Surfactant Protein-A Inhibits Mycoplasma-Induced Dendritic Cell Maturation through Regulation of HMGB-1 Cytokine Activity

Mycoplasma pneumoniae is recognized as one of the most common causes of community-acquired pneumonia and >50% of chronic stable asthmatics have evidence of airway infection with M. pneumoniae (1, 2). M. pneumoniae are atypical bacteria that form strong attachments to ciliated airway epithelial cells where they release oxidative products that can cause airway tissue damage and contribute to exacerbations in chronic asthmatics (3). Infections with M. pneumoniae may persist, with mild symptoms, for several weeks with manifestations in the upper and lower respiratory tract.

Because M. pneumoniae is primarily an extracellular pathogen that invades and resides in the respiratory tract, it has the potential to encounter pulmonary surfactant proteins that are produced by alveolar type II cells, Clara cells, and submucosal glands of the respiratory tract. Indeed, studies have demonstrated that surfactant protein A (SP-A) binds M. pneumoniae through disaturated phosphatidylglycerols and through a specific surface-binding protein, MPN372 (4, 5), and limits the growth of M. pneumoniae in vitro (5). SP-A also helps to maintain airway homeostasis and reduce hyperresponsiveness by curtailing an overly ambitious proinflammatory immune response during M. pneumoniae infection in mice in vivo (6).

Several immune functions have been ascribed to SP-A, including inhibition of T cell proliferation, augmentation of pathogen phagocytosis by acting as an opsonin, and modulation of chemotaxis and cytokine production (reviewed in Ref. 7). An additional role for SP-A was established in mediating adaptive immune responses through interactions with dendritic cells (DCs). For example, SP-A binds to DCs and negatively regulates their maturation in vitro, thereby reducing their T cell allostimulatory ability (8). The consequences of this interaction during an infection in vivo, as well as the mechanism by which SP-A modulates DC functional maturation, have not been defined. Therefore, using mice deficient in SP-A, we tested the hypothesis that SP-A regulates recruitment, activation, and maturation of adaptive immune cells in response to M. pneumoniae by regulating expression of the endogenous stress factor high-mobility group box-1 (HMGB-1), which, if released in the context of infection, can activate DCs and lead to their maturation (9).

We report in this article that M. pneumoniae infection leads to increased numbers of exudative macrophages and DCs in the lung parenchyma, a response that is augmented by the absence of SP-A. Likewise, the total number and activation state of DCs that have migrated to the mediastinal draining lymph nodes during the acute phase (3 d) of infection are also increased in the absence of SP-A. Additionally, elevated numbers of activated T and B cells in the lungs and mediastinal lymph nodes (MLNs), as well as M. pneumoniae-specific IgG in the serum, are observed in mice lacking SP-A 9 d after M. pneumoniae infection. Treatment with glycyrrhizin, a specific extracellular inhibitor of the potent proinflammatory cytokine HMGB-1, protects the DCs in the MLNs of mice lacking SP-A from M. pneumoniae-induced maturation, suggesting that SP-A inhibits M. pneumoniae-induced DC maturation by regulating HMGB-1 cytokine activity.

M. pneumoniae (catalog no. 15531) from American Type Culture Collection (Manassas, VA) was grown in SP4 broth (REMEL, Lenexa, KS) at 35°C until adherent. M. pneumoniae concentration was determined by plating serial dilutions of M. pneumoniae on pleuropneumonia-like organism agar plates (REMEL). Colonies were counted under ×10 magnification on plates after incubation for 14 d. For in vivo infection, adherent M. pneumoniae were washed by centrifuging at 6000 rpm for 5 min and resuspended in sterile saline for infection at a concentration of 1 × 10⁸ M. pneumoniae/50 μl inoculum. M. pneumoniae burden was assessed, as previously described, by plating bronchoalveolar lavage (BAL) or by RT-PCR using primers against M. pneumoniae-specific P1-adhesin gene relative to the housekeeper cyclophilin (6).

An inbred strain of SP-A–deficient mice was generated by disrupting the murine gene encoding SP-A by homologous recombination, as previously described (10). SP-A–null mice were backcrossed for 12 generations with wild-type (WT) C57BL/6 background mice from Charles River Laboratories (Wilmington, MA). WT C57BL/6 mice used as controls were purchased from Charles River Laboratories and bred in-house to account for any possible effects of environmental conditions. All mice used in experiments were age (8–12 wk) and sex (males) matched. Protocols were approved by the Institutional Animal Care and Use Committee at Duke University.

Mice 8–12 wk of age were anesthetized via i.p. injections of a 12% ketamine (100 mg/ml) and 5% xylazine (20 mg/ml) mix (10 μl/g body weight). Mice were infected with 50 μl sterile saline or 50 μl ∼1 × 10⁸ M. pneumoniae units in sterile saline by intranasal instillation. Some groups of mice received i.p injections of glycyrrhizin (10 mg/kg body weight) 2 h prior to M. pneumoniae infection and 24 h postinfection, to neutralize HMGB-1 cytokine activity, as previously described (11).

