Abstract
When lungs of experimental animals are repeatedly challenged with Ag, pulmonary inflammation wanes via unknown mechanisms. We hypothesized that changes in the balance of lung cytokines are responsible for immune down-regulation to repeated Ag challenge. We used intratracheal (IT) challenge of primed C57BL/6 mice with SRBC and on various days after single (1IT) or triple (3IT) challenge counted lung inflammatory cells and measured whole-lung cytokine mRNA and protein concentrations using RT-PCR and ELISA. We found that lung lymphocyte numbers and parenchymal lung inflammation decreased significantly at days 6 and 9 after final Ag challenge in 3IT mice compared with 1IT mice. Lungs of 3IT mice showed the following changes in relative mRNA expression: an earlier peak in IL-10, decreased IL-1β, and a change from a Th2 response in 1IT mice to a Th1 response in 3IT mice (with pronounced increases in IL-12, IL-18, and IFN-γ and decreased IL-4, IL-13, and IL-5). Similar types of changes were seen in whole-lung protein concentrations for TNF-α, IL-10, IL-12 p40, IFN-γ, and IL-4. Additionally, mRNA expression of the endothelial selectins CD62E and CD62P decreased and lung lymphocyte apoptosis increased in the 3IT group. Thus, physiologic down-regulation of the pulmonary immune response to repeated Ag exposure is characterized by increased anti- and decreased proinflammatory cytokines that accompanies Th1 polarization. Similar mechanisms may act to minimize chronic lung inflammation in the majority of normal humans who do not develop progressive lung pathology when repeatedly exposed to inhaled or aspirated environmental Ags.
The pathogenesis of common immunologic lung diseases is known or suspected to involve inappropriately regulated pulmonary immune responses to inhaled Ags. Because many of the causative Ags are ubiquitous or unidentified, Ag avoidance cannot be relied upon as the sole therapeutic strategy. Current therapies for immunologic lung diseases are only partially effective and in many cases are toxic and expensive. To develop novel immunomodulatory therapies, the mechanisms that successfully regulate physiologic pulmonary immune responses must be defined in greater detail. Definition of these mechanisms cannot rely solely on extrapolation from in vitro studies but instead ultimately requires analysis of immune responses in the lungs of intact animals or patients.
To this purpose, we and others have analyzed the response of Ag-primed animals of a variety of species to intratracheal (IT)4 challenge with the prototypic particulate T cell Ag, the SRBC (1). The most prevalent model system uses inbred mice, predominantly of the C57BL/6 or A/J strains, which are high responders to SRBC (2, 3, 4). This experimental model system has been used to analyze the anatomic and molecular mechanisms of lymphocyte recruitment to lung parenchyma (5, 6, 7), the cytokine requirements for airways hyperresponsiveness (AHR) (8), and the role of neuropeptides in development of lung inflammation (9). Lung inflammation in this model system depends crucially on recruitment of functional CD4+ cells, which enter the lung via both VLA-4-dependent and selectin-dependent pathways (10, 11). Net lung lymphocyte accumulation depends minimally on in situ lymphocyte proliferation, but is associated with prominent lung eosinophilia (12). We have observed previously that inflammatory cell recruitment is accompanied by transient angiitis, alveolar macrophage activation, and in situ maturation of specific Ab-secreting cells; however fibrosis is not seen (5, 13). Thus, this experimental model system shares many of the characteristics of relevant human immunologic lung diseases. Importantly, however, the response is self-limiting within 3 wk (3, 5).
A key question in pulmonary immunology is why a minority of individuals develop progressive lung inflammation to ubiquitous inhaled agents that are ignored by the immune systems of most exposed individuals. One potential clue to this puzzle comes from animal models that indicate that the pulmonary immune response typically decreases spontaneously during repeated pulmonary Ag challenge of normal animals (14, 15, 16, 17). Thus, normal animals appear to mimic the natural response of most healthy humans to inhalation of noninjurious Ags. The molecular basis for this waning response during continued Ag exposure remains unknown despite considerable study. Evidence against a variety of mechanisms has been provided. The normally antiproliferative effect of alveolar macrophages on activated lymphocytes in vitro is not decreased on repeated challenge of rabbits or mice with Micropolyspora faeni (18, 19). Peripheral lymphocyte tolerance does not occur using multiple IT challenges of M. faeni in rabbits or Thermoactinomyces vulgaris Ag in mice (20, 21), and T suppresser activity is not enhanced on repeated challenge of mice with T. vulgaris Ag (21). However, because T cells can differentiate in vivo into immune or suppressor cells depending upon the nature and dosage of Ag (22, 23, 24), oral tolerance cannot be totally eliminated as an explanation for the waning response to multiple challenge.
