Transgenic Overexpression of Interleukin (IL)-10 in the Lung Causes Mucus Metaplasia, Tissue Inflammation, and Airway Remodeling via IL-13-dependent and -independent Pathways*

To address the complex chronic effector properties of interleukin (IL)-10, we generated transgenic mice in which IL-10 was overexpressed in the lung. In these mice, IL-10 inhibited endotoxin-induced tumor necrosis factor production and neutrophil accumulation. IL-10 also caused mucus metaplasia, B and T cell-rich inflammation, and subepithelial fibrosis and augmented the levels of mRNA encoding Gob-5, mucins, and IL-13. In mice bred to have null mutations of IL-13, IL-4Rα, or STAT-6, transgenic IL-10 did not induce mucus metaplasia but did induce inflammation and fibrosis. IL-10 was also a critical mucin regulator of virus-induced mucus metaplasia. Thus, IL-10, although inhibiting lipopolysaccharide-induced inflammation, also causes mucus metaplasia, tissue inflammation, and airway fibrosis. These responses are mediated by multiple mechanisms with mucus metaplasia being dependent on and the inflammation and fibrosis being independent of an IL-13/IL-4Rα/STAT-6 activation pathway.

IL-10 1 was originally described as a product of type 2 T cells that inhibited interferon-␥ (IFN-␥) production by type I murine T cell clones. Subsequent studies demonstrated that IL-10 is produced by a wide variety of cells including monocytes, macrophages, eosinophils, and bronchial epithelial cells (1,2). They also highlighted the impressive immunosuppressive, anti-inflammatory, and tissue-protective effects of this cytokine (3)(4)(5)(6). This included studies that demonstrated that IL-10 decreases tissue inflammation and injury in animal models of a variety of pulmonary and extrapulmonary human disorders (5,(7)(8)(9) and that IL-10-deficient mice have Th1 polarized immune responses and develop severe inflammatory colitis resulting from chronic stimulation by enteric antigens (10). In an attempt to capitalize on these anti-inflammatory and protective properties, IL-10 is being used in clinical trials to treat patients with inflammatory bowel disease, rheumatoid arthritis, and psoriasis and chronic IL-10 administration has been proposed as a treatment for pulmonary disorders such as cystic fibrosis and asthma (2,(11)(12)(13).
In contrast to its anti-inflammatory properties, IL-10 also has well defined pro-inflammatory and immunostimulatory effector functions. They include the ability to stimulate thymocyte, T cell, B cell, and mast cell proliferation, differentiation, and cytokine elaboration and to activate endothelial cells (14 -18). These and other properties of IL-10 likely contribute to its ability to augment transplant graft rejection (19), contribute to the pathogenesis of autoimmunity and anti-tumor immunity (20 -22), and decrease survival and microbe clearance in models of specific bacterial pneumonias (23,24). Although IL-10 has been proposed to be a useful therapeutic agent that can be chronically administered to patients with inflammatory disorders, the chronic effects of IL-10 in the lung and other organs have not been adequately defined.
The complexities of IL-10 are nicely illustrated in its role in the pathogenesis and treatment of asthma. A number of lines of evidence suggest that IL-10 has a predominantly anti-inflammatory effect in this setting. This includes in vitro studies that demonstrate that IL-10 inhibits Th2 cell cytokine elaboration (9,13,25), IgE production (26), eosinophil survival (27), and antigen-induced eosinophilic inflammation and tumor necrosis factor (TNF) production (9,25,28). Similarly, some in vivo studies have demonstrated a relative deficiency in the production and/or accumulation of IL-10 in asthmatics (12,29,30). This led to the belief that decreased IL-10 production contributes to the chronic inflammation that is characteristic of asthma and that rIL-10 could be a useful therapeutic in this disorder (6,13,25). In contrast, an equally cogent body of data suggests that IL-10 can contribute to the pathogenesis of this disorder. This is based on studies that demonstrate that IL-10 augments antigen-induced eosinophilic inflammation, airway hyperresponsiveness, airway mucus metaplasia, and IL-5 elaboration in acute asthma models (9,31,32) and studies demonstrating that IL-10 mRNA can be detected in elevated quantities in asthmatics at base line and after segmental antigen challenge (9,(33)(34)(35). The degree to which chronic IL-10 can, by itself, induce or exacerbate inflammatory, mucus, and fibrotic responses comparable to those in asthmatic airways has, however, not been investigated.
