Proinflammatory responses generated by T helper type 1 (Th1) cells may contribute significantly to immune-mediated lung injury. We describe a murine model of Th1 cell-induced lung injury in which adoptive transfer of alloreactive Th1 cells produces pulmonary inflammation characterized by mononuclear cell vasculitis, alveolitis, and interstitial pneumonitis. To investigate the link between activation of Th1 cells in the lung and inflammatory cell recruitment, we characterized cytokine and chemokine mRNA expression in Th1 cells activated in vitro and in lung tissue after adoptive transfer of Th1 cells. Activated Th1 cells per se express mRNA for interferon (IFN)-γ and several members of the tumor necrosis factor family as well as the C-C chemokine receptor-5 ligands regulated on activation normal T cells expressed and secreted and macrophage inflammatory protein-1α and -1β. Additional chemokine genes were induced in the lung after Th1 cell administration, most notably IFN-γ-inducible protein (IP-10) and monokine induced by IFN-γ (MIG). Remarkable increases in IP-10- and MIG-immunoreactive proteins were present in inflammatory foci lung and identified in macrophages, endothelium, bronchial epithelium, and alveolar structures. The findings suggest that IFN-γ-inducible chemokines are an important mechanism for amplifying inflammation initiated by Th1 cells in the lung.
- T helper type 1 and type 2 cells
- in vivo animal models
the influx of cells to the lung during inflammatory and immune responses is dictated by a series of signals including chemokines, cytokines with chemotactic activity that are known to direct cell migration to inflammatory sites. The chemokines can be subdivided into families on the basis of the position of NH2-terminal cysteine residues. The α- and β-chemokine families are among the best characterized. The α-chemokine family has one amino acid separating the first two cysteine residues (C-X-C), whereas the β-chemokines have no intercalating amino acids (C-C chemokines). The amino acid sequence has important functional implications. α-Chemokines with a glutamic acid-leucine-arginine residue preceding the C-X-C sequence are chemotactic for neutrophils [e.g., interleukin (IL)-8], whereas those lacking this sequence [such as interferon (IFN)-γ-inducible protein (IP-10) and monokine induced by IFN-γ (MIG)] act predominantly on lymphocytes. The β-chemokines [including monocyte chemotactic protein-1, macrophage inflammatory protein (MIP)-1α, MIP-1β, and released on activation normal T cells expressed and secreted (RANTES)] attract monocytes, eosinophils, and lymphocytes with variable selectivity (17, 23). The selective attraction of leukocytes in response to chemokines may be determined by the expression of specific chemokine receptors on their cell surface. For example, neutrophils express receptors for C-X-C chemokines (e.g., CXCR1 and CXCR2) but not for C-C chemokines. Monocytes express receptors for C-C chemokines. Specific chemokine receptors are also associated with subsets of T cells. CXCR3, the receptor for IP-10 and MIG, is highly expressed on Th1 cells but not on Th2 cells (24).
Expression of specific chemokine molecules helps determine the pattern of inflammatory response in tissues. A recent study (14) has begun to elucidate the importance of chemokines in a number of immune and inflammatory lung diseases and the possible relationship of specific chemokines to Th1- and Th2-induced responses. In asthma, a disease in which Th2 cells are important effectors, a number of chemokines including eotaxin, MIP-1α, and monocyte chemotactic proteins are increased. These molecules appear to have a key role in governing the recruitment of specific inflammatory cells, including eosinophils, to the lung (3, 11, 30). The chemokines induced by Th1 effector cells are less well understood, although there is evidence to suggest induction of the C-C chemokine IP-10 in diseases with Th1 cell involvement such as sarcoid and tuberculosis (1, 22).