We have previously found that termination of the pulmonary immune response to a single IT SRBC challenge does not require CD8+ T cells, but does involve considerable lymphocyte apoptosis within the alveolar space and lung interstitium (13, 25). We also found that, at the single time-point tested, the magnitude of lung inflammation and lymphocyte accumulation after three IT Ag challenges is markedly less than after a single IT Ag challenge (25). Previous experiments showed that multiple i.p. priming exposures to SRBC did not appreciably alter lung lymphocyte numbers after single IT challenge (12), excluding the trivial explanation of exhaustion of the systemic immune response or systemic tolerance. These findings suggested that a variation of our well-characterized model system could be studied to understand the mechanism of this previously described physiologic down-regulation seen in many animal models of pulmonary hypersensitivity. The goal of this study was to further characterize the down-regulation of lung inflammation and of lymphocyte accumulation on repeated IT Ag challenge, which we hypothesized would result from decreased production of proinflammatory cytokines within the lungs. We found that repeated Ag challenge was associated with up-regulation of antiinflammatory IL-10 as well as a down-regulation of E- and P-selectin mRNA. These changes were accompanied by a shift from a predominantly Th2 cytokine response to a distinct Th1 response.
Materials and Methods
Animals
Pathogen-free female C57BL/6J mice were obtained at 7–8 wks of age from Charles River Laboratory (Wilmington, MA). Mice were housed in the Animal Care Facility at the Ann Arbor Veterans Affairs Medical Center, which is fully accredited by the American Association for Accreditation of Laboratory Animal Care. This study complied with the National Institutes of Health “Guide for the Care and Use of Laboratory Animals” (Department of Health, Education, and Welfare Publication No. (National Insitutes of Health) 80-23). Mice were fed standard animal chow (Rodent Lab Chow 5001; Purina, St. Louis, MO) and chlorinated tap water ad libitum. Mice were used at 8–14 wks of age.
Monoclonal Abs
The following mAbs were purchased from PharMingen (San Diego, CA): RM4-4 (anti-murine CD4; rat IgG2b), 53-6.72 (anti-murine CD8; rat IgG2b), 1D3 (anti CD19), RM2-5 (anti CD2), 7D4 (anti-murine CD25, rat IgM), 145-2C-11 (hamster anti-murine CD3), and H1.2F3 (anti-murine CD69, hamster IgG). Monoclonal Abs were either directly FITC conjugated or biotinylated; in the latter case staining was visualized using streptavidin-PE (PharMingen). Isotype-matched irrelevant control Abs (107.3 (anti-TNP control mouse IgG1, κ), anti-TNP hamster IgG, and rat IgM, κ (PharMingen)) were tested simultaneously to exclude nonspecific staining.
Induction of pulmonary immune response
SRBC (sheep 4151) (Colorado Serum, Denver, CO) were washed three times in normal saline before use. Mice were Ag primed by i.p. injection with 1 × 108 SRBC in 0.5 ml normal saline. Beginning 2 wks later, mice were IT challenged with 5 × 108 SRBC in 50 μl normal saline as previously described (7) either once (1IT group) or three times with 1 wk between challenges (3IT group) (Fig. 1⇓). Two control experiments were performed (Fig. 1⇓). In control A, 4 wk elapsed between i.p. priming and IT challenge. This experimental group was used to control for the difference between the 1IT and 3IT groups in the time from initial priming to final IT challenge. In control B, the first two of the 3IT challenges used the same volume of normal saline rather than SRBC. This experimental group was used to control for the repeated immunization procedures to which the 3IT group was subjected. In a single additional control experiment not depicted in Fig. 1⇓, SRBC-primed mice were IT challenged a single time with 50 μl saline containing India ink (Pelikan, Hannover, Germany), and lungs were harvested 6 days later. This group was used to analyze the effect of the immunization procedure at the site of Ag deposition.
Experimental protocols. The horizontal axis denotes time in days, synchronized such that the day of the final IT challenge is day zero for each group. Arrows above the line denote Ag administration by the intraperitoneal (IP) or intratracheal (IT) route, whereas arrows below the line indicate days on which groups of mice were euthanized for assay.
Sample collection
At various times from 0–14 d after the final IT Ag challenge, mice were deeply anesthetized with pentobarbital (80 mg/kg i.p.) and killed by exsanguination and induction of bilateral pneumothoraces. Lungs were then processed in one of three ways. First, to obtain lung inflammatory cells, some lungs were perfused via the right heart using normal saline, and then bronchoalveolar lavage (BAL) and total and differential cell counts were performed as previously described (3). Aliquots of these cells were stained with mAbs and analyzed on a FACScan flow cytometer (Becton Dickinson, Mountain View, CA) as previously described (26), except that CellQuest software (Becton Dickinson) was used. Second, for histologic analysis some lungs were inflated using 1 ml 10% neutral-buffered formalin, and then the entire thoracic contents were dissected free and fixed by immersion in 10% formalin in PBS for 18–24 h. Parasagittal sections through these fixed lungs were then cut, embedded in paraffin, and sectioned at 5 μm thickness. The slides, each consisting of sections of both lungs from an individual mouse, were stained with hematoxylin and eosin and with Masson’s trichrome stains. Third, for samples to be analyzed for total lung cytokine mRNA or protein production, lungs that were perfused but not lavaged were carefully dissected to exclude any extrapleural lymphatic tissue. These lungs, separated from the mainstem bronchi at the medial pleural surface, were snap frozen in liquid nitrogen and stored at −70°C until processed as described below.
Apoptosis measurement
27) and exclude propidium iodide.
Isolation of RNA
Lungs were homogenized in 2 ml of TRIzol reagent (Life Technologies, Grand Island, NY), and RNA was isolated as described in the TRIzol protocol. RNA was quantitated spectrophotometrically. The integrity of individual RNA samples was confirmed by electrophoresing aliquots on a 2% agarose gel containing 0.5 μg/ml ethidium bromide and observing 28S and 18S rRNA bands. RNA samples were stored at −70°C.
