CCR2 and CCR6, but Not Endothelial Selectins, Mediate the Accumulation of Immature Dendritic Cells within the Lungs of Mice in Response to P
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
Dendritic cells (DC) migrate from sites of inflammation to lymph nodes to initiate primary immune responses, but the molecular mechanisms by which DC are replenished in the lungs during ongoing pulmonary inflammation are unknown. To address this question, we analyzed the secondary pulmonary immune response of Ag-primed mice to intratracheal challenge with the particulate T cell-dependent Ag sheep erythrocytes (SRBC). We studied wild-type C57BL/6 mice and syngeneic gene-targeted mice lacking either both endothelial selectins (CD62E and CD62P), or the chemokine receptors CCR2 or CCR6. DC, defined as non-autofluorescent, MHC class II+CD11cmod cells, were detected in blood, enzyme-digested minced lung, and bronchoalveolar lavage fluid using flow cytometry and immunohistology. Compared with control mice, Ag challenge increased the frequency and absolute numbers of DC, peaking at day 1 in peripheral blood (6.5-fold increase in frequency), day 3 in lung mince (20-fold increase in total DC), and day 4 in bronchoalveolar lavage fluid (55-fold increase in total DC). Most lung DC expressed CD11c, CD11b, and low levels of MHC class II, CD40, CD80, and CD86, consistent with an immature myeloid phenotype. DC accumulation depended in part upon CCR2 and CCR6, but not endothelial selectins. Thus, during lung inflammation, immature myeloid DC from the bloodstream replace emigrating immature DC and transiently increase total intrapulmonary APC numbers. Early DC recruitment depends in part on CCR2 to traverse vascular endothelium, plus CCR6 to traverse alveolar epithelium. The recruitment of circulating immature DC represents a potential therapeutic step at which to modulate immunological lung diseases.
Introduction
Pulmonary immune response initiation depends crucially on dendritic cells (DC),4 the potent APC essential to activate naive T cells (1, 2, 3). Within unperturbed lungs, immature DC (iDC) reside in airway epithelium, alveolar septae, and the connective tissue surrounding pulmonary vessels (4), but are rarely found within the alveolar spaces. iDC are specialized for Ag uptake and express only low surface amounts of MHC class II and costimulatory molecules. In response to inflammatory stimuli, DC increase expression of these molecules and become functional APC (5, 6). Simultaneously, DC up-regulate CCR7 and migrate to draining lymph nodes under the influence of the CCR7 ligands CCL19 and CCL21 (7, 8, 9). Migration of DC to lymph nodes is rapid, peaking within 24–48 h of Ag challenge (10, 11, 12, 13).
This well-described efflux of lung DC to the draining lymph node implies that peripheral tissues must become depleted of APC during inflammation, unless they are replaced. Although it is generally accepted that DC are replenished in lungs from peripheral blood precursors, direct evidence of such recruitment is limited (14, 15, 16, 17, 18, 19), and the potential molecular mechanisms controlling it are incompletely understood. Knowledge of such mechanisms could prove vital in the development of new therapeutics targeting immune-mediated lung diseases such as asthma and chronic obstructive pulmonary disease.
The goal of this study was to determine the kinetics with which DC are recruited to the interstitial and alveolar compartments of the lung during the response to lung inflammation, and to analyze the molecular requirements for their appearance. We used a model of CD4 T cell-dependent lung inflammation induced by a single intratracheal (i.t.) challenge of Ag-primed mice, using the particulate Ag SRBC (20). This model was chosen for its ability to induce an acute, self-limited inflammatory response with a clearly definable onset and well-established kinetics. Previous studies have proven this model system useful in defining the molecular mechanisms of lung leukocyte recruitment (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33). Following i.t. Ag challenge, we assessed the presence of DC within the peripheral blood, the lung interstitium, and the alveolar spaces. We used gene-targeted mice to examine two classes of molecules that might mediate DC recruitment to the lungs. The first were the endothelial selectins, E-selectin (CD62E) and P-selectin (CD62P). In this model system, the endothelial selectins are markedly up-regulated (27) and are essential for recruitment of some lymphocyte subsets (30, 32). iDC bind E- and P-selectins under shear stress (15), making this adhesive system a plausible requirement for their recruitment. The second class of potential recruitment molecules were the chemokine receptors CCR2 and CCR6, which have recently been implicated in DC recruitment to the lung (19, 34, 35), tonsils (36, 37), and injured skin (36, 37, 38). Our results demonstrate that, in response to a single Ag challenge, the frequency of DC within peripheral blood increased rapidly, followed by a marked accumulation of DC within the lung interstitium and subsequently in the alveolar spaces. We further show that this process was independent of endothelial selectins, but that deficiency of chemokine receptor expression altered the efficacy with which DC were recruited into either the interstitial or alveolar compartments.
Materials and Methods
Animals
Specific pathogen-free inbred female C57BL/6 mice purchased from Charles River Laboratories were used except as specified. Three strains of gene-targeted mice, each on a C57BL/6J background, were bred locally. Each strain has been described previously: E- and P-selectin double knockouts (E–P–) (39), chemokine receptor 2 gene knockouts (CCR2–/–) (40), and chemokine receptor 6 gene knockouts (CCR6–/–) (41). Mice were used in experiments at 8–14 wk of age. Mice were housed in the Animal Care Facility at the Ann Arbor Veterans Affairs Health System, which is fully accredited by the American Association for Accreditation of Laboratory Animal Care. Mice were provided standard animal chow and chlorinated tap water ad libitum. This study complied with the 1996 National Academy of Sciences Guide for the Care and Use of Laboratory Animals (www.nap.edu/readingroom/books/labrats/) and followed a protocol approved by the Animal Care Subcommittee of the local Institutional Review Board.
Monoclonal Abs
The following mAbs purchased from BD Pharmingen were used: RM4-4 (anti-murine CD4, rat IgG2b), 53-6.72 (anti-murine CD8, rat IgG2b), M1/70 (anti-murine CD11b, rat IgG2b), HL3 (anti-murine CD11c, hamster IgG1), 2.4G2 (anti-murine CD16/CD32 Fc block, rat IgG2b), 3/23 (anti-murine CD40, rat IgG2a), RA-36B2 (anti-murine CD45R/B220, rat IgG2a), 30-F11 (anti-murine CD45, rat IgG2b), 16-10A1 (anti-murine CD80, hamster IgG2), GL1 (anti-murine CD86, rat IgG2a), AF6-120.1 (anti-murine I-Ab MHC class II, mouse IgG2a), and RB6-8C5 (anti-murine Ly6G Gr-1, rat IgG2b). mAbs were primarily conjugated with FITC, biotin, or PE; biotinylated Abs were visualized using streptavidin-PerCP (BD Pharmingen). Isotype-matched irrelevant control mAbs (BD Pharmingen) were tested simultaneously in all experiments.
Experimental design
In all experiments, a secondary pulmonary immune response was induced using the particulate Ag SRBC. Mice were primed by i.p. injection with 1 x 108 SRBC (Colorado Serum) in 0.5 ml of normal saline. Two to 3 wk later, mice were challenged by the i.t. route with 5 x 108 SRBC in 50 μl of normal saline as previously described (20, 42). Control mice that received only priming by the i.p. route are referred to as day 0 mice. At various later times, tissues were harvested, and three types of experiments were performed. First, to analyze DC phenotypes, single-cell mononuclear suspensions were isolated from the alveolar spaces by bronchoalveolar lavage (BAL), from the lung interstitium by enzymatic digestion of perfused lung tissue following BAL, and from peripheral blood. These preparations were stained with mAbs directed against MHC class II, CD11c, and other receptors, and were analyzed by flow cytometry. We identified DC as non-autofluorescent, MHC class II+CD11cmod cells (3, 10, 12, 15, 16, 43), and lung macrophages (M) as MHC class IIlowCD11chigh cells (33, 44). Second, to determine the microanatomic location of DC within the lung interstitium, we performed one- and two-color immunohistochemistry on frozen sections. Third, to determine the contribution of endothelial selectins, CCR2, or CCR6 to DC recruitment, the pulmonary immune response was assessed using gene-targeted mice lacking those molecules.
