Prostaglandin E₂ deficiency uncovers a dominant role for thromboxane A₂ in house dust mite-induced allergic pulmonary inflammation

Abstract

Prostaglandin E₂ (PGE₂) is an abundant lipid inflammatory mediator with potent but incompletely understood anti-inflammatory actions in the lung. Deficient PGE₂ generation in the lung predisposes to airway hyperresponsiveness and aspirin intolerance in asthmatic individuals. PGE₂-deficient ptges^−/− mice develop exaggerated pulmonary eosinophilia and pulmonary arteriolar smooth-muscle hyperplasia compared with PGE₂-sufficient controls when challenged intranasally with a house dust mite extract. We now demonstrate that both pulmonary eosinophilia and vascular remodeling in the setting of PGE₂ deficiency depend on thromboxane A₂ and signaling through the T prostanoid (TP) receptor. Deletion of TP receptors from ptges^−/− mice reduces inflammation, vascular remodeling, cytokine generation, and airway reactivity to wild-type levels, with contributions from TP receptors localized to both hematopoietic cells and tissue. TP receptor signaling ex vivo is controlled heterologously by E prostanoid (EP)₁ and EP₂ receptor-dependent signaling pathways coupling to protein kinases C and A, respectively. TP-dependent up-regulation of intracellular adhesion molecule-1 expression is essential for the effects of PGE₂ deficiency. Thus, PGE₂ controls the strength of TP receptor signaling as a major bronchoprotective mechanism, carrying implications for the pathobiology and therapy of asthma.

Prostaglandins (PGs) derive from the metabolism of arachidonic acid by cyclooxygenase (COX)-1 and COX-2 (1). COX-derived PGH₂ is converted to one of five PGs [PGD₂, PGE₂, PGF₂α, PGI₂, or thromboxane A₂ (TXA₂)] by distinct terminal synthases. Specific G protein-coupled receptors (2) mediate the effects of PGs in vivo, including roles in regulation of vascular tone, hemostasis, hematopoiesis, epithelial cell turnover, cancer, and inflammation (1). Although COX inhibitors control inflammatory pain (3) and may prevent vascular disease and certain cancers (1), these reagents can also interfere with the production of PGs needed for homeostatic functions and cause unwanted side effects (1, 4).

PGE₂ is one of the most functionally versatile PGs. Three terminal synthases can convert PGH₂ to PGE₂ (5–7). Of these, microsomal PGE₂ synthase (mPGES)-1 is the dominant enzyme responsible for converting COX-2–derived PGH₂ to PGE₂ during inflammation (8). Four E prostanoid (EP) receptors (9) mediate the effects of PGE₂. PGE₂ is pathogenetic in models of autoimmune diseases (10) and inflammatory arthritis (8), but is bronchoprotective (11) and anti-inflammatory (12, 13) in certain models of allergic lung disease. In patients with atopic asthma, inhalation of PGE₂ blocks bronchoconstriction in response to challenge with inhaled specific allergen (14) and attenuates the recruitment of eosinophils and basophils to the airway. Reduced PGE₂ levels in the setting of sputum eosinophilia correlate with airway hyperresponsiveness (AHR) (15), a cardinal feature of asthma. Aspirin-exacerbated respiratory disease (AERD), a severe variant of asthma, is associated with deficient COX-2–derived PGE₂ production (16–19) and reduced expression of the EP₂ receptor (20, 21), as well as marked tissue eosinophilia and bronchoconstrictive responses to the administration of nonselective COX inhibitors. These findings suggest potential therapeutic applications of PGE₂ in asthma and AERD if the mechanisms responsible for the homeostatic functions of PGE₂ in asthma can be fully defined.