RT-PCR detection of cytokine mRNA
Isolated RNA was reverse transcribed to DNA as follows. To 20 μg RNA was added 24.4 μl of mix A (403 μl RNase inhibitor (Life Technologies; 10 U/ml) and 280 μl of random primers (Life Technologies, 3 μg/ml)), which was heated at 70°C for 2 min in a programmable thermocycler (Perkin-Elmer 9600, Norwalk, CT). The temperature was reduced to 4°C and 68.9 μl of mix B (812 μl of double distilled water, 300 μl of 0.1 M DTT, 75 μl of 20 mM each dNTP, 602 μl of 5× first-strand buffer, and 140 μl of 200 U/μl Moloney murine leukemia virus reverse transcriptase (all from Life Technologies)) was added. RNA was transcribed in this mixture for 1 h at 42°C, and then the reaction was stopped by a 5-min incubation at 95°C.
The DNA was subjected to PCR as follows. In a thin-walled PCR tube, 5 μl of sample was added to 20 μl of PCR buffer containing 5 nM of each dNTP, 50 nM MgCl2, 1.0 U of Taq DNA polymerase (all from Life Technologies), and 0.15 ng each of sense and anti-sense primers. Amplification was then performed in a thermocycler as follows: 5 min at 95°C, followed by up to 35 cycles of 15 s each at 95°C, 20 s at 58°C, 30 s at 72°C. After cycling, there was a DNA extension period of 6 min at 72°C. The amplified DNA was analyzed by electrophoresis on 1.5% agarose containing 0.5 μg/ml ethidium bromide.
The primer sequences used are defined in Table I⇓. Primers were designed based upon nucleotide sequences downloaded from the National Center for Biotechnology Information data bank, using primer design software (Primer 2; Scientific & Educational Software, State Line, PA). All primers were prepared by the University of Michigan DNA Core Facility (Ann Arbor, MI). PCR conditions and cycle number were defined for each cytokine primer pair such that a linear relationship between input RNA and final PCR product (defined as OD signal) was obtained. Positive and negative controls were included in each assay to establish that only cDNA PCR products were detected and that none of the reagents were contaminated with cDNA or extraneous PCR products. Authenticity of reaction products was verified by Southern analysis using a chemiluminescent detection system (Amersham Life Science, Little Chalfont, U.K.).
Primer sequences used for PCR amplification
To quantitate cDNA bands, the ethidium bromide-stained agarose gels were photographed using Polaroid 667 film, scanned using a Scan Jet IIcx (Hewlett Packard, Palo Alto, CA), and analyzed on a Macintosh 8100/80AV computer using the public domain software program NIH Image (version 1.60; developed at the National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/). Results were expressed as a ratio of OD signal for a given cytokine to that for GAPDH, and fold increases were calculated relative to unchallenged mouse lung.
ELISA
Lungs frozen at −80°C were homogenized in 1 ml of PBS and centrifuged at 1500 rpm. The supernatant was aliquoted and stored at −80°C. These samples were analyzed for IFN-γ, IL-4, TNF-α, IL-12 p40, and IL-10 using PharMingen kits according the manufacturer’s instructions. OD readings of samples were converted to picograms per milliliter using standard curves generated with varying concentrations of recombinant cytokine supplied with the kit. The limit of detection was 31.2 pg/ml for each assay.
Statistical analysis
Data were expressed as mean ± SEM. Statistical calculations were performed using Statview (Abacus Concepts, Berkeley, CA) on a Macintosh PowerPC G3 computer. Differences between the groups of mice in continuous ratio scale data at individual time points after final IT challenge were evaluated using the unpaired Student t test (28). Differences in cytokine levels between the groups of mice were evaluated using Mann-Whitney nonparametric unpaired analysis. Significant differences were defined as p < 0.05.
Results
Lung inflammation wanes on repeated IT Ag challenge
Lung lymphocyte numbers peaked at day 6 after single Ag challenge and then progressively decreased, in agreement with previous data in this model system. Compared with the 1IT group, lymphocyte numbers were significantly decreased in the 3IT group on days 6 and 9 after final IT challenge, but not during early inflammation (day 3) or on day 14, when inflammation in both groups was largely resolved (Fig. 2⇓). Histologic analysis showed a decrease in the extent but no discernible change in the anatomic distribution of inflammatory cell accumulation after 3IT (Fig. 3⇓, A–D). Inflammatory cell accumulation in both groups was chiefly associated with the bronchovascular bundles and pulmonary veins, as previously reported (5, 7). Examination of trichrome-stained sections did not reveal any evidence of fibrosis. Histologic analysis of the lungs of mice challenged solely with saline containing India ink to detect the site of Ag deposition did not disclose inflammatory cell accumulation (Fig. 3⇓, E and F). Thus, kinetic analysis indicated that, in this model system as in other models (14, 15, 16, 17, 18, 20, 21), repeated intrapulmonary Ag deposition led to a diminution of the lung inflammatory cell response.
