Cigarette smoke promotes dendritic cell accumulation in COPD; a Lung Tissue Research Consortium study

Slides of formalin-fixed, paraffin-embedded specimens were obtained from twenty-four patients - 8 controls (these control subjects would have been classified as GOLD stage 0, or at risk subjects based on the 2001 GOLD guidelines [8], but would now be classified as not having spirometric criteria of COPD based on a post-bronchodilator FEV₁/FVC ratio > 70% and % predicted FEV1 > 80; 1 subject had an FEV1 of 78% predicted but an FEV₁/FVC ratio > 70% [9]), 8 with moderate (post-bronchodilator FEV₁/FVC ratio < 70%; FEV₁ 50-80% predicted), and 8 with severe COPD (post-bronchodilator FEV₁/FVC ratio < 70%, FEV₁ < 50%) based on spirometry. All tissue samples were obtained from the National Institutes of Health funded Lung Tissue Research Consortium (LTRC: http://​www.​ltrcpublic.​com - a detailed account of the LTRC protocol used is available to the public at http://​www.​ltrcpublic.​com/​PRO_​NOV_​2009.​pdf). Asthma and other causes of obstructive lung disease were excluded in all subjects based on history, physical examination, and spirometric data. Tissue sections were de-paraffinized in 3 changes of xylene (5 minutes each), hydrated in 2 changes of 100% ethanol (10 minutes each), and subsequently hydrated in 2 changes of 95% ethanol (10 minutes each). Antigen unmasking was performed in a pressure cooker and with 10 mM sodium citrate buffer, pH 6.0. Endogenous peroxidase activity was quenched by placing slides in a 1% hydrogen peroxide solution in methanol for 30 minutes. Tissue slides were then washed (1× Dako Wash Buffer) and immunostaining performed using a commercially available VectaStain Elite ABC (Vector Labs) kit according to the manufacturer's instructions. Tissues were stained using CD1a (Dako; marker for Langerhans cells) and CD83 (Dako; marker for mature dendritic cells). Images of the stained tissue were digitally captured at a 10× magnification using a NanoZoomer system (Bacus laboratories, Olympus America). The percent area of lung tissue positively stained for CD1a and CD83 were quantitatively determined using IHCscore software (Bacus lab, Olympus America), as recently described [10]. Image capture and quantitative determination of tissue staining for CD1a and CD83 were performed by one the investigators (JL) in a blinded manner (control vs COPD). The area of tissue positively stained was estimated relative to the area of total lung tissue present on the slide to determine the extent of Langerhans cell and mature dendritic cell infiltration for each sample. Relevant demographic and physiologic lung function variables for the study population were obtained from LTRC. Correlations were performed between quantitative surface marker expression (CD1a and CD83) and cumulative tobacco exposure (pack years of cigarette smoked), as well as physiologic lung function measurements.

Quantitative real-time RT-PCR was performed to determine the relative expression of Langerhans cell restricted CD207 (langerin) gene transcripts in COPD and control tissues. For qPCR analysis, the CFX96 Real-Time PCR detection System was utilized (Bio-Rad). RNA was extracted from whole lung frozen tissue using a Qiagen RNA isolation kit. RNA was extracted only from frozen tissue specimens, and all extractions were performed in batch format in the institutional Advanced Genomics Technology Center. Samples with an RNA Integrity Number (RIN: determined by Agilent Bioanalyzer software) < 7.5 were excluded from the analysis. To make cDNA for qPCR analysis, SuperScript^® III Reverse Transcriptase was employed (Invitrogen). To quantify the final cDNA PCR products, SYBR^® Green PCR Master Mix was used. Conditions for the PCR reactions were as follows: 50°C for 2 min, 95°C for 2 min, followed by 40 cycles of 95°C for 15 sec, 60°C for 30 sec, and 72°C for 30 sec. During the 72°C temperature, analysis of the SYBR fluorophore for quantification was conducted. Relative expression levels of CD207 mRNA were calculated by normalizing to the level of GAPDH mRNA by using comparative threshold cycle (ct) method, in which fold difference = 2-(Δct of target gene-Δct of reference). Primers for amplification of CD207 mRNA were 5'-GTGGACAAACAGAACATCTCCCTC-3' and 5'-GACCTTTCAGCAACTGGACATTG-3'. GAPDH transcript was amplified with primers 5'-CGGTATTTGGTCGTATTGGGC-3' and 5'-TGGAAGATGGTGATGGGATTTC-3'. An identical approach was used to determine the relative expression levels of osteopontin mRNA using the following primers: 5'-GCAGTGATTTGCTTTTGCCTCC-3' and 5'-CTTTCGTTGGACTTACTTGGAAGG-3'.