An extract (Df) of the house dust mite Dermatophagoides farinae contains clinically relevant protease allergens, as well as adjuvants (glycans, endotoxin) that elicit sensitization through the airway mucosa, resulting in eosinophilic airway inflammation and a mixed T helper (Th)2, Th1, and Th17 pulmonary immune response (22). PGE₂-deficient mice with a targeted deletion in the gene encoding mPGES-1 (ptges^−/− mice) develop increased bronchovascular eosinophilia compared with WT C57BL/6 controls when challenged intranasally with a low dose (3 μg) of Df (12), revealing a bronchoprotective effect of endogenous PGE₂. This effect is restored by administration of agonists of the EP₁, EP₂, and EP₃ receptors. Df-challenged ptges^−/− mice also develop extensive remodeling of the small- to medium-sized pulmonary arterioles (12), a feature observed with sustained pulmonary Th2 responses in WT mice induced by inhaled allergens (23). TXA₂, a platelet-derived PG, activates endothelial-dependent leukocyte recruitment (24), causes constriction and proliferation of pulmonary vascular smooth muscle (25, 26), and amplifies airway reactivity to methacholine (MCh) in humans (27). Because of its prominent pulmonary vascular and airway effects, we speculated that TXA₂ might mediate the pathologic effects of PGE₂ deficiency in the lung. We now demonstrate that virtually all functional and immunologic perturbations resulting from PGE₂ deficiency in Df-treated mice are prevented by the deletion of tpr encoding the T prostanoid (TP)α and TPβ receptors for TXA₂. PGE₂ mediates bronchoprotection in part by controlling TXA₂ effector pathways, suggesting that pharmacologic blockade of TP receptors may prove helpful for the treatment of asthma variants in which local PGE₂-mediated homeostasis is deficient.

Results and Discussion

Dust Mite-Mediated Bronchovascular Inflammation and Vascular Remodeling in ptges^−/− Mice Depends on TP Receptor Signaling.

Because the vascular remodeling observed in Df-challenged ptges^−/− mice suggested a potential contribution from TP receptors, we eliminated TP receptors in ptges^−/− mice by mating them with tpr^−/− C57BL/6 mice (28). As expected, the intranasal administration of Df (3 μg) elicited mild inflammation in WT C57BL/6 mice, and severe inflammation in ptges^−/− mice (12). Ptges^−/− mice demonstrated higher total bronchoalveolar lavage (BAL) fluid cell counts, eosinophils, macrophages, lymphocytes, and neutrophils relative to WT controls (Fig. 1A). Df-challenged ptges^−/− mice had slightly more Periodic acid-Schiff (PAS)-staining goblet cells and cellular infiltration of their bronchovascular structures than did the WT controls (Fig. 1B, Top, as shown in Fig. 1C). Deletion of tpr from the ptges^−/− mice reduced all of these features to levels indistinguishable from or below those of WT controls (Fig. 1). Findings in WT controls did not differ from those in the tpr^−/− single KO mice (Fig. 1), although single KO mice showed reduced eosinophilic inflammation in experiments where the dose of Df was increased to 10 μg (Fig. S1).

Fig. 1.

Deletion of TP receptors eliminates the potentiation of pulmonary inflammation in PGE₂-deficient mice. Mice were challenged intranasally with either Df (3 μg) or saline on days 0, 3, 7, 10, 14, and 17, and were killed on day 18. (A) The numbers of total cells, macrophages, neutrophils, lymphocytes, and eosinophils recovered from the BAL fluid of saline-challenged and Df-challenged mice of the indicated genotypes are displayed. (B) Quantitative analysis of goblet cell numbers (Upper, based on PAS staining) and cellular infiltration of bronchovascular bundles (Lower, based on H&E staining). (C) Representative samples of lungs stained with PAS (Upper) and H&E (Lower) showing bronchovascular bundles for each genotype. Results in A and B are mean ± SD from 15 to 20 mice per group from four different experiments. BM, basement membrane. (Scale bars, 100 μm.)