Tissue collection
Mice were deeply anesthetized with pentobarbital (80 mg/kg, i.p.) and killed by exsanguination and induction of bilateral pneumothoraces. Individual mice were used either to collect cells for flow cytometry or for immunohistochemistry. BAL was performed as previously described (42). After BAL, lungs were then perfused via the right heart using PBS containing 0.5 mM EDTA until pulmonary vessels were grossly clear. Lungs were excised, minced finely, and incubated in RPMI 1640 containing 10% FBS (Invitrogen Life Technologies), 30 μg/ml DNase-1, and 150 U/ml collagenase (both from Worthington Biochemical) for 35 min at 37°C with constant gentle swirling. Next, the lung fragments were filtered through a 70-μm mesh (BD Biosciences). RBC were lysed using 110 mM NH4Cl. This population is referred to as "lung mince" and is enriched for cells located within the pulmonary interstitium. PBMC were harvested from mice that were systemically anticoagulated using 300 U of heparin (Elkins-Sinn). Under deep anesthesia, 0.5 ml of blood was removed by right heart puncture. Blood was layered over Lympholyte-M (Cedarlane Laboratories), and centrifuged at 1100 x g for 25 min at room temperature. Cells at the interface were harvested, and residual RBC were lysed using 110 mM NH4Cl. All cell preparations were washed twice in PBS before use for Ab staining.
For immunohistochemical analysis, lungs were perfused but not lavaged. The trachea was cannulated with PE50 tubing (Clay-Adams), and the lungs were inflated first with 1 ml of air followed by 1 ml of an 80:20 mixture of RPMI 1640 (Invitrogen Life Sciences) to OCT (TissueTek; Sakura Finetek). Lungs were placed into a cryomold and covered with OCT, immersed in liquid N2 for 35 s, and then wrapped in aluminum foil and stored overnight at –20°C. The next day, lung sections were cut at 5-μm thickness using a refrigerated microtome (Cryocut 1900; Reichart Jung) and placed on Superfrost/Plus microscope slides (Fisher). After air drying to improve adherence, slides were placed in a sealed slide box and stored at –20°C until used.
For PCR, lungs were perfused but not lavaged, and immediately stored in RNAlater (Ambion). Lungs were homogenized using a Tissue Tearor (Biospec Products), and total RNA was prepared using a RiboPure Isolation kit (Ambion). Contaminating DNA was removed with DNA-free (Ambion). Samples were reverse transcribed using a Retroscript kit (Ambion).
Ab staining and flow-cytometric analysis
Staining, including blockade of FcRs, and analysis by flow cytometry were performed as described previously (25, 44). Data were collected on a FACScan flow cytometer using CellQuest software (both from BD Immunocytometry Systems) and analyzed using FlowJo software (Tree Star). At least 10,000 cells were analyzed per sample. Initial gates were set based on light scatter characteristics to eliminate debris, red cells, and cell clusters. Starting with BAL samples from normal mice, care was taken to prospectively identify alveolar M (AM), which are known to be CD11chigh cells (33, 44) (see Fig. 1, larger gate on the upper left in each panel). Identical M gates were used for lung mince samples. We then identified lung DC as non-autofluorescent, MHC class II+CD11cmod cells (3, 10, 12, 15, 16, 43) (see Fig. 1, smaller gate on the lower right in each panel). Inclusion of autofluorescent cells among lung DC was minimized by use of elliptical gates offset from the diagonal observed in samples stained using isotype control mAbs, which are essentially identical with unstained samples (our unpublished observation). Thereafter, cytometer parameters and gate position were held constant during analysis of all samples. The percentage of DC obtained from flow cytometry was used to calculate the total number of DC from each tissue by multiplying the frequency of DC by the total cell count for that sample. Three-color analysis was performed by the addition of a third Ab (receptor-specific or isotype control) to cells stained with MHC class II and CD11c. Cells located within the DC gate were then analyzed for expression of the third marker of interest. In some experiments, plasmacytoid DC (pDC) were identified as CD11cmodGR-1mod cells (45).
Immunohistochemistry
Slides were brought to room temperature and placed in cold acetone for 10 min. After air drying, tissue sections were circled using a diamond pen. Slides were immersed in methanol/H2O2 for 15 min, and then in PBS for 5 min. All subsequent incubations were conducted in a humidity chamber at room temperature; reagents and Abs were diluted in PBS without cations containing 0.1 g/dl BSA. Ags were visualized by either of two methods: immunogold staining, as previously described (27), or two-color staining using consecutive peroxidase and alkaline phosphatase reactions. For two-color staining, slides were incubated in hamster serum (150 μl in 10 ml PBS without BSA) for 30 min; excess liquid was removed carefully, and biotinylated anti-CD11c or isotype control mAb (1/50 dilution) was added for 1 h. After washing in PBS for 5 min, slides were incubated with Universal ABC Peroxidase Complex (Vector Laboratories) for 30 min, and with AEC peroxidase substrate (Vector Laboratories) for 15 min, with interval 5-min washes in PBS. Next, slides were blocked using mouse serum for 30 min, incubated with mouse anti-MHC class II or isotype control (1/500 dilution) for 1 h, and then incubated with alkaline phosphatase complex for 30 min and alkaline phosphatase substrate (black reaction product) for 15 min, with interval 5-min washes in PBS. Following a rinse in running tap water, slides were counterstained in hematoxylin, rinsed in tap water, and coverslipped using aqueous mounting medium (Biomeda).
RT-PCR
Quantitative real-time RT-PCR was performed using TaqMan chemistry and primers for the chemokines CCL2 (MCP-1), CCL7 (MCP-3), CCL20 (MIP-3), and the housekeeping gene GAPDH obtained from Applied Biosystems. cDNA conversion, amplification, and data analysis were performed on a Mx3000P real-time PCR system computerized cycler from Stratagene, using a TaqMan Universal PCR Master mix (Applied Biosystems). Cycle parameters were as follows: a denaturation step at 95°C for 10 min, followed by up to 40 cycles composed of 15-s denaturation at 95°C and 1-min annealing at 60°C. For analysis, the fluorescence values of the threshold cycle were collected at the end of the annealing step from each reaction. The threshold cycle values obtained from GAPDH amplification were used to normalize chemokine quantification. Data are expressed as relative increase of specific mRNA as measured by fluorescence intensity of the treated samples compared with the untreated control sample, which was used as calibrator.
Statistical analysis
All data were expressed as mean ± SEM. Continuous ratio scale data were evaluated by unpaired Student t test (for comparison between two samples) or by ANOVA (for multiple comparisons) with post hoc analysis by either Fisher protected least significant difference or by two-tailed Dunnett test, which compares treatment groups to a specific control group (46). Statistical calculations were performed on a Macintosh G4 computer using the Statview 4.0 (Abacus Concepts). Statistical difference was accepted at p < 0.05.
Results
DC accumulated in the lung in response to particulate Ag-induced inflammation
Flow-cytometric analysis was used to assay two anatomic lung compartments, the lung interstitium (lung mince) and the alveolar spaces (BAL), during the development of Ag-induced lung inflammation. Care was taken to identify resident AM, because these highly autofluorescent cells contaminate lung mince preparations even after intensive BAL, and unlike M in other tissues, constitutively express high levels of CD11c (33, 44). Within samples from both compartments, we identified lung M as MHC class IIlowCD11chigh cells (33, 44) (Fig. 1, d0, larger elliptical gates) and DC as MHC class II+CD11cmod cells (3, 10, 12, 15, 16) (Fig. 1, d0, smaller elliptical gates). In unchallenged mice, a distinct population of DC was detectable in both anatomic compartments (lung mince, 6.2 ± 0.1 x 104; BAL, 5.5 ± 0.9 x 103; mean ± SEM cells/mouse; n = 5 mice), although their numbers were much smaller than those of lung M, especially in BAL. At day 0, DC comprised 1.9 ± 0.1% of cells in the lung mince and 0.8 ± 0.1% of cells in BAL, in good agreement with basal numbers in previous studies of mice (10, 18).
