Pulmonary Expression of Oncostatin M (OSM) Promotes Inducible BALT Formation Independently of IL-6, Despite a Role for IL-6 in OSM-Driven Pulmonary Inflammation

The gp130 family of cytokines, including IL-6, IL-11, LIF, IL-31, and oncostatin M (OSM), plays various roles in inflammation, hematopoiesis, and immune responses (1, 2). Receptors for this family of cytokines share a common subunit, gp130, which complexes with a variety of cell surface and soluble receptor chains that provide specificity for the separate ligands (2). Functions of this family of cytokines in mucosal immunity are complex and overlapping due to the combinatorial expression of shared and specific receptor subunits on various cell types. Previous studies indicate shared but also private and sp. act. of OSM among gp130 cytokines (2, 3). However, the regulatory roles of OSM in lung mucosal immunity are currently unclear.

OSM is a 26-kDa extracellular proinflammatory glycoprotein that promotes connective tissue remodeling, chemokine expression, and infiltration of inflammatory granulocytes, lymphocytes, and myeloid cells in animal models (4–8). Lung OSM levels are elevated in patients with severe asthma (9), allergic rhinitis (10), and idiopathic pulmonary fibrosis (IPF) (11). Inflammatory mononuclear cells are a major source of OSM (12), whereas its specific receptors, consisting of gp130 and an OSM-specific receptor-β subunit (3), are broadly expressed by structural cells, such as fibroblasts, osteoblasts, smooth muscle cells, and endothelial cells.

Although the effects of OSM on structural cells in various systems have been described, fewer studies have examined its effects on hematopoietic immune cells. Activated myeloid and lymphoid cells secrete OSM (3, 12), and the presence of mononuclear cells producing OSM in vivo correlates with neutrophil influx during early stages of inflammation (13). Additionally, myeloid dendritic cells (DCs) express OSM receptors and respond to OSM by differentiating into potent APCs (14). Transgenic overexpression of OSM stimulates extrathymic T cell differentiation, expansion of memory T cells (15), accumulation of immature B cells, and production of circulating autoantibodies (16).

The prototype gp130 family member IL-6 has effects on ectopic lymphoid tissue development in rodent lungs. IL-6 overexpression (along with the IL-6R) promotes the formation of inducible BALT (iBALT) (17), a tertiary lymphoid structure that contains large B cell aggregates, surrounded by T cells and maintained by DCs (18, 19). Working together with classical lymphoid tissues, iBALT helps control respiratory pathogens (20). The presence of iBALT is associated with various lung inflammatory conditions, including severe asthma, chronic obstructive pulmonary disease (18), and lung complications of rheumatoid arthritis (21). iBALT has been detected in the lungs of mice infected with virus (20) and mycobacteria (22), or exposed neonatally to the bacterial product LPS (23), although the precise pathways by which each of these conditions result in iBALT might not be identical (24). In any case, the function of OSM in iBALT formation and B cell responses during respiratory infection remains to be understood.

In this study, we examined the role of OSM in iBALT formation and activation of B cell lymphocyte populations using an adenoviral vector expressing murine OSM (Ad-mOSM). This approach allows us to investigate transient OSM transgenic expression in the context of viral infection in mouse lungs. Because OSM has been demonstrated to markedly induce IL-6 expression (25), we further assessed the biological effects of Ad-mOSM and control vectors in wild-type (WT) C57BL/6 and IL-6^−/− mice. We observed that Ad-mOSM induced significant inflammation, B cell activation, and iBALT formation, particularly in the lung parenchyma. Although inflammation in the airways was markedly reduced in IL6^−/− mice, B cell accumulation, activation, and iBALT formation in the lung tissue was independent of IL-6.

C57BL/6 (WT) and IL-6–deficient mice (IL-6^−/−, C57BL/6 background, 6–8 wk old) were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were maintained under specific pathogen-free conditions in an access-restricted area, on a 12-h light-dark cycle, with food and water provided ad libitum. The Animal Research Ethics Board of McMaster University approved all experiments. Endotracheal administration of indicated amounts (PFU) of control adenovirus vector (Addel70) or Ad-mOSM has been previously described (25).