Kinetics of lung lymphocyte accumulation. Lung lymphocytes were collected by BAL of SRBC-primed C57BL/6 mice at various days after final IT SRBC challenge. 1IT, ▪; 3IT, ⋄. Absolute lymphocyte numbers were determined as the product of total BAL cell count (by hemocytometer) and lymphocyte percentage on differential cell count of hematoxylin and eosin-stained preparations. Data represent mean ± SEM of at least three mice assayed in two to five experiments per time point; ∗, p < 0.05, unpaired Student t test.
Histologic evidence of waning lung inflammation on repeated IT Ag challenge. A–D, Representative sections of lungs of SRBC-primed C57BL/6 mice 6 days after final IT challenge of 1IT (A and B) or 3IT (C and D) mice. A, Note dense cellular accumulation in a predominantly perivenous and periarterial location. B, High-power view of perivenous infiltrate, showing mononuclear cell predominance and activated endothelial cells (arrow). C, Markedly less dense cellular accumulation in 3IT mouse. D, High-power view of mononuclear cell infiltrate adjacent to venule. E and F, Sections of lungs of SRBC-primed C57BL/6 mice 6 days after IT challenge with saline containing India ink to mark sites of Ag deposition. E, Low-power view showing virtual absence of inflammation induced by saline challenge alone. F, High-power view of alveoli containing macrophages filled with India ink particles (arrows). Note absence of inflammatory cell accumulation. Hematoxylin and eosin stain; magnifications: A, C, and E, ×100; B, D, and F, ×1000 (oil immersion).
In control experiments that compared timing of the initial i.p. Ag priming at 2 wk (Fig. 1⇑, 1⇑ IT) or 4 wk (Fig. 1⇑, Control A) before IT SRBC injection, there were no significant differences in numbers of total BAL inflammatory cells, macrophages, or lymphocytes (data not shown). There were also no differences in BAL cell numbers or types when comparing experiments with 1IT challenge and 3IT challenge where the first two injections were saline (Fig. 1⇑, Control B) (data not shown). These results indicate that multiple surgeries or variations in timing of i.p. injections were not the causes of differences between 1IT- and 3IT-challenged mice in lung lymphocyte numbers.
Lung lymphocyte phenotypes change minimally after repeated Ag challenge
Flow cytometric analysis of lung lymphocyte phenotypes showed significant differences only at day 9 after final IT challenge, when 3IT mice had a decreased percentage of CD19+ cells and increased percentage of CD8+ cells (Fig. 4⇓). Overall, there was a trend toward an increasing fraction of CD4+ and toward a decreasing fraction of CD19+ cells in the 3IT group, both with time after challenge and relative to the 1IT group, but these differences did not attain statistical significance. We and others have previously reported that a large fraction of lymphocytes in the lungs of immunized mice are negative for most subpopulation markers (3, 29). In the current study, the percentage of such null cells (defined as CD2+, CD4−, CD8−, and CD19− cells within light-scatter gates characteristic of lymphocytes) did not differ between the two groups. Thus, no systematic change in the distribution of lung lymphocyte subsets between the two groups was consistently seen over the course of the immune response.
Lung lymphocyte phenotypes. Phenotypes of BAL lymphocytes, collected as described in Fig. 2⇑, were determined by direct immunofluorescence staining and flow cytometry. 1IT, dark hatching, 3IT, light hatching. Data are expressed as the percentage of each phenotype at various days after final IT challenge. Note differences in scales. Each point represents the mean and SEM of four to eight individual mice assayed in two experiments for days 6 and 9 or three samples of pooled cells from at least six mice each for day 14. ∗, p < 0.05, unpaired Student t test.
To investigate the degree of T cell activation in the two treatment groups, in some experiments CD4+ and CD8+ cells were double stained with mAb against CD25 or CD69 (Table II⇓). In all but two of the comparisons, a greater percentage of both CD4+ cells and CD8+ cells were positive for these activation receptors in the 3IT group, although the difference was statistically significant for only five of the 12 conditions tested.
Lung T cell activation receptor expression increases on repeated Ag challenge
Down-regulation of lung inflammation on repeated Ag challenge is characterized by a decrease in proinflammatory cytokines and an increase in IL-10
To determine which cytokines may be involved in down-regulation of the immune response in 3IT-challenged mice, we determined relative levels of whole-lung cytokine mRNA by RT-PCR and of selected whole-lung cytokine protein concentrations by ELISA. In the remainder of this paper, significant differences between 1IT and 3IT mice are only mentioned if results of 1IT as well as control A and control B mice are similarly significantly different from those of 3IT mice. In the 1IT groups, relative mRNA levels of the proinflammatory cytokine IL-1β was actually elevated at the time of initial IT challenge, perhaps reflecting residual changes from previous i.p. priming (Fig. 5⇓). IL-1β levels were significantly higher in the 1IT and control groups relative to 3IT at day 1. TNF-α mRNA levels were greater in 1IT and control groups than 3IT levels at early time points but did not reach significance. Relative mRNA levels of the antiinflammatory cytokine IL-10 peaked earlier in 3IT mice, at day 1 compared with days 3–4 in 1IT and control (Fig. 5⇓). Relative mRNA levels of IL-6 were not significantly different between 3IT and 1IT and control groups at any time point.