Pulmonary arteriolar remodeling is not a feature of asthma in humans, but develops in sensitized WT mice treated with long-term or high-dose allergen (23, 29). In our previous study (12), ptges^−/− mice developed significant increases in the numbers of α-smooth muscle actin (α-SMA)⁺ cells and the thickness of the α-SMA⁺ cell layer in lung sections at doses of dust mite allergen ∼10-fold lower than those used to elicit this response in WT mice in other reports (29). Administration of selective EP₁, EP₂ (especially), and EP₃ agonists protected ptges^−/− mice from remodeling (12). In the present study, the Df-induced vasculopathy in the ptges^−/− mice was completely eliminated by the introduction of the tpr^−/− allele (Fig. 2). Treatment of the ptges^−/− mice with intraperitoneal injections of the selective TP receptor antagonist SQ 29,548 each day (2 mg/kg of body weight) during the period of Df challenge reduced total BAL cell and eosinophil counts (Fig. S2A), bronchovascular inflammation, goblet cell metaplasia, and vascular remodeling (Fig. S2B). These data strongly indicate that TP receptors are dominant effectors of pulmonary inflammation and its pathophysiologic consequences when PGE₂ is depleted.

Fig. 2.

TP receptors mediate Df-induced remodeling of pulmonary arterioles in ptges^−/− mice. Paraffin-embedded lungs from the saline- or Df-challenged mice were stained with an Ab for α-SMA and counterstained with H&E. (A) Thickness of α-SMA–staining layer (Left) and numbers of α-SMA⁺ cells per millimeter of basement membrane (BM) (Right) in the lungs of the indicated genotypes. (B) Representative images for each genotype are shown. Arrows show arterioles. Data in A are mean ± SD from 15 to 20 mice per group from four separate experiments. (Scale bar, 100 μm.)

Deletion of TP Receptors Eliminates the Increment in Airway Reactivity Observed in Df-Challenged PGE₂-Deficient Mice.

AHR to MCh is a feature of asthma that is potentiated after allergen challenge of atopic humans. Although inhaled PGE₂ blocks allergen-induced potentiation of AHR (14), inhaled TXA₂ causes bronchoconstriction and potentiates AHR (27). Although exogenously administered PGE₂ and TXA₂, respectively, dampen and potentiate airway reactivity to MCh in mice (11, 30), the roles of endogenous PGs in controlling allergen-induced AHR are less clear. Indeed, a recent study reported that ptges^−/− mice showed reduced MCh reactivity compared with WT controls in a model of pulmonary disease where mice were immunized systemically to ovalbumin (31). In contrast to that study, Df-challenged ptges^−/− mice displayed greater AHR than did Df-challenged WT mice over the entire range of MCh doses (6.25–100 mg/mL). Deletion of the tpr allele from the ptges^−/− mice restored MCh reactivity following Df challenge to WT levels (Fig. 3). Thus, endogenous PGE₂ is strongly bronchoprotective against AHR in our model, and does so at least in part by suppressing the pathogenetic contribution from TP receptors. Tpr deletion also slightly increased basal reactivity to the highest doses of MCh in naive, PGE₂-sufficient mice (Fig. 3).

Fig. 3.

Amplification of MCh reactivity in Df-challenged ptges^−/− mice is TP receptor-dependent. Saline- or Df-challenged mice of the indicated genotypes were anesthetized and mechanically ventilated. Airway resistance was monitored during incremental inhalation challenge with MCh. Data are from 10 mice per group from two separate experiments. *P < 0.05 vs. Df-treated WT group, tpr^−/−, and DKO groups. ^#P < 0.05 PBS-treated tpr^−/− group vs. PBS-treated WT group.

Deletion of TP Receptor Signaling Attenuates Th2-Type Recall Responses to Df.