Bronchoalveolar lavage (BAL) fluid was collected after instilling lungs with 0.4 ml ice-cold PBS (Invitrogen, Carlsbad, CA) twice. Total cell numbers were determined using a hemocytometer. BAL cytospins were prepared for differential cell counts and stained with Hema-3 (Biochemical Sciences, Swedesboro, NJ). Five hundred cells in each cytospin slide were examined to identify and count macrophages, lymphocytes, neutrophils, and eosinophils. To detect cytokine levels in BAL samples (stored at −80°C before use), ELISA (purchased from R&D Systems) and Luminex bead assays were used according to the protocol recommended by the manufacturer.

Lung mononuclear cell suspensions were generated by mechanical mincing and collagenase digestion. Debris were removed by passage through 45-μm screen-size nylon mesh, and cells were resuspended in PBS containing 0.3% BSA (Invitrogen, Burlington, ON, Canada) or in RPMI 1640 supplemented with 10% FBS (Sigma-Aldrich, Oakville, ON, Canada), 1% l-glutamine, and 1% penicillin/streptomycin (Invitrogen, Burlington, ON, Canada). A quantity amounting to 1 × 10⁶ lung mononuclear cells was washed once with PBS/0.3% BSA and stained with primary Abs directly conjugated to fluorochromes for 30 min at 4°C. A total of 10⁵ live events was acquired on an LSR II (BD Biosciences, San Jose, CA) flow cytometer, and the data were analyzed with FlowJo analysis software (Tree Star, Ashland, OR). Side scatter and forward scatter parameter internalization of 7-aminoactinomycin D (eBiosciences, San Jose, CA) by dead cells were used to define live cell and lymphocyte gates. All Abs were purchased from BD Biosciences (San Jose, CA) or eBiosciences, unless otherwise stated. The following Abs were used for flow cytometric analysis: PE-cy5–conjugated anti-CD69, PerCP-cy5.5–conjugated anti-CD11c, allophycocyanin-conjugated anti-MHC class II, Alexa Fluor 700–conjugated anti-CD86, allophycocyanin-cy7–conjugated anti-CD45, and Pacific Blue–conjugated anti-CD3. Qdot605-conjugated anti-CD4 and Qdot655-conjugated anti-B220 were purchased from Invitrogen. V500-conjugated anti-CD8 was purchased from BD Biosciences.

The entire lung was fixed at 30 cm H₂0 pressure in 10% formalin for histological assessment. Lungs were embedded in paraffin blocks, and 4-μm–thick cross-sections were cut and stained with H&E to assess lung inflammation. Immunofluorescent staining was performed, as previously described (26). For immunofluorescence analyses, slides were hydrated in PBS and blocked for 30 min at 25°C with Fc block (10 μg/ml) and 5% (v/v) normal donkey serum in PBS. Endogenous biotin was blocked with a sequential avidin–biotin incubation step (Sigma-Aldrich). After blockade, the slides were incubated overnight at 25°C with rat anti-B220 (clone RA3-6B2; BD Pharmingen) to visualize B cells, anti–CD21-CD35 (clone 7E9; BioLegend, San Diego, CA) in combination with rat anti-mouse follicular DC (FDC; clone FDCM1; BD Pharmingen) to detect FDC, goat anti-CD3ε to stain T cells (clone M-20; Santa Cruz Biotechnology, Santa Cruz, CA), goat anti–proliferating cell nuclear Ag (PCNA) from Santa Cruz Biotechnology (clone C-20) to detect proliferating cells, and peanut agglutinin (PNA; L7381; Sigma-Aldrich) to detect germinal center B cells. Fluorescently labeled secondary Abs were incubated for 3 h at room temperature. Finally, the slides were incubated with streptavidin conjugated to either Alexa Fluor 594 or Alexa Fluor 488 and counterstained with ProLong Gold antifade with DAPI from Invitrogen. All sections were viewed with a Zeiss Axioplan 2 microscope. Images were recorded with a Zeiss AxioCam HR digital camera.