Kinetics of total lung mRNA expression of pro- and antiinflammatory cytokines. Total lung RNA was extracted from SRBC-primed C57BL/6 mice on various days after final challenge. RNA was reverse-transcribed, and aliquots of cDNAs from those lungs were amplified by PCR for expression of IL-1β, TNF-α, IL-10, and IL-6 as described in Materials and Methods. Photographs of UV-transilluminated ethidium bromide-stained agarose gels were digitized. 1IT, ▪; 3IT, ⋄ and dark line; control A, ○; control B, ▵. Results are expressed as a ratio of OD signal for a given cytokine to that for GAPDH, and fold increases were calculated relative to normal mouse lung. Note the difference in scale between individual cytokines. Each point represents the mean and SEM of at least three mice. ∗, p < 0.05, unpaired nonparametric Mann-Whitney test comparing groups at the same time point.
At the protein level, TNF-α was significantly increased in 1IT mice compared with 3IT mice at days 1, 6, and 9 after final IT Ag challenge (Fig. 6⇓). The protein data for IL-10 also indicated a stronger early response in 3IT mice although the levels were significantly higher at later time points as well (Fig. 6⇓). In combination, these results indicate that the down-regulation of immune response in 3IT-challenged mouse lungs is characterized by a decrease in proinflammatory cytokines and an increase in IL-10.
Changes in whole-lung cytokine protein concentrations on repeated IT Ag challenge. Lungs of SRBC-primed C57BL/6 mice were harvested at various days after final IT Ag challenged, homogenized, and assayed by ELISA. 1IT, ▪; 3IT, ⋄; N, normal mice. Each point represents the mean and SEM of at least three mice assayed individually. ∗, p < 0.05, unpaired nonparametric Mann-Whitney test comparing groups at the same time point.
Down-regulation of lung inflammation on repeated Ag challenge is associated with polarization to a type 1 response
To define the nature of the T cells involved in the reduction of the immune response in 3IT mice, we examined expression of mRNA of cytokines associated with Th1 (IFN-γ, IL-12 p40, IL-12 p35, and IL-18) (Fig. 7⇓) and Th2 (IL-4, IL-5, and IL-13) responses (Fig. 8⇓). In 1IT and both groups of control mice, aside from an isolated relative mRNA expression of IL-12 p35 at baseline, levels of both species of IL-12 and of IL-18 were minimal throughout the response to IT Ag challenge. IFN-γ mRNA was increased only at day 1 in 1IT and control mice, and then fell below the levels found in the 3IT group. Unlike this minimal evidence of a Th1 response in the 1IT and control groups, there was a striking predominance of relative mRNA expression of the Th2 cytokines in these groups: IL-4 and IL-13 peaking at day 3 and IL-5 having high relative levels on days 1–3 after IT challenge. By contrast, the 3IT group showed markedly elevated expression of relative mRNA levels of all four Th1 cytokines, and especially of IL-18. Conversely, there was little expression of any of the three Th2 cytokines in the 3IT group.
Total lung Th1 mRNA cytokine expression on repeated IT Ag challenge. Total lung RNA was extracted from SRBC-primed C57BL/6 mice on various days after final challenge. RNA was reverse-transcribed, and aliquots of cDNAs from those lungs were amplified by PCR. Symbols are as in Fig. 5⇑. Each point represents the mean and SEM of at least three mice in which the ratio of target cytokine gene signal to GADPH signal was determined and expressed relative to the ratio for unchallenged normal lung. Note the difference in scale between individual cytokines. ∗, p < 0.05, unpaired nonparametric Mann-Whitney test comparing groups at the same time point.
Total lung Th2 mRNA cytokine expression on repeated IT Ag challenge. Total lung RNA was extracted from SRBC-primed C57BL/6 mice on various days after final challenge. RNA was reverse-transcribed, and aliquots of cDNAs from those lungs were amplified by PCR. Symbols are as in Fig. 5⇑. Each point represents the mean and SEM of at least three mice in which the ratio of target cytokine gene signal to GADPH signal was determined and expressed relative to the ratio for unchallenged normal lung. Note the difference in scale between individual cytokines. ∗, p < 0.05 unpaired nonparametric Mann-Whitney test comparing groups at the same time point.
Expression of the important T cell cytokines IL-2 and IL-15 was also examined. IL-2 mRNA levels peaked sharply in 1IT mice at day 1 and then fell throughout the remainder of the response (Fig. 7⇑). Significantly higher relative mRNA levels of IL-2 were seen at day 1 in 1IT and control groups compared with 3IT mice. IL-15 is member of the four-helix bundle cytokine family that mediates its function via the β- and γ-chains of the IL-2 receptor and that shares many of IL-2’s T cell activation and chemoattractant properties (30). Levels of IL-15 mRNA were increased early after Ag challenge in all groups, with no significant differences between them (Fig. 7⇑).
Whole-lung protein levels of selected cytokines analyzed by ELISA agreed with the patterns detected by mRNA analysis. The Th1 cytokines IL-12 p40 and IFN-γ showed significantly elevated concentrations at many time points in the 3IT mice relative to the 1IT group (Fig. 9⇓). Whole-lung cytokine protein concentrations of IL-4 agreed with the mRNA data, with a distinct peak of activity on day 4 in the 1IT mice and lower concentrations in the 3IT group (Fig. 9⇓). Taken together, these data demonstrate a polarization of the response from a mixed response with features of a Th2 response after one Ag challenge to a distinctly Th1 response after repeated IT Ag challenge.