Ag challenge of SRBC-primed C57BL/6 mice by the i.t. route induced prompt recruitment of large numbers of mononuclear cells and granulocytes into the lungs, in agreement with previous data in this model system (20, 25, 27, 30, 31, 32, 42, 47, 48). In lung mince, total cell numbers peaked on day 3 postchallenge (1.4 ± 0.1 x 107 cells) (mean ± SEM; 5–11 mice pooled in five separate experiments), a >4-fold increase compared with mice that received only i.p. priming (day 0) (3.3 ± 0.4 x 106 cells). In BAL, total cell numbers peaked on day 4 postchallenge (5.8 ± 0.9 x 106 cells), a >8-fold increase compared with day 0 (6.9 ± 0.7 x 105 cells).
Maintaining gates constant throughout analysis (Fig. 1) permitted the accurate quantification of DC at various time points during the course of lung inflammation (Fig. 2). The frequency of DC among leukocytes in lung mince increased markedly, peaking at 8.5 ± 1.0% at day 3 (Fig. 2A). Absolute numbers of DC in lung mince were increased by day 1 postchallenge and peaked on day 3 at 1.2 ± 0.2 x 106 cells/mouse, a 20-fold increase from baseline (Fig. 2B). DC prevalence in BAL did not change at day 1, but then increased at day 4 to a peak of 5.7 ± 0.8% of lung leukocytes (Fig. 2C). Absolute numbers of DC in BAL peaked on day 4 at 3.1 ± 0.4 x 105 cells/mouse, a 55-fold increase from baseline (Fig. 2D). DC numbers in both tissue compartments were returning toward baseline by day 7 postchallenge. In contrast, numbers of cells within the M gate did not increase significantly in either tissue over the observed time course, except for a 2.5-fold increase in BAL only at day 7 (data not shown). There was a marked increase in both compartments in autofluorescent cells outside these fluorescence-defined gates. Analysis of BAL using vital staining has recently shown that these cells are monocytes (33).
Discussion
This study defines for the first time the kinetics and anatomic distribution of DC recruitment from the bloodstream into the lungs of mice during a secondary pulmonary immune response. A single i.t. Ag challenge induced sequential increases in DC within three anatomic compartments. DC first increased in frequency in peripheral blood (peaking on day 1 at a 6.5-fold increase), followed by increases in the total lung DC numbers, initially within the interstitium (peaking on day 3 at a 20-fold increase), and then within the alveolar spaces (peaking on day 4 at a 55-fold increase). DC localized to foci of active inflammation within peribronchovascular aggregates, assuming a stellate pattern interdigitating with mononuclear cells. Recruited lung DC were of an immature myeloid DC phenotype. A second major outcome of our study is the demonstration that DC accumulation within the lungs was not impaired in mice lacking the endothelial selectins CD62E and CD62P. Third, in contrast, DC accumulation was markedly impaired in both the lung interstitium and BAL in CCR2–/– mice, and in the alveolar space only in CCR6–/– mice. These novel findings have important implications for the development of protective pulmonary immune responses and immunologically mediated lung diseases.
This analysis of DC recruitment depended on key features of our experimental design. The priming needed for this CD4-dependent secondary response is temporally remote from the challenge and spatially remote from the lungs (20, 42), so lung DC are unlikely to be perturbed at baseline. Use of a single challenge allowed us to time precisely the onset of DC recruitment, which proved to be rapid, in agreement with results of bronchial challenge in humans (16, 17). These two features differ from experimental model systems that require repeated airway challenge before analysis of lung DC (10). Our challenge Ag, SRBC, is a nonreplicating particulate that deposits in the airways and distal airspaces (42). Unlike microbial pathogens, SRBC neither penetrates epithelial barriers nor secretes mediators capable of altering leukocyte recruitment. The inflammation induced by i.t. challenge of a primed mouse is vigorous but self-limited. It localizes in the peribronchovascular bundles, the preferential site of lung inflammation in response to Ags or infectious agents (reviewed by Pabst and Tschernig (52)). These features, together with the availability of knockout technology, make this murine model system ideal for the analysis of DC recruitment during immune lung inflammation.
This detailed kinetic and compartmental analysis yields several novel observations. First, the finding that Ag challenge of the lungs induced a change in peripheral blood frequency of DC within 1 day implies rapid response to one or more signals released by the earliest stages of the response to i.t. challenge. This response could occur by release of CD31high, Ly-6C– precursors from the bone marrow (51), but might also reflect mobilization of iDC marginated on endothelial surfaces. The peak change in DC frequency among PBMC precedes influx into the lungs of all but the earliest recruited leukocytes, and we have previously shown that Ag priming alone does not induce discernable lung inflammation (21, 42). Thus, this signal is likely a circulating factor released by endogenous lung cells. The identity of this signal is uncertain but merits study as a means of mobilizing DC precursors for use in immunotherapeutic strategies
Second, individual chemokines are produced during lung inflammation with a differential timing that correlates well with the observed kinetic of lung DC accumulation and that likely regulates recruitment into specific anatomic locations. There was rapid and robust increase in mRNA by real-time PCR for the CCR2 ligands CCL2 and CCL7, both of which peaked at day 1 and remained elevated before declining toward baseline by day 7. This rapid onset implies that these chemokines are also produced by endogenous lung cells. In contrast, production of mRNA for CCL20, the only known chemokine ligand of CCR6, was delayed. Human bronchial epithelial cells have been shown to produce CCL20 in response to inflammation (53). However, these kinetics also suggest production by recruited T cells, which peak in number at day 7 after challenge (42). Hence, full development of pulmonary immune responses likely depends on the sequential secretion of specific chemokines by distinct cell types.
To migrate from the bloodstream in response to chemokines, leukocytes must traverse the vascular endothelium. In most organs, this process requires initial rolling on the endothelial selectins CD62E and CD62P. Our novel finding that DC recruitment to the lung interstitium and alveoli is independent of endothelial selectins differs markedly from recruitment of lymphocytes in this model system (30, 32). Interestingly, we previously found that monocyte recruitment to the lungs is also independent of endothelial selectins (32). Selectin independence by DC could result from several possibilities, which are not mutually exclusive. DC transmigration could depend on cytoskeletal changes resulting in arrest not requiring endothelial selectins, as is the case of neutrophil transmigration of normal pulmonary capillary endothelium and of noncapillary lung endothelium in response to pneumococcus (54). DC might also use alternative mechanisms that mediate rolling under flow conditions, such as 4 integrins (55) or the stalked chemokine CX3CL1 (fractalkine) (56, 57). The differences in our findings from those of Pendl et al. (15), who showed partial endothelial selectin dependence in the recruitment of iDC to inflamed skin, most likely result from the disparity in tissues studied. However, it should also be noted that their study used bone marrow-derived iDC, whereas we examined recruitment of unmanipulated DC. Our findings predict that therapeutic agents based on blocking selectins (58, 59) will fail to impair DC recruitment to the lungs.
Endothelial transmigration and other steps involved in localization to the peribronchovascular space depend crucially on CCR2. This chemokine receptor has also been implicated in monocyte recruitment to the lung in response to LPS or CCL2 (60, 61). Thus, our data could indicate either that newly recruited lung DC share this requirement with monocytes, or that DC are derived from monocytes, a possibility neither we nor previous studies have specifically tested. These data on CCR2 dependence agree with recent findings in two models of granulomatous lung injury (19, 34), and extend them from the response to microbial products to nontoxic Ag deposited in the airspaces.