Frozen serum samples were used to determine the neutralizing Ab titer. Serial dilutions of sera were added to Ad-LacZ, and the ability of serum Abs to block the infection of Hela cells with Ad-LacZ was evaluated with a colorimetric assay 24 h postinfection. The titer is expressed as the serum dilution that produced 50% of maximal inhibition of colorimetric conversion of substrate xGAL by Hela cells infected with Ad-LacZ.

Data were analyzed using IBM SPSS Statistics version 18.0 software (Chicago, IL) and expressed as mean ± SEM. A minimum of four and usually five animals per group was analyzed individually in experiments, and the results shown represent one of at least two experiments, each of which showed the same trend and statistically significant differences (with p < 0.05) in the observations emphasized in this study. We assessed significance (p < 0.05) using the SPSS Univariate General Linear Model, and one- or two-way ANOVA was followed by multiple t tests.

We and others have previously highlighted the ability of IL-6 overexpression to induce iBALT in the rodent lung (17, 27). IL-6 is well known to induce B cell expansion and stimulate Ab production (1, 3); however, the role of the IL-6/gp130 family member OSM in B cell expansion and function is less clear. To test whether pulmonary delivery of transgenic OSM could induce iBALT formation, we endotracheally administered Ad-mOSM or empty vector (Ad-del70) to C57BL/6 mice and examined their lungs 7 and 14 d later. We found numerous lymphoid aggregates in lung parenchymal tissue of Ad-mOSM–treated mice (Fig. 1A, top panels), whereas control (Addel70-treated) lungs had scarce or no detectable inflammatory cell infiltrates [as previously observed (8, 25)] (Fig. 1B). We performed immunofluorescence on Ad-mOSM–infected lungs and found that the lymphoid aggregates were mainly composed of large aggregates of B220⁺ B cells, many of which expressed PCNA (Fig. 1A, left lower panel, red and white stains) and bound PNA, a phenotype consistent with germinal center B cells (Fig. 1A, bottom panels). We found a few CD3⁺ T cells on the edge of the follicles (Fig. 1A, bottom middle panel, red stain) and CD21⁺CD35⁺FDCM1⁺ FDC in the center of the B cell follicles (Fig. 1A, bottom right panel, red stain). These data indicate that pulmonary OSM overexpression promotes iBALT formation.

To assess whether the Ad-mOSM–associated B cell activation and iBALT formation resulted in enhanced Ab responses, we measured neutralizing Ab to adenovirus vector (see 2Materials and Methods) in the serum, 24 d after primary infection and 7 d after a secondary infection (day 35) with Addel70. We found that Ad-mOSM–treated (but neither naive nor Ad-del70–treated) animals had markedly increased titres of neutralizing Ab at day 24 (Fig. 1C). We also found that the neutralizing titer increased a further 10-fold within 7 d after challenge with empty vector on day 28 (Fig. 1C). In sharp contrast, challenge of naive or Addl70-treated animals did not increase neutralizing Ab to detectable levels.

We next used flow cytometry to examine accumulation of B220⁺ B cells and their activation status (CD69 and CD86 expression). Using the gating strategy outlined in Supplemental Fig. 1, we found that mOSM expression elicited a 2- to 3-fold increase in the numbers of B220⁺ B cells evident in the lung on day 7 and day 14 postinfection (Fig. 2A). As expected, numbers of B220⁺ B cells in Ad-del70–infected animals were similar to uninfected controls. Ad-mOSM treatment increased the number of CD69-expressing (CD69^high) and CD86-expressing (CD86^high) B cells (Fig. 2B). CD69 and CD86 expression on B cells, examined by mean fluorescence intensity, was significantly increased in Ad-mOSM–treated mice compared with Ad-del70–treated mice. In addition, marked eosinophilia was observed in lung tissue at 7 and 14 d after Ad-mOSM administration (Fig. 2C). These data demonstrated that OSM transgenic expression in lung induces B cell activation and accumulation of both B cells and eosinophils.