Kinetics of whole-lung Th1 and Th2 cytokine protein concentrations on repeated IT Ag challenge. Lungs of SRBC-primed C57BL/6 mice were harvested at various days after final IT Ag challenged, homogenized, and assayed by ELISA. Note differences in scales. 1IT, ▪, 3IT, ⋄. N, normal mice. Each point represents the mean and SEM of at least three mice assayed individually. ∗, p < 0.05, unpaired nonparametric Mann-Whitney test comparing groups at the same time point.
Potential determinants of lung lymphocyte accumulation
We next sought to examine several potential determinants of net lung lymphocyte accumulation. The endothelial selectins E-selectin (CD62E) and P-selectin (CD62P) are cell adhesion molecules that initiate recruitment of type 1 T cells to sites of inflammation (31). Expression of these molecules is induced by TNF-α and IL-1β and is transcriptionally regulated. We have previously found that expression of these selectins markedly increases after IT Ag challenge in this model system (7) and that lymphoblasts unable to bind to these molecules (due to genetic targeting of fucosyltransferase VII, the rate-limiting enzyme in synthesis of leukocyte ligands for the selectins) have significantly decreased short-term recruitment to the lung (11). In the current study, analysis of relative mRNA levels by RT-PCR showed slightly decreased transcription of the genes for E-selectin and P-selectin in the 3IT group compared with 1IT and control groups (Fig. 10⇓).
Expression of endothelial selectin genes decreases after repeated IT Ag challenge. Total lung RNA was extracted from SRBC-primed C57BL/6 mice on various days after final challenge. Symbols are as in Fig. 5⇑. Each point represents the mean and SEM of at least three mice in which the ratio of target cytokine gene signal to GADPH signal was determined and expressed relative to the ratio for unchallenged normal lung. ∗, p < 0.05, unpaired nonparametric Mann-Whitney test comparing groups at the same time point.
Another mechanism that could decrease steady-state lymphocyte numbers is increased death by apoptosis. We have previously shown in this experimental model system that lung lymphocyte apoptosis contributes to the decrease in lung inflammation with time after Ag challenge, and that mice genetically lacking Fas (CD95) or its ligand do not down-regulate lung inflammation to repeated Ag challenge (25). Analysis using annexin V binding and propidium exclusion showed a statistically significant increase in BAL lymphocyte apoptosis at day 6 in 3IT mice compared with 1IT mice (Fig. 11⇓). Thus, increased lymphocyte apoptosis within the lungs also appears to contribute to the physiologic down-regulation of lung inflammation during repeated IT Ag challenge.
Lung lymphocytes apoptosis. Apoptosis of BAL lymphocytes, collected as described in Fig. 2⇑, was determined by annexin V-FITC staining and flow cytometry. 1IT, dark hatching, 3IT, light hatching. Each bar represents the mean and SEM of four to seven individual mice for days 6 and 9 or five samples of pooled cells from at least six mice each for day 14. ∗, p < 0.05, unpaired Student t test.
Discussion
The major finding of this study is that repeated IT challenge of primed mice with particulate Ag results in prompt, marked, and coordinated changes in whole-lung expression of multiple cytokines. Triple IT Ag challenge was associated with decreased proinflammatory and increased antiinflammatory cytokines and a distinct switch to a Th1 response. In our studies, cytokine changes occurred at least in part at the level of gene transcription and were reflected in parallel changes in whole-lung expression of the five cytokine proteins examined. There was also a small decrease in the transcription of the endothelial selectins, which are necessary for one pathway of lung lymphocyte recruitment (7, 11). As demonstrated in Figs. 5⇑, 7⇑, 8⇑, and 10⇑, these changes occurred by day 3 after the third Ag exposure, while the magnitude and duration of lung inflammation, as shown in Fig. 2⇑, was still only mildly decreased below that seen with single Ag challenge, whereas the major decrease in inflammation occurred at day 6. Hence, it is plausible that the cytokine changes are a cause, rather than effect, of down-regulation. These novel findings provide a mechanistic explanation for the diminished lung inflammation on repeated intrapulmonary Ag exposure, which has been described in numerous animal model systems (14, 15, 16, 17, 18, 20, 21).
Murine model systems have become widely used in the study of pulmonary immunity, immunopathology, and host defense (32, 33, 34). For this purpose, SRBC Ag, used since the infancy of cellular immunology because it induces such a vigorous specific immune response in mice, has several advantages as a model Ag. SRBC are nonreplicating particles of a size that can be delivered reproducibly to the distal airspaces of the lungs. In primed mice, such challenge induces time- and Ag-dose-dependent lung inflammation (2, 9). By contrast, in the absence of Ag priming, challenge with even the large Ag doses used in the current study (near the maximum that can be administered without excessive numbers of mouse deaths during anesthesia) induce very little lung inflammation (35). Thus, although SRBC is not itself a clinically relevant allergen, it is likely the cytokine changes we found will be seen with other Ags that also do not directly activate the innate immune response or cause tissue damage. The lung inflammation seen in SRBC-challenged mice most closely resembles the human lung disease hypersensitivity pneumonitis, which is, however, a CD8-predominant condition resulting from chronic inhalation of organic Ags. The distribution and extent of peribronchovascular mononuclear cell infiltration in the SRBC model system is also similar to that seen in mice challenged with such Ags as OVA or Asperigillus (17, 36, 37), but features of asthma, such as epithelial cell damage or mucus hypersecretion, are not observed in the SRBC model in C57BL/6 mice.