The lungs of mice were perfused with 10 ml PBS and then lavaged with PBS containing 0.1 mM EDTA (warmed to 37°C). Analysis of cytokines and chemokines present in the cell-free BAL of infected and uninfected mice was carried out by multiplex cytokine analysis (Luminex technology; Invitrogen, Carlsbad, CA). Lungs were removed, minced with a razor blade, and resuspended in 5 ml HBSS (containing calcium and magnesium) with 1.0 mg/ml collagenase A and 0.2 mg/ml DNase I. The cell suspensions were incubated for 1 h with shaking (200 rpm) at 37°C for enzymatic digestion to occur. Lung digests were then filtered through 40-μm strainers. Remaining RBCs in the digests or BAL were lysed with Gey’s lysis solution (0.83% NH₄Cl, 0.1% KHCO₃). The suspended cells were layered on top of a 4.0% solution of iodixanol (Optiprep; Axis-Shield, Norton, MA), placed above a 14.5% iodixanol solution, and centrifuged at 600 × g (no brake) for 20 min at room temperature. Low-density cells were isolated from the 4–14.5% interface and used for FACS analysis. Cells were tristained with allophycocyanin-labeled anti-CD11c, FITC- (or PE-) labeled anti-MHC class II, and PE-anti–CD80 or PE-anti–CD86.

MLNs collected from each mouse were placed in individual wells of a six-well plate containing 5 ml PBS with 5% FCS, 1.0 mg/ml collagenase A, and 0.2 mg/ml DNase I. Lymph nodes were minced using a scalpel and incubated at 37°C for ∼35 min. Digestion was stopped by adding 1 ml 120 mM EDTA in PBS and incubating for an additional 5 min at room temperature. Lymph node digests were pushed through a 40-μm strainer to obtain single-cell suspensions. Some lymph node digests were centrifuged through a density gradient (Nycodenz; Axis-Shield) to enrich for DCs. RBCs were lysed with Gey’s solution (0.83% NH₄Cl, 0.1% KHCO₃). Cells were then resuspended in HBSS with 5% FCS, 2 mM EDTA, and 100 U/ml penicillin-streptomycin and stained for FACS analysis. Cells were tristained with allophycocyanin-labeled anti-CD11c, FITC- labeled anti-MHC class II, PE-labeled anti-CD86, and PE-Texas Red–labeled anti-CD80.

Flow cytometry was performed in the Duke Human Vaccine Institute Flow Cytometry Core Facility, which is supported by the National Institutes of Health Award AI-51445. Initially, cells were examined by forward scatter versus side scatter to separate those smaller nongranular lymphoid populations from the larger granular myeloid populations. The percentage of cells within each of these gates was used to calculate the total number of lymphoid cells and myeloid cells examined from the total number of cells obtained from the lung digests, as counted on a hemacytometer. Based on similar gating strategies published by Pecora et al. (12) and Lin et al. (13), alveolar macrophages were defined as CD11b^neg-low/CD11c^mid/MHCII^mid, DCs were defined as CD11b^high/CD11c^high/MHCII^high, and inflammatory lung macrophages (also known as exudate macrophages) were defined as CD11b^high/CD11c^neg-mid/MHCII^mid.

Bone marrow-derived DCs (BMDCs) were generated as described previously by Inaba et al. (14) and as modified by Brinker et al. (15). Briefly, marrow from the tibia, femur, and humerus of mice was harvested and cultured in RPMI 1640 supplemented with 5% FCS, antibiotics, and 50 μM 2-ME plus 5% GM-CSF–conditioned media for 6 d. Loosely attached cells were harvested and negatively selected with biotinylated Gr-1 Abs (BD Pharmingen, San Diego, CA) and streptavidin paramagnetic microbeads (Miltenyi Biotec, Auburn, CA). DCs were matured for 24 h in 24-well plates at 5 × 10⁵ cells/ml in the presence of 100 ng/ml LPS (serotype 055:B5), as previously described (8).

Undifferentiated THP-1 cells (American Type Culture Collection) were maintained in RPMI 1640 with supplements (10% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, and streptomycin-penicillin) at 37°C at 5% CO₂. Clonetics normal human bronchial epithelial cells (NHBEs) were purchased from Lonza (Basel, Switzerland) and grown to ∼80% confluence in bronchial epithelial growth media (Clonetics) with supplements and growth factors (bovine pituitary extract, hydrocortisone, human epidermal growth factor, epinephrine, insulin, triiodothyronine, transferrin, gentamicin/amphotericin-B, and retinoic acid), at which point cells were placed in bronchial epithelial basal medium (no supplements) for stimulation experiments. THP-1 cells were resuspended in RPMI 1640 (no supplements). DCs, NHBEs, and THP-1 cells were resuspended in the respective medium in the presence or absence of purified human SP-A (50 μg/ml), which was purified as described (16), and ∼5 × 10⁵ cells were seeded per well into 12-well plates. After 2 h, M. pneumoniae was added at a concentration of 10 M. pneumoniae CFU/cell (5 × 10⁶/well) to some wells, whereas others received saline as a control buffer for the M. pneumoniae. Some sample wells received neutralizing anti-TLR2 Ab (No. 16-9024; eBioscience, San Diego, CA) or isotype control prior to stimulation. After M. pneumoniae stimulation for ∼16 h, the cell-free supernatant was collected, and pelleted cells were lysed for Western blot analysis.