To further understand the chronic effects of IL-10 in the lung, we generated and characterized mice in which IL-10 was chronically overexpressed in the airway. These studies demonstrate that IL-10, although inhibiting lipopolysaccharide-induced tissue inflammation and TNF production, simultaneously induces a B and T cell-rich inflammatory response, mucus metaplasia, and airway remodeling. They also demonstrate that IL-10-induces these responses via multiple mechanisms, with the mucus response and the inflammatory and fibrotic responses being mediated by IL-13-dependent and -independent pathways, respectively.

EXPERIMENTAL PROCEDURES
Production and Identification of Transgenic Mice-Construct pKS-CC10-rtTA-hGH, which contains the 2.3-kb rat CC10 promoter (a gift from Dr. J. Whitsett, University of Cincinnati, Cincinnati, OH), reverse tetracycline transactivator (rtTA), and human growth hormone (hGH) intronic and polyadenylation sequences was prepared as described previously by our laboratory (36). Plasmid pcD(Salpha)-F115, which contains the murine IL-10 cDNA, was obtained from the American Type Culture Collection (ATCC no. 68027) The IL-10 cDNA in this plasmid was then amplified using the following PCR primers: 5Ј-CACTAAGCT-TGCCACAAAGC-3Ј and 5Ј-TCCACAGGATCCGTGTTTTAGC-3Ј. The PCR amplification product was then subcloned into pKS-CC10-rtTA-hGH replacing the rtTA with IL-10. This yielded the construct pKS-CC10-IL-10-hGH. After the fidelity of the junction areas and the IL-10 cDNA was confirmed by sequencing, the construct (Fig. 1A) was digested, isolated, purified, dialyzed, and microinjected as previously described (36,37).
The presence or absence of the transgene in the resulting animals and their progeny was determined using tail DNA and Southern blot analysis with 32 P-labeled murine IL-10 cDNA as a probe or PCR. The primers 5Ј-TGCTATGCTGCCTGCTCTTA-3Ј and 5Ј-TCATTTCCGATA-AGGCTTGG-3Ј were used for PCR genotyping. The following PCR protocol was employed: 95°C for 3 min, 30 cycles of 95°C for 1 min, 58°C for 1 min and 72°C for 1 min, and a final extension at 72°C for 10 min.
Bronchoalveolar Lavage (BAL) and Quantification of IL-10 -Mice were euthanized, the trachea was isolated by blunt dissection, and small caliber tubing was inserted and secured in the airway. Three sequential 0.75-ml volumes of PBS with 0.1% bovine serum albumin were then instilled and gently aspirated and pooled. BAL fluid samples were centrifuged, and supernatants were stored at Ϫ70°C until used. The levels of IL-10 were determined with a commercial ELISA as per instructions from the manufacturer (R&D Systems, Inc., Minneapolis, MN).
RPA assays were performed using the RiboQuant kit purchased from BD PharMingen (San Diego, CA). These assays were performed according to instructions provided by the manufacturer.
Immunofluorescence Staining and Confocal Microscopy-To analyze the inflammatory cells in IL-10 transgenic mice, we stained frozen tissue sections with FITC-or PE-labeled antibodies against cell surface markers. In these experiments lungs were perfused, inflated, and fixed overnight at 4°C with paraformaldehyde lysine periodate solution (0.001 M periodate, 0.075 M lysine, 1% paraformaldehyde, pH 7.4). Cryoprotection was accomplished by consecutive 20-min incubations in graded cold sucrose solutions (10, 20, and 30% in 0.1 M phosphate buffer, pH 7.4). The lungs were then inflated with a 40% optimal cutting temperature solution diluted in PBS; cryomolds were prepared with optimal cutting temperature; and 7-m tissue sections were prepared on microscopic glass slides, air-dried at room temperature, and stored at Ϫ80°C for further use. For immunofluorescence staining, the slides were washed and rehydrated with TN buffer (20 mM Tris-HCl, 0.5 M NaCl, pH 7.4) for 5 min, and blocked with 100 l of 3% bovine serum albumin in PBS for 1 h. PE-and FITC-labeled antibodies were directly added onto the samples and incubated overnight at 4°C. PE-labeled anti-mouse CD3 (145-2C11), CD4 (GK 1.5), CD45/B220 (RA3-6B2), and I-A b (A␣ b ) (25-9-3) antibodies and FITC-labeled anti-mouse CD3, CD4, CD45/220, and CD8a (Ly-2) antibodies were purchased from BD PharMingen and used for single or double staining with a combination of PE-and FITC-conjugated antibodies. The slides were then washed three times with TN buffer, mounted, evaluated, and photographed using the fluorescent settings on a Zeiss LSM 510 confocal laser scanning inverted microscope (Axiovert 100M, Zeiss). The laser was set at 488 nm for green (FITC) fluorescence and 568 nm for red (PE) fluorescence.