To test whether transient transgenic OSM expression stimulated additional components of the adaptive immune system in the infected lungs, we examined the accumulation and activation of DC, CD4⁺, and CD8⁺ T cells. We found that endotracheal administration of Ad-mOSM, but not control vector, induced a significant increase in the number of CD4⁺ T cells in the lung at 7 and 14 d (Fig. 3A). Although the numbers of CD8⁺ T cells were higher in Ad-mOSM–treated lungs, a statistically significant increase was not detected between the experimental groups (Fig. 3B). However, at both 7 and 14 d post–Ad-mOSM treatment, there was a statistically significant increase in CD69 surface expression and the numbers of activated (CD69-expressing), pulmonary CD4⁺ and CD8⁺ T cells, compared with the control Ad-del70 vector-treated group (Fig. 3A, 3B). Thus, in addition to B cell expansion, T cell accumulation and activation were increased as a result of pulmonary transgenic OSM expression.

Given that DCs are necessary for iBALT maintenance in response to viral infections (19), we next examined whether OSM transgenic expression could stimulate the accumulation and activation of DC, in parallel to iBALT formation. We found that Ad-mOSM treatment elicited a significant increase in the total number of both CD11c^high MHC class II^high DCs and activated CD86⁺ CD11c^high MHC class II^high DCs (Fig. 3C), whereas the number of DCs in control animals was similar to that in uninfected animals. Additionally, the frequency of CD86⁺ DCs and the level of expression of CD86 on DCs were significantly increased in lungs from Ad-mOSM–infected animals, compared with those from Ad-del70–infected or uninfected animals. These data demonstrate that mature, activated DCs accumulate in the lungs in response to transgenic OSM expression, a finding that could be potentially linked to participation of DCs in the formation of OSM-induced iBALT.

Because OSM is a potent inducer of IL-6 in vitro and in mouse lungs in vivo (25, 28), OSM effects on B cells in this system could be due to IL-6 induction. To determine the role of IL-6 in lymphocyte accumulation, we endotracheally administered Ad-mOSM or control vectors to the lungs of WT and IL-6^−/− mice (C57BL/6 background) and evaluated DC, B cell, and T cell accumulation and activation. Surprisingly, IL-6 deficiency did not abrogate OSM-induced B cell accumulation and activation and, in fact, resulted in slightly increased numbers of total and CD69⁺ B cells compared with those in WT mice (Fig. 4A). We also found that OSM-induced accumulation of CD4⁺ and CD8⁺ T cells was normal in IL-6^−/− mice (Fig. 4B, 4C). Similarly, IL-6 deficiency affected neither OSM-induced increase in the number of CD69⁺-activated CD4⁺ T cells nor the increased expression of CD69 on CD4⁺ T cells (Fig. 4B). However, there was a modest decrease in CD69⁺ CD8⁺ T cells in lungs of IL-6^−/− mice, compared with lungs of WT mice (Fig. 4C). CD69 expression on B cells and CD4 T cells, examined by mean fluorescence intensity, was maintained at normal levels in IL-6^−/− mice (panels on right for Fig. 4A–C). In addition, we found that IL-6 deficiency did not impair the accumulation of DCs in Ad-mOSM–treated mice and that DC activation, as measured by CD86 expression, was even greater in the absence of IL-6 (Fig. 4D). Together, these data indicate that OSM is capable of stimulating DC, B cell, and T cell lymphocyte accumulation and activation in the lungs in the absence of IL-6, highlighting the in vivo biological functions of OSM executed independently of IL-6. The accumulation of eosinophils in the lung homogenates induced by Ad-mOSM was significantly reduced in IL-6–deficient mice compared with IL-6^−/− at day 7, and, although showing some reduction at day 14, the difference was not statistically significant (Fig. 4E).