In contrast to the interstitium, DC must also cross an epithelial barrier to enter the alveoli, the site where inhaled Ag first deposits. Our data indicate that CCR6 and it sole chemokine ligand, CCL20, are important mediators of this process. Pulmonary epithelial tight junctions provide the majority of resistance to fluid transudation, and significantly, we have previously shown using monastral blue dye that this Ag challenge does not induce fluid leakage (21). Hence, leukocyte transmigration into the alveoli implies an active cooperation between alveolar epithelial cells and DC. CCR6 contributes to accumulation of the closely related Langerhans cell in the skin in both normal (62) and pathologic conditions (63, 64). Our study concentrated on recruitment of DC during lung inflammation, and did not evaluate effector functions. However, it is likely that the bronchial hyperresponsiveness and peribronchial eosinophilia seen in this model (23, 47) would be reduced, as reported in CCR6–/– mice (65). Importantly, the preserved recruitment of T cells (D. J. Linderman, J. Sonstein, S. W. Chensue, and J. L. Curtis, manuscript in preparation), a sensitive marker of Ag priming in the model system, argues against defective immunization, in agreement with previous data (65).
Collectively, our kinetic analysis provides in vivo evidence for a compartmental model in which DC navigate from bloodstream into tissues due to distinct temporal and spatial patterns of chemokine ligand/receptor expression. Specifically, we propose that CCR2 mediates early recruitment into lung interstitium, whereas CCR6 more centrally drives the later transit across epithelia into the airways. DC cultured from human peripheral blood sequentially express CCR2 before CCR6 (37). Chemokine expression is spatially segregated in resected human tonsils; a CCR2 ligand localized to vascular endothelium, whereas CCL20 expression was restricted to epithelial cells bordering the exterior (36, 37). Our results using gene-targeted mice imply that identical temporal and spatial patterns occur during Ag-driven lung inflammation.
The majority of newly recruited DC identified in our study were myeloid, based upon their expression of CD11b and absence of B220, and immature, as shown by low levels of costimulatory molecules. Lung DC did not express CD4 or CD8 homodimers, which form the basis for additional splenic DC subsets. A predominately myeloid lung DC phenotype has been found in several previous studies in both mice and humans (13, 66, 67, 68). However, these data contrast with the more mature phenotype seen after repetitive challenge with aerosolized Ag (10, 12, 69). Our model of particulate Ag challenge also found relatively few lung pDC, despite the use of two distinct methods to identify them, another difference from models using repetitive challenge with aerosolized Ag (45). These disparate findings could result from differences in the number of times the lungs are stimulated with Ag, the physical form (particulate vs aerosolized), or the time point at which DC were analyzed.
Mechanisms of lung DC accumulation in this model system other than recruitment are unlikely. We previously showed that interstitial lung mononuclear cells proliferate minimally in this model and that numbers of lung mononuclear cells did not decrease when cell division was blocked transiently using hydroxyurea (47). These data argue against a major component of in situ DC proliferation within the lungs. Given the clear difference we found in receptor expression and level of autofluorescence, it is unlikely that AM were mistakenly identified as DC. Moreover, our data agree closely with a recent study that used a similar approach to distinguish AM from DC, based on differences in autofluorescence and relative expression of CD11c and MHC class II (43). In that study, cell sorting of enzymatically digested lungs of normal C57BL/6 mice verified the marked differences in morphological, phenotypic, and Ag-presenting capacities of the two populations (43). We cannot exclude the possibility that part of the observed lung DC expansion in SRBC-challenged wild-type mice resulted from rapid differentiation of monocytes after arrival in the lungs. Although in vitro differentiation of monocytes into DC requires 5–7 days in medium containing GM-CSF plus IL-4 or IL-13 (70, 71), the process is accelerated in vitro by transendothelial migration (70, 72). How rapidly monocytes might differentiate into DC in the lungs is uncertain and will be an important goal of future studies. A population of CD11c–, CXCR1low, CCR2+ monocytes has recently been shown to be recruited to the inflamed peritoneum, where a large fraction differentiated to express CD11c and MHC class II within 18 h (57). Although recruitment was interpreted to be the major mechanism for changes in cell number in that study, a separate analysis of lung inflammation in response to Ag-coated beads identified an additional defect in maturation (i.e., up-regulation of MHC class II and CD40) in CD11c+ cells in CCR2–/– mice (35). The differences in the methods used to identify DC precludes direct comparison of those results with the current data.
In summary, we provide data that lung mucosal recognition of previously encountered Ags rapidly induces both local chemokine production and mobilization of immature myeloid DC precursors from the bone marrow. DC are recruited into the lungs in a process that is independent of endothelial selectins but which depends in part on CCR2 for entry into the peribronchovascular space, and on CCR2 and CCR6 for epithelial transmigration into the alveoli. These novel findings imply that DC recruitment from the bloodstream into the lung could be targeted therapeutically to control lung inflammation within specific anatomic compartments in such immunologically mediated diseases as asthma and chronic obstructive pulmonary disease.
Acknowledgments
We thank all of the members of the Ann Arbor Veterans Affairs Research Enhancement Award Program in Pulmonary Immunology for helpful suggestions and discussions, and Joyce O’Brien for secretarial support.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by a Career Development Award, Merit Review funding, and a Research Enhancement Award Program grant from the Biomedical Laboratory R&D Service, Department of Veterans Affairs; and by RO1 HL56309 and T32 HL07749 from the U.S. Public Health Service.
2 Portions of this work have been presented previously at the International Conference of the American Thoracic Society, Seattle, WA, May 19, 2003, and at the Keystone Symposium on "Dendritic Cells at the Center of Innate and Adaptive Immunity: Eradication of Pathogens and Cancer and Control of Immunopathology," Vancouver, BC, Canada, February 5, 2005; and have been published in abstract form (73 ).
3 Address correspondence and reprint requests to Dr. Jeffrey L. Curtis, Pulmonary and Critical Care Medicine Section (111G), Department of Veterans Affairs Medical Center, 2215 Fuller Road, Ann Arbor, MI 48105-2303. E-mail address: jlcurtis@umich.edu
4 Abbreviations used in this paper: DC, dendritic cell; iDC, immature DC; pDC, plasmacytoid DC; mod, moderate; i.t., intratracheal; BAL, bronchoalveolar lavage; M, macrophage; AM, alveolar M.
Received for publication September 28, 2004. Accepted for publication May 3, 2005.
References
Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, B. Pulendran, K. Palucka. 2000. Immunobiology of dendritic cells. Annu. Rev. Immunol. 18: 767-811.
Havenith, C. E., P. P. van Miert, A. J. Breedijk, R. H. Beelen, E. C. Hoefsmit. 1993. Migration of dendritic cells into the draining lymph nodes of the lung after intratracheal instillation. Am. J. Respir. Cell Mol. Biol. 9: 484-488.
Xia, W., C. E. Pinto, R. L. Kradin. 1995. The antigen-presenting activities of Ia+ dendritic cells shift dynamically from lung to lymph node after an airway challenge with soluble antigen. J. Exp. Med. 181: 1275-1283.
Holt, P. G., S. Haining, D. J. Nelson, J. D. Sedgwick. 1994. Origin and steady-state turnover of class II MHC-bearing dendritic cells in the epithelium of the conducting airways. J. Immunol. 153: 256-261.
Cella, M., F. Sallusto, A. Lanzavecchia. 1997. Origin, maturation and antigen presenting function of dendritic cells. Curr. Opin. Immunol. 9: 10-16
Banchereau, J., R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392: 245-252.
Sallusto, F., P. Schaerli, P. Loetscher, C. Schaniel, D. Lenig, C. R. Mackay, S. Qin, A. Lanzavecchia. 1998. Rapid and coordinated switch in chemokine receptor expression during dendritic cell maturation. Eur. J. Immunol. 28: 2760-2769.