Characterization of the Effects of IL-10 on Lipopolysaccharide (LPS)induced Inflammation-BAL was performed, as described above, on transgene (Ϫ) and transgene (ϩ) mice before and 6 h after the intratracheal administration of 50 l of LPS (Escherichia coli serotype 0.55:B5, 100 ng/ml) (Sigma) or appropriate vehicle control. Total cell and neutrophil recovery were then assessed. BAL TNF levels were evaluated using a commercial ELISA (R&D, Inc.) as per instructions from the manufacturer.
PCR Cloning of Mouse Muc-4 cDNA-To be able to characterize the expression of the Muc-4 gene in transgenic and control mice, the nucleotide sequence of murine Muc-4 needed to be defined. To accomplish this we designed the PCR primers based on the regions that were conserved in both the human Muc-4 and rat ascites sialoglycoprotein-2 sequences (41,42). The sense 5Ј-TCACTGGTAACCGCTGCTTC-3Ј and antisense 5Ј-CATCCTGGGGGCTGGTAGAC-3Ј primers encompassed the 3Ј-flanking region of human Muc-4 and the 5Ј-coding region of rat ascites sialoglycoprotein-2. After reverse transcription and 30 cycles of PCR amplification with a 60°C annealing temperature, the 513-bp amplified product was purified and cloned into a TA cloning vector (pCR-II, Invitrogen). The inserted cDNA was sequenced, and the nucleotide and corresponding protein sequences were submitted to Gen-Bank TM (accession no. AF218819).
Calculation of Histologic Mucus Index-The histologic mucus index (HMI) provides a measurement of the percentage of epithelial cells that are PAS (ϩ) per unit airway basement membrane. It is calculated from PAS-stained sections as described previously by our laboratories (38,39).
Slot Blotting and Immunodetection of Mucins in the BAL Fluid-To quantitate the levels of mucins in BAL fluids from transgene (ϩ) and transgene (Ϫ) mice, 0.1 ml of BAL fluid was slot blotted onto nitrocellulose membranes using a Minifold II slot blot apparatus (Schleicher & Schuell) according to the protocol provided by the manufacturer. After air-drying, the membrane was blocked with 5% skim milk in TTBS (0.1% Tween 20, 20 mM Tris-Cl, 500 mM NaCl) for 2 h and washed three times with TTBS. The membrane was then incubated overnight at 4°C with a polyclonal antiserum against Mucin-1 (C-20; Santa Cruz Inc., Santa Cruz, CA), a monoclonal antibody against Mucin-2 (Ccp58, Santa Cruz Inc.), or a monoclonal antibody against Mucin-5AC (45M1; Neo-Markers, Union City, CA). After washing with TTBS, the membranes were incubated for 1 h at room temperature with horseradish peroxidase-conjugated anti-mouse or anti-goat immunoglobulin (Ig)-G (Pierce). Immunoreactive mucins were detected using a chemiluminescent procedure (ECL Plus Western blotting detection system, Amersham Biosciences) according to instructions from the manufacturer.
In Situ Hybridization-In situ hybridization was undertaken as previously described (43). The mouse Gob-5 gene probe contained the segment from the mouse Gob-5 sequence between base pairs 1027 and 1465 (GenBank TM accession no. NM_011577).
Respiratory Syncytial Virus (RSV) Sensitization and Challenge-Mice were sensitized intradermally, at the base of the tail, with 5 ϫ 10 5 plaque-forming units of vaccinia virus (VV) expressing secreted RSV G glycoprotein (a gift from Dr. G. W. Wertz, University of Alabama at Birmingham, Birmingham, AL) or with control vaccinia virus expressing ␤-galactosidase (vac-lac; a gift from Dr. B. Moss, National Institutes of Health, Bethesda, MD). Six weeks later the mice were challenged with 1 ϫ 10 7 plaque-forming units (in 0.1 ml) of live RSV virus A2 strain, which was administered intranasally and aspirated into the lung. Four days after challenge the mice were sacrificed, their lungs were removed, and mucin gene expression was assessed with RT-PCR analysis as described. Comparisons were made of the levels of mucin gene expression in (a) wild-type (WT) mice and IL-10 (Ϫ/Ϫ) mice (Jackson Laboratories), (b) WT mice treated with 0.2 mg of either a neutralizing monoclonal antibody against IL-10 (clone JES5.2A5, Genetics Institute, Cambridge MA) or an IgG1 isotype control (American Type Culture Collection), and (c) WT mice treated with the IL-13 antagonist IL-13 R␣ 2 -Fc (a gift from Dr. Sandy Goldman, Wyeth Inc., Cambridge, MA) or an appropriate Ig control. These antibody interventions were administered on days Ϫ1, 0, and ϩ 1 with respect to the live virus challenge.