The significant increase in lung IL-10 production seen in the 3IT group suggests one potential unifying mechanism for down-regulation of lung inflammation to repeated Ag challenge. IL-10 is a 178-aa glycoprotein expressed as noncovalently linked homodimers that appears to act as a naturally occurring antiinflammatory cytokine (38). IL-10 inhibits the release of IL-1β, TNF-α, macrophage inflammatory protein-1α, and monocyte chemoattractant protein-1 and prevents Ag-specific Th1 cell proliferation and cytokine production both directly and by reducing the Ag-presenting capacities of monocytes (39, 40). However, to what degree defective production of counterregulatory antiinflammatory cytokines such as IL-10 participates in pathogenesis of lung disease is controversial. Initially identified as a Th2 T cell product, IL-10 is now recognized to be produced by a host of cell types, including alveolar macrophages and bronchial epithelial cells (41, 42). The potential role of IL-10 to suppress lung inflammation has been shown using knockout mice in murine models of allergic bronchopulmonary aspergillosis and hypersensitivity pneumonitis (37, 43), in Pseudomonas aeruginosa-induced mouse lung damage during repeated challenge (44), and in a rat model of lung response to quartz (45). Because endogenous lung IL-10 levels increase early during multiple IT SRBC challenge, our results suggest that IL-10 may be playing a role in the down-regulation of immune response to repeated particulate Ag challenge. Confirmation of that possibility awaits further experimentation. However, recent studies indicate that IL-10 may not be required for down-regulation of mouse lung inflammation associated with CpG DNA treatment of LPS-induced injury (46), with blockade of the chemokine C10 in experimental allergic bronchopulmonary aspergillosis (47), or during resolution of replicative Legionella pneumophila infection (48). Hence, the importance of IL-10 in down-regulation of lung inflammation appears to vary with the experimental model system.
The finding that IL-12 p40 mRNA and protein expression increased on repeated Ag challenge is significant because IL-12 plays an important role in both innate resistance and Ag-specific immunity to intracellular bacterial, fungal, viral, and protozoan pathogens. Moreover, a variety of protocols that down-regulate allergic manifestations against soluble Ags, including infections with Mycobacterium tuberculosis or adenovirus, administration of unmethylated CpG oligodeoxynucleotides, and allergen-injection immunotherapy, both up-regulate IL-12 and switch T cells toward a Th1 phenotype (49, 50, 51, 52). Endogenous production of IL-12 is essential for the resistance of some murine strains to OVA-induced AHR (53). Low-dose intranasal IL-12 decreased Ag-induced AHR and lung eosinophilia in an OVA model of asthma in BALB/c mice (54). Interestingly, Hofstra and colleagues found in another OVA model in BALB/c mice that IL-12 plus IL-18, but not IL-12 alone, was required to significantly inhibit Ag-induced lung eosinophilia, AHR, and elevated serum IgE (55). Therefore, lung inflammation involving predominance of Th2 cytokines followed by recovery associated with up-regulation of Th1 cytokines appears to be well established in response to soluble Ags such as OVA. However, many Ags of clinical significance are of a particulate nature, and with these results are not as consistent. For example, human and animal studies of hypersensitivity pneumonitis, a granulomatous inflammatory lung disease caused by inhalation of particulate Ag (most commonly thermophilic actinomycetes) indicate that Th1 cytokines predominate during infection (56, 57). In contrast, particulate Asperigillus fumigatus Ags elicited a Th2 response in BALB/c mice (58, 59). Single IT SRBC challenge of primed A/J mice induced AHR and lung eosinophilia in a CD4 T cell-dependent fashion reminiscent of human asthma (4). In that model system, repeated systemic administration of exogenous IL-12 at the time of a second IT SRBC challenge abolished AHR and lung eosinophilia, increased IFN-γ, and decreased IL-4 and IL-5 expression; these effects of IL-12 were only partially dependent on IFN-γ (8). The current study extends these results by showing spontaneous production within the lungs of IL-12, IL-18, and IFN-γ in response to repeated particulate Ag challenge.
The dose of SRBC used in this study, although seemingly large, is in the same range used by others measuring immune response to particulate cellular Ag (4, 35, 60, 61). In contrast to soluble Ags, generation of pulmonary immune responses to particulates typically requires large Ag doses, presumably reflecting evolutionary pressure for alveolar macrophages to eliminate particles without invoking lung inflammation. Nevertheless, the immune response to particulates merits investigation due to the large quantities of particulates to which even normal individuals are repeatedly exposed focally through aspiration (62) and more insidiously by inhalation (63). With regard to the effect of dose on Th2/Th1 polarization, published results are contradictory. Some studies have indicated that low-dose soluble Ag preferentially supports IL-4 over IFN-γ production (64, 65, 66, 67, 68, 69). A recent study using transgenic T cells bearing a single TCR indicated that over the long term, high doses of high-affinity peptides led selectively to IFN-γ-secreting cells whereas IL-4- and IL-5-secreting cells predominated with moderate doses (70). In contrast, with intact organisms such as Leishmania major, Trichuris muris, and Schistosoma mansoni, a greater degree of Ag stimulation induced a Th2 response, whereas single immunizations or low-level infection favored Th1 cytokines (71, 72, 73). The cytokine patterns seen in our 1IT and control groups are consistent with results of a recent study that found that a high dose of SRBC (4 × 108 cells) given i.v. to naive CBA/J mice induces either a mixed Th1/Th2 or a predominantly Th2 response, in contrast to the Th1 response seen with lower doses (4 × 105 or 4 × 106 cells) (61). Because Th1/Th2 polarization can also be influenced by other variables including strength of TCR signal (65, 74), amount of costimulation (75), nature of cytokines present (76), and T cell responsiveness to particular cytokines (77), it is currently difficult to predict accurately the effect of a given immunization protocol. The results of the current study are important because they provide additional information about the pulmonary immune response to nonreplicating particulate Ag, and because they imply that route of Ag administration should also be considered.