For in vitro studies, aqueous CSE was prepared from Kentucky research cigarettes 3RF4 as recently described [11]. The nicotine levels in the CSE preparations were measured in the Mayo institutional clinical laboratory using liquid chromatography-tandem mass spectrometry: the nicotine content measured in 1% CSE was 174 ng/ml. Cellular experiments were conducted with CSE concentrations <3% because of viability studies that demonstrated cellular toxicity as determined by the XTT assay and AnnexinV/propridium iodide staining with CSE preparations >3%.

Human monocytes were isolated from buffy coats obtained from healthy non-smoking adult blood donors, and monocyte-derived dendritic cells were generated using a 6 day differentiation protocol with recombinant human IL-4 and GM-CSF as previously described [11, 12]. Maturation was induced by overnight culture with 100 ng/ml lipopolysaccharide [LPS from E. coli; Sigma] unless otherwise specified.

Semi-quantitative determination of cellular protein levels were obtained by immunoblotting as recently described [12]. Human monocyte-derived dendritic cells were plated at a concentration of 1 × 10⁶/ml in complete media [RPMI, 10% fetal bovine serum] and GM-CSF/IL-4 as previously described [12]. Cigarette smoke extract and LPS were added to the cells at the time points indicated. Protein lysates were prepared using RIPA buffer [150 mM NaCl, 1.0% Igepal CA-630, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris] as recently described [12]. Antibodies used in immunoblotting included Bcl-xL (Cell Signaling # 2762), HO-1 (Stressgen # SPA-896), and CCR7 (eBioscience # 14-9977-82).

Human monocyte-derived dendritic cells were treated for 18 hours with 100 ng/ml of LPS and 50 ng/ml of IFN-γ, with or without 2% CSE, or 2% CSE and 1 mM NAC. After this period of activation, dendritic cells were washed in RPMI, and equal numbers (0.5 × 10⁶) were placed in 0.5 ml of RPMI without serum and assayed for migration towards CCL21 placed in the lower chamber of a transwell plate. The lower chambers of Transwell plates (5.0-μm pore size; Corning, Acton, MA) were filled with 0.5 ml of complete media (RPMI with 10% fetal calf serum) supplemented with 10 ng/ml CCL21. Human dendritic cells placed in the upper chambers of the Transwell plates were allowed to migrate to the lower chamber for 4 hours at 37°C in 5% CO2. The total numbers of migrated dendritic cells from the lower chambers were determined by flow cytometry (60-second counts).

All murine experiments described were approved by the institutional animal use and care committee. To further expand on our in vitro studies, we conducted in vivo studies to determine the effect of chronic cigarette smoking on lung and systemic dendritic cell survival. Mice were chronically exposed to cigarette smoke generated from a Teague TE-2 system, as recently described [12]. This smoking chamber is a manually-controlled cigarette smoking machine that produces a combination of side-stream and mainstream cigarette smoke in a chamber, which is then transported to a collecting and mixing chamber where varying amounts of air is mixed with the smoke mixture. In this model, mice were exposed to regulated concentrations of cigarette smoke generated from 2 cigarettes every 10 minutes for a total of 3 hours/day, 5 days/week for a total of 4-6 weeks. Following 4-6 weeks in the smoking chamber, mice were sacrificed, blood was removed by right heart puncture and submitted for nicotine analyses, and dendritic cells from lung and splenic tissues isolated as recently described [12].

Statistical differences were considered to be significant if P was <0.05. For the lung tissue analysis, data is expressed as means and comparisons between means were performed using the Mann Whitney U test. Correlations between lung tissue characteristics and clinical parameters were performed using the Spearman test. For the in vitro studies with dendritic cells, data is expressed as means and standard error, while comparison between multiple means was done with ANOVA. Statistical analysis was performed using SigmaPlot5. Densitometry was performed using Image J software http://​rsbweb.​nih.​gov/​ij/​.