Statistics-Values are expressed as means Ϯ S.E. As appropriate, groups were compared by analysis of variance with Scheffe's procedure post hoc analysis, Student's t test, or nonparametric assessments (Wilcoxon's rank sum, Mann-Whitney U test) using StatView software for the Macintosh (Abacus Concepts Inc., Berkeley, CA).

RESULTS
Generation of Transgenic Mice-To generate transgenic mice in which IL-10 is selectively expressed in the lung, constructs containing the CC10 promoter and murine IL-10 were prepared (Fig. 1A), standard microinjection was undertaken, tail DNA was isolated, and the presence or absence of IL-10 transgenic sequences was determined via Southern blot analysis and PCR. Four animals contained the appropriate transgenic construct. Independent lines were subsequently generated by breeding these founders with C57BL/6 mice, and the appropriateness of gene expression in each line was evaluated. Qualitatively, similar phenotypes were seen on all four lines. The details of these phenotypes are described below.
Characterization of Gene Expression in Transgenic Mice-To determine whether IL-10 was appropriately targeted to the lung, we quantitated the levels of BAL IL-10 and compared the levels of IL-10 mRNA in the lungs and extrapulmonary organs from transgene (ϩ) and transgene (Ϫ) mice. The BAL fluid from transgene (ϩ) line 2 animals contained modest amounts of IL-10 (100 -200 pg/ml). Higher levels of IL-10 were noted in BAL fluids from transgene (ϩ) line 5 (400 -600 pg/ml), line 6 (1.7-1.9 ng/ml), and line 7 (1.9 -2.3 ng/ml) animals (Fig. 1B). IL-10 was not detected in the BAL fluid from transgene (Ϫ) littermate control animals (data not shown). In accord with these findings, mRNA encoding IL-10 was readily detected in RNA from lungs from line 6 and 7 mice, was detected at lower levels in the RNA from line 5 and 2 mice, and was undetectable in pulmonary tissues from transgene (Ϫ) animals (data not shown). In addition, transgene-induced IL-10 mRNA was not noted and histologic abnormalities were not appreciated on hematoxylin and eosin analysis of a variety of extrapulmonary tissues from transgene (ϩ) animals ( Fig. 1C and data not shown). This demonstrates that the CC10 promoter selectively targeted IL-10 to the lungs of transgene (ϩ) animals.
Effect of IL-10 on LPS-induced Inflammation and TNF Production-Comparisons were subsequently undertaken of the responses elicited by LPS in transgene (Ϫ) and transgene (ϩ) mice. BAL neutrophils were only intermittently detected, and TNF was not detected in BAL from transgene (Ϫ) and (ϩ) animals in the absence of LPS administration (data not shown). A profound BAL neutrophilia and impressive levels of BAL TNF were noted 6 h after LPS administration to transgene (Ϫ) mice (Fig. 2). IL-10 inhibited these responses in an impressive fashion. Transgene (ϩ) line 6 and 7 mice had levels of BAL TNF Histologic Effects of IL-10 -The lungs from transgene (ϩ) and transgene (Ϫ) animals were indistinguishable on gross examination, and differences in alveolar size or gross airway morphology were not noted on light microscopic exam ( Fig. 3 and data not shown). In contrast to the transgene (Ϫ) littermates, however, lungs from transgene (ϩ) mice manifest a mononuclear inflammatory response around small and large airways and nearby vascular structures (Fig. 3B). This response was most pronounced in the mice producing high levels of IL-10 for long periods of time (3-4 months) (data not shown). It was, however, readily apparent in line 2 mice at all time points (data not shown). Immunohistochemistry and confocal analysis demonstrated that these infiltrates contained increased numbers of CD4ϩ T cells and B220(ϩ) and MHC II (ϩ) B cells (Fig. 3, C-E). Enhanced staining with antibodies against CD8 was not appreciated (data not shown). Congo red and Alcian blue stains also did not reveal increased numbers of eosinophils or mast cells, respectively (data not shown).