Cell-free lavages from uninfected and M. pneumoniae-infected WT and SP-A^−/− mice; stimulated DC, THP-1, or NHBE-free supernatants; and DC, THP-1, or NHBE lysates were denatured by heating at 95°C for 5 min in buffer containing DTT prior to gel electrophoresis on 12.5% polyacrylamide gels (Protogel, National Diagnostics, Atlanta, GA). For cell-free BAL analysis, 25 μl from 1-ml lavages was loaded for each sample. For the DC, THP-1, and NHBE-free supernatants, ∼25 μl from a total well volume of 250 μl was loaded for each sample, and 25 μl was loaded from 250 μl of the cell lysates. Gels were transferred onto nitrocellulose membrane, blocked with 5% block (milk in TBS), and labeled with an HMGB-1 Ab (ab12029 or ab65003; Abcam, Cambridge, MA) overnight at 1:1500. The secondary Ab, HRP-conjugated anti-mouse IgG or anti-rabbit IgG (Cell Signaling Technology, Beverly, MA), was used at 1:10,000 for 1 h. The SP-A Ab used was rabbit anti-sheep SP-A (cross-reacts with mouse SP-A), which was used with an anti-rabbit HRP secondary Ab. GAPDH Ab (MAB374) was purchased from Millipore (Bedford, MA). SuperSignal West Femto (Thermo Scientific, Rockford, IL) was used to visualize protein expression.

Cytokines and chemokines were analyzed by Luminex technologies using a mouse 20-plex kit (Invitrogen); GM-CSF was analyzed by ELISA (eBioscience), according to the manufacturer’s instructions. M. pneumoniae-specific IgG was analyzed by ELISA, as described previously (17). Briefly, sonicated heat-killed M. pneumoniae dissolved in coating buffer (0.05 M carbonate-bicarbonate buffer [pH 10]) was used at a concentration of 10 μg/ml and was incubated overnight at 4°C. Wells were then blocked for 1 h with 2% BSA/PBS/0.05% Tween-20 mix, after which the serum samples were diluted 1:10,000 in 1% BSA/PBS buffer and added from uninfected and M. pneumoniae-infected mice. A biotin-goat anti-mouse IgG Ab (Jackson ImmunoResearch Laboratories, West Grove, PA) and streptavidin-HRP (BD Pharmingen) were used for the detection. Samples were analyzed at 450 nm absorbance on the FLUOstar Optima plate reader (BMG Labtech, Offenburg, Germany).

Images of the lung were taken on a Nikon Eclipse 50i light microscope at ×10 (aperture 0.3) or ×40 (aperture 0.75) at room temperature by digital photography of bright-field images using a Nikon Infinity 2 camera (Nikon, Melville, NY). Infinity Capture software was used for image acquisition, and Adobe Photoshop (Adobe Systems, San Jose, CA) was used for color-contrast settings and figure presentation.

All data measurements were analyzed with Prism software (GraphPad, San Diego, CA), first to determine whether data were distributed normally, followed by the t test to determine significance. Data sets with significant variance between comparison groups were analyzed by the t test using the Welch correction, as assessed in Prism.

To assess the role of SP-A in regulating immune cell responses in vivo, the cells present in the lungs due to M. pneumoniae infection and the resulting inflammation were analyzed in WT and SP-A–null mice. Although the total numbers of myeloid cells present in the lung digests were not significantly different in the infected SP-A–null mice compared with infected WT mice, the cellular composition, which was analyzed by flow cytometry, showed differences in certain cell populations. Inflammatory monocytes and PMNs increased in number in response to M. pneumoniae challenge, although levels were equivalent in the lungs of WT and SP-A–null mice (data not shown). Among the other cell populations, DCs were examined from the lung digests of infected and uninfected mice. Lung DCs were defined based on their pattern of side scatter versus forward scatter and high levels of expression of CD11c and MHC class II cell surface markers (12, 18). Very few DCs (<1% of total cells) were present in the lung digests of the saline-treated mice (Fig. 1A). After M. pneumoniae infection, the number of DCs observed in WT mice was only slightly elevated over those observed in the saline-treated mice. However, the number of cells observed within the DC gate was significantly increased in the SP-A–null mice (Fig. 1A).

The numbers of macrophages (MHCII^intCD11c⁺) isolated from lungs of saline-treated mice were not significantly different between the WT control and the SP-A–null mice. However, 72 h after M. pneumoniae infection, there were significantly more macrophages (including resident and inflammatory) in both groups of infected mice compared with saline-treated animals in each group (Fig. 1B). Although the number of macrophages increased in the infected WT mice by 2-fold, the number of macrophages in the lungs of the infected SP-A–null mice increased almost 4-fold over their saline controls and was significantly greater in comparison with the infected WT mice (p < 0.01).

Expression of the cell-surface marker CD11b was also assessed on macrophage populations in the M. pneumoniae-infected mice to differentiate inflammatory newly recruited macrophages (CD11b⁺), also known as exudate macrophages, which had migrated into the lung during pulmonary infection from the resident alveolar macrophages (CD11b⁻) (13, 19). Similar to the alveolar macrophages analyzed in the BAL fluid of saline-treated mice, very few of the tissue macrophages expressed CD11b in the saline-treated mice, suggesting that those were primarily resident macrophages (Fig. 1B). However, CD11b expression was significantly elevated on macrophages in the WT and SP-A–null mice after M. pneumoniae infection, suggesting that these were predominantly newly recruited or differentiated inflammatory macrophages. Additionally, there were significantly more CD11b⁺ macrophages in M. pneumoniae-infected mice deficient in SP-A compared with infected WT mice, suggesting that SP-A is important in inhibiting the influx of these inflammatory macrophages into the lung during M. pneumoniae infection.