Allergen-induced eosinophilic inflammation, pulmonary vascular remodeling, and AHR in mice are all characteristic signatures of Th2-type cytokines (23, 32). Df elicits a mixed Th1, Th2, and Th17-type pulmonary immune response in mice based on cytokine profiles of restimulated lung-draining parabronchial lymph node (PLN) cells. Because both PGE₂ and TXA₂ can exert complex effects on T cells depending on the model studied (33, 34), we measured cytokines generated by Df-restimulated PLN cells from each strain. The cells from Df-treated ptges^−/− mice generated more IL-4, IL-5, IL-13, and IL-17A than did the cells from WT animals, but produced less IFN-γ (Fig. S3). This finding is consistent with previous reports that PGE₂ promotes IFN-γ production in a Th1-dependent model of contact hypersensitivity (10). Because IFN-γ counteracts Th2 cell development, the lower level of IFN-γ generated by the ptges^−/− PLN cells could amplify Th2 cytokine production occurring in our model. Deletion of the tpr allele reduced the levels of Th2 cytokines in both PGE₂-deficient and -sufficient genetic backgrounds (Fig. S3), and also reduced the levels of serum IgE and specific IgG1 (Fig. S4). Treatment with SQ 29,548 also reduced Th2 cytokine generation, although this was significant only for the PLNs from the ptges^−/− mice (Fig. S2C). Thus, the pathologic and physiologic features of our model that are enhanced and abrogated by the absences of mPGES-1 and TP receptors could in part reflect the influences of PGE₂ and TXA₂, respectively, on cytokine production by T cells. The reduced levels of allergen-specific IgE and IgG1 could explain some of the protective effects of TP receptor deletion.

Absence of PGE₂ Does Not Increase TP Receptor-Dependent Pathology by Enhancing TXA₂ Formation.

COX-2 and mPGES-1 ensure increased PGE₂ production with inflammation (35). Deletion of mPGES-1 can therefore alter the synthesis of other PGs through shunting of the COX-2–derived PGH₂ to other terminal synthases (36). We monitored the generation of all PGs by the cells of enzymatically dispersed lung cells from all strains. Df challenge induced a marked (∼threefold) increase in PGE₂ generation by the lung cells of WT mice that were incubated ex vivo with arachidonic acid (50 μM), and increased the productions of PGF₂α, PGI₂, PGD₂, and TXA₂ more modestly (Fig. S5). Consistent with our previous study (12), the deletion of ptges impaired the conversion of arachidonic acid to PGE₂ by the lung cells from Df-challenged mice, without altering conversion to other PGs (Fig. S5), but the deletion of TP receptors had no effect.

Because TXA₂ synthesis was not enhanced in the absence of mPGES-1, we suspected that the amplification of TP receptor-dependent contributions in this strain reflected deficient cross-regulation of TP receptor signaling by EP receptors. TP receptors in the lung are constitutively serine-phosphorylated by other receptors that maintain cAMP-dependent PKA activity, limiting the strength of TP receptor signaling (37, 38). TP receptors can also be phosphorylated and desensitized by PKC (38). As noted above, EP₁, EP₂, and EP₃ agonists all protect ptges^−/− mice from vascular remodeling (12), and the EP₂ agonist also blocked bronchovascular inflammation and BAL fluid eosinophilia. Narumiya and colleagues showed a homeostatic role for EP₃ receptor signaling in a model of ovalbumin-induced airway disease (39). EP₂ and EP₄ receptors activate PKA (9), whereas EP₁ and EP₃ receptors can activate PKC (40, 41). We studied cross-regulation ex vivo in platelets and human umbilical vein endothelial cells (HUVEC), both of which express TP (42, 43) and EP receptors (42). The TP receptor-selective agonist U-46619 dose-dependently induced CD62P expression by platelets (Fig. 4 A and B). The platelets from tpr^−/− mice showed no response to U-46619 (Fig. 5B). Pretreatment with an EP₁ receptor agonist (DI-004) suppressed TP-dependent induction of CD62P (Fig. 4C), whereas agonists of EP₂ (AE1-259-01), EP₃ (AE-248), and EP₄ receptors showed no effect below 10 μM. The effect of the EP₁ agonist was reversed by treatment of the samples with the PKC inhibitors Go6976 and Calphostin C (Fig. 4D). U-46619 induced the expression of intercellular adhesion molecule (ICAM)-1 by HUVEC (Fig. 4 E and F). This induction was blocked by pretreatment with AE1-259–01 (0.1–10 μM) or the stable PGE₂ analog 16, 16-diPGE₂ (10 μM), but not with the other EP receptor selective agonists (Fig. 4 E and F). The EP₂ receptor-dependent blockade of ICAM-1 induction was reversed by the PKA inhibitor H89 (Fig. 4 E and F) but not by the PKC inhibitor Go6976 (Fig. 4E). Thus, TP receptor signaling on both hematopoietic and nonhematopoietic cells can be desensitized by EP₁ and EP₂ receptors through respective PKC- and PKA-dependent signaling pathways, consistent with the hypothesis that PGE₂ controls allergic pulmonary inflammation by maintaining heterologous desensitization of TP receptors.