Trichrome stains were also used to compare the fibrotic responses in the lungs from transgene (Ϫ) and (ϩ) animals. A small amount of collagen could be appreciated in and near the airway wall, and loosely packed collagen could be appreciated in the bronchovascular bundles of transgene (Ϫ) mice. In contrast, enhanced collagen deposition was readily appreciated in the subepithelial region and, to a lesser extent, the adventitia of the small and medium-sized airways of the transgenic animals (Fig. 3, panels F-I). This accumulation was readily apparent in all four transgenic lines. It was, however, both doseand time-dependent because it was most prominent in animals that expressed high levels of IL-10 (lines 6 and 7) for prolonged periods of time (3-4 months) (data not shown). When viewed in combination, these studies demonstrate that IL-10 is a potent in vivo stimulator of inflammation and subepithelial airway fibrosis.
Effects of IL-10 on Airway Mucus and Mucin and Gob-5 Gene Expression-To define the effects of IL-10 on airway mucus, we compared the PAS and Alcian blue staining and levels of mucin and Gob-5 gene expression in transgene (Ϫ) and transgene (ϩ) mice. At all time points, PAS-and/or Alcian blue-staining cells could not be appreciated in the airways of the transgene (Ϫ) animals (HMI values, 0 -1). In contrast, PAS-and Alcian bluestaining cells were prominent in the airways of the transgene (ϩ) animals (Fig. 4A). HMI calculations demonstrated that this response was present in all four transgenic lines. The most prominent alterations were noted in line 6 and 7 transgene (ϩ) animals (HMI between 65 Ϯ 3 and 75 Ϯ 5, p Ͻ 0.01). As can be seen in Fig. 4B, IL-10 simultaneously increased the levels of mRNA encoding Muc-5ac, Muc-4, and Muc-2 but not Muc-1. Mucus hypersecretion was also readily apparent in the BAL immunoblot evaluations (Fig. 4C). In accord with the important role of the calcium-regulated chloride channel Gob-5 in mucus metaplasia (46,47), Gob-5 mRNA was also prominently induced in IL-10 transgenic animals. This induction was seen in RT-PCR studies (Fig. 4B) and in in situ hybridization evaluations, which demonstrated that Gob-5 mRNA accumulated selectively in airway epithelial cells (Fig. 4D).
IL-10 Regulation of Th2 Cytokines-Previous studies from our laboratories demonstrated that Th2 cytokines are potent inducers of mucus metaplasia in the airway (48). Thus, studies were undertaken to determine whether IL-4, IL-5, IL-13, or IL-9 were induced by IL-10 in our transgenic mice. Transcripts encoding Th2 cytokines were not detected in total lung RNA from transgene (Ϫ) animals. They were, however, readily apparent in lungs from antigen-sensitized and -challenged WT animals (Fig. 5). As shown in Fig. 5, increased levels of mRNA encoding IL-13 were detected in RNA from IL-10 transgenic mice. This stimulation was at least partially IL-13-specific because IL-10 did not stimulate the accumulation of mRNA encoding IL-4, IL-5, or IL-9 in a similar fashion (Fig. 5).
Mucus-regulating Effects of IL-10 in a Viral Modeling System-To confirm that endogenous IL-10 is an important regulator of in vivo mucus responses, we compared the mucin gene expression in WT and IL-10 null (IL-10 (Ϫ/Ϫ)) mice that had been sensitized with RSV G glycoprotein and challenged with live RSV. This model was chosen because previous studies from our laboratories demonstrated increased mucus and IL-13 production in this system (49,50). In these experiments, significant increases in Muc-5ac gene expression were seen in viruschallenged wild-type animals that had been sensitized with G glycoprotein (Fig. 7). This mucus response was IL-13-dependent because treatment with the IL-13 antagonist IL-13R␣ 2 -Fc abrogated RSV-induced increases in mucin gene expression in this system (Fig. 7A). It also appeared to depend on IL-10 production because comparable levels of induction of mucin gene expression were not seen after virus challenge of G glycoprotein-sensitized IL-10 (Ϫ/Ϫ) animals and the administration of neutralizing antibodies against IL-10 during the viral challenge phase of this protocol abrogated Muc-5ac mRNA induction (Fig. 7, B and C). Interestingly, these effects were not a result of the ability of IL-10 to inhibit IFN-␥ production because similar levels of IFN-␥ gene expression were noted in comparisons of G glycoprotein-sensitized and virus-challenged IL-10 (ϩ/ϩ) and (Ϫ/Ϫ) mice (data not shown). Thus, these studies demonstrate that IL-10 is an important inducer of mucin gene expression