100%. The severity and extent of involvement were then combined to yield the global arthritis score for each joint (maximum possible score, 32).

Peptide synthesis The peptide 139 –151 from proteolipid peptide (PLP) was assembled on a Gilson multiple peptide synthesizer (Middleton, WI) using F-moc amino acids. The peptide resin was treated with trifluoroacetic acid-water-thioanisole-ethanedithiol (100/5/5/2.5) for 2 h, and the cleaved, deprotected peptide was purified by reversed-phase chromatography on a Dynamax C-8 column. The final product was shown to have the expected m.w. by mass spectrometry on a Bio-Ion 20 instrument (Bio-Ion, Uppsala, Sweden).

Induction of experimental autoimmune encephalomyelitis (EAE) EAE was induced using a protocol similar to that previously described (21). B10.S-H2(S)SgMcdJ mice were immunized by s.c. injection at two sites in the abdominal flanks on day 0 with PLP 139 –151 peptide (125 mg) and 300 mg M. tuberculosis H37RA in 200 ml of a 1:1 mixture PBS and IFA. Mice were then injected i.p. with 400 ng pertussis toxin diluted in PBS immediately following the peptide injection. The animals received a second injection of 400 ng pertussis toxin i.p. at 48 h postimmunization. Animals were treated with OM (100 ml, 100 mg/ml, PBS, i.p.) or control diluent (PBS/BSA) on days 4 –7 and 12–18. All mice were examined daily for neurological signs of disease. Disease was evaluated as previously described (22) using the following scale: 0, no abnormality; 1, floppy tail with mild hind limb weakness; 2, floppy tail with moderate hind limb weakness; 3, hind limb paresis with or without mild forelimb weakness; 4, hind leg paralysis with or without moderate forelimb weakness; 5, quadriplegia; 6, dead or moribund requiring sacrifice.

Lymph node cell stimulation assay Animals were immunized with PLP peptide and treated with OM or control diluent as described above. At day 18, animals (5/group) treated with OM or control diluent were sacrificed, and the inguinal and axillary lymph nodes were removed. The nodes from each animal were pooled, a single cell suspension was prepared, and the red cells were lysed. Cells were plated at 500,000 cells per well in media (200 ml, RPMI 1640, 10% FBS, 10 mM HEPES, 50 mM 2-ME) to which was added PLP peptide at various concentrations. Following culture at 37°C for 3 days, proliferation was measured by addition of [3H]thymidine (1 mC/well, 6.7 Ci/mmol; DuPont NEN, Boston, MA) incorporation for 24 h. Media and cells were transferred and washed onto glass fiber filters using a Tomtec cell harvester, air dried, aqueous scintillation fluid was added, and the filters were counted in a Betaplate scintillation counter (LKB Wallac, Gaithersburg, MD).

Immune response to a T cell-dependent Ag BALB/c mice were injected with 1 3 108 SRBC i.v. (100 ml, PBS) or KLH (250 mg, 100 ml, PBS) i.p. on day 0 then treated with OM (30 mg, i.v., days 0 –10), PBS (100 ml, i.v., days 0 –10), or anti-CD154 mAb MR1 (23) (200 mg i.v., days 0, 2, and 4). Mice were bled at 7-day intervals and assayed for titers to SRBC or KLH by ELISA as previously described (24).

Results Regulation of OM expression following leukocyte activation OM is produced by both activated T cells and macrophages (11, 16), therefore we examined the temporal expression of OM in the 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-g (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

FIGURE 1. Expression of OM after leukocyte activation. A, Time course of OM production from human monocytes following activation. Peripheral blood monocytes were activated by treatment with E. coli LPS. Media was sampled at the indicated times following activation, and the levels of cytokines were quantified by ELISA. Shown are the levels of TNF-a and OM 6 SD at the indicated times after activation. B, A time course of OM production from human T lymphocytes following activation. Peripheral blood T lymphocytes were activated by treatment with anti-CD3 and anti-CD28 Abs. Following activation, media were sampled at the indicated times, and the levels of secreted lymphokines were quantified by ELISA. Shown are the concentrations of IFN-g, IL-2, and OM (6SD) in the media at the indicated times after activation.

proinflammatory stimuli. Treatment of peripheral blood monocytes with E. coli-derived LPS resulted in a rapid induction of TNF-a (Fig. 1A) (25, 26). Secreted TNF-a levels in the media peaked at 2 h postactivation and then declined over the next 46 h. Analysis of the same supernatants demonstrated that OM was produced significantly later than TNF-a. Increased OM levels were first detectable at 24 h post-LPS (.10 pg/ml), then continued to rise over the next 24 h, reaching a maximum of 1000 pg/ml at 48 h. These results demonstrate that OM secretion from monocytes is significantly delayed following cell activation with a pro-inflammatory stimulus. In T cells, induction of IL-2 and other activation markers rapidly follows receptor-mediated signaling (27). As shown in Fig. 1B, activation of human peripheral blood T cells by the cross-linking of Abs to CD28 and CD3 rapidly induced IL-2 to