Fig. 4.

EP₁ and EP₂ receptors control TP receptor signaling in platelets and endothelial cells, respectively. (A) Cytofluorographic detection of CD62P expression by platelets identified in whole blood after 15 min of stimulation with or without the selective TP receptor agonist U-46619 (10 μM). (B) Effects of various doses of U-46619 on CD62P expression by platelets in the blood of WT or tpr^−/− mice. (C) Effect of various concentrations of the EP₁ receptor agonist DI-004 on TP receptor-mediated CD62P expression by blood platelets. (D) Effect of the PKC inhibitors Go6976 and Calphostin C on the suppression of TP-induced platelet CD62P expression by DI-004. Results in B–D are the mean ± SD from three separate experiments, with the induction of CD62P induced by U-46619 alone (10 μM) set at 100% for each experiment in C. (E) Effects of various EP receptor agonists on the induced expression of ICAM-1 by HUVEC cells stimulated with U-46619. The PKA inhibitor H89 was added to the indicated samples that were stimulated with U-46619 (10 μM) in the presence of the maximally effective dose (1 μM) of the EP₂ receptor agonist AE1-259–01. Data are mean ± range from two separate experiments. (F) A representative histogram from one of the two experiments is shown.

Fig. 5.

PGE₂ deficiency amplifies the contribution of TP-dependent ICAM-1 induction. (A) Double immunofluorescent stains of bronchovascular bundles in the lungs of saline (Upper) and Df-challenged mice of the indicated genotypes. Paraffin-embedded lung sections were stained for α-SMA (red fluorescence, to identify pulmonary arterioles, as indicated by arrows) and ICAM-1 (green fluorescence), and counterstained with Hoechst stain to detect the nuclei (blue). Note that ICAM-1 localizing primarily to alveolar walls. Images are 400×. (B) Confocal imaging (1,200×) of the lung of a Df-challenged ptges^−/− mouse double-stained for ICAM-1 (green fluorescence) and the endothelial cell marker PECAM-1 (red fluorescence). ICAM-1 localizes to PECAM-1⁻ alveolar epithelial cells (white arrowhead), and to PECAM-1⁺ endothelial cells (white arrows). Some additional PECAM-1⁺ cells are negative for ICAM-1 (red arrow). (C) Western blot of proteins extracted from the lungs of the indicated strains showing inducible expression of ICAM-1 with Df treatment and the effect of TP receptor deletion. Results are shown for representative mice (Upper), and densitometry is shown for groups of five to seven mice (Lower).

TP Receptors Expressed by both Resident and Hematopoietic Cells Contribute to the Enhanced Pulmonary Response to Allergen in the Setting of PGE₂ Deficiency.