Sallusto, F., A. Lanzavecchia. 1999. Mobilizing dendritic cells for tolerance, priming, and chronic inflammation. J. Exp. Med. 189: 611-614.(John J. Osterholzer*,, Th)
Dendritic cells (DC) migrate from sites of inflammation to lymph nodes to initiate primary immune responses, but the molecular mechanisms by which DC are replenished in the lungs during ongoing pulmonary inflammation are unknown. To address this question, we analyzed the secondary pulmonary immune response of Ag-primed mice to intratracheal challenge with the particulate T cell-dependent Ag sheep erythrocytes (SRBC). We studied wild-type C57BL/6 mice and syngeneic gene-targeted mice lacking either both endothelial selectins (CD62E and CD62P), or the chemokine receptors CCR2 or CCR6. DC, defined as non-autofluorescent, MHC class II+CD11cmod cells, were detected in blood, enzyme-digested minced lung, and bronchoalveolar lavage fluid using flow cytometry and immunohistology. Compared with control mice, Ag challenge increased the frequency and absolute numbers of DC, peaking at day 1 in peripheral blood (6.5-fold increase in frequency), day 3 in lung mince (20-fold increase in total DC), and day 4 in bronchoalveolar lavage fluid (55-fold increase in total DC). Most lung DC expressed CD11c, CD11b, and low levels of MHC class II, CD40, CD80, and CD86, consistent with an immature myeloid phenotype. DC accumulation depended in part upon CCR2 and CCR6, but not endothelial selectins. Thus, during lung inflammation, immature myeloid DC from the bloodstream replace emigrating immature DC and transiently increase total intrapulmonary APC numbers. Early DC recruitment depends in part on CCR2 to traverse vascular endothelium, plus CCR6 to traverse alveolar epithelium. The recruitment of circulating immature DC represents a potential therapeutic step at which to modulate immunological lung diseases.
Introduction
Pulmonary immune response initiation depends crucially on dendritic cells (DC),4 the potent APC essential to activate naive T cells (1, 2, 3). Within unperturbed lungs, immature DC (iDC) reside in airway epithelium, alveolar septae, and the connective tissue surrounding pulmonary vessels (4), but are rarely found within the alveolar spaces. iDC are specialized for Ag uptake and express only low surface amounts of MHC class II and costimulatory molecules. In response to inflammatory stimuli, DC increase expression of these molecules and become functional APC (5, 6). Simultaneously, DC up-regulate CCR7 and migrate to draining lymph nodes under the influence of the CCR7 ligands CCL19 and CCL21 (7, 8, 9). Migration of DC to lymph nodes is rapid, peaking within 24–48 h of Ag challenge (10, 11, 12, 13).
This well-described efflux of lung DC to the draining lymph node implies that peripheral tissues must become depleted of APC during inflammation, unless they are replaced. Although it is generally accepted that DC are replenished in lungs from peripheral blood precursors, direct evidence of such recruitment is limited (14, 15, 16, 17, 18, 19), and the potential molecular mechanisms controlling it are incompletely understood. Knowledge of such mechanisms could prove vital in the development of new therapeutics targeting immune-mediated lung diseases such as asthma and chronic obstructive pulmonary disease.
The goal of this study was to determine the kinetics with which DC are recruited to the interstitial and alveolar compartments of the lung during the response to lung inflammation, and to analyze the molecular requirements for their appearance. We used a model of CD4 T cell-dependent lung inflammation induced by a single intratracheal (i.t.) challenge of Ag-primed mice, using the particulate Ag SRBC (20). This model was chosen for its ability to induce an acute, self-limited inflammatory response with a clearly definable onset and well-established kinetics. Previous studies have proven this model system useful in defining the molecular mechanisms of lung leukocyte recruitment (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33). Following i.t. Ag challenge, we assessed the presence of DC within the peripheral blood, the lung interstitium, and the alveolar spaces. We used gene-targeted mice to examine two classes of molecules that might mediate DC recruitment to the lungs. The first were the endothelial selectins, E-selectin (CD62E) and P-selectin (CD62P). In this model system, the endothelial selectins are markedly up-regulated (27) and are essential for recruitment of some lymphocyte subsets (30, 32). iDC bind E- and P-selectins under shear stress (15), making this adhesive system a plausible requirement for their recruitment. The second class of potential recruitment molecules were the chemokine receptors CCR2 and CCR6, which have recently been implicated in DC recruitment to the lung (19, 34, 35), tonsils (36, 37), and injured skin (36, 37, 38). Our results demonstrate that, in response to a single Ag challenge, the frequency of DC within peripheral blood increased rapidly, followed by a marked accumulation of DC within the lung interstitium and subsequently in the alveolar spaces. We further show that this process was independent of endothelial selectins, but that deficiency of chemokine receptor expression altered the efficacy with which DC were recruited into either the interstitial or alveolar compartments.
Materials and Methods
Animals
Specific pathogen-free inbred female C57BL/6 mice purchased from Charles River Laboratories were used except as specified. Three strains of gene-targeted mice, each on a C57BL/6J background, were bred locally. Each strain has been described previously: E- and P-selectin double knockouts (E–P–) (39), chemokine receptor 2 gene knockouts (CCR2–/–) (40), and chemokine receptor 6 gene knockouts (CCR6–/–) (41). Mice were used in experiments at 8–14 wk of age. Mice were housed in the Animal Care Facility at the Ann Arbor Veterans Affairs Health System, which is fully accredited by the American Association for Accreditation of Laboratory Animal Care. Mice were provided standard animal chow and chlorinated tap water ad libitum. This study complied with the 1996 National Academy of Sciences Guide for the Care and Use of Laboratory Animals (www.nap.edu/readingroom/books/labrats/) and followed a protocol approved by the Animal Care Subcommittee of the local Institutional Review Board.
Monoclonal Abs
The following mAbs purchased from BD Pharmingen were used: RM4-4 (anti-murine CD4, rat IgG2b), 53-6.72 (anti-murine CD8, rat IgG2b), M1/70 (anti-murine CD11b, rat IgG2b), HL3 (anti-murine CD11c, hamster IgG1), 2.4G2 (anti-murine CD16/CD32 Fc block, rat IgG2b), 3/23 (anti-murine CD40, rat IgG2a), RA-36B2 (anti-murine CD45R/B220, rat IgG2a), 30-F11 (anti-murine CD45, rat IgG2b), 16-10A1 (anti-murine CD80, hamster IgG2), GL1 (anti-murine CD86, rat IgG2a), AF6-120.1 (anti-murine I-Ab MHC class II, mouse IgG2a), and RB6-8C5 (anti-murine Ly6G Gr-1, rat IgG2b). mAbs were primarily conjugated with FITC, biotin, or PE; biotinylated Abs were visualized using streptavidin-PerCP (BD Pharmingen). Isotype-matched irrelevant control mAbs (BD Pharmingen) were tested simultaneously in all experiments.
Experimental design
In all experiments, a secondary pulmonary immune response was induced using the particulate Ag SRBC. Mice were primed by i.p. injection with 1 x 108 SRBC (Colorado Serum) in 0.5 ml of normal saline. Two to 3 wk later, mice were challenged by the i.t. route with 5 x 108 SRBC in 50 μl of normal saline as previously described (20, 42). Control mice that received only priming by the i.p. route are referred to as day 0 mice. At various later times, tissues were harvested, and three types of experiments were performed. First, to analyze DC phenotypes, single-cell mononuclear suspensions were isolated from the alveolar spaces by bronchoalveolar lavage (BAL), from the lung interstitium by enzymatic digestion of perfused lung tissue following BAL, and from peripheral blood. These preparations were stained with mAbs directed against MHC class II, CD11c, and other receptors, and were analyzed by flow cytometry. We identified DC as non-autofluorescent, MHC class II+CD11cmod cells (3, 10, 12, 15, 16, 43), and lung macrophages (M) as MHC class IIlowCD11chigh cells (33, 44). Second, to determine the microanatomic location of DC within the lung interstitium, we performed one- and two-color immunohistochemistry on frozen sections. Third, to determine the contribution of endothelial selectins, CCR2, or CCR6 to DC recruitment, the pulmonary immune response was assessed using gene-targeted mice lacking those molecules.