Levels of TNF-a 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 pathogenesis of many of these diseases (5, 28). We initially asked if TNF-a production following challenge of mice with LPS was affected by OM. Following i.v. injection of LPS (1 mg), control animals had the expected rapid increase of TNF-a with maximal levels of 50 ng/ml being measured 1 h postinjection. OM inhibited the induction of TNF-a expression in a dose-dependent manner when administered concurrently with LPS (Fig. 2A). Maximal inhibition of TNF-a (;75%) occurred at a dose of 1 mg 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-a induction (Fig. 2B). Because IL-6 is regulated in vivo by both OM and TNF-a (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-a production (29) and OM blocked the production of TNF-a. Increased levels of IL-6 diminished 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 TNF-a levels. IL-1 levels were also measured in these experiments and were not significantly altered by OM administration (data not shown). We next asked whether the increases in IL-6 expression seen with OM in combination with LPS were the result of a synergistic or additive effect between OM and TNF-a produced by LPS. Combinations of OM (10 mg) with TNF-a (1 mg or 10 mg) produced significantly higher levels of IL-6 (more than sevenfold) than each cytokine alone and at greater levels than would be expected from just an additive effect (Fig. 2C). In vitro studies using human or murine macrophages isolated from either peripheral blood or peritoneal exudates were unable to demonstrate a direct effect of OM on the production of TNF-a or IL-6. Also, macrophages isolated from mice treated with OM produced similar levels of soluble TNF-a following LPS stimulation in vitro to normal controls (data not shown). These findings suggest that the effects of OM on TNF-a production in vivo are indirect. The levels of IL-10, another antiinflammatory cytokine that inhibits TNF-a production (10), and the proinflammatory cytokine IL-1b were not significantly changed in mice treated with OM and LPS compared with mice receiving LPS alone (data not shown). In conclusion, OM is able to inhibit TNF-a production while augmenting the normal feedback loop of IL-6 expression. The alteration in cytokine production was accompanied by a reduced lethality of OM-treated mice following exposure to LPS. Animals were injected with increasing doses of LPS (25–200 mg) either alone or with OM. OM (10 mg) was administered i.p. 4, 2, and 1 h prior and at the time of LPS injection. As shown in Table

5550

OM ATTENUATES INFLAMMATION

FIGURE 2. LPS-induced TNF-a regulation by OM. A, Dose-response of TNF-a 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-a and IL-6. Shown are the mean concentrations of cytokine (6SD). B, Effect of timing of OM administration on TNF-a inhibition. Animals were treated with a fixed dose of OM (1 mg, i.v.) at the designated time points before LPS administration and cytokines measured as in A. Shown are the mean concentrations of cytokine (6SD) 1 h after LPS injection. C, Combined effects of OM and TNF-a on IL-6 production. Mice were injected with OM only, TNF-a only, or coinjected with OM and TNF-a combined at the indicated doses. IL-6 levels (6SD) in the blood 1 h after administration are shown.

I, OM treatment increased the maximum tolerated dose of LPS in mice twofold when compared with animals treated with control diluent. Suppression of inflammation and joint destruction in a mouse model of rheumatoid arthritis To further investigate the antiinflammatory properties of OM following inflammatory stimuli, its effects were studied in an Abinduced model of rheumatoid arthritis (19). In this model, inflammation occurs in the absence of a primary immune response, allowing one to distinguish between two immunoregulatory pathways, immune response and inflammation, which are often interdependent and therefore difficult to separate experimentally. To induce joint inflammation, animals received a mixture of four mAbs to collagen type II (1 mg each) followed 72 h later by LPS (25 mg) (19). This protocol induces a severe arthritis ;24 h following LPS injection. OM treatment was initiated on day 4, after joint inflammation was clearly established, and continued for 7 days thereafter (10 mg, i.v.). As shown in Fig. 3A, the severity of joint inflammation was significantly reduced in OM-treated mice compared with control animals when assessed for the incidence and severity of arthritis (20). Nine of 10 control animals were afflicted with arthritic injury by day 6 and had a median score of

Table I. Effect of OM on the survival of mice following LPS treatmenta Treatment (mg LPS)

Control Diluent (no. alive/group total)

OM Treatment (no. alive/group total)

200 100 50 25

1/15 1/30 12/20 10/10

3/15 19/20 10/10 10/10

a BALB/c mice were injected with OM (10 mg, i.p., 4, 2, 1, and 0 (coinjected) h) before LPS challenge. Shown is the number of mice surviving to day 7 after LPS treatment.