The function of the pulmonary DC network is altered dramatically during inflammatory conditions as the result of signals produced by pulmonary epithelium and myofibroblasts of the airways that attract immature DCs from the bloodstream into areas of pulmonary challenge (20). To determine whether any such chemokine or cytokine mediators were differentially expressed in response to M. pneumoniae infection in the absence of SP-A, BAL fluid was harvested after 3 d of infection and analyzed. As shown in Table I, neither MCP-1 nor MIP-1α, factors known to be chemotactic for immature DCs (21, 22), was detectable in BAL fluid of saline-treated mice. However, in BAL from M. pneumoniae-infected animals lacking SP-A, MCP-1 was detected at 15.6 pg/ml (not detectable in infected WT mice), and MIP-1α was detected at 21.9 pg/ml (compared with 9.7 pg/ml in infected WT mice). The amount of GM-CSF, a factor known to enhance the differentiation of monocytes into immunostimulatory DCs in the lung vasculature, was increased in BAL from WT and SP-A–null M. pneumoniae-infected mice compared with the untreated control sample.

Analysis of cell-surface markers from the lymphoid panel revealed a dramatic influx of CD3⁺ lymphocytes (mean, 1.6 × 10⁶ per total lung cells) into the lung tissue of SP-A^−/− mice after 3 d of M. pneumoniae infection compared with their respective saline controls (p < 0.001) or infected WT mice (p < 0.001) (Fig. 2A). Interestingly, the number of CD3⁺ lymphocytes present in the M. pneumoniae-infected WT mice at day three (mean = 0.7 × 10⁶) was not significantly elevated over noninfected mice (mean = 0.7 × 10⁶). T lymphocytes were also analyzed 9 d postinfection to determine whether the inflammatory response was diminished, maintained, or augmented over levels detected after 3 d of infection. Despite no significant T cell infiltration in WT mice 3 d postinfection, they did present with significantly enhanced numbers of CD3⁺ T lymphocytes in the lungs 9 d postinfection compared with noninfected controls (Fig. 2A). Additionally, CD3⁺ T cells detected in the lungs of SP-A–null mice after 9 d of infection were significantly increased over the numbers present at 3 d of infection in mice lacking SP-A (p < 0.001), as well as over the numbers present in the WT mice after 9 d of infection (p < 0.001). Interestingly, the T cell chemoattractant monokine induce by IFN-γ was also significantly increased in M. pneumoniae-infected mice lacking SP-A compared with infected WT mice (Table I).

Those cells expressing CD3 were further analyzed by flow cytometry for their surface expression of CD4 and CD8. M. pneumoniae-infected SP-A^−/− mice had significantly elevated numbers of CD3⁺CD4⁺ and CD3⁺CD8⁺ lymphocytes compared with M. pneumoniae-infected WT mice (Fig. 2B). The cell-surface expression of CD25, which plays dual roles in lymphocyte differentiation and activation/proliferation, on CD4⁺ T cells was also significantly greater in the SP-A^−/− M. pneumoniae-infected mice compared with infected control mice (data not shown).

To determine the activation status of those T lymphocytes present in the lung after 9 d of infection, we analyzed CD69 expression, a very early activation Ag, on CD3⁺CD4⁺ and CD3⁺CD8⁺ T cells by flow cytometry. CD69 was not upregulated in WT mice 9 d postinfection on CD4⁺ or CD8⁺ populations. However, there was significant expression of CD69 on CD4⁺ and CD8⁺ cells in the SP-A–null mice compared with infected WT mice and saline controls (Fig. 2C). In support of these findings, expression of IL-12, which is involved in the differentiation of T cells and is associated with T cell activation, was also significantly increased in the BAL of infected SP-A mice compared with infected WT mice (Table I).

T cells present in the MLNs after 9 d of M. pneumoniae infection were also examined. There were no appreciable increases in the total number of CD4⁺ or CD8⁺ T cells recovered in MLNs from the infected WT mice. However, in M. pneumoniae-infected SP-A–deficient mice, there were significantly more CD4⁺ and CD8⁺ T cells compared with infected WT mice (Fig. 2E). The state of T cell activation was determined by the presence of CD25 and CD69 on the cell surface. The numbers of CD8⁺ activated cells were quite low in all groups of mice examined (Fig. 2F). However, as shown in Fig. 2F, the numbers of CD4⁺ T cells expressing these activation markers were significantly increased in the MLNs of infected SP-A–deficient mice compared with infected WT mice and saline control mice.

Mycoplasma burden in SP-A–deficient mice was previously described in our studies after 3 d of infection, which showed that although there was no difference in M. pneumoniae burden from BAL fluid at day 3 of infection, significantly more M. pneumoniae were colonizing the airway by binding to epithelial cells (6). Although WT and SP-A–deficient mice were clearing M. pneumoniae by day 9, interestingly, only 20% of WT mice tested positive for M. pneumoniae in the BAL compared with >80% of the SP-A–deficient mice (Fig. 3A). Likewise, M. pneumoniae burden in association with lung tissue, as determined by RT-PCR, was also significantly increased in SP-A–deficient mice at day 9 of infection (Fig. 3B). Although some mycoplasmas are believed to be mitogenic for T cells, we did not find a correlation between M. pneumoniae burden and T cell number in our system (Fig. 3C).