There are probably multiple causes for the observed decrease in lung lymphocyte accumulation in the 3IT group. Based on previous findings in this model system, decreased expression of endothelial selectins and increased lymphocyte apoptosis would both be anticipated to have major effects on steady-state lymphocyte accumulation and hence the severity and duration of lung inflammation. It is likely that lymphocyte retention within lung parenchyma is also decreased after repeated Ag challenge; however, we have not yet measured this parameter. We also did not measure local lymphocyte proliferation within the lungs, reasoning that it would be difficult to account for significant net changes between the groups based on further reduction of the already very low rate of proliferation previously seen in the 1IT protocol (12).
Several points about our results merit discussion in relationship to previous studies using this experimental model system. Kaltreider and colleagues demonstrated expression of cytokine proteins (IL-2, IL-4, IL-6, and IFN-γ) as a function of time after a single IT SRBC challenge (9). Because that study analyzed concentrated BAL fluid, whereas ours analyzed total lung digests, the results are not directly comparable. However, both studies found peak IL-4 production at 3–4 days after challenge. Our finding of decreased lymphocyte and total leukocyte numbers in the 3IT group differ from results of Denis and Bisson (35). In addition, they found striking fibrosis in mice challenged three times, which was not observed in our studies. There are several methodological differences between our study and theirs, including control groups (they compared 3IT only to primed mice without IT challenge and to unprimed mice after single IT challenge), interval between IT challenges (10-day rather than 7-day intervals as is our studies), and method of cytokine analysis (cytokines were analyzed in the supernatants of purified BAL T cells stimulated in vitro with SRBC, whereas we measured protein levels in unstimulated lung cell digests). Despite these methodological differences, both the Denis and Bisson and our study found that protein concentrations of IFN-γ increased on repeated Ag challenge, consistent with evolution of a Th1 response. In agreement with our findings, most other animal models of repeated stimulation using exogenous Ag have found both decreased inflammation and distinct difficulty in inducing fibrosis (14, 15, 16, 18, 20, 21).
In summary, this study demonstrates that, relative to single Ag challenge, repeated IT Ag challenge of primed mice is associated with significantly decreased lymphocyte accumulation and significant changes in total lung production of mRNA and proteins. These changes are down-regulation of inflammatory cytokines, up-regulation of the antiinflammatory cytokine IL-10, as well as striking Th1 polarization. These cytokine changes are associated with decreased transcription of the genes for the inducible endothelial selectins and by increased lung lymphocyte apoptosis. These results are important because similar mechanisms may act to minimize chronic lung inflammation in the majority of normal humans who do not develop progressive lung pathology when repeatedly exposed to inhaled or aspirated particulate environmental allergens and Ags.
Acknowledgments
We thank Dr. Richard E. Goodman (Monsanto, St. Louis, MO) for the primer and probe sequences for GAPDH, GM-CSF, IL-2, IL-4, IFN-γ, and IL-12 p35; Dr. Lloyd Stoolman for recombinant human IL-2; Angela Preston, Kelly Warmington and Drs. James M. Beck, Paul Christensen, Gary Huffnagle, and Bethany Moore for helpful suggestions; Joyce O’Brien for secretarial support; Michael Hormuth and Carolyn White for assistance with the photomicrographs; and Drs. Beck and Moore for critiquing the manuscript.
Footnotes
-
↵1 This work was supported by research funds and a REAP center from the Department of Veterans Affairs, Grant RO1 HL56309 from the U.S. Public Health Service, and by the University of Michigan Medical School Student Biomedical Research Fund (to G.D.S.). J.L.C. is a Career Investigator of the American Lung Association of Michigan.
-
↵2 Portions of these data were presented at the International Scientific Conference of the American Thoracic Society (San Diego, CA; April 27, 1999) and have been published in abstract form (1999 Am. J. Respir. Crit. Care Med. 159:A660).
-
↵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: E-mail address: jlcurtisumich.edu
-
↵4 Abbreviations used in this paper: IT, intratracheal; 1IT, group of mice which received a single Ag challenge; 3IT, group of mice which received three Ag challenges at weekly intervals; AHR, airways hyperresponsiveness; BAL, bronchoalveolar lavage.
- Received June 28, 1999.
- Accepted February 3, 2000.
- Copyright © 2000 by The American Association of Immunologists
References
In this issue