Chen et al. (4) and Clark et al. (6) recently described a murine model of Th1 cell-induced lung injury in which adoptive transfer of cloned alloreactive Th1 cells produces selective pulmonary inflammation. The alloreactive Th1 cells recognize Ly5, an antigen expressed exclusively on hematopoietic cells. Two forms of Ly5 exist in mice, Ly5a and Ly5b. Adoptive transfer of Ly5a-specific Th1 cells into Ly5a mice but not into Ly5b mice produces lung inflammation characterized by mononuclear cell vasculitis, alveolitis, and interstitial pneumonitis. The adoptively transferred cells are preferentially localized in lung (7). However, the inflammatory cell response is composed mainly of recipient-derived cells (6). Localization of Th1 cells and the mechanisms by which they initiate an inflammatory response in lung are incompletely understood.
We used this model of Th1-induced lung injury to investigate the link between activation of Th1 cells and inflammatory cell recruitment to the lung, specifically with regard to expression of chemokines. We characterized the cytokines and chemokines produced by in vitro activated Th1 cells per se as well as chemokine expression in lung tissue after Th1 cell transfer. Our findings implicate both Th1 cell-derived chemokines (i.e., MIP-1α, MIP-1β, and RANTES) and IFN-γ-inducible chemokines expressed in the lung subsequent to Th1 cell activation. In lung tissue, both IP-10 and the closely related chemokine MIG were markedly induced and were associated immunohistochemically with inflammatory foci. These findings suggest that inducible chemokine pathways may represent an important mechanism for amplifying the inflammatory response initiated by activated Th1 cells.
MATERIALS AND METHODS
C57BL/6 (Ly5b) and Ly5a mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were housed in microisolator cages under specific pathogen-free conditions with free access to sterile water and chow.
T cell clone and adoptive transfer.
T cell clones specific for the Ly5a allele were developed and maintained in culture as previously described (4). In brief, Ly5b mice were immunized with 13-mer Ly5a peptides. CD4+ Th cells specific for the 13-mer peptides were elicited and cloned by limiting dilution. The T cell clone used in this study was of the Th1 phenotype and produced IFN-γ and tumor necrosis factor (TNF)-α. The cells were maintained by periodic stimulation with Ly5a peptide in the presence of congenic irradiated splenocytes and were maintained in the presence of IL-2 (10 U/ml). Activated cells were harvested 1 day after stimulation, and resting cells were harvested 14 days after stimulation.
In adoptive transfer studies, 107 Ly5a-specific Th1 cells were administered to Ly5a or Ly5b (control) recipient by tail vein injection.
RNase protection assay.
To analyze RNA in mouse lungs after administration of Th1 cells, Ly5a and Ly5b (control) mice (3 mice/group) were killed at intervals after adoptive transfer of cells. The lungs were harvested and snap-frozen in liquid nitrogen. Total RNA was extracted from the lungs and also from Th1 cells by the TRIzol method (Life Technologies, Grand Island, NY). Gene transcripts were detected by RNase protection assay (RPA) according to the manufacturer's directions (Riboquant, PharMingen, San Diego, CA). In brief, 32P-labeled complementary RNA transcripts were generated from a DNA probe set including templates for mouse cytokines or chemokines and housekeeping proteins [L32 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH)]. RNA samples (10 μg) were hybridized to the labeled transcripts for 16 h at 56°C. The nonhybridized RNA was digested by a mixture of RNase A and T1. The hybridized RNA was resolved on a 4.75% polyacrylamide gel. The32P-labeled bands were visualized by exposing phosphorimager plates (Molecular Dynamics, Sunnyvale, CA) for 18 h. The plates were scanned with a phosphorimager, and the relative quantities of hybridized probe were determined with ImageQuant software (Molecular Dynamics). The signal intensity for chemokine mRNA bands was divided by the signal intensity for L32 or GAPDH in the same lane to control for variation in the quantity of RNA loaded.