To determine the extent to which TP receptors on hematopoietic cells and on resident tissue cells were important in our model, we generated chimeric ptges^−/− mice lacking TP receptors on nonbone marrow-derived cells by transferring ptges^−/− bone marrow into lethally irradiated double-KO mice. We also generated chimeric ptges^−/− mice lacking TP receptors on bone marrow-derived cells by transferring double-KO bone marrow into irradiated ptges^−/− mice. After 10 wk, the chimeric mice were challenged with either Df or saline. Control ptges^−/− mice that were reconstituted with ptges^−/− bone marrow developed Df-induced increases in total BAL fluid cell numbers, eosinophils, bronchovascular cellular infiltrates, and goblet cell metaplasia in response to Df. In contrast, double-KO mice reconstituted with double-KO bone marrow showed little response to Df challenge on all parameters (Fig. S6 A–D). Mice lacking TP receptors only on the hematopoietic compartment and mice lacking TP receptors only on the tissue compartment demonstrated Df-induced BAL fluid eosinophilia, bronchovascular inflammation, and PLN cytokine production that were intermediate between the ptges^−/− and double-KO mice reconstituted with marrow from their corresponding respective identical littermates (Fig. S6). Peripheral blood cell counts were similar among the various groups (WBC counts ranging from 2.15 ± 0.43 to 3.07 ± 1.01 × 10³ cells/μL, mean ± SD) and no different from the nonirradiated age-matched WT mice (2.77 ± 0.55 × 10³ cells/μL). Thus, the functions of TP receptors on hematopoietic cells and nonhematopoietic cells are synergistic in this model, potentially reflecting the contributions of TP receptors to platelet-dependent priming of granulocyte adhesiveness (24, 44, 45) and enhancement of endothelial adhesion molecule expression (24), respectively. Both of these contributions may be held in check by PGE₂.

PGE₂ Deficiency Augments ICAM-1–Dependent Pulmonary Leukocyte Recruitment in Vivo.

We next sought to directly determine whether TP receptors regulated leukocyte recruitment to the lung in the setting of PGE₂ deficiency by controlling the expression of ICAM-1. Double immunofluorescent staining with anti–ICAM-1 and anti–α-SMA (to identify the bronchovascular structures) revealed ICAM-1 staining in the WT and ptges^−/− strains that tended to be more intense after challenges with Df (Fig. 5A, green fluorescence). Confocal images of lung double-stained with Abs against ICAM-1 and platelet endothelial adhesion molecule-1 (PECAM-1) revealed strong ICAM-1 signals on the PECAM-1⁻ alveolar epithelium, with additional colocalization to PECAM-1⁺ endothelial cells (Fig. 5B). Df treatment strongly up-regulated ICAM-1 protein expression in the lungs of WT mice, and more strongly in the ptges^−/− mice (Fig. 5C). Remarkably, Df treatment failed to up-regulate ICAM-1 protein in the lungs of tpr^−/− and double-KO mice (Fig. 5C).

To determine whether the TP receptor-driven ICAM-1 expression observed in the ptges^−/− mice accounted for the pathologic features observed in this strain, we treated WT and ptges^−/− mice with intraperitoneal injections of blocking anti–ICAM-1 Abs or isotype controls during the period of Df challenge. ICAM-1 blockade markedly decreased the numbers of total cells and eosinophils in the BAL fluids from the ptges^−/− mice and in the WT controls (Fig. S7A), and also reduced the bronchovascular inflammation, goblet cell metaplasia, and arteriolar remodeling (Fig. S7B, as shown in Fig. S7C). Anti–ICAM-1 treatment reduced the numbers of cells recovered from the PLNs of both strains of mice, and reduced the levels of IL-4, IL-5, and IL-13, but not IL-17A generated by the cells from both strains (Fig. S8). Thus, EP and TP receptor signaling pathways respectively dampen and induce ICAM-1 expression in the lung, controlling susceptibility to the pathologic and physiologic changes resulting from Df. The lack of PGE₂ likely dysregulates the control of TP receptor-dependent ICAM-1 expression as a major phenotypic component of the ptges^−/− strain.