We next enumerated iBALT structures in the lung tissue of WT and IL-6^−/− mice treated with Ad-mOSM or control adenoviruses, 2 wk after adenoviral vector delivery. We found that mice treated with Ad-mOSM, but not Ad-del70, developed lymphocytic cell aggregates in lung parenchymal tissue in both WT and IL-6^−/− mice (Fig. 5A). We observed a diffuse inflammatory cell infiltrate in the airspaces of Ad-mOSM–treated WT mice. At higher magnification (Fig. 5B), inflammatory cells (examples of eosinophils indicated by arrowheads) were located in the alveolar interstitium. In contrast, accumulation of inflammatory cells in airspaces and the alveolar interstitium was less evident in Ad-mOSM–treated IL-6^−/− mice. This is consistent with the analysis of lung tissue homogenates performed by flow cytometry (Figs. 2C, 4E).

Immunofluorescence analysis demonstrated that Ad-mOSM, but not the Ad-del70 control, induced iBALT formation in both WT and IL-6^−/− mice. iBALT was characterized by the presence of large B cell follicles containing PCNA⁺ proliferating B cells and CD21⁺CD35⁺FDCM1⁺ FDC (Fig. 5C). Total number of Ad-mOSM–induced ectopic lymphoid follicles was similar in WT and IL-6^−/− mice. The total area covered by ectopic lymphoid follicles in the lungs of Ad-mOSM–treated IL-6^−/− mice was actually larger (Fig. 5D).

Given that Ad-mOSM–mediated iBALT formation and B and T cells were activated in an IL-6–independent fashion, we next assessed whether IL-6 played a role in OSM-induced eosinophil accumulation and chemokine expression. To test this possibility, we collected bronchoalveolar lavage (BAL) from Ad-mOSM– or Ad-del70–treated WT and IL-6^−/− mice at day 7 postinfection and enumerated inflammatory cells in cytospin preparations. As shown in Fig. 6, numbers of macrophages, lymphocytes, neutrophils, and eosinophils were significantly increased in Ad-mOSM–treated mice, compared with those in Ad-del70–treated and uninfected animals. However, the numbers of neutrophils, eosinophils, and lymphocytes, but not macrophages, were significantly decreased in the BAL of Ad-mOSM–treated IL-6^−/− mice compared with those in WT mice. Decreases were also observed in the percentages of these cell types in the BAL fluid (Supplemental Fig. 2). Thus, IL-6 deficiency attenuates recruitment of innate inflammatory cells (eosinophils, neutrophils) to the airway alveolar spaces in response to OSM.

Impaired recruitment of inflammatory cells to the airways of IL-6^−/− mice suggested that IL-6 may be controlling the local expression of inflammatory chemokines. To test this possibility, we next quantified the concentration of eotaxin-2, MCP-1, and KC as well as IL-6 in BAL fluid (Fig. 6B). Consistent with changes observed in the total number of eosinophils, macrophages, and neutrophils, we found that the levels of eotaxin-2, MCP-1, and KC were elevated in BAL fluid of Ad-mOSM–treated WT mice, compared with Ad-mOSM–treated IL-6^−/− mice. Although IL-6 was induced by Ad-mOSM in the BAL fluid of WT mice, it was not detectable in the BAL fluid of IL-6^−/− mice. These data clearly suggest that IL-6 is selectively involved in OSM-mediated inflammatory cell trafficking in the airways, but not in OSM-induced iBALT formation.