in this viral system. DISCUSSION Inflammation in the lung (and other visceral structures) eliminates microbial pathogens and initiates healing responses after tissue injury. However, inflammation can also contribute to organ dysfunction and, if severe, organ failure. As a result, homeostatic systems have evolved that ensure that inflammation is not initiated after trivial exposures or injuries and that inflammatory responses are appropriately down-regulated when they are no longer required. IL-10 has been assumed to play a major role in the latter regulatory responses. This is based on studies that demonstrate that IL-10 is constitutively elaborated and prevents inflammation in normal healthy pulmonary airways and that endogenous IL-10 is induced and controls inflammatory responses initiated by infectious, particulate and antigenic stimuli (2,8,9,25,29,51). Similarly, exogenously administered recombinant IL-10 is protective and/or controls the inflammation in Pseudomonas pneumonia, silicosis, immune complex lung injury, and ischemia-reperfusion lung injury and a deficiency of IL-10 production may contribute to the pathogenesis of the chronic inflammation seen in diseases such as cystic fibrosis (2,7,8,11,51,52). As a consequence of these findings, it has been assumed that IL-10 can be used to control chronic inflammatory pulmonary disorders. However, IL-10 also has pro-inflammatory tissue effects and detrimental effects in models of transplant rejection and infectious pneumonias (19,23,24), and it can contribute to the pathogenesis of autoimmune and inflammatory disorders (53)(54)(55). These studies suggest that IL-10 may contribute to disease pathogenesis and may not always ameliorate tissue inflammation. To address the mechanisms responsible for these complex effector functions and define the mechanisms by which IL-10 can contribute to or ameliorate inflammatory responses in the lung, we generated overexpression transgenic mice in which IL-10 was expressed in the murine airway at levels comparable with or below those seen in relevant animal models, human biologic fluids, and rIL-10-treated patients (9, 33, 56 -61). Studies of these mice demonstrated that IL-10 has anti-inflammatory effects in the setting of endotoxin challenge. They also demonstrated, for the first time, that IL-10 simultaneously induces mucus metaplasia, enhances mucin gene expression and mucin elaboration, and causes a T and B cell-rich inflammatory response and airway fibrosis. Finally, they demonstrate that IL-10 induces IL-13 in the lung and that the effects of IL-10 are mediated via IL-13-dependent and -independent pathways.
To address the complex effector properties of IL-10, we evaluated the effects of our transgene in naive and inflamed lungs. In accord with reports in the literature (3,5,25), these studies revealed potent anti-inflammatory effects of IL-10. IL-10 simultaneously induced a peribronchiolar and perivascular tissue inflammatory response, which, by confocal microscopy, was made up largely of T cells and B cells. This finding is also in accord with studies demonstrating that transgenic IL-10 causes lymphocytic inflammation in the pancreas (62) and salivary gland (53) and that IL-10 stimulates T and B cell proliferation, stimulates T cell chemotaxis, and inhibits T cell apoptosis (17,63,64). One could interpret these studies to demonstrate that IL-10 simultaneously exerts pro-and antiinflammatory effects when chronically expressed in the lung. Although this may be true, we feel it is likely an overly simplistic interpretation. Instead, we believe that our studies and the literature suggest that IL-10 has complex effector properties that can vary depending on the properties of the ongoing inflammation, injury, and repair response and the timing, dose, and localization of the cytokine. In addition, the inflammation and anti-inflammatory effects of IL-10 can be tied together in a cause and effect fashion if some or all of the cells that are found in these pulmonary infiltrates are inhibitory T regulatory 1 (Tr1) cell populations that can be induced in the presence of chronic IL-10 elaboration (4). Preliminary experiments from our laboratory have, however, failed to reveal evidence for Tr1 cells in these infiltrates. 2 Additional experimentation will be required to define the mechanisms of these responses, the effector capacities of the infiltrating lymphocytes in the IL-10 transgenic mice and the relationships between the different effector properties of this cytokine.