Cells that expressed the cell surface marker B220 were also significantly increased in mice lacking SP-A, including the saline-treated and infected mice, at 3 and 9 d postinfection (Fig. 4A). The number of B cells did not increase significantly in WT mice during the infection. However, B cells in the lungs of SP-A–null mice were significantly increased over their saline-treated controls and at each time point examined compared with infected WT controls (Fig. 4A). Although there were significantly more B220⁺ cells in saline-treated SP-A–null mice, the total number of B220⁺ cells considered to be activated, as determined by coexpression of IgM and CD69, was equivalent in the saline-treated groups (Fig. 4B). There also was no increase in the total number of activated cells (B220⁺IgM⁺CD69⁺) isolated from M. pneumoniae-infected WT mice after 9 d of infection compared with their saline controls. In striking contrast, M. pneumoniae-infected SP-A–null mice had significantly greater numbers of activated cells isolated from the lungs after 9 d of infection compared with their saline controls and infected WT mice at the same day of infection (Fig. 4B).

B cells were also analyzed from the MLNs of infected mice after 9 d of infection. The total numbers of B220⁺ cells harvested from the MLNs were unchanged in infected WT mice compared with saline-treated WT mice. However, significantly more B220⁺ cells were collected from infected SP-A–deficient mice compared with saline-treated control mice and infected WT mice (Fig. 4C). Additionally, the number of B220⁺ activated cells was significantly increased in infected WT and SP-A–deficient mice (Fig. 4D). However, the total number of activated B cells was significantly enhanced in infected SP-A–deficient mice compared with infected WT mice.

Because the M. pneumoniae-infected SP-A–deficient mice displayed a strikingly enhanced B cell response compared with infected WT mice, we sought to determine whether M. pneumoniae-specific IgG Abs were produced by day 9 of the pulmonary infection. Serum was collected from WT and SP-A–null mice that were infected with M. pneumoniae, and M. pneumoniae-specific IgG was determined by ELISA. The amount of M. pneumoniae-specific IgG is expressed as the fold increase over the baseline levels measured from the serum of noninfected mice. As shown in Fig. 4E, M. pneumoniae-specific IgG levels were significantly elevated in the serum of infected SP-A–null mice compared with infected WT mice.

Given that increased numbers of activated CD4⁺ T cells were discovered in the MLNs of infected SP-A–deficient mice at the later time point (day 9), we examined the activation state of DCs in the MLNs at an earlier time point (day 3), because their early activation state could directly influence the conversion of naive T cells into activated T cells at the later time point. WT and SP-A–null mice were instilled with M. pneumoniae or saline, and the MLNs were harvested after 3 d. Lymph nodes were enzymatically digested to aid in the release of DCs, a single-cell suspension was obtained, and the cells were stained and analyzed by flow cytometry. There are at least two populations of CD11c⁺ cells that can be differentiated in the lymph nodes. The more mature population expressing higher levels of MHC class II and CD86 are predicted to be the DCs that have migrated from the lungs (23, 24). M. pneumoniae-infected SP-A–null mice had a greater percentage of MHCII^hiCD11c⁺ cells in the MLNs than infected WT mice (data not shown). The percentage of cells that were MHCII^hiCD11c⁺, which was determined by gating the flow-cytometry measurements, was used to calculate the total number of MHCII^hiCD11c⁺ cells based on hemacytometer counts from individual lymph node preparations. As shown in Fig. 5A, the total number of MHCII^hi CD11c⁺ cells was also significantly greater in infected mice lacking SP-A, suggesting that, in the absence of SP-A, more DCs in the infected pulmonary environment migrate into the draining lymph nodes. The number of DCs in the MLNs and lungs at day 9 were back to levels observed in saline-treated animals; likewise, the lymphocytic inflammation had subsided by day 15 in WT and SP-A^−/− mice (data not shown). Expression of the cell-surface markers associated with DC maturation, CD80 and CD86, were also analyzed. Because the total number of MHCII^hiCD11c⁺ cells was higher in the infected SP-A–deficient mice, CD80 and CD86 levels are shown as the fold increase in expression within the DC populations. CD80 and CD86 cell-surface expression was significantly greater on the MHCII^hi CD11c⁺ cells of the SP-A–null mice compared with cells harvested from the infected WT mice (Fig. 5B).

Previous studies from our laboratory demonstrated a role for SP-A in inhibiting LPS-induced DC maturation in vitro (8). Our in vivo studies examining DC maturation in M. pneumoniae–infected SP-A–null mice compared with control WT mice also supported a role for SP-A in limiting DC maturation in response to M. pneumoniae. To determine whether SP-A directly carries out this protective role in response to a clinically relevant pulmonary pathogen, DC maturation in response to M. pneumoniae was examined in the presence or absence of SP-A. BMDCs were plated, and some samples were preincubated with SP-A, after which live M. pneumoniae were added to the DCs, and M. pneumoniae-induced maturation in the presence or absence of SP-A was determined after 24 h of culture by flow cytometry. MHCII^hiCD11c⁺ cells were analyzed for expression of the maturation marker CD86.