Accumulation of matured myeloid dendritic cells (defined as CD83+), or Langerhans-type dendritic cells (defined as CD1a+) in lung biopsy specimens, was evaluated in 24 subjects from the LTRC database. Subject characteristics are shown in Table 1. As expected, the % predicted FEV₁ and FEV₁/FVC ratio were significantly lower in patients with COPD when compared with control subjects. Rather than count specific numbers of cells, quantitative measures of expression for the dendritic cell receptors, CD83 and CD1a, were estimated from electronically scanned entire control and COPD tissue slides, and the relative intensity of expression for either marker were determined using image capture analysis. A statistically significant increase in CD83+ staining cells - a cell surface receptor expressed by matured dendritic cells - was detected in COPD tissues compared to control tissues (Figure 1A, Mann Whitney U p = 0.049). The extent of CD83+ staining correlated with cumulative tobacco consumption defined by pack years of cigarettes smoked (R 0.398, p = 0.026), but did not correlate with % predicted FEV₁ (R -0.194, p = 0.182) or % predicted DLCO (R -0.084, p = 0.348). The extent of CD83 staining on tissue slides also correlated with intensity of CD1a expression (R 0.566, p = 0.002). Quantitative total tissue expression of CD1a did not show a statistically significant difference between control and COPD tissue (Figure 1B, Mann Whitney U p = 0.301). The extent of CD1a staining on tissue slides did not correlate with cumulative pack years smoked (R 0.305, p = 0.073) or % predicted FEV₁ (R -0.241, p = 0.128), but inversely correlated with the % predicted DLCO (R -0.350, p = 0.047).

As the extent of CD1a positive staining in COPD tissue appeared to show a higher trend in comparison with control tissue, we determined the relative amounts of CD207 mRNA as an alternative approach to determine whether smokers with COPD have increased numbers of Langerhans cells infiltrating lung tissue. To determine this, RNA was extracted from 12 frozen lung biopsy samples available from the tissue samples used in the immunohistochemical approach described above (the remaining samples could not be used as the extracted RNA was not of adequate quality), and an additional 22 frozen lung samples obtained through the LTRC. All samples used in this analysis had RIN > 7.5. The 34 samples analyzed included 7 controls without COPD (mean % predicted FEV₁ 91.3, range 84-103) and 27 COPD subjects (mean % predicted FEV₁ 38.3, range 22-75). All patients were either current or former smokers. A statistically significant increase in CD207 mRNA was found in lung tissue with COPD compared to lung tissue without COPD; Figure 1C, mean/SE CD207 expression 1.335 ± 0.627 vs 0.389 ± 0.128 respectively; p = 0.048.

Recent studies have shown that cigarette smoke induces osteopontin production by lung epithelial cells and macrophages [13, 14]. Osteopontin is involved in the regulation of Th-1 immunity and autoimmunity, is chemotactic and anti-apoptotic to dendritic cells, and when over-expressed in murine lungs, induces dendritic cell recruitment [15, 16]. Analysis of osteopontin mRNA levels in the 7 control and 27 COPD tissues used to determine CD207 mRNA levels, did not show a significant difference between controls and COPD (0.188 ± 0.14 vs 0.297 ± 0.15 respectively, p = 0.17). Although osteopontin mRNA levels were not significantly different between the control and COPD groups, osteopontin mRNA levels in the 27 tissue specimens with COPD correlated with CD207 mRNA (R = 0.6529, p = 0.001). Taken together these data indicate that COPD lung tissue is infiltrated by greater numbers of both cells expressing the Langerhans cell gene CD207, and matured CD83+ dendritic cells. The correlation between osteopontin and CD207 mRNA levels suggests a potential link between osteopontin regulation and Langerhans cell infiltration in COPD.