Tissue collection
Mice were deeply anesthetized with pentobarbital (80 mg/kg, i.p.) and killed by exsanguination and induction of bilateral pneumothoraces. Individual mice were used either to collect cells for flow cytometry or for immunohistochemistry. BAL was performed as previously described (42). After BAL, lungs were then perfused via the right heart using PBS containing 0.5 mM EDTA until pulmonary vessels were grossly clear. Lungs were excised, minced finely, and incubated in RPMI 1640 containing 10% FBS (Invitrogen Life Technologies), 30 μg/ml DNase-1, and 150 U/ml collagenase (both from Worthington Biochemical) for 35 min at 37°C with constant gentle swirling. Next, the lung fragments were filtered through a 70-μm mesh (BD Biosciences). RBC were lysed using 110 mM NH4Cl. This population is referred to as "lung mince" and is enriched for cells located within the pulmonary interstitium. PBMC were harvested from mice that were systemically anticoagulated using 300 U of heparin (Elkins-Sinn). Under deep anesthesia, 0.5 ml of blood was removed by right heart puncture. Blood was layered over Lympholyte-M (Cedarlane Laboratories), and centrifuged at 1100 x g for 25 min at room temperature. Cells at the interface were harvested, and residual RBC were lysed using 110 mM NH4Cl. All cell preparations were washed twice in PBS before use for Ab staining.
For immunohistochemical analysis, lungs were perfused but not lavaged. The trachea was cannulated with PE50 tubing (Clay-Adams), and the lungs were inflated first with 1 ml of air followed by 1 ml of an 80:20 mixture of RPMI 1640 (Invitrogen Life Sciences) to OCT (TissueTek; Sakura Finetek). Lungs were placed into a cryomold and covered with OCT, immersed in liquid N2 for 35 s, and then wrapped in aluminum foil and stored overnight at –20°C. The next day, lung sections were cut at 5-μm thickness using a refrigerated microtome (Cryocut 1900; Reichart Jung) and placed on Superfrost/Plus microscope slides (Fisher). After air drying to improve adherence, slides were placed in a sealed slide box and stored at –20°C until used.
For PCR, lungs were perfused but not lavaged, and immediately stored in RNAlater (Ambion). Lungs were homogenized using a Tissue Tearor (Biospec Products), and total RNA was prepared using a RiboPure Isolation kit (Ambion). Contaminating DNA was removed with DNA-free (Ambion). Samples were reverse transcribed using a Retroscript kit (Ambion).
Ab staining and flow-cytometric analysis
Staining, including blockade of FcRs, and analysis by flow cytometry were performed as described previously (25, 44). Data were collected on a FACScan flow cytometer using CellQuest software (both from BD Immunocytometry Systems) and analyzed using FlowJo software (Tree Star). At least 10,000 cells were analyzed per sample. Initial gates were set based on light scatter characteristics to eliminate debris, red cells, and cell clusters. Starting with BAL samples from normal mice, care was taken to prospectively identify alveolar M (AM), which are known to be CD11chigh cells (33, 44) (see Fig. 1, larger gate on the upper left in each panel). Identical M gates were used for lung mince samples. We then identified lung DC as non-autofluorescent, MHC class II+CD11cmod cells (3, 10, 12, 15, 16, 43) (see Fig. 1, smaller gate on the lower right in each panel). Inclusion of autofluorescent cells among lung DC was minimized by use of elliptical gates offset from the diagonal observed in samples stained using isotype control mAbs, which are essentially identical with unstained samples (our unpublished observation). Thereafter, cytometer parameters and gate position were held constant during analysis of all samples. The percentage of DC obtained from flow cytometry was used to calculate the total number of DC from each tissue by multiplying the frequency of DC by the total cell count for that sample. Three-color analysis was performed by the addition of a third Ab (receptor-specific or isotype control) to cells stained with MHC class II and CD11c. Cells located within the DC gate were then analyzed for expression of the third marker of interest. In some experiments, plasmacytoid DC (pDC) were identified as CD11cmodGR-1mod cells (45).
Immunohistochemistry
Slides were brought to room temperature and placed in cold acetone for 10 min. After air drying, tissue sections were circled using a diamond pen. Slides were immersed in methanol/H2O2 for 15 min, and then in PBS for 5 min. All subsequent incubations were conducted in a humidity chamber at room temperature; reagents and Abs were diluted in PBS without cations containing 0.1 g/dl BSA. Ags were visualized by either of two methods: immunogold staining, as previously described (27), or two-color staining using consecutive peroxidase and alkaline phosphatase reactions. For two-color staining, slides were incubated in hamster serum (150 μl in 10 ml PBS without BSA) for 30 min; excess liquid was removed carefully, and biotinylated anti-CD11c or isotype control mAb (1/50 dilution) was added for 1 h. After washing in PBS for 5 min, slides were incubated with Universal ABC Peroxidase Complex (Vector Laboratories) for 30 min, and with AEC peroxidase substrate (Vector Laboratories) for 15 min, with interval 5-min washes in PBS. Next, slides were blocked using mouse serum for 30 min, incubated with mouse anti-MHC class II or isotype control (1/500 dilution) for 1 h, and then incubated with alkaline phosphatase complex for 30 min and alkaline phosphatase substrate (black reaction product) for 15 min, with interval 5-min washes in PBS. Following a rinse in running tap water, slides were counterstained in hematoxylin, rinsed in tap water, and coverslipped using aqueous mounting medium (Biomeda).
RT-PCR
Quantitative real-time RT-PCR was performed using TaqMan chemistry and primers for the chemokines CCL2 (MCP-1), CCL7 (MCP-3), CCL20 (MIP-3), and the housekeeping gene GAPDH obtained from Applied Biosystems. cDNA conversion, amplification, and data analysis were performed on a Mx3000P real-time PCR system computerized cycler from Stratagene, using a TaqMan Universal PCR Master mix (Applied Biosystems). Cycle parameters were as follows: a denaturation step at 95°C for 10 min, followed by up to 40 cycles composed of 15-s denaturation at 95°C and 1-min annealing at 60°C. For analysis, the fluorescence values of the threshold cycle were collected at the end of the annealing step from each reaction. The threshold cycle values obtained from GAPDH amplification were used to normalize chemokine quantification. Data are expressed as relative increase of specific mRNA as measured by fluorescence intensity of the treated samples compared with the untreated control sample, which was used as calibrator.
Statistical analysis
All data were expressed as mean ± SEM. Continuous ratio scale data were evaluated by unpaired Student t test (for comparison between two samples) or by ANOVA (for multiple comparisons) with post hoc analysis by either Fisher protected least significant difference or by two-tailed Dunnett test, which compares treatment groups to a specific control group (46). Statistical calculations were performed on a Macintosh G4 computer using the Statview 4.0 (Abacus Concepts). Statistical difference was accepted at p < 0.05.
Results
DC accumulated in the lung in response to particulate Ag-induced inflammation
Flow-cytometric analysis was used to assay two anatomic lung compartments, the lung interstitium (lung mince) and the alveolar spaces (BAL), during the development of Ag-induced lung inflammation. Care was taken to identify resident AM, because these highly autofluorescent cells contaminate lung mince preparations even after intensive BAL, and unlike M in other tissues, constitutively express high levels of CD11c (33, 44). Within samples from both compartments, we identified lung M as MHC class IIlowCD11chigh cells (33, 44) (Fig. 1, d0, larger elliptical gates) and DC as MHC class II+CD11cmod cells (3, 10, 12, 15, 16) (Fig. 1, d0, smaller elliptical gates). In unchallenged mice, a distinct population of DC was detectable in both anatomic compartments (lung mince, 6.2 ± 0.1 x 104; BAL, 5.5 ± 0.9 x 103; mean ± SEM cells/mouse; n = 5 mice), although their numbers were much smaller than those of lung M, especially in BAL. At day 0, DC comprised 1.9 ± 0.1% of cells in the lung mince and 0.8 ± 0.1% of cells in BAL, in good agreement with basal numbers in previous studies of mice (10, 18).