As described previously, addition of SP-A to nonstimulated DCs leads to decreased expression of maturation markers compared with media alone. These results were repeated as controls and are in agreement with published reports (8) (Fig. 5C). Addition of M. pneumoniae to DCs resulted in increased expression of CD86 compared with those DCs receiving media alone or those preincubated with SP-A alone (no M. pneumoniae stimulation). Interestingly, those samples that had been preincubated with SP-A prior to stimulation with M. pneumoniae were significantly protected from the upregulation of the maturation marker CD86 compared with the M. pneumoniae-stimulated DCs that were treated with vehicle only (Fig. 5C).

HMGB-1, although most commonly associated with necrotic cells and used as a marker of tissue damage, has more recently been identified as a potent proinflammatory cytokine released by activated macrophages and monocytes in the lung during acute inflammation (9, 25, 26). Because HMGB-1 cytokine activity was shown to be integral for DC maturation and is necessary for Ag presentation leading to the activation of T cells (27), we sought to determine whether levels of HMGB-1 were elevated in M. pneumoniae-infected SP-A–null mice. BAL fluid from uninfected and M. pneumoniae-infected mice was collected, and the presence of HMGB-1 was examined by Western blot analysis. Viability of the cells recovered in BAL was assessed using trypan blue and microscopy; no significant differences in cell death were noted between groups. As shown in Fig. 6A, levels of HMGB-1 in BAL of WT and SP-A–null mice were similar in the uninfected mice. However, levels of HMGB-1, albeit increased in infected WT mice over uninfected controls, were dramatically increased in infected mice lacking SP-A (Fig. 6A).

To determine whether the heightened HMGB-1 secretion and its potential cytokine activity observed in the absence of SP-A could be influencing the maturation of DCs in the M. pneumoniae-infected mice, we used glycyrrhizin, a direct inhibitor of HMGB-1 cytokine activity (28). Cells from the MLNs were collected and analyzed from mice that were infected with M. pneumoniae and treated with vehicle (saline) and were compared with those that were infected with M. pneumoniae and treated with glycyrrhizin. The total number of cells collected from the MLNs, which were enriched for DCs by gradient centrifugation, of uninfected WT and SP-A–null mice were not significantly different. However, the total number of cells collected from the MLNs of the WT and SP-A–deficient M. pneumoniae-infected mice were elevated over their uninfected controls (data not shown). When the DCs (MHCII^hiCD11c⁺) isolated from the MLNs were analyzed for the presence of the maturation marker CD86, again we detected significantly more cells expressing CD86 in the infected mice lacking SP-A that were treated with vehicle compared with WT mice (Fig. 6B). In contrast, SP-A–null mice that were treated with the HMGB-1 inhibitor, glycyrrhizin, had significantly fewer DCs expressing CD86 compared with those receiving vehicle (Fig. 6B), indicating that SP-A can inhibit M. pneumoniae-induced DC maturation by modulating HMGB-1 cytokine activities.

To determine whether exogenously added SP-A inhibits the release of HMGB-1 from M. pneumoniae-activated cells into the culture supernatant, additional experiments were carried out in vitro using a human acute monocytic cell line (THP-1) as well as NHBEs. The levels of HMGB-1 in the supernatant from THP-1 cells treated with saline or SP-A alone were below the level of detection by Western blot analysis. However, when cells were infected with 10 M. pneumoniae CFU per cell, secreted HMGB-1 was readily detected in the supernatants after 16 h of stimulation (Fig. 7A). Incubation of the cells prior to infection with exogenous human SP-A (50 μg/ml) inhibited secretion of HMGB-1 into the culture supernatant. Importantly, cellular viability (>85%) was not significantly altered in stimulated conditions, as determined by trypan blue exclusion and by lactate dehydrogenase assays (Supplemental Fig. 1).

Although monocytes are the most likely source of the secreted HMGB-1 present in the inflamed pulmonary environment, human epithelial cells might also produce HMGB-1 upon stimulation. Therefore, we also examined the ability of SP-A to regulate HMGB-1 release from M. pneumoniae-stimulated NHBEs. Similar to observations with the THP-1 cells, levels of HMGB-1 were also undetectable in samples with the addition of saline or SP-A alone. However, stimulation of cells with 10 M. pneumoniae CFU per cell also induced HMGB-1 secretion from NHBEs, whereas the addition of exogenous SP-A attenuated this response in M. pneumoniae-stimulated samples (Fig. 7B).

Although SP-A could be acting directly on DCs to inhibit M. pneumoniae-induced maturation, it could also be functioning by binding to M. pneumoniae, which is known to occur through a binding protein (MPN372) and surface lipids (4, 5), and, thereby, protecting the cells from interactions with M. pneumoniae. SP-A inhibited DC maturation when immature (Fig. 8A) or mature (Fig. 8B) DCs were preincubated with human SP-A. To determine which aspect of these DC–SP-A–M. pneumoniae interactions was key to limiting DC maturation, we preincubated DCs with SP-A and washed away unbound SP-A prior to the addition of M. pneumoniae. In parallel, we precoated M. pneumoniae with SP-A and washed away unbound SP-A prior to adding it to DCs. Interestingly, preincubation of DCs with SP-A prior to stimulation did not lead to reduced levels of HMGB-1 secretion, in fact, HMGB-1 levels were still quite high (Fig. 8C). However, in the parallel experiment, M. pneumoniae that had been precoated with SP-A was not able to elicit HMGB-1 secretion from DCs in contrast with M. pneumoniae that was not coated with SP-A (Fig. 8C).