3.8 6 1.35. In contrast, only one OM-treated animal had macroscopic evidence of disease with a score .1 and the median arthritic score was 0.4 6 1. At day 11 postinduction of disease, the animals were sacrificed, and the rear limbs were subjected to microscopic examination of disease. Shown in Fig. 3B are representative histological sections of joints from OM-treated mice and those treated with control diluent. Histological examination showed that treatment with OM completely inhibited the influx of inflammatory cells seen in control animals and prevented the tissue damage associated with a severe inflammatory reaction. The inflammation and tissue injury was quantitated. Nine of 10 control animals had severe inflammation and tissue injury including pannus formation, connective tissue destruction and erosion of cartlidge and bone, with an average score of 26.9 6 16.2. The OMtreated mice had histological measures of inflammation and injury consistent with macroscopic evidence, and this group had significantly better score of 2.4 6 7.6 ( p , 0.001). In two additional independent studies, similar efficacy was seen and the cessation of OM treatment at day 7 was not followed by a delayed onset of inflammation in animals monitored for an additional 14 days (data not shown). OM inhibits EAE in the absence of immunosuppression To further evaluate the ability of OM to suppress the inflammatory process, its effects were studied in a murine model of multiple sclerosis, EAE. This model shares many inflammatory components that are key to the destruction of the neural sheath and disease progression in multiple sclerosis (30) and to joint destruction in rheumatoid arthritis. In contrast to the arthritis model, T cells are responsible for the inflammatory stimulus in EAE (31). Further, the model also provided a means to independently measure the effects of OM on the immune and inflammatory components of the disease. Groups of susceptible mice (B10.S-H2 strain) were immunized with a peptide from a myelin sheath protein, PLP. The peptide contained amino acids 139 –151 of PLP and has been previously


5551 of damage in the histological examination correlated with the symptoms of disease when quantitated (OM 0.10 1 0.10, Control 1.60 1 0.31; p # 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. Treatment of the cells with the immunogen induced a concentrationdependent increase in proliferation with no significant differences seen between the control and OM-treated animals (Fig. 5A). Consistent with this, there was no significant difference in levels of circulating Abs to the PLP peptide found in control and OMtreated 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 celldependent 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 positive control. No significant difference was seen in animals that received OM (10 mg) 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 anticollagen mAbs. At 72 h after mAbs injection, an i.v. boost of 25 mg LPS was given to accelerate the progression of disease. Joint and limb inflammation was apparent within 24 h. Treatment with OM (10 mg/day, i.v.) or control diluent of animals began 24 h following LPS and continued until day 10. Arthritic disease was assessed as described in the Materials and Methods. (Representative limbs for each score are shown in the insert.) Shown are the median arthritic scores of control (F) and OM-treated (E) animals 6 SD (1, p # .005; and p, p # .001). B, Representative histology of a hind articular joints at day 11 following initiation of Ab-induced arthritis from animals treated with OM or control diluent (Pos. Control).

demonstrated to be encephalogenic (32). Immunization with the peptide (in IFA and 3 mg/ml M. tuberculosis) was followed by two injections with pertussis toxin as previously described (32). Animals were treated with OM (10 mg, i.p.) or control diluent i.p. on days 4 –7 and days 12–18. At day 11 postimmunization, control animals began to exhibit neurological symptoms of the disease, particularly paralysis. By day 15, 9 of 10 animals in the control group had succumbed to the disease, and the median score was 4 (Fig. 4A). In contrast, no animal that received treatment with OM showed overt signs of the disease in the 18 days following initiation of the disease-inducing immunization. Inhibition of the inflammation associated with this disease was confirmed by histological examination of the brain and spinal cord. Control animals treated with diluent had lesions typical of EAE (Fig. 4B). The majority of the infiltrate were mononuclear cells (mainly T cells and smaller numbers of macrophages) and a few granulocytes. Infiltration involved the meninges with extension in a perivascular, white matter orientation. In contrast, inflammatory infiltrate was not detectable in the histology of OM-treated animals. The extent

Discussion The temporal expression of OM by activated T cells and macrophages follows the production of cytokines associated with the initiation phase of host defense IL-2 and TNF-a, respectively. The role IL-2 may play in the subsequent induction of OM by T cells was not addressed in these studies. However the kinetics of cytokine expression are consistent with the recent findings that the murine OM gene is inducible by IL-2 (33), and the idea that delayed OM production may represent a regulatory function involved in a feedback mechanism following an initial response. The inhibition of endotoxin-induced TNF-a production by OM further supports a role for OM in an attenuation phase following an inflammatory stimulus rather than in the initiation or effector phases. The significance of this attenuation was demonstrated as OM increased the survival of endotoxin-treated animals. Interestingly, patients with septic shock have elevated OM levels (34). The efficacy of OM treatment is also demonstrated in murine disease models of rheumatoid arthritis and multiple sclerosis. Studies of both rheumatoid arthritis and multiple sclerosis have clearly demonstrated the importance of TNF-a in disease progression, and in models of each disease a variety of TNF-a antagonists have proven efficacious (35). Therefore, it is reasonable that inhibition of TNF-a by OM would protect from other inflammatory stimuli, such as in autoimmune disease, in addition to LPS. Others have reported that commercially available OM, expressed in bacteria, is proinflammatory and induces the expression of adhesion molecules on endothelial cells in vitro, and an inflammatory infiltrate in vivo (36). Using highly purified mammalian protein or yeast derived protein, we have previously studied the effects of OM on endothelial cells in vitro (16, 37) and injected protein in vivo in five species (17)