Ag challenge of SRBC-primed C57BL/6 mice by the i.t. route induced prompt recruitment of large numbers of mononuclear cells and granulocytes into the lungs, in agreement with previous data in this model system (20, 25, 27, 30, 31, 32, 42, 47, 48). In lung mince, total cell numbers peaked on day 3 postchallenge (1.4 ± 0.1 x 107 cells) (mean ± SEM; 5–11 mice pooled in five separate experiments), a >4-fold increase compared with mice that received only i.p. priming (day 0) (3.3 ± 0.4 x 106 cells). In BAL, total cell numbers peaked on day 4 postchallenge (5.8 ± 0.9 x 106 cells), a >8-fold increase compared with day 0 (6.9 ± 0.7 x 105 cells).
Maintaining gates constant throughout analysis (Fig. 1) permitted the accurate quantification of DC at various time points during the course of lung inflammation (Fig. 2). The frequency of DC among leukocytes in lung mince increased markedly, peaking at 8.5 ± 1.0% at day 3 (Fig. 2A). Absolute numbers of DC in lung mince were increased by day 1 postchallenge and peaked on day 3 at 1.2 ± 0.2 x 106 cells/mouse, a 20-fold increase from baseline (Fig. 2B). DC prevalence in BAL did not change at day 1, but then increased at day 4 to a peak of 5.7 ± 0.8% of lung leukocytes (Fig. 2C). Absolute numbers of DC in BAL peaked on day 4 at 3.1 ± 0.4 x 105 cells/mouse, a 55-fold increase from baseline (Fig. 2D). DC numbers in both tissue compartments were returning toward baseline by day 7 postchallenge. In contrast, numbers of cells within the M gate did not increase significantly in either tissue over the observed time course, except for a 2.5-fold increase in BAL only at day 7 (data not shown). There was a marked increase in both compartments in autofluorescent cells outside these fluorescence-defined gates. Analysis of BAL using vital staining has recently shown that these cells are monocytes (33).
Discussion
This study defines for the first time the kinetics and anatomic distribution of DC recruitment from the bloodstream into the lungs of mice during a secondary pulmonary immune response. A single i.t. Ag challenge induced sequential increases in DC within three anatomic compartments. DC first increased in frequency in peripheral blood (peaking on day 1 at a 6.5-fold increase), followed by increases in the total lung DC numbers, initially within the interstitium (peaking on day 3 at a 20-fold increase), and then within the alveolar spaces (peaking on day 4 at a 55-fold increase). DC localized to foci of active inflammation within peribronchovascular aggregates, assuming a stellate pattern interdigitating with mononuclear cells. Recruited lung DC were of an immature myeloid DC phenotype. A second major outcome of our study is the demonstration that DC accumulation within the lungs was not impaired in mice lacking the endothelial selectins CD62E and CD62P. Third, in contrast, DC accumulation was markedly impaired in both the lung interstitium and BAL in CCR2–/– mice, and in the alveolar space only in CCR6–/– mice. These novel findings have important implications for the development of protective pulmonary immune responses and immunologically mediated lung diseases.
This analysis of DC recruitment depended on key features of our experimental design. The priming needed for this CD4-dependent secondary response is temporally remote from the challenge and spatially remote from the lungs (20, 42), so lung DC are unlikely to be perturbed at baseline. Use of a single challenge allowed us to time precisely the onset of DC recruitment, which proved to be rapid, in agreement with results of bronchial challenge in humans (16, 17). These two features differ from experimental model systems that require repeated airway challenge before analysis of lung DC (10). Our challenge Ag, SRBC, is a nonreplicating particulate that deposits in the airways and distal airspaces (42). Unlike microbial pathogens, SRBC neither penetrates epithelial barriers nor secretes mediators capable of altering leukocyte recruitment. The inflammation induced by i.t. challenge of a primed mouse is vigorous but self-limited. It localizes in the peribronchovascular bundles, the preferential site of lung inflammation in response to Ags or infectious agents (reviewed by Pabst and Tschernig (52)). These features, together with the availability of knockout technology, make this murine model system ideal for the analysis of DC recruitment during immune lung inflammation.
This detailed kinetic and compartmental analysis yields several novel observations. First, the finding that Ag challenge of the lungs induced a change in peripheral blood frequency of DC within 1 day implies rapid response to one or more signals released by the earliest stages of the response to i.t. challenge. This response could occur by release of CD31high, Ly-6C– precursors from the bone marrow (51), but might also reflect mobilization of iDC marginated on endothelial surfaces. The peak change in DC frequency among PBMC precedes influx into the lungs of all but the earliest recruited leukocytes, and we have previously shown that Ag priming alone does not induce discernable lung inflammation (21, 42). Thus, this signal is likely a circulating factor released by endogenous lung cells. The identity of this signal is uncertain but merits study as a means of mobilizing DC precursors for use in immunotherapeutic strategies
Second, individual chemokines are produced during lung inflammation with a differential timing that correlates well with the observed kinetic of lung DC accumulation and that likely regulates recruitment into specific anatomic locations. There was rapid and robust increase in mRNA by real-time PCR for the CCR2 ligands CCL2 and CCL7, both of which peaked at day 1 and remained elevated before declining toward baseline by day 7. This rapid onset implies that these chemokines are also produced by endogenous lung cells. In contrast, production of mRNA for CCL20, the only known chemokine ligand of CCR6, was delayed. Human bronchial epithelial cells have been shown to produce CCL20 in response to inflammation (53). However, these kinetics also suggest production by recruited T cells, which peak in number at day 7 after challenge (42). Hence, full development of pulmonary immune responses likely depends on the sequential secretion of specific chemokines by distinct cell types.
To migrate from the bloodstream in response to chemokines, leukocytes must traverse the vascular endothelium. In most organs, this process requires initial rolling on the endothelial selectins CD62E and CD62P. Our novel finding that DC recruitment to the lung interstitium and alveoli is independent of endothelial selectins differs markedly from recruitment of lymphocytes in this model system (30, 32). Interestingly, we previously found that monocyte recruitment to the lungs is also independent of endothelial selectins (32). Selectin independence by DC could result from several possibilities, which are not mutually exclusive. DC transmigration could depend on cytoskeletal changes resulting in arrest not requiring endothelial selectins, as is the case of neutrophil transmigration of normal pulmonary capillary endothelium and of noncapillary lung endothelium in response to pneumococcus (54). DC might also use alternative mechanisms that mediate rolling under flow conditions, such as 4 integrins (55) or the stalked chemokine CX3CL1 (fractalkine) (56, 57). The differences in our findings from those of Pendl et al. (15), who showed partial endothelial selectin dependence in the recruitment of iDC to inflamed skin, most likely result from the disparity in tissues studied. However, it should also be noted that their study used bone marrow-derived iDC, whereas we examined recruitment of unmanipulated DC. Our findings predict that therapeutic agents based on blocking selectins (58, 59) will fail to impair DC recruitment to the lungs.
Endothelial transmigration and other steps involved in localization to the peribronchovascular space depend crucially on CCR2. This chemokine receptor has also been implicated in monocyte recruitment to the lung in response to LPS or CCL2 (60, 61). Thus, our data could indicate either that newly recruited lung DC share this requirement with monocytes, or that DC are derived from monocytes, a possibility neither we nor previous studies have specifically tested. These data on CCR2 dependence agree with recent findings in two models of granulomatous lung injury (19, 34), and extend them from the response to microbial products to nontoxic Ag deposited in the airspaces.