Total cellular RNA was isolated from aliquots of frozen right lung by a modification of the method of Chirgwin et al. (5) with CsCl density gradient centrifugation (28). Mouse MIG cDNA (kindly provided by J. Farber, National Institutes of Health, Bethesda, MD) was labeled with a random-prime labeling kit (Promega, Madison, WI) with [α-32P]dCTP (NEN Life Science Products, Boston, MA) (9, 10).
A cDNA probe to 28S rRNA (kindly provided by Dr. Luisa Iruela-Arispe, University of Washington, Seattle, WA) was used as an internal control for RNA loading (19). Northern analysis for MIG mRNA was performed as previously described (29). Briefly, total cytoplasmic RNA (10 μg/lane) from each animal was resolved by electrophoresis through a 1% agarose-formaldehyde gel (25) and transferred to a nylon membrane (0.45-mm pore size; Nytran, Schleicher and Schuell, Keene, NH). The membrane was hybridized with the MIG cDNA probe and washed in 0.1× saline-sodium citrate (SSC; 1× SSC is 150 mM NaCl and 15 mM sodium citrate, pH 7.0)-0.1% (wt/vol) SDS at 55°C for 15 min. The32P-labeled bands were visualized and quantified as described in RNase protection assay. The membrane was stripped of the MIG cDNA probe in 0.01× SSC-0.01% (wt/vol) SDS at 100°C for 20 min (repeated twice). The stripped membrane was hybridized with a 28S cDNA probe as an internal control for RNA loading of the gel as previously described (29). The signal intensity for each MIG mRNA band was normalized to the signal intensity for 28S rRNA in the same lane.
Immunohistochemical staining of IP-10 and MIG in the lung.
Ly5a and Ly5b (control) mice were killed 1 and 3 days after adoptive transfer of Th1 cells. Lungs of untreated mice were also examined. The lungs were fixed in 4% paraformaldehyde (Sigma, St. Louis, MO) and then embedded in paraffin. Sections were deparaffinized in Histoclear (National Diagnostics, Atlanta, GA) and rehydrated in graded ethanol. To facilitate staining, sections were incubated in 1× PBS with 1% saponin (Sigma) for 1 h. Subsequent steps except antigen retrieval were performed in this solution. Antigen retrieval was performed by steaming sections for 20 min in 0.01 M citrate buffer, pH 4. Sections were incubated in 0.3% H2O2 to quench endogenous peroxide activity and then incubated in 10% normal rabbit serum before staining. For immunostaining, sections were incubated with a 1:100 dilution of goat antibody to murine IP-10 (R&D Systems, Minneapolis, MN), a 1:100 dilution of goat antibody to murine MIG (R&D Systems), or a goat IgG control (Jackson ImmunoResearch, West Grove, PA). Biotinylated rabbit anti-goat IgG (Vector, Burlingame, CA) was used as a secondary antibody and visualized by the ABC method (Vector) followed by the chromogen diaminobenzidine (Sigma). The sections were counterstained with methyl green.
Cytokine and chemokine expression of Th1 cells.
Using RPAs, we characterized cytokine and chemokine mRNA production by resting and in vitro stimulated Th1 cells. After in vitro stimulation, cells produced mRNA for IFN-γ and TNF family members, characteristic of Th1 cells (Fig.1 A). The cells did not express detectable mRNA for cytokines such as IL-4, IL-5, IL-10, or IL-13 that are typically associated with type 2 responses (Fig.1 B). When hybridized against a panel of chemokines, the resting cells constitutively expressed mRNA only for the housekeeping proteins L32 and GAPDH. After in vitro activation, the cells expressed mRNAs for RANTES, MIP-1β, and MIP-1α and low levels of T cell activation gene 3 (TCA3; Fig. 1 C).
Chemokine expression in the lung.