Th cells and their cytokine products are critical for the generation of protective Ab responses. Because we previously observed that mOSM triggered the production of Th2-associated cytokines (25), we assessed whether Th2 cytokine expression induced by OSM was dependent on IL-6. Protein levels in BAL fluid from control or Ad-mOSM–infected WT or IL-6^−/− mice were examined 7 d after virus inoculation. As shown in Fig. 7A, Ad-mOSM induced a significant increase in the Th2 cytokines IL-4, IL-5, and IL-10, as well as in the Th1 cytokines TNF-α and IL-12. In contrast, OSM did not significantly increase IFN-γ protein levels in BAL fluid (data not shown). With the exception of IL-4, all the cytokines tested were produced in an IL-6–dependent manner.

Expression of homeostatic chemokines (HC), including CXCL13, CCL19, and CCL21, is generally associated with the formation and organization of iBALT structures (29). Moreover, the expression of CCL20 is associated with the formation of mucosal lymphoid tissues and the recruitment of DCs (30). Thus, we assessed the expression of HC in lungs of WT and IL-6^−/− mice treated with the adenoviral vectors. We observed the B cell chemokine CXCL13 and DC chemokine CCL20 were elevated in lungs from Ad-mOSM–treated but not Addel70-treated WT mice (Fig. 7B). CCL19 levels were either at or below the limit of detection of the ELISA (data not shown). Finally, CCL21 protein levels remained unchanged or decreased in response to OSM (Fig. 7B). Production of CXCL13 and CCL20, induced by Ad-mOSM inoculation, was markedly reduced in IL-6^−/− mice, suggesting that other chemokines or cytokines may participate in the recruitment and organization of iBALT structures in the absence of IL-6.

Because OSM-induced IL-4 upregulation was independent of IL-6 production (Fig. 7), we assessed whether the IL-4/IL-13 signaling molecule STAT6 was required for Ad-mOSM–induced iBALT formation (Fig. 8). Pictures of H&E-stained lung tissue (top panels) show the presence of lymphocytic aggregates in STAT6^−/− mice in response to Ad-mOSM administration. These ectopic lymphoid structures contained B220⁺ cells, some of which were actively proliferating (PCNA⁺B220⁺), as well as FDC in the center of the B cell follicles (middle panels, lower panels, a higher magnification of the same images). Morphometric analysis of ectopic lymphoid follicles shows a considerable increase in the number of lymphoid follicles in both WT and STAT6^−/− mouse lungs in response to Ad-mOSM, but not Addel70 administration (Fig. 8B). Collectively, these data (Figs. 5, 8) indicate that OSM-induced iBALT formation occurs independently of both IL-6 and STAT6.

In this study, we showed that transient, transgenic expression of OSM in mouse lungs using adenovirus vector elicited marked accumulation of activated B cells at days 7 and 14 postinfection and promoted iBALT formation, particularly in lung parenchyma. The Ad-mOSM vector induced increases in the number and activation of DC, CD4⁺, and CD8⁺ T cells, although to a lesser extent than those changes in B cells. These changes were associated with elevated levels of HC that support iBALT formation and most likely influenced production of neutralizing Ab to adenovirus. Importantly, none of these effects were seen in mice receiving control adenovirus. The results suggested that, in the context of respiratory infection, elevated levels of mOSM induce a profound B cell response and iBALT formation in C57BL/6 mice. We have also observed similar lymphocyte aggregate formation in BALB/c mice upon endotracheal administration of Ad-mOSM (data not shown), suggesting these observations are not peculiar to the C57BL/6 strain. The elevation of neutralizing anti-adenovirus Ab titer could be associated with the iBALT formation, and Ad-mOSM appears to act with an adjuvant-like effect in this scenario, boosting the response to a subsequent lung challenge (of empty vector). The neutralizing Abs are most likely directed against the penton base and/or fiber protein that are important in infection of cells by the virus. Whether Abs to other Ags or potential pathogens can be produced using this system awaits further analysis.