In contrast to the interstitium, DC must also cross an epithelial barrier to enter the alveoli, the site where inhaled Ag first deposits. Our data indicate that CCR6 and it sole chemokine ligand, CCL20, are important mediators of this process. Pulmonary epithelial tight junctions provide the majority of resistance to fluid transudation, and significantly, we have previously shown using monastral blue dye that this Ag challenge does not induce fluid leakage (21). Hence, leukocyte transmigration into the alveoli implies an active cooperation between alveolar epithelial cells and DC. CCR6 contributes to accumulation of the closely related Langerhans cell in the skin in both normal (62) and pathologic conditions (63, 64). Our study concentrated on recruitment of DC during lung inflammation, and did not evaluate effector functions. However, it is likely that the bronchial hyperresponsiveness and peribronchial eosinophilia seen in this model (23, 47) would be reduced, as reported in CCR6–/– mice (65). Importantly, the preserved recruitment of T cells (D. J. Linderman, J. Sonstein, S. W. Chensue, and J. L. Curtis, manuscript in preparation), a sensitive marker of Ag priming in the model system, argues against defective immunization, in agreement with previous data (65).
Collectively, our kinetic analysis provides in vivo evidence for a compartmental model in which DC navigate from bloodstream into tissues due to distinct temporal and spatial patterns of chemokine ligand/receptor expression. Specifically, we propose that CCR2 mediates early recruitment into lung interstitium, whereas CCR6 more centrally drives the later transit across epithelia into the airways. DC cultured from human peripheral blood sequentially express CCR2 before CCR6 (37). Chemokine expression is spatially segregated in resected human tonsils; a CCR2 ligand localized to vascular endothelium, whereas CCL20 expression was restricted to epithelial cells bordering the exterior (36, 37). Our results using gene-targeted mice imply that identical temporal and spatial patterns occur during Ag-driven lung inflammation.
The majority of newly recruited DC identified in our study were myeloid, based upon their expression of CD11b and absence of B220, and immature, as shown by low levels of costimulatory molecules. Lung DC did not express CD4 or CD8 homodimers, which form the basis for additional splenic DC subsets. A predominately myeloid lung DC phenotype has been found in several previous studies in both mice and humans (13, 66, 67, 68). However, these data contrast with the more mature phenotype seen after repetitive challenge with aerosolized Ag (10, 12, 69). Our model of particulate Ag challenge also found relatively few lung pDC, despite the use of two distinct methods to identify them, another difference from models using repetitive challenge with aerosolized Ag (45). These disparate findings could result from differences in the number of times the lungs are stimulated with Ag, the physical form (particulate vs aerosolized), or the time point at which DC were analyzed.
Mechanisms of lung DC accumulation in this model system other than recruitment are unlikely. We previously showed that interstitial lung mononuclear cells proliferate minimally in this model and that numbers of lung mononuclear cells did not decrease when cell division was blocked transiently using hydroxyurea (47). These data argue against a major component of in situ DC proliferation within the lungs. Given the clear difference we found in receptor expression and level of autofluorescence, it is unlikely that AM were mistakenly identified as DC. Moreover, our data agree closely with a recent study that used a similar approach to distinguish AM from DC, based on differences in autofluorescence and relative expression of CD11c and MHC class II (43). In that study, cell sorting of enzymatically digested lungs of normal C57BL/6 mice verified the marked differences in morphological, phenotypic, and Ag-presenting capacities of the two populations (43). We cannot exclude the possibility that part of the observed lung DC expansion in SRBC-challenged wild-type mice resulted from rapid differentiation of monocytes after arrival in the lungs. Although in vitro differentiation of monocytes into DC requires 5–7 days in medium containing GM-CSF plus IL-4 or IL-13 (70, 71), the process is accelerated in vitro by transendothelial migration (70, 72). How rapidly monocytes might differentiate into DC in the lungs is uncertain and will be an important goal of future studies. A population of CD11c–, CXCR1low, CCR2+ monocytes has recently been shown to be recruited to the inflamed peritoneum, where a large fraction differentiated to express CD11c and MHC class II within 18 h (57). Although recruitment was interpreted to be the major mechanism for changes in cell number in that study, a separate analysis of lung inflammation in response to Ag-coated beads identified an additional defect in maturation (i.e., up-regulation of MHC class II and CD40) in CD11c+ cells in CCR2–/– mice (35). The differences in the methods used to identify DC precludes direct comparison of those results with the current data.
In summary, we provide data that lung mucosal recognition of previously encountered Ags rapidly induces both local chemokine production and mobilization of immature myeloid DC precursors from the bone marrow. DC are recruited into the lungs in a process that is independent of endothelial selectins but which depends in part on CCR2 for entry into the peribronchovascular space, and on CCR2 and CCR6 for epithelial transmigration into the alveoli. These novel findings imply that DC recruitment from the bloodstream into the lung could be targeted therapeutically to control lung inflammation within specific anatomic compartments in such immunologically mediated diseases as asthma and chronic obstructive pulmonary disease.
Acknowledgments
We thank all of the members of the Ann Arbor Veterans Affairs Research Enhancement Award Program in Pulmonary Immunology for helpful suggestions and discussions, and Joyce O’Brien for secretarial support.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by a Career Development Award, Merit Review funding, and a Research Enhancement Award Program grant from the Biomedical Laboratory R&D Service, Department of Veterans Affairs; and by RO1 HL56309 and T32 HL07749 from the U.S. Public Health Service.
2 Portions of this work have been presented previously at the International Conference of the American Thoracic Society, Seattle, WA, May 19, 2003, and at the Keystone Symposium on "Dendritic Cells at the Center of Innate and Adaptive Immunity: Eradication of Pathogens and Cancer and Control of Immunopathology," Vancouver, BC, Canada, February 5, 2005; and have been published in abstract form (73 ).
3 Address correspondence and reprint requests to Dr. Jeffrey L. Curtis, Pulmonary and Critical Care Medicine Section (111G), Department of Veterans Affairs Medical Center, 2215 Fuller Road, Ann Arbor, MI 48105-2303. E-mail address: jlcurtis@umich.edu
4 Abbreviations used in this paper: DC, dendritic cell; iDC, immature DC; pDC, plasmacytoid DC; mod, moderate; i.t., intratracheal; BAL, bronchoalveolar lavage; M, macrophage; AM, alveolar M.
Received for publication September 28, 2004. Accepted for publication May 3, 2005.
References
Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, B. Pulendran, K. Palucka. 2000. Immunobiology of dendritic cells. Annu. Rev. Immunol. 18: 767-811.
Havenith, C. E., P. P. van Miert, A. J. Breedijk, R. H. Beelen, E. C. Hoefsmit. 1993. Migration of dendritic cells into the draining lymph nodes of the lung after intratracheal instillation. Am. J. Respir. Cell Mol. Biol. 9: 484-488.
Xia, W., C. E. Pinto, R. L. Kradin. 1995. The antigen-presenting activities of Ia+ dendritic cells shift dynamically from lung to lymph node after an airway challenge with soluble antigen. J. Exp. Med. 181: 1275-1283.
Holt, P. G., S. Haining, D. J. Nelson, J. D. Sedgwick. 1994. Origin and steady-state turnover of class II MHC-bearing dendritic cells in the epithelium of the conducting airways. J. Immunol. 153: 256-261.
Cella, M., F. Sallusto, A. Lanzavecchia. 1997. Origin, maturation and antigen presenting function of dendritic cells. Curr. Opin. Immunol. 9: 10-16
Banchereau, J., R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392: 245-252.
Sallusto, F., P. Schaerli, P. Loetscher, C. Schaniel, D. Lenig, C. R. Mackay, S. Qin, A. Lanzavecchia. 1998. Rapid and coordinated switch in chemokine receptor expression during dendritic cell maturation. Eur. J. Immunol. 28: 2760-2769.
Sallusto, F., A. Lanzavecchia. 1999. Mobilizing dendritic cells for tolerance, priming, and chronic inflammation. J. Exp. Med. 189: 611-614.(John J. Osterholzer*,, Th)