Our study suggests that the long recognized homeostatic function of PGE₂ in allergic pulmonary inflammation (12–14) and airway reactivity (15, 46) involves cross-regulation of TP receptor function. EP receptors “fine-tune” the strength of signaling through TP receptors by activating PKA and PKC. Consequently, TP-dependent effects become magnified in the setting of inflammation if the capacity to up-regulate PGE₂ generation (or to signal through EP receptors) is impaired. Deficient PGE₂ may potentiate AHR in the setting of eosinophilic inflammation (15, 47). Defective COX-2–dependent production of PGE₂ and deficient expression of EP₂ receptors have both been documented in AERD (19, 21, 48), and might account for the marked eosinophilic airway inflammation and high basal levels of leukotriene E₄ typical of this syndrome. The lack of a contribution from COX-2 to the generation of homeostatic PGE₂ may explain the fact that COX-2–selective agonists are well-tolerated in AERD (49). Because aspirin potently inhibits COX-1 but only weakly inhibits COX-2 (50), most asthmatic individuals likely tolerate aspirin because of the persistent bronchoprotection provided by “aspirin resistant” COX-2/mPGES-1–derived PGE₂. Conversely, the inability to maintain COX-2/mPGES-1–derived PGE₂ when COX-1 is inhibited could explain the characteristic surge in cysteinyl leukotriene production and bronchoconstriction in response to challenges with low-dose (30–160 mg) aspirin in individuals with AERD. Basal COX-1–derived TXA₂ formation is unimpaired in AERD (45). Subsequently, the interference with TXA₂ generation (51) could explain part of the apparently paradoxical therapeutic benefit of aspirin treatment in patients with AERD after they are clinically desensitized (52). Existing orally active TP receptor agonists have shown moderate efficacy in reducing AHR or sputum eosinophilia in small controlled studies of individuals with asthma (53–55). Such antagonists may have major efficacy as treatments for asthma in individuals stratified for lower levels of pulmonary PGE₂ production or deficient EP receptor signaling, such as those with AERD (21).

Methods

Mice.

C57BL/6 ptges^−/− mice were from Satoshi Uematsu (Osaka University, Osaka, Japan). C57BL/6 tpr^−/− mice were provided by Thomas Coffman (Duke University, Durham, NC). ptges^−/− males were intercrossed with tpr^−/− females to obtain double-KO (ptges/tpr^−/−) mice. The mice and the WT controls were housed at Charles River. All studies were conducted on 6- to 8-wk-old male mice and were approved by the Animal Care and Use Committee of the Dana-Farber Cancer Institute. Df challenges, cellular and standard histologic analyses, restimulation of PLN cells, measurements of Abs, and GC-MS analyses were all performed as described previously (12, 56).

Bone Marrow Transfer.

Five-week-old mice were lethally irradiated with 1,200 rads (12 Gy) in two split doses, 4 h apart. Within 24 h of irradiation, 1 × 10⁷ bone marrow cells were infused via the tail vein into sex-matched irradiated mice in 200 μL of PBS. For the first 8 wk after the injection of bone marrow cells, chimeric mice drank water supplemented with enrofloxacin (Baytril; Bayer Health Care). Ten weeks after the injection, mice were exposed to six doses of saline or Df and killed 24 h after the last challenge. Peripheral blood cell counts were performed before the beginning of treatment to ensure full engraftment.

ICAM-1 Blocking.

Twenty-four hours before each Df challenge, mice were injected intraperitoneally with 50 μg of a neutralizing Ab against mouse ICAM-1 (clone YN1/1.7.4; Biolegend) or rat IgG2b-κ isotype control Ab (Biolegend). Mice were treated with 3 μg of Df twice weekly for 3 wk.

Statistical Analysis.

Data are expressed as mean ± SD. Statistical analyses were performed using ANOVA. Differences were considered significant when P values were less than 0.05.