The function of OSM in human lung inflammation and disease is not yet clear, and our present studies suggest that OSM might be involved in iBALT formation in human disease. The BALT was first identified many years ago (31) and is considered part of the common mucosal immune system. Inducible BALT forms in response to infection or inflammation (18), and a number of human pathologies are associated with iBALT formation, including lung complications of rheumatoid arthritis (RA) (21). Interestingly, elevated OSM has been detected in RA synovial tissues and fluids (32), where there is formation of ectopic lymphoid tissues. However, it is not known whether RA patients with lung complications show elevated OSM in lung tissue. Lungs from patients with IPF display increased iBALT formation (33), coinciding with elevated OSM levels in BAL fluid (11). Another interesting observation is the formation of iBALT in lung tissue from asthmatic patients (34) and detection of OSM in BAL of patients suffering severe asthma (9). iBALT has been detected in chronic smokers (34), raising the question of whether such effects might be explained by production of OSM by activated macrophages. Although OSM has been detected in neutrophils and eosinophils, activated macrophages and T cells are the major cell sources of high amounts of OSM (as seen in RA synovial fluid). Altogether, these cells could add to the load of active OSM in inflammatory lung tissue. Therefore, our observations support the rationale for further study of the contribution of OSM in lung disease.

Previous work shows that sustained transgenic overexpression of IL-6 together with its soluble receptor in the lung induces iBALT formation primarily in subepithelial areas (17). In contrast, following Ad-mOSM administration, we primarily observed iBALT in the lung parenchyma. In Ad vector systems, encoded cytokines are typically expressed for a brief period of 4–12 d (27), and, although expression is relatively high, these transient systems allow examination of potential effects of cytokines with less concern than long-term transgenic systems. The current lack of effective systems for quantitatively measuring mouse OSM protein precludes our ability to determine OSM levels in our system or other mouse models of inflammation or disease, and this is a limitation in our study. In previous work using Ad vector but expressing GM-CSF in mouse lung, GM-CSF levels in BAL ranged from pg/ml to peak levels at day 4 (∼8 ng/ml) and then declined by day 12 (35). Analysis of OSM in human BAL samples of IPF patients was in the 100 pg/ml range (11), whereas OSM levels in induced sputum of severe asthmatic patients included values at >1 ng/ml (9). Such studies on clinical samples may or may not capture peaks of OSM detectable in human tissues in context of disease progression. Thus, we believe our study supports further investigation into the potential role of OSM overexpression in lung disease. In addition, our observations in the mouse are not restricted to Ad vector–induced OSM effects, because Mozaffarian et al. (11) have found effects in eosinophil accumulation or extracellular matrix accumulation after 10 d of administration of recombinant OSM protein (2 μg daily) that resemble the observations we have made in our model of pulmonary delivery of Ad-OSM (8, 25). Loss of function studies in future work would further clarify the role of OSM in models of lung disease.

One difference in our studies and those using constitutive long-term transgenes may be due to the timing of expression. Previous work shows that overexpression of human IL-6 (with soluble IL-6R) by Ad vectors in rat lungs leads to a transient lymphocytic alveolitis (36), whereas the overexpression of rat IL-6 by Ad vector leads to iBALT formation, primarily located in subepithelial locations (27). In contrast to this previous work, the iBALT in our Ad-mOSM system in this study was remarkably pronounced in the lung parenchyma, supporting the concept of separate mechanisms of OSM from those exerted by IL-6 overexpression.

Despite the ability of IL-6 to generate iBALT (17), we found that iBALT formation and B cell accumulation occurred in the lungs of Ad-OSM–treated IL-6^−/− mice, consistent with the IL-6–independent effect of OSM on iBALT formation. In contrast, pulmonary inflammation, including cell infiltration into bronchoalveolar spaces as well as the elevated expression of inflammatory chemokines (eotaxin-2, KC, and MCP-1) and cytokines (IL-5, IL-10, TNF-α, and IL-12), were all highly dependent on IL-6. The accumulation of eosinophils in the lung tissue, as measured by flow cytometry of lung homogenates in Figs. 2C and 4E, and in alveolar interstitium in sample histology sections (Fig. 5B), suggests that eosinophilic infiltrates occurred in the interstitial lung compartments as well as in the alveolar spaces (as measured in BAL in Fig. 6) due to Ad-mOSM in WT mice. Detection of markedly lower eosinophils in the lung homogenates at day 7 (Fig. 4E), histological sections (Fig. 5B), and BAL (Fig. 6) of IL-6^−/− mice after Ad-OSM administration suggests that IL-6 participates in recruitment of eosinophils to both the alveolar lumen and lung tissue compartments. In contrast, IL-6 deficiency had no detectable effect on the formation/organization of iBALT in the lung tissue. Thus, the IL-6^−/− deficiency was reflected in reduced inflammation in lungs due to Ad-mOSM in this system, and these results collectively demonstrate that OSM acts through IL-6–dependent and independent pathways to mediate its effects in the mouse lung.