Because M. pneumoniae is known to work almost exclusively through TLR2, additional experiments were conducted to determine whether this receptor was a key receptor in M. pneumoniae-induced HMGB-1 secretion from these cells. A TLR2-neutralizing Ab or an isotype control was incubated with DCs prior to stimulation with M. pneumoniae. Interestingly, inhibition of TLR2 resulted in no detectable HMGB-1 secretion upon M. pneumoniae stimulation (Fig. 8D).

Our findings describe a novel role for SP-A in limiting M. pneumoniae -induced DC maturation via inhibition of HMGB-1 cytokine activity. The absence of SP-A in mice during M. pneumoniae infection leads to increased numbers, as well as the activation state, of APCs in the lung and draining lymph nodes during the acute phase of infection and, consequently, increased numbers of activated T and B cells later during the course of infection. These findings are consistent with reports that describe SP-A as a vital component of the pulmonary innate immune system that limits inflammation and inhibits LPS-induced DC maturation in vitro. In contrast, studies have only recently begun to focus on the extracellular cytokine activity of HMGB-1 as an important mediator of inflammation that is actively secreted from stimulated myeloid cells and induces DC activation. Although no association between HMGB-1 and SP-A has been described, several phenotypic parameters measured in M. pneumoniae-infected SP-A–null mice suggested that SP-A may play a role in regulating HMGB-1 extracellular cytokine activity. Using a specific inhibitor of HMGB-1 cytokine activity (glycyrrhizin), our research describes a role for SP-A in regulating HMGB-1 activity during M. pneumoniae pulmonary infection.

M. pneumoniae infection of mice lacking SP-A induced a dramatic and significant increase in the total number of CD3⁺ T cells present in the lung and MLNs at the early time point (3 d) and the later time point (9 d) compared with infected WT mice. M. pneumoniae likely is not acting as a mitogen in this response because the levels of M. pneumoniae detected at 9 d of infection is significantly decreased in the lungs of WT and SP-A–null mice compared with the burden at 3 d, whereas the numbers of T cells are continuing to increase (data not shown). Cell-surface markers on T cells indicative of activation were also significantly increased in the lung and MLNs of infected mice lacking SP-A. Additionally, compared with infected WT mice, SP-A–null M. pneumoniae–infected mice had ∼4-fold more BAL IL-12, which is involved in T cell differentiation.

Although one of the many known roles of SP-A is to inhibit T cell proliferation and attenuate the initial Ca²⁺ spike (29), an additional likely explanation for the increase in activated T cells in infected SP-A–null mice can be ascribed to the role of SP-A in DC maturation. Previous work from our laboratory determined that SP-A inhibits basal and LPS-induced DC maturation in vitro. Therefore, it was important to determine whether this role would be carried out in vivo in response to a respiratory pathogen, such as M. pneumoniae. More DCs were found in the lung of M. pneumoniae-infected mice that lacked SP-A, and more functionally mature DCs were detected in the MLNs, where they would act as potent T cell stimulators. Additionally, MCP and MIP-1, factors known to be chemotactic for immature DCs (21, 22), were present in BAL fluid of infected mice at much higher levels in mice lacking SP-A, suggesting that more DCs migrated into the lung from the bloodstream of M. pneumoniae-infected mice in the absence of SP-A as the result of greater chemokine production. The amount of GM-CSF, a factor vital to enhancing the differentiation of monocytes into immunostimulatory DCs in the lung vasculature, was significantly increased in BAL from WT and SP-A–null M. pneumoniae-infected mice. Collectively, these findings suggest that, in the absence of SP-A, more DCs infiltrate the airways as the result of increases in chemoattractant signals; however, because GM-CSF is increased comparably in infected WT and SP-A–null mice, it is not likely that the increase in GM-CSF accounts for the difference in DC maturation observed.

Because differential increases in GM-CSF in SP-A–null mice were not observed, we investigated the possibility that other mediators may be responsible for the SP-A–mediated inhibition of DC maturation. Recent studies suggested that endogenous stress factors, such as HMGB-1, released during infection can aid in maturing DCs that will then further contribute to the initiation and maintenance of an immune response against an invading pathogen (9). We thought that HMGB-1 was a good candidate for mediating the SP-A–dependent modulation of DCs and T cells because HMGB-1 can be actively secreted from activated monocytes and macrophages, is actively involved in regulating the maturation of DCs (30), and is thought to be necessary for the proliferation and polarization of naive CD4⁺ T cells (27). Indeed, we found that the level of HMGB-1 was dramatically increased in the BAL of M. pneumoniae-infected mice when SP-A was absent. There was no indication of increased cell death in these mice, as determined by cell staining from the BAL and lung tissue, suggesting that increased HMGB-1 was released from activated, but not necrotic, airway monocytes or macrophages.