We speculated that enhanced dendritic cell survival may be a potential mechanism by which the observed increase in dendritic cell numbers occurs in COPD. To determine whether cigarette smoke promotes dendritic cell survival, immature human monocyte-derived dendritic cells were generated and subsequently incubated with increasing concentrations of freshly generated CSE (0.01 to 1%). After 24 hours of incubation in a standard tissue culture chamber, cellular viability was determined using the XTT assay. A statistically significant increase in cellular viability was observed when dendritic cells were incubated with CSE concentrations >0.1% (Figure 2A, one-way ANOVA p = 0.001; Dunnett's post-test for 0.1% CSE vs 0 and 1% CSE vs O = <0.05 and <0.001 respectively). To further characterize whether cigarette smoke promotes dendritic cell survival, we conducted complimentary studies on lung and splenic dendritic cells (defined as CD11c+ cells) extracted from mice exposed to cigarette smoke for ≥4 weeks in a Teague smoking chamber as previously described [11, 12]. Lung and spleen dendritic cells were purified by magnetic sorting [12], and adjusted to a concentration of 1 × 10⁶/ml in media containing 5 ng/ml of murine GM-CSF. Viability of immature and LPS-matured lung and splenic dendritic cells from cigarette smoke exposed mice and control mice was performed using XTT assay following 24 hours of incubation in a culture chamber. Consistent with the observed effect on human dendritic cells, a statistically significant increase in viability was observed in both murine lung and systemic (spleen) dendritic cells from cigarette smoke exposed mice, in comparison with control mice (Figure 2B; 2-way ANOVA p = 0.007 for smoking effect: Figure 2C; 2-way ANOVA p < 0.001 for smoking effect). Thus, chronic exposure to cigarette smoke in vivo enhances ex vivo survival of both immature and LPS-matured lung and systemic dendritic cells. Cigarette smoke and CSE contain thousands of chemicals: to determine whether the nicotine component of cigarette smoke or CSE enhances dendritic cell survival in vitro, human dendritic cells were incubated with nicotine for up to 48 hours and viability measured with the XTT assay. Incubation of dendritic cells with nicotine concentrations that include those described in the circulation of active smokers [17, 18], and those measured in CSE, failed to augment dendritic cell survival in vitro (Figure D and E, p < 0.05 by ANOVA).

The observation that both CSE and cigarette smoke exposure enhance dendritic cell survival, led us to speculate that cigarette smoke constituents activate pro-survival or anti-apoptotic factors in dendritic cells. To test this, human dendritic cells were incubated with CSE, and cellular levels of known pro-survival factors were determined using immunoblotting. Following variable periods of incubation with CSE, protein lysates were prepared, and the cellular levels of 2 key cell survival proteins previously described to be activated in smokers macrophages [19] - heme-oxygenase-1 (HO-1), and Bcl-xL - were determined. Figure 3A shows the induction of both HO-1 and Bcl-xL protein levels in whole cell lysates of dendritic cells incubated with CSE. The induction of HO-1, an inducible cytoprotective cellular sensor of oxidative stress, occurs early and peaks by 8 hours following acute stimulation. In contrast, the induction of Bcl-xL, an anti-apoptotic member of the bcl family of proteins that promotes dendritic cell survival in vitro [20] and in vivo [21], accumulates after 24-48 hours of incubation with CSE (Figure 3A). To determine whether cigarette smoke induced oxidative stress is responsible for induction of dendritic cell HO-1, human dendritic cells (1 × 10⁶/ml media) were incubated with complete media in the presence or absence of freshly-prepared 2% CSE, and in the presence or absence of either 2.5 mM NAC as an inhibitor of oxidative stress, or 10 μg/ml dexamethasone. NAC and dexamethasone where added 60 minutes prior to stimulation with CSE. Following 6 hours of incubation with CSE and relevant inhibitors of oxidative stress (NAC) or inflammatory gene transcription (dexamethasone), whole cell lysates were prepared, and equal amounts of protein from each sample (50 μg) were separated on a 12% gel and immunoblotting performed for HO-1. As illustrated in Figures 3B and 3C, CSE-induced upregulation of HO-1 was completely suppressed by NAC, but unaffected by dexamethasone, implicating oxidative stress as the most likely inducer of HO-1 protein in cigarette smoke stimulated dendritic cells. In parallel experiments, we determined whether induction of Bcl-xL protein by CSE was primarily induced by oxidative constituents. To determine this, whole cell protein lysates were prepared from human dendritic cells incubated in the presence or absence of 2.5 mM NAC and stimulated with 2% CSE for 48 hours. As shown in Figure 3D, whereas CSE induced Bcl-xL protein, nicotine had virtually no effect on endogenous Bcl-xL levels. Dendritic cells pre-incubated with NAC failed to upregulate Bcl-xL protein levels, indicating oxidative stress as a predominant mechanism by which CSE induces Bcl-xL protein in dendritic cells.