To determine whether additional chemokines were expressed in the lung during inflammatory responses generated by alloreactive Th1 cells, we examined chemokine expression in the lung 1, 24, and 48 h after adoptive transfer of 107 Ly5a-specific Th1 cells into Ly5a and Ly5b (control) mice. We identified constitutive expression of RANTES and small increases in mRNAs for lymphotactin, MIP-1β, and monocyte chemotactic protein-1 (Fig. 2). However, Ly5a mice had a 12- to 13-fold increase in expression of IP-10 at 24 and 48 h compared with that in control (Ly5b) mice (P < 0.05 at 24 and 48 h by one-way ANOVA; Fig.3).
MIG mRNA was determined by Northern analysis at 3 and 6 h and 1, 3, 5, and 7 days after adoptive transfer of 107Ly5a-specific cells into Ly5a and Ly5b (control) mice (Fig.4). Levels of MIG mRNA were elevated fourfold at 24 h and were still elevated at 7 days in Ly5a mice compared with those in Ly5b mice (P < 0.05 at 5 and 7 days by one-way ANOVA; Fig. 5).
IP-10 and MIG expression in lung.
To determine the in vivo cellular expression of IP-10 protein and the closely related chemokine MIG, we performed immunohistochemistry on lungs of untreated, Ly5a, and Ly5b (control) mice after adoptive transfer of Th1 cells. The pattern of staining for both IP-10 (Fig. 6) and MIG (Fig. 7) was similar. Sections from normal untreated mice had no detectable staining for either IP-10 or MIG. Sections stained with a goat IgG control antibody were also negative. Ly5b (control) mice had absent or very low levels of staining. One day after adoptive transfer of Th1 cells to Ly5a mice, there was a markedly increased expression of both IP-10 and MIG in the vascular endothelium and perivascular mononuclear cells. IP-10- and MIG-immunoreactive proteins were also clearly visualized in airway epithelia. Three days after adoptive transfer, expression of both chemokines was more widespread, with pronounced staining of vascular endothelium and alveolar septa, although there was little staining of airway epithelium at this time. IP-10 and MIG expression was also prominent in perivascular inflammatory cells and alveolar macrophages.
This study shows a marked induction of chemokines in the lung after adoptive transfer and activation of alloreactive Th1 cells. The link between Th1 cell activation and recruitment of mononuclear cells appears to involve T cell-derived cytokines and chemokines as well as cytokine-inducible chemokines in the lung. The activated T cells per se express the characteristic Th1 cytokine IFN-γ as well as TNF family members, transforming growth factor-β1, and macrophage migration inhibitory factor. The Th1 cells also express the CCR5 ligands, RANTES, MIP-1α, and MIP-1β. Additional chemokines are induced in the lung after Th1 cell administration. Among those analyzed by RPAs, IP-10 was the most notable, with >10-fold increases in steady-state mRNA levels 1 day after Th1 cell transfer. Similar large increases in MIG mRNA were detected by Northern analysis. Remarkable increases in IP-10- and MIG-immunoreactive proteins were also present in a spatially restricted pattern in inflammatory foci and were specifically identified in alveolar macrophages, vascular endothelium, bronchial epithelium, and alveolar structures. This inducible chemokine pathway may represent an important mechanism for amplifying the inflammatory cell response initiated by activated Th1 cells.
IP-10 and MIG are related members of the C-X-C family of cytokines. Both were identified as a consequence of the dramatic induction of their genes in monocytic cells treated with IFN-γ (8,18). In addition to monocytes and macrophages, IP-10 and MIG are inducible in endothelial cells (20) and bronchial epithelial cells in vitro (26). Unlike other C-X-C chemokines, IP-10 and MIG have no activity on neutrophils but target activated T cells. Their common receptor, CXCR3, is expressed at high levels on activated Th1 cells, whereas Th2 cells express low levels (21, 24). Thus expression of both the inducible chemokines IP-10 and MIG and their chemokine receptor, CXCR3, may play a role in controlling tissue-specific migration of effector Th1 cells.