To determine whether the increased HMGB-1 present in the lungs after M. pneumoniae infection was responsible for the increased DC maturation observed in the absence of SP-A, we used an inhibitor of HMGB-1 cytokine activity, glycyrrhizin. Glycyrrhizin, a product produced by the licorice plant, has been shown to bind HMGB-1 directly and block its extracellular functions (28, 31). Taking advantage of the ability of glycyrrhizin to functionally inhibit HMGB-1 cytokine activity, we were able to examine the maturation state of those DCs that had migrated to the MLNs following M. pneumoniae infection and determine whether HMGB-1 was a key mediator in this process. In mice lacking SP-A, more mature DCs were again observed in the M. pneumoniae-infected vehicle-treated mice compared with infected vehicle-treated infected WT mice. However, in mice lacking SP-A treated with glycyrrhizin, the level of DC maturation of migrated cells was significantly reduced to those levels measured in the infected WT mice. These findings suggest that SP-A inhibits DC maturation in vivo in response to M. pneumoniae infection, at least in part by limiting HMGB-1 extracellular cytokine activity, which can directly influence initiation of an adaptive immune response.

Although we used cell-surface markers most commonly used to distinguish macrophages from DCs, we acknowledge that phenotypic distinction of these two populations of APCs is increasingly nebulous. Although the total numbers of phenotypically activated DCs were significantly increased in M. pneumoniae-infected SP-A–null mice, the percentage of activated T lymphocytes was similar. This profile suggests that the increase in APCs may be due to increases in the pool of activated macrophages rather than differentiated DCs, the latter of which should induce activation of T cells. Therefore, more detailed studies were conducted using a purified population of BMDCs in which we found that exogenous SP-A added at physiologic levels could inhibit DC maturation induced by M. pneumoniae stimulation. However, functional assays examining Ag-specific lymphocyte proliferation should be conducted to strengthen our understanding of the inhibitory role of SP-A in DC differentiation, independent of any macrophage participation.

Additional experiments were carried out in vitro using a human monocyte cell line (THP-1) and human bronchial epithelial cells to determine whether exogenously added SP-A inhibits the release of HMGB-1 from M. pneumoniae-activated cells into the culture supernatant. The levels of HMBG-1 in the supernatant from THP-1 cells treated with saline or SP-A alone were below the level of detection by Western blot analysis. However, when cells were infected with 10 M. pneumoniae CFU per cell, secreted HMGB-1 was detected in the supernatants, as was described for human monocytes (32). Incubation of the cells prior to infection with exogenous human SP-A inhibited secretion of HMGB-1 into the supernatant in the human monocyte cells and bronchial epithelial cells. Importantly, cellular viability and membrane permeability were not significantly altered in stimulated conditions, indicating that the increased HMGB-1 was not from increased cell death or membrane leakage within those samples.

The amount of HMGB-1 detected in BAL from infected SP-A–null mice is increased compared with infected WT mice, which suggests that SP-A may regulate the amount of HMGB-1 secreted from the activated cells. Although it is possible that SP-A directly binds extracellular HMGB-1, coimmunoprecipitation experiments with BAL fluid from infected WT mice showed no detectable binding, whereas HMGB-1 and SP-A were readily observed. Further studies determined that the interaction and binding of SP-A to M. pneumoniae were vital components in limiting M. pneumoniae-induced HMGB-1 secretion from cultured DCs. Additionally, when a neutralizing Ab was used to inhibit TLR2, M. pneumoniae was unable to elicit HMGB-1 secretion from these cells. Taken together, these findings suggest that the binding of SP-A to M. pneumoniae is critical in curtailing HMGB-1 secretion from DCs by restricting the interaction of M. pneumoniae with its primary receptor, TLR2.

M. pneumoniae is known to colonize the respiratory tract where it initiates a cascade of immune-response amplification, including proliferation of lymphocytes, proinflammatory cytokine release, and production of Igs. Several studies reported increased IgG serum levels in M. pneumoniae-infected individuals (33, 34). Our findings showed that, in the absence of SP-A, more M. pneumoniae-specific IgG was present in the serum of infected mice compared with infected WT mice. The production of IgG Abs is predominantly associated with the secondary immune response. This finding further supports an indirect role for SP-A in limiting the advancement of an adaptive immune response by regulating the initiation of the Ab response to M. pneumoniae that could be a direct result of increased DC maturation and migration in the innate phase of the response.

In summary, our findings support an inhibitory role for SP-A in M. pneumoniae-induced DC maturation, which is a key step in the initiation of an immune response. Additionally, our studies showed that one mechanism by which SP-A inhibits M. pneumoniae-induced DC maturation is by regulating HMGB-1 secretion in vivo and in vitro in human cells. Previous studies showed that the receptor for advanced glycation end products (RAGE) and HMGB-1 are required for maturation of human DCs (30). Because SP-A inhibits DC maturation in the presence and absence of stimulation, further experiments testing whether SP-A interferes with HMGB-1 binding via RAGE and limiting DC maturation through this interaction may be of interest. These findings might be of value in designing therapies, because HMGB1, as well as other proinflammatory ligands, and RAGE are present in myriad acute and chronic inflammatory diseases, such as sepsis, diabetes, atherosclerosis, and renal failure (35). Additionally, chronic inflammation and inflammatory diseases in the lung often result in epithelial damage and airway remodeling, therefore, blocking HMGB-1 by pharmacological interventions or potentially with surfactant treatment may alleviate lung damage.

We thank Charles Giamberardino for technical assistance, Pamela Hesker for HMGB-1 Ab advice, and Sambuddho Mukherjee and Amy Pastva for helpful discussions.

Disclosures The authors have no financial conflicts of interest.