Urokinase plasminogen activator receptor attenuates allergen-induced eosinophil migration and airway hyperresponsiveness

Abstract

Allergic asthma is a widespread disease of the airway stemming from the actions of multiple cell types, including eosinophils and epithelial cells. The urokinase plasminogen activator receptor (uPAR) is a membrane bound protein that can contribute to the activation and mobilization of leukocytes and is present at increased levels in asthmatics. However, its role in allergic asthma remains poorly understood. Here, we used multiple mouse strains and different models of allergic airway disease to study the function of uPAR in the pathogenesis of this disease. Plaur, the gene encoding uPAR, was rapidly induced following allergic sensitization through the airway, and again following subsequent allergen challenge. Plaur-deficient mice displayed both increased numbers of eosinophils and heightened airway hyperresponsiveness (AHR) in multiple models of allergic asthma. Mice selectively lacking Plaur in eosinophils also had more robust eosinophilia than did WT mice, and eosinophils lacking Plaur displayed increased activity in an ex vivo assay of chemokine-dependent migration. However, those mice did not have increased AHR compared with WT mice. Conversely, although mice selectively lacking Plaur in lung epithelial cells did not have increased inflammation compared with wild type (WT) mice, they displayed heightened AHR. These findings suggest that uPAR controls both airway inflammation and AHR, but through distinct mechanisms. Targeting uPAR might have therapeutic potential for treating inflammation and AHR in asthma.

Introduction

Allergic asthma is a complex disease stemming from inflammatory responses to a variety of recurrent exposures including environmental allergens and adjuvants (1). Physiological manifestations include a decline in lung function stemming from mucus hypersecretion and airway hyperresponsiveness (AHR). Infiltration of numerous cell types, including T cells, eosinophils and neutrophils are also essential to asthma pathogenesis. Upon allergen recognition, allergen-specific CD4⁺ T helper type (Th)2 lymphocytes release interleukin (IL)-4, -5, and -13, which promote IgE class switching, eosinophil migration and activation, and AHR, respectively (2). Concurrently, allergen-specific Th17 cell production of IL-17 leads to the recruitment of neutrophils and exacerbation of AHR (3).

In contrast to the proinflammatory activities of effector cell types such as Th2 and Th17 cells, other cell types and mechanisms limit the extent of inflammation and decline in lung function. The best-known example of this are regulatory T cells (Tregs), which produce anti-inflammatory cytokines, such as Transforming growth factor (TGF)-β or IL-10 (4, 5). In addition to these secreted molecules, other proteins act in a cell intrinsic manner to limit the proinflammatory activities of leukocytes. Interestingly, some anti-inflammatory molecules are also associated with coagulation, including activated protein C (APC) and anti-thrombin (AT) (6). APC directly inhibits chemoattractant mediated chemotaxis of neutrophils (7) and eosinophils (8), and interferes with the production of pro-inflammatory cytokines (9), whereas anti-thrombin inhibits pro-inflammatory cytokines and adhesion molecule (6). This close relationship between coagulation and inflammation might have evolved to help coordinate clot formation and resolution with inflammation and wound repair, respectively.

Another molecule that bridges coagulation and inflammation is the urokinase plasminogen activator receptor (uPAR), which is encoded by the Plaur gene. uPAR binds urokinase plasminogen activator (uPA), which, like tissue plasminogen activator (tPA), can convert plasminogen to plasmin to accelerate clot resolution (10). However, uPAR can also signal independently of uPA, in part by acting through vitronectin (VTN) and various cell surface receptors, including integrins (11). A variety of pro-inflammatory activities have been attributed to uPAR, including abilities to promote cell migration, adhesion, proliferation, and angiogenesis (10, 12). How these various activities contribute to the progression of different diseases is an active area of research.

Several lines of evidence suggest that uPAR might function in allergic asthma. For example, elevated levels of uPAR have been detected at elevated concentrations in the sputum and bronchial epithelium of asthmatic patients (13–16). Further, some polymorphisms of Plaur are associated with human asthma and lung function decline (17). However, despite these associations, it has been unclear whether the increased levels of uPAR causally contribute to inflammation or if they are a consequence of this multifactorial disease. It is also unknown whether uPAR associates with particular forms of asthma, or whether blocking uPAR might have therapeutic benefit. To address these questions, we used two different models of allergic asthma to compare responses of Plaur-deficient (Plaur^−/−) and wild-type (WT) mice. In each of these models, the absence of uPAR led to increased allergic airway inflammation and heightened AHR. We also found that although Plaur is expressed in multiple cell types, it acts in eosinophils in a cell autonomous manner to suppress their migration in vivo and ex vivo. Paradoxically, Plaur expression in eosinophils was dispensable for suppression of AHR. Instead, expression of this gene in the lung epithelium was required for AHR suppression. These results suggest that agonism of uPAR may be a broadly effective means to limit asthma severity.

Methods

Mice

C57BL/6J, Plaur^−/− (B6.129P2-Plaur^tm1Jld/J), Plau^−/− (B6.129S2-Plau^tm1Mlg/J) and OT-II T cell receptor (TCR) transgenic (B6.Cg (TcraTcrb) 425Cbn/J) mice were purchased from Jackson Laboratories, Bar Harbor, ME. Conditionally mutant Plaur^fx/fx mice were generated at the Gene Editing and Mouse Model Core Facility of the National Institute of Environmental Health Sciences (NIEHS) / National Institutes of Health (NIH) by inserting loxP sites into intron 1–2 and intron 6–7 of B6129F1 embryonic stem cells (G4; 129S6/SvEvTac x C57BL/6Ncr). A detailed description of the methods used to generate Plaur^fx/fx mice can be found in Supplementary Methods and Figure E1. Mice were bred and housed in specific pathogen-free conditions at the NIEHS and were used between 6 and 12 wk of age. Male mice were used in all the analysis unless otherwise indicated. All experiments were conducted in accordance with guidelines provided by the Institutional Animal Care and Use Committee at the NIEHS.

Animal models of asthma.

The models of asthma used here have been described in detail previously (18). Briefly, mice were sensitized to ovalbumin (OVA) by two oropharyngeal (o.p.) administrations, 1 wk apart, of 50 μg LPS-free OVA (Worthington Biomedical, Lakewood, NJ) mixed with an adjuvant in a final volume of 50 μl with sterile PBS as the diluent. Unless otherwise stated, the adjuvants used were 100 ng