In our model of Th1-mediated lung inflammation, IP-10 and MIG are rapidly induced, presumably as the direct result of IFN-γ production from the in vivo activated Th1 cells. Expression of these chemokines in the endothelium and alveolar macrophages may serve to amplify the inflammatory response by recruitment of additional Th1 cells and other inflammatory cells such as natural killer cells that express CXCR3 (16, 21) and monocytes that migrate in response to recombinant IP-10 (31). The response of monocytes specifically to IP-10 is less well established, and other chemokines such as CCR5 ligands may play a role in recruitment of these cells.
In addition to chemotactic activity, IP-10 and MIG may influence adhesion of T lymphocytes in vascular endothelium. IP-10 induces T cell adherence to recombinant adhesion molecules and to extracellular matrix proteins (15). In vitro stimulated endothelial cells express IP-10 and MIG, and adhesion of IL-2-activated T lymphocytes to these endothelial cells is inhibited by an antibody to CXCR3 (20). Our experiments clearly establish that IP-10 and MIG expression is inducible on vascular endothelium in vivo. Th1 cells thus may induce the pulmonary vascular endothelium to perform new effector functions by elaboration of chemokines and recruitment of leukocytes. In this way, pulmonary vascular endothelial cells may participate as effector cells in Th1 cell-induced lung inflammation.
We also observed induction of IP-10 and MIG expression on bronchial epithelial cells in vivo. However, increased chemokine expression was not uniformly associated with the presence of inflammatory cells. Additional signals such as vascular adhesion molecules present on activated endothelium might be required to elicit migration of leukocytes to sites of chemokine expression. Alternatively, a more sustained signal may be required to induce airway inflammation. We noted that airway expression of MIG and IP-10 was transient compared with expression by endothelial cells and macrophages. A previous study (26) had reported increased IP-10 and MIG expression in bronchial epithelium associated with chronic diseases of airways (i.e., sarcoidosis and tuberculosis).
IP-10 and MIG have been associated with a number of clinical disorders in which Th1 cells have also been implicated. Soon after the cloning of IP-10, its expression was noted in delayed-type hypersensitivity responses in the skin (12). IP-10 and MIG have also been associated with inflammatory responses of immune-mediated diseases such as multiple sclerosis (27), rheumatoid arthritis (21) and graft rejection (13). mRNAs for IP-10 and MIG in the lung were increased by experimental infection withToxoplasma gondii in mice (2). IP-10 has been detected in granulomas and bronchoalveolar lavage fluid from patients with sarcoid (1). Expression of both IP-10 and MIG mRNAs was found to be increased in bronchoalveolar lavage fluid cells and bronchial epithelial cells from patients with tuberculosis (26). These observations support the idea that these IFN-γ-inducible chemokines may contribute to the further recruitment of specific lymphocyte subsets to the lung.
In summary, our study establishes the association of Th1 cell activation and the induction of IP-10 and MIG expression in the lung in vivo. In addition, we show a strong temporal and spatial correlation of these chemokines with inflammatory foci in the lung induced by Th1 cells. Both endothelial cells and cells of monocyte/macrophage lineage appear to be significant cellular sources of these inducible chemokines. Further in vivo studies are needed to understand how these chemokines contribute to immune-mediated lung injury and to explore the potential for modifying the inflammatory response by neutralization of this pathway. Importantly, our studies reveal a marked redundancy of related chemokine molecules and convey the very strong likelihood that approaches to inhibiting these pathways will need to be aimed at neutralizing the effects of multiple molecules.
We thank Andrew Tager, Mary Beauchamp, Andrew Elston, and Caroline Sawe for technical advice and assistance and Heather Peake for assistance with preparation of the manuscript.
This work was supported by National Heart, Lung, and Blood Institute Grants R01-HL-55200 and K32-HL-07237.
Address for reprint requests and other correspondence: J. G. Clark, Pulmonary and Critical Care Medicine, Mailstop D3-190, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave. North, Seattle, WA 98109 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society