Mucus is a viscoelastic gel that coats and lubricates the epithelium of mucosal surfaces and protects against infectious and other environmental insults. Low levels of mucus are produced in the normal lung, and mucus metaplasia with mucus hypersecretion are characteristic features of airways disorders such as asthma, chronic bronchitis, and cystic fibrosis. A variety of inflammatory mediators, including the Th2 cytokines IL-4, IL-5, IL-13, and IL-9, are believed to contribute to the pathogenesis of the mucus abnormalities in these disorders (37,(65)(66)(67). The role that IL-10 plays in these responses has not been thoroughly evaluated. In addition, the literature that exists is controversial, with investigators reporting that a deficiency of IL-10 does not alter or decreases mucus responses in animal models of airway inflammation (31,32). Our studies demonstrate, for the first time, that chronic IL-10 production induces mucus metaplasia with goblet cell hyperplasia, enhanced neutral and acidic mucus accumulation, enhanced mucin and Gob-5 gene expression, and mucus hypersecretion. They are also the first to demonstrate that endogenous IL-10 in necessary for the maximal induction of mucin gene expression 2 C. G. Lee and J. A. Elias, unpublished observation. In panel B challenges were undertaken in mice that received neutralizing antibodies against IL-10 (anti-IL-10 (ϩ)) or isotype controls (anti-IL-10 (Ϫ)). The levels of Muc-5ac gene expression were evaluated via RT-PCR and are expressed as a relative ratio compared with the levels of ␤-actin mRNA (*, p Ͻ 0.05 compared with WT IL-10 (ϩ/ϩ) mice treated with VV containing the soluble RSV G glycoprotein). A representative experiment using the neutralizing antibodies against IL-10 is illustrated in panel C.
in an IL-13-dependent in vivo virus challenge system. These findings suggest that IL-10 can contribute to the pathogenesis of the mucus responses that are seen at sites of chronic inflammation. They also suggest that Gob-5 may be a critical intermediary of this response because Gob-5 expression induces mucus metaplasia in the murine airway (46). This may be particularly important in inflammatory bowel disease, lung cancer, Sjogren's syndrome, and asthma (see below), where increased levels of mucus production and/or mucin gene expression are known to co-exist with the exaggerated production of IL-10 (33,34,55,68).
Tissue fibrosis can represent a healing and repair response and/or a manifestation of disease pathogenesis. From our perspective, these pathways are not mutually exclusive and can be differentiated, in part, by characterizing the effects of the mediators of fibrosis on local tissue inflammation. Mediators involved in healing and repair (such as TGF-␤ and IL-11 (Ref. 38)) have been shown to stimulate fibrosis while decreasing inflammation. In contrast, mediators that are not involved in healing might not alter or could increase local inflammatory responses. IL-10 has well documented anti-inflammatory properties. Our studies demonstrate that IL-10, while inhibiting LPS-induced inflammation, simultaneously generates tissue fibrosis. In contrast to previous studies (51), they also demonstrate that the fibrotic effects of IL-10 are seen in the absence of known exogenous fibrogenic stimuli. These findings have allowed us to hypothesize that, under appropriate circumstances, IL-10 production is a manifestation of local healing and repair. Our demonstration that IL-10 stimulates TGF-␤ 1 would further support this contention. Our findings also suggest that the overproduction of IL-10 can contribute to the pathogenesis of human fibrotic disorders. This may be particularly relevant to the airway remodeling in asthma, the fibrosis in interstitial lung diseases such as silicosis, and the salivary gland scarring seen in Sjogren's syndrome. These studies do not, however, support the concept that IL-10 is anti-fibrotic, as has been proposed by other investigators (69). In circumstances where anti-fibrotic effects of IL-10 are seen, our data would suggest that the decrease in fibrosis is a result of the ability of IL-10 to diminish tissue inflammation and inflammation-induced injury and not direct effects of IL-10 on matrix deposition or structural cell proliferation.
Asthma is characterized by Th2 cytokine excess, chronic inflammation, and varying degrees of mucus metaplasia and subepithelial fibrosis. The role of IL-10 in the asthmatic diathesis, however, is controversial. Early studies suggested that IL-10 had important anti-inflammatory effects in models of asthma (9,25,28) and that a deficiency in IL-10 production contributed to the chronic inflammation in the airways of patients with this disorder (12,29,30). This led to the belief that IL-10 administration would be an appropriate therapeutic intervention in these patients (12,29,30). In contrast, other investigators demonstrated that IL-10 can augment asthmalike pathologic responses in acute animal modeling systems (9,31,32) and that IL-10 production is increased in patients with asthma at base line and after segmental antigen challenge (33,34). In addition, RSV infection is the most common cause of lower respiratory infection in children and correlates with the development of asthma in later life (70). In RSV-infected children, the capacity of stimulated mononuclear cells to produce IL-10 has been shown to correlate with recurrent wheezing on follow-up (71). These studies suggest that IL-10 can contribute to disease pathogenesis in asthma. When viewed in combination with our findings, they also call into question the appropriateness of IL-10 as a therapy for these individuals because the IL-10 could worsen the inflammatory, mucus, fibrotic, and physiologic alterations seen in these individuals. A relative IL-10 deficiency has also been reported in cystic fibrosis, and exogenous rIL-10 has been proposed as a treatment for patients with this disorder (2,11). Similar concerns about this therapeutic approach, however, need to be raised because mucus abnormalities, inflammation, and tissue fibrosis are all characteristic features of cystic fibrosis tissue pathology.