In addition to augmented survival, we speculated that cigarette smoke may promote retention and accumulation of dendritic cells in the lung by diminishing migration to secondary lymph nodes. We focused our analysis on CCR7 expression, since deficiency of CCR7 results in impaired or absent migration of dendritic cells in mice [22‐24], and CCR7 was previously reported to be suppressed in lung dendritic cells of smokers with COPD [25]. The expression of CCR7 by immature and LPS-matured human monocyte-derived dendritic cells was determined following overnight incubation of dendritic cells cultured in the presence of either 2% CSE or 1000 ng/ml nicotine. To determine whether CSE-induced oxidative stress mediates suppression of CCR7 following LPS activation, a control group of dendritic cells were pretreated with 1 mM NAC for 60 minutes prior to the addition of 2% CSE and LPS. As expected, LPS robustly induced the expression of surface CCR7 (Figure 4A). The addition of 2% CSE to the culture media, but not 1000 ng/ml nicotine, resulted in significant reduction of surface CCR7 expression (Figure 4A). Pre-incubation of dendritic cells with NAC prior to the addition of CSE resulted in complete abrogation of the inhibitory effect of CSE on LPS-mediated dendritic cell CCR7 expression, implying that oxidative stress is responsible for suppression of CCR7 (Figure 4A). Identical findings were observed when the effect of CSE on whole cellular CCR7 levels was determined using immunoblotting (Figure 4B). In parallel to the flow cytometric findings, dendritic cells incubated overnight with CSE and LPS had substantially lower cellular levels of CCR7 compared to dendritic cells activated with LPS alone (Figure 4B). In addition, pre-incubation of dendritic cells with 1 mM NAC 60 minutes prior to the addition of CSE and LPS resulted in restoration of LPS-induced whole cellular CCR7 levels (Figure 4B), suggesting that rather than inhibiting translocation of preformed CCR7 to the surface of dendritic cells, CSE inhibits de novo formation of cellular CCR7 following LPS activation. To directly determine whether CSE suppresses migration of dendritic cells, immature and LPS-activated human dendritic cells were incubated with or without CSE for 18 hours. Half a million dendritic cells under different conditions were washed and resuspended in 0.5 ml RPMI without serum and placed in cell culture inserts. The inserts were subsequently placed in a 12-well tissue culture plate containing 0.5 ml of RPMI with 10% fetal calf serum and 10 ng/ml of CCL21, the respective ligand for CCR7. After 4 hours of incubation in a tissue culture chamber, the number of cells that migrated from the insert to the lower chamber was quantified. As expected, few immature dendritic cells migrated from the upper chamber to the lower chamber, while dendritic cells matured with LPS demonstrated a marked increase in migratory capacity (1745 ± 6.3 vs 10665 ± 106.1 cells; p < 0.001 with 1-way ANOVA and Bonferroni's Multiple Comparison Test). Surprisingly, dendritic cells activated with LPS in the presence of CSE did not demonstrate a reduction in migratory capacity (Figure 4C). In contrast, migration of LPS-activated dendritic cells conditioned with CSE towards the CCR7 ligand was augmented (Figure 4C; 10665 ± 106.1 vs 14385 ± 551.5 cells for LPS-activated dendritic cells and CSE-conditioned LPS-activated dendritic cells respectively; p < 0.001 with 1-way ANOVA and Bonferroni's Multiple Comparison Test). Although CSE did not diminish migration of LPS-activated dendritic cells towards CCL21, a statistically significant reduction in migration was observed in LPS-activated dendritic cells conditioned with 1000 ng/ml nicotine (Figure 4C; 10665 ± 106.1 vs 8205 ± 275.8 cells for LPS-activated dendritic cells and nicotine conditioned LPS-activated dendritic cells respectively; p < 0.01 with 1-way ANOVA and Bonferroni's Multiple Comparison Test).