Our previous studies show that Ad-mOSM induces Th2 cytokine production and eosinophil accumulation in the lungs (8) and that these effects are dependent on the signaling molecule STAT6 (25). Our results are consistent with previous observations (25), including upregulation of IL-4, IL-5, and IL-10, but not IFN-γ. Surprisingly, both IL-12 and TNF were produced in the lung of Ad-mOSM–treated mice, suggesting that OSM overexpression does not exclusively induce Th2 skewing in the C57BL/6 mouse lung. These results further show that IL-6 is required for optimal production of inflammatory chemokines and eosinophil accumulation in C57BL/6 mice. IL-6 deficiency did not affect IL-4 concentration in BAL (Fig. 6), suggesting that there could be activation of the STAT-6 signaling pathway. However, we observed iBALT formation in STAT6^−/− mice upon Ad-mOSM administration (Fig. 8). This indicated that STAT6 signaling, typically activated by the classic Th2 cytokine IL-4 or IL-13, is not required for iBALT formation elicited by Ad-mOSM. Collectively, the results suggest a unique function for OSM in inducing B cell activation and parenchymal iBALT formation.

HC CXCL13, CXCL19, CCL21, and CCL20 are expressed in iBALT structures and together regulate iBALT formation, organization, and function (29). In addition, mOSM upregulates C-CL21 expression, a potent DC chemoattractant (37, 38) in the skin. After examining BAL levels of HC at day 7 postinfection (Fig. 7), we observed increases in CXCL13 and CCL20 protein concentration, a reduction in CCL21 protein levels, and low/undetectable levels of CCL19. Reduced CCL21 expression has been previously observed after pulmonary infection with influenza (20). This suggests that OSM overexpression is preferentially inducing production of the B cell chemokines CXCL13 and CCL20. Unexpectedly, none of these chemokines was elevated in IL-6^−/− mice by Ad-mOSM (Fig. 7), thus suggesting the participation of nonclassical molecules on iBALT formation. Alternatively, we might have missed earlier peaks of HC expression that could have been sufficient for supporting the initial waves of iBALT formation in the IL-6^−/− animals. The precise mechanism by which mOSM supports iBALT formation in either WT or IL-6^−/− mice requires further study and might involve CD4⁺ follicular Th cells, which are known to facilitate B cell follicle expansion, through expression of IL-21 (39). Overall, our results suggest that mOSM induces iBALT formation independently of IL-6, occurs in the presence of low levels of classic HC (CCL19, CCL20, CCL21, CXCL13), and in the absence of STAT6 signaling. Collectively, these findings support a unique role for OSM in lung mucosal immunity and inflammation. OSM-mediated formation of iBALT in our system suggests that elevated production of OSM in human diseases such as IPF, RA, or asthma could explain the formation of iBALT and ectopic lymphoid structures, which are most likely linked to the perpetuation of chronic inflammation and the consequent pathological tissue damage.

We gratefully acknowledge the expert technical support of Jane Ann Smith and Rebecca Rodrigues and assistance from Sean Lauber, Jessica Guerette, and David Schnittker. We also thank Dr. Mark McDermott for critical reading of the manuscript.

The authors have no financial conflicts of interest.