IL-13 is a pleiotropic 12-kDa cytokine that is produced in large quantities by appropriately stimulated CD4ϩ Th2 cells and lesser quantities by Th1 cells. It has a variety of proinflammatory effects that are relevant to asthma and other Th2-dominated inflammatory disorders, including the ability to induce IgE production and endothelial cell VCAM-1 expression. Studies from our laboratory and others have also demonstrated that IL-13 is a potent inducer of eosinophilic tissue inflammation, mucus metaplasia, subepithelial fibrosis, and Gob-5 gene expression (37,47,72). In keeping with the similarities between these IL-13 effector functions and the phenotype of our IL-10 transgenic mice, studies were undertaken to determine whether IL-13 played a role in mediating the IL-10 phenotype. These studies demonstrated, for the first time, that IL-10 stimulates IL-13 production in vivo. Our studies of the progeny of crosses of transgenic mice and IL-13 null mice also demonstrated that the IL-13 that is produced plays a critical role in the induction of the mucus metaplasia and the enhanced mucin and Gob-5 gene expression, but is not critical for the inflammatory or fibrotic responses in these animals. Our demonstration that the IL-10-induced mucus response was completely abrogated whereas the inflammatory and fibrotic responses were not altered in crosses of IL-10 transgenic mice and mice with null mutations of IL-4R␣ or STAT-6 provides additional support for the importance of IL-13 in mediating the IL-10 phenotype and insights into the signaling pathway that it uses in this setting. Interestingly, IL-10 did not stimulate other Th2 cytokines and IL-10-induced mucus metaplasia was not abolished in the absence of IL-4. When viewed in combination, these genetic manipulations demonstrate that IL-10 production selectively induces IL-13 elaboration, which causes mucus transformation of the airway via an IL-4R␣-and STAT-6-dependent mechanism. Because IL-13 is a key mediator in asthma pathogenesis and IL-13 antagonism is a therapeutic goal in this disorder (73), the finding that IL-10 induces IL-13 in vivo further supports the need to be cautious with approaches that utilize IL-10 to control this and other chronic Th2-dominated inflammatory disorders.
Our studies demonstrate that IL-10 induces tissue fibrosis in the murine airway. We previously demonstrated that IL-13 is a potent stimulator of tissue fibrosis and that this effect is mediated, at least in part, via the ability of IL-13 to stimulate the production and activate TGF-␤ 1 (43). As a result, we hypothesized that the fibrogenic effects of IL-10 might be mediated via an IL-13/STAT-6/IL-4R␣-dependent pathway. To our surprise, this did not prove to be the case because IL-10-induced tissue fibrosis was not abrogated in the absence of IL-13, STAT-6, or IL-4R␣. To gain insights into the mechanisms that might contribute to this fibrotic response, studies were undertaken to determine whether IL-10 stimulated the production of TGF-␤ 1 in the murine lung. These studies demonstrated that IL-10 stimulated the accumulation of TGF-␤ 1 mRNA and latent TGF-␤ 1 protein in transgene (ϩ) animals. In keeping with our histologic findings and genetic manipulations, this inductive response was mediated via an IL-13/STAT-6-independent pathway. These observations allow for the speculation that IL-10 induces tissue fibrosis, at least in part, via the induction of TGF-␤ 1 . Additional investigations in which IL-10-dependent fibrotic responses are treated with TGF-␤ 1 antagonists, how-ever, will be required to thoroughly evaluate this possibility.
In summary, these studies demonstrate that IL-10 causes a complex phenotype when chronically overexpressed in the lung. While inhibiting LPS-induced inflammation and TNF production, IL-10 simultaneously caused a T and B cell-rich inflammatory response, subepithelial fibrosis, and mucus metaplasia with goblet cell hyperplasia, mucin hypersecretion, and enhanced mucin and Gob-5 gene expression. They also demonstrate that IL-10 induces IL-13 production in vivo and that this induction is responsible for the mucus, but not the inflammatory and fibrotic effects of IL-10 in this setting. Endogenous IL-10 was also shown to be necessary for maximal Muc-5ac gene expression in a virus sensitization and challenge modeling system. Thus, IL-10 may contribute to the pathogenesis of inflammatory, fibrotic, and mucus responses at sites of pathology. The inflammatory, fibrotic, mucus, and IL-13-stimulating effects of IL-10 must be taken into account when balancing the risks and benefits of chronic IL-10 therapy for inflammatory disorders in the lung and other organs.