5552


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 mg/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 Materials and Methods. Shown are the median scores of control (F) and OM-treated (E) animals 6 SD. B, EAE-related histopathologic changes. OM, Representative histopathology of meninges between the mesencephalon (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 these findings. These differences may reflect differences in the sources and purity of the proteins used. The mechanisms by which OM inhibits TNF-a remain to be elucidated. No direct effects of OM on TNF-a production could be demonstrated on macrophages either treated ex vivo or isolated from OM-treated animals. IL-6 has been found to inhibit TNF-a production in vivo (34) and is induced by OM (17). However, both the inhibition of TNF-a by OM in IL-6-deficient mice (P. Wallace, unpublished results) and the kinetics of IL-6 induction in the present study make it unlikely that IL-6 is necessary for TNF-a inhibition when LPS and OM are coadministered. However, IL-6 may participate in the sustained effects of OM when LPS administration was delayed. In addition to the inhibition of TNF-a, collateral antiinflammatory properties of OM are likely to contribute to the effects in these disease models. Cell culture studies have shown that OM can block and modify the response to IL-1, the agent provocateur

most closely associated with TNF-a in mediating tissue damage at sites of inflammation (38). In vivo, inflammatory stimuli resulting in production of IL-1 initiate a cascade of effectors including IL-8 and GM-CSF that amplify the inflammatory response by recruiting, expanding, and activating inflammatory cells (39). While OM has no apparent effect on IL-1 production, Richards et al. have demonstrated that gene expression of GM-CSF and IL-8 induced by treatment of synovial fibroblasts with IL-1 are suppressed by cotreatment with OM in a dose-dependent manner (39). One of the primary chemoattractants of neutrophils, IL-8 also stimulates the production of neutrophil peroxide and the exocytosis of tissue degradative granules at sites of inflammation (40). Coincident with suppression of GM-CSF and IL-8, OM acted synergistically with IL-1 to induce expression of IL-6 and tissue inhibitor of metalloproteinase-1 (TIMP-1) in these same cells. Interestingly, the ability of OM to synergize with IL-1 to suppress inflammatory cytokine expression and induce expression of IL-6 was not paralleled by other members of this cytokine family (39).


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 (6SD) 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 3 108 SRBC i.v. or 250 mg KLH i.p. on day 0 then treated with OM (30 mg, i.v., days 0 –10), PBS (100 ml i.v., days 0 –10), or anti-CD154 mAb MR1 (200 mg 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 (6SD) of groups of five mice.

OM treatment in vivo may also act indirectly on the function of IL-1 and TNF-a by inducing a constellation of protein antagonists. The acute phase proteins serum amyloid A (SAA) and a-1 glycoprotein are produced locally following tissue injury to minimize damage proximal to the site of injury. In addition, systemic release of cytokines results in an acute phase response by the liver to down-regulate the inflammatory response and reestablish homeostasis. These proteins are normally produced by adult mammals in response to tissue injury and/or infection (41) and in a normal, self-limiting process are induced by IL-1 and TNF-a themselves. Administration of a-1 glycoprotein protects animals from TNF-a-induced lethality (42). SAA and a-1 glycoprotein produced during the acute phase are thought to decrease inflammation by sequestering circulating IL-1 and decreasing TNF-a expression, respectively (22, 43, 44). We have previously demonstrated that administration of OM can up-regulate the expression of SAA and a-1 glycoprotein in vivo in both mice and in nonhuman primates (17). Corticosteroids are also potent inhibitors of proinflammatory cytokines, including IL-1, IL-8, and TNF-a. OM in combination with IL-1 stimulates the hypothalamus-pituitary-adrenal axis to secrete corticosterone (45), providing an additional mechanism whereby it can feedback to attenuate inflammation. Infiltration of inflammatory cells into the articular synovium or the CNS during acute and chronic inflammation results in tissue damage. The secretion of reactive oxygen intermediates, and destructive proteases including neutrophil elastase, cathepsins, and matrix metalloproteinases by these activated cells degrade connective tissue and cartilage in rheumatoid arthritis and the neural sheath in multiple sclerosis (46). In addition to attenuating the cytokines that stimulate secretion of these proteases, in vitro studies have demonstrated that OM is capable of inducing a spectrum of protease inhibitors. The acute phase proteins induced by OM also include two major serine proteinase inhibitors: a1-proteinase inhibitor (a1-Pi) and anti-chymotrypsin. a1-Pi, the primary inhibitor of neutrophil elastase, is secreted from lung epithelial cells and synovial fibroblasts stimulated by OM (C. Richards, unpublished observations). In comparison, LIF and IL-6 have little or no effect on expression of a1-Pi. Interestingly, the stimulation of a1-Pi by OM is greatly enhanced in the presence of IL-1. Anti-chymotryp-

sin, an inhibitor of cathepsin G and other chymotrypsin-like enzymes secreted during inflammation, also inhibits superoxide generation by activated neutrophils (47). Anti-chymotrypsin is produced following OM treatment of hepatic and numerous nonhepatic human cells (48). Its expression by epithelial cells in response to treatment with OM is synergistically increased by cotreatment with OM and corticosteroid (48). In vivo, the action of collagenase and gelatinase are regulated by the relative level of their cognate inhibitor, TIMP-1 (49). OM increased the expression of TIMP-1 by synovial fibroblast and did so more effectively than other members of the IL-6 cytokine family (50). Similarly, OM stimulated the production of TIMP-1 from human articular chondrocytes and cartilage explant culture more effectively than IL-11, LIF, or IL-6 (51). Expression of TIMP-1 from fibroblasts in response to OM occurs with no effect on matrix metalloproteinase (MMP) levels, resulting in a net decrease in MMP activity (50). The effects of OM on a1-Pi, antichymotrypsin, and TIMP-1 expression are greatly enhanced in the presence of IL-1 (52), again suggesting that OM works in consort with proinflammatory molecules as part of an antiinflammatory feedback loop. Based on the numerous cell types responsive to OM, it is reasonable to expect this synergistic feedback could occur at sites of inflammation. The synergy of OM with proinflammatory mediators is also seen for the induction of IL-6. Although the role of IL-6 in inflammation remains controversial (53), in vivo IL-6 induces IL-1 receptor antagonist and soluble TNF receptor p55 that attenuate inflammation (54). IL-6 is a key inducer of protease- and cytokine-inhibitors that reduce inflammation and initiate healing. Its protective effects are also inferred from the failure of IL-6-deficient mice to repair tissue and recover from inflammation, to attenuate TNF-a, or produce proteins that limit damage at sites of injury (55). In this light, the ability of OM to enhance IL-6 production is consistent with a role for the protein in tissue repair. The synergistic interaction of OM and TNF-a on increased IL-6 expression supports the concept that OM activity is enhanced in the presence of this proinflammatory cytokine at sites of injury and inflammation. This synergy between pro- and antiinflammatory cytokines is not unique to OM

5554 and TNF-a, as OM combined with IL-1 also yields a similar enhancement of IL-6 (45). As described above, the production of acute phase proteins is also maximized by the combined presence of OM and inflammatory mediators. We have demonstrated that expression of OM from activated T cells and macrophages is temporally delayed and increases coincident with a decline in expression of TNF-a and IL-2. We, and others, have demonstrated synergy between OM and inflammatory cytokines in suppression of inflammatory mediators, and we have shown herein that, administered systemically, the molecule is efficacious in three different models of acute disease with common proinflammatory mediators. The recent cloning of the murine OMspecific receptor has called into question the interaction of the human protein in murine studies (56). However, cytokine production and the generation of acute phase proteins also occur from human cells treated with the human protein. Similarly, we have also established that OM functions in nonhuman primates to inhibit LPS-induced TNF-a production and up-regulate both IL-6 (P. M. Wallace and A. F. Wahl, unpublished observations) and acute phase proteins (17). In the context of an inflammatory cycle, initiators such as TNF-a and IL-1, which promote inflammatory cell activation and secretion of chemoattractants and proteinases, would remain maximal at the peak of inflammatory response. Local expression of OM in the presence of these activators would then potentiate a return to homeostasis as the proinflammatory mediators are suppressed. Taken together, these in vivo findings suggest that OM participates in attenuating the inflammatory responses and restoring normal homeostasis following tissue injury and/or infection. The ability of OM to enhance the negative feedback of proinflammatory cytokine production, in addition to inhibiting their biological effects, distinguishes the therapeutic potential of this molecule from those of individual cytokine antagonists such as IL-1 receptor antagonist or anti-TNF-a soluble receptor. Moreover, some key antiinflammatory activities of OM such as its direct effect on fibroblasts and epithelial cells and its induction of protease inhibitors are not seen with other antiinflammatory cytokines such as IL-6, IL-10, or IL-11. Such OM-specific responses may reflect the tissue distribution of the recently characterized OM-specific receptor complex being coupled to a specific set of OM-inducible genes (57). These data indicate OM could act therapeutically in the treatment of inflammatory diseases by regulating the spectrum of events that comprise a natural feedback loop to return active inflammation to homeostasis.


6. 7.

8.

9. 10. 11.

12.

13. 14.

15.

16. 17.

18.

19.

20.

21.

22. 23.

24.

Acknowledgments We thank Dr. Mohammed Shoyab, Dr. Najma Malik, and Dr. Stephen McAndrew for their endeavors in protein production and helpful discussions. We also thank the Animal Facility of BMS-PRI (Seattle, WA), without whose diligent efforts these studies would not have been possible. Marcia Hanson made critical contributions to many aspects of the project, and we thank Debbie Baxter for preparation of the manuscript.

References 1. Brennan, F. M., R. N. Maini, and M. Feldmann. 1985. Cytokine expression in chronic inflammatory disease. Br. Med. Bull. 51:368. 2. Maini, R. N. 1995. A perspective on anti-cytokine and anti-T cell-directed therapies in rheumatoid arthritis. Clin. Exp. Rheum. 13:35. 3. Breedveld, F. C., and P. A. van der Lubbe. 1995. Monoclonal antibody therapy of inflammatory rheumatic diseases. Br. Med. Bull. 51:493. 4. Garcia, I., Y. Miyazaki, K. Araki, M. Araki, R. Lucas, G. E. Grau, G. Milon, Y. Belkaid, C. Montixi, W. Lesslauer, and P. Vassalli. 1995. Transgenic mice expressing high levels of soluble TNF-R1 fusion protein are protected from lethal septic shock and cerebral malaria, and are highly sensitive to Listeria monocytogenes and Leishmania major infections. Eur. J. Immunol. 25:2401. 5. Baker, D., D. Butler, B. J. Scallon, J. K. O’Neill, and M. Feldmann. 1994. Control of established experimental allergic encephalomyelitis by inhibition of tumor

25.

26.

27.

28.

29.

30. 31. 32.

necrosis factor (TNF) activity within the central nervous system using monoclonal antibodies and TNF receptor-immunoglobulin fusion proteins. Eur. J. Immunol. 24:2040. Burger, D., and J.-M. Dayer. 1995. Inhibitory cytokines and cytokine inhibitors. Neurology 45:39. Lorenz, H.-M., C. Antoni, T. Valerius, R. Repp, M. Gru¨nke, N. Schwerdtner, H. Nu¨blein, J. Woody, J. R. Kalden, and B. Manager. 1996. In vivo blockade of TNF-a by intravenous infusion of a chimeric monoclonal TNF-a antibody in patients with rheumatoid arthritis. J. Immunol. 156:1646. Davidsen, B., O. H. Nielsen, B. Vainer, and I. Kirman. 1996. The immunological network: novel approaches to the treatment of Crohn’s disease. Exp. Opin. Invest. Drugs 5:555. Raine, C. S. 1995. Multiple sclerosis: TNF revisited, with promise. Nat. Med. 1:211. de Vries, J. E. 1995. Immunosuppressive and anti-inflammatory properties of interleukin 10. Ann. Med. 27:537. 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 histiocytic lymphoma cells. Proc. Natl. Acad. Sci. USA 83:9739. Richards, C. D., and M. Shoyab. 1993. The role of oncostatin M in the acute phase response. In Acute Phase Proteins: Molecular Biology, Biochemistry and Clinical Applications. A. Mackiewicz, I. Kushner, and H. Baumann, eds. CRC Press, Cleveland, Ohio, p. 321. Kishimoto, T., S. Akira, M. Narazaki, and T. Taga. 1995. Interleukin-6 family of cytokines and gp130. Blood 86:1243. Bruce, A. G., T. M. Rose, P. S. Linsley, and P. M. Wallace. 1993. Oncostatin M. In Human Cytokines: Handbook for Basic and Clinical Research. B. B. Aggarwal and J. U. Gutterman, eds. Blackwell Science, Oxford, U.K., p. 361. Horowitz, M. C., and J. A. Lorenzo. 1996. Local regulators of bone: IL-1, TNF, lymphotoxin, interferon-g, IL-8, IL-10, IL-4, the LIF/IL6 family, and additional cytokines. In Principles of Bone Biology. J. P. Bilezikian, L. A. Raisz, and G. A. Rodan, eds. Academic Press, San Diego, CA, p. 687. Brown, T. J., J. M. Rowe, J. W. Liu, and M. Shoyab. 1991. Regulation of IL-6 expression by oncostatin M. J. Immunol. 147:2175. Wallace, P. M., J. F. MacMaster, J. R. Rillema, K. A. Rouleau, M. B. Hanson, S. A. Burstein, and M. Shoyab. 1995. In vivo properties of Oncostatin M. Ann. NY Acad. Sci. 762:42. Malik, N., D. Graves, M. Shoyab, and A. F. Purchio. 1992. Amplification and expression of heterologous oncostatin M in Chinese hamster ovary cells. DNA Cell Biol. 11:453. Terato, K., D. S. Harper, M. M. Griffiths, D. L. Hasty, X. J. Ye, M. A. Cremer, and J. M. Seyer. 1995. Collagen-induced arthritis in mice: synergistic effect of E. coli lipopolysaccharide bypasses epitope specificity in the induction of arthritis with monoclonal antibodies to type II collagen. Autoimmunity 22:137. Wooley, P. H., D. L. Whalen, A. E. Berger, K. A. Richard, D. G. Aspar, and N. D. Staite. 1993. The effect of an interleukin-1 receptor antagonist protein on type II-induced arthritis and antigen-induced arthritis in mice. Arthritis Rheum. 36:1305. Yu, M., J. M. Johnson, and V. K. Tuohy. 1996. A predictable sequential determinant spreading cascade invariably accompanies progression of experimental autoimmune encephalomyelitis: a basis for peptide-specific therapy after onset of clinical disease. J. Exp. Med. 183:1777. Engler, R. 1995. Acute-phase proteins in inflammation. C. R. Seances Soc. Biol. Fil. 189:563. 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 thymusdependent humoral immunity. II. Prolonged suppression of the humoral immune response by an antibody to the ligand for CD40, gp39. J. Exp. Med. 178:1567. 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-dependent antigen following CTLA4Ig treatment. J. Immunol. 154:5885. Michie, H. R., K. R. Manogue, D. R. Spriggs, A. Revhaug, S. O’Dwyer, C. A. Dinarello, A. Cerami, S. M. Wolff, and D. W. Wilmore. 1988. Detection of circulating tumor necrosis factor after endotoxin administration. N. Engl. J Med. 318:1481. 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 LPS and TNF. Cytokine 7:291. Kownatzki, E., A. Kapp, and S. Uhrich. 1986. Novel neutrophil chemotactic factor derived from human peripheral blood mononuclear leukocytes. Clin. Exp. Immunol. 64:214. Williams, R. O., M. Feldmann, and R. Maini. 1992. Anti-tumor necrosis factor ameliorates joint disease in murine collagen-induced arthritis. Proc. Natl. Acad. Sci. USA 89:9784. Chapman, P. B., T. J. Lester, E. S. Casper, J. L. Gabrilove, G. Y. Wong, S. J. Kempin, P. J. Gold, S. Welt, R. S. Warren, and H. F. Starnes. 1987. Clinical pharmacology of recombinant tumor necrosis factor in patients with advanced cancer. J. Clin. Oncol. 5:1942. Martin, R., and H. McFarland. 1996. Experimental immunotherapies for multiple sclerosis. Semin. Immunopathol. 18:1. Zamvil, S. S., and L. Steinman. 1990. The T lymphocyte in experimental allergic encephalomyelitis. Annu. Rev. Immunol. 8:579. Greer, J. M., R. A. Sobel, A. Sette, S. Southwood, M. B. Lees, and V. K. Kuchroo. 1996. Immunogenic and encephalitogenic epitope clusters of myelin proteolipid protein. J. Immunol. 156:371.

The Journal of Immunology 33. Yoshimura, A., M. Ichihara, I. Kinjyo, M. Moriyama, N. G. Copeland, D. J. Gilbert, N. A. Jenkins, T. Hara, and A. Miyajima. 1996. Mouse oncostatin M: an immediate early gene induced by multiple cytokines through the JAKSTAT5 pathway. EMBO J. 15:1055. 34. Aderka, D., J. Le, and J. Vilcek. 1989. IL-6 inhibits lipopolysaccharide-induced tumor necrosis factor production in cultured human monocytes, U937 cells, and in mice. J. Immunol. 143:3517. 35. Mori, L., S. Iselin, G. De Libero, and W. Lesslauer. 1996. Attenuation of collagen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1-treated and TNFR1-deficient mice. J. Immunol. 157:3178. 36. Modur, V., M. J. Feldhaus, A. S. Weyrich, D. L. Jicha, S. M. Prescott, G. A. Zimmerman, and T. M. McIntyre. 1997. Oncostatin M is a proinflammatory mediator: in vivo effects correlate with endothelial cell expression of inflammatory cytokines and adhesion molecules. J. Clin. Invest. 100:158. 37. Brown, T. J., J. M. Rowe, M. Shoyab, and P. Gladstone. 1990. Oncostatin M: a novel regulator of endothelial cell properties. In Molecular Biology of Cardiovascular System: UCLA Symposium on Molecular and Cellular Biology. R. R. Schneider, ed. Wiley-Liss, New York, p. 195. 38. Dayer, J. M., and D. Burger. 1994. Interleukin-1, tumor necrosis factor and their specific inhibitors. Eur. Cytokine Netw. 5:563. 39. Richards, C. D., C. Langdon, F. Botelho, T. J. Brown, and A. Agro. 1996. Oncostatin M inhibits IL-1-induced expression of IL-8 and granulocyte-macrophage colony-stimulating factor by synovial and lung fibroblasts. J. Immunol. 156:343. 40. Baggiolini, M., and I. Clark-Lewis. 1992. Interleukin-8, a chemotactic and inflammatory cytokine. FEBS Lett. 307:97. 41. Steel, D. M., and A. S. Whitehead. 1994. The major acute phase reactants: Creactive protein, serum amyloid P component and serum amyloid A protein. Immunol. Today 15:81. 42. Libert, C., P. Brouckaert, and W. Fiers. 1994. Protection by a1-acid glycoprotein against tumor necrosis factor-induced lethality. J. Exp. Med. 180:1571. 43. Yonemura, Y., M. Kawakita, T. Masuda, K. Fujimoto, and K. Takatsuki. 1993. Effect of recombinant human interleukin-11 on rat megakaryopoiesis and thrombopoiesis in vivo: comparative study with interleukin-6. Br. J. Haematol. 84:16. 44. Shainkin-Kestenbaum, R., G. Berlyne, S. Zimlichman, H. R. Sorin, M. Nyska, and A. Danon. 1991. Acute phase protein, serum amyloid A, inhibits IL-1- and TBF-induced fever and hypothalamic PGE2 in mice. Scand. J. Immunol. 34:179. 45. Benigni, F., G. Fantuzzi, S. Sacco, M. Sironi, P. Pozzi, C. A. Dinarello, J. D. Sipe, V. Poli, M. Cappelletti, G. Paonessa, D. Pennica, N. Panayotatos, and P. Ghezzi. 1996. Six different cytokines that share GP130 as a receptor subunit,

5555

46. 47.

48.

49.

50.

51.

52. 53. 54.

55.

56.

57.

induce serum amyloid A and potentiate the induction of Interleukin-6 and the activation of the hypothalamus-pituitary-adrenal axis by Interleukin-1. Blood 87: 1851. Opdenakker, G., and J. Van Damme. 1994. Cytokine-regulated proteases in autoimmune diseases. Immunol. Today 15:103. Kilpatrick, L., J. L. Johnson, E. B. Nickbarg, Z.-M. Wang, T.F. Clifford, M. Banach, B. S. Cooperman, S. D. Douglas, and H. Rubin. 1991. Inhibition of human neutrophil superoxide generation by a1-antichymotrypsin. J. Immunol. 146:2388. Cichy, J., J. Potempa, R. K. Chawla, and J. Travis. 1995. Stimulatory effect of inflammatory cytokines on a1-antichymotrypsin expression in human lung-derived epithelial cells. J. Clin. Invest. 95:2729. Edwards, D. R., P. P. Beaudry, T. D. Laing, V. Kowal, P. A. Leco, and M. S. Lim. 1996. The roles of tissue inhibitors of metalloproteinases in tissue remodelling and cell growth. Int. J. Obes. Relat. Metab. Disord. 20:9. Richards, C. D., M. Shoyab, T. J. Brown, and J. Gauldie. 1993. Selective regulation of metalloproteinase inhibitor (TIMP-1) by oncostatin M in fibroblasts in culture. J. Immunol. 150:5596. Nemoto, O., H. Yamada, M. Mukaida, and M. Shimmei. 1996. Stimulation of TIMP-1 production by oncostatin M in human articular cartilage. Arthritis Rheum. 39:560. Cichy, J., J. Potempa, R. K. Chawla, and J. Travis. 1995. Regulation of a1antichymotrypsin synthesis in cells of epithelial origin. FEBS Lett. 359:262. Scholz, W. 1996. Interleukin 6 in diseases: cause or cure? Immunopharmacology 31:131. Tilg, H., E. Trehu, M. B. Atkins, C. A. Dinarello, and J. W. Mier. 1994. Interleukin-6 (IL-6) as an anti-inflammatory cytokine: induction of circulating IL-1 receptor antagonist and soluble tumor necrosis factor receptor p55. Blood 83: 1113. Fattori, E., M. Cappelletti, P. Costa, C. Sellitto, L. Cantoni, M. Carelli, R. Faggioni, G. Fantuzzi, P. Ghezzi, and V. Poli. 1994. Defective inflammatory response in interleukin 6-deficient mice. J. Exp. Med. 180:1243. Lindberg R. A., T. S.-C. Juan, A. A. Welcher, Y. Sun, R. Cupples, B. Guthrie, and F. A. Fletcher. 1998. Cloning and characterization of a specific receptor for mouse oncostatin M. Mol. Cell. Biol. 18:3357. Mosley, B., I. C. De, D. Friend, N. Boiani, B. Thoma, L. S. Park, and D. Cosman. 1996. Dual oncostatin M (OSM) receptors: cloning and characterization of an alternative signaling subunit conferring OSM-specific receptor activation. J. Biol. Chem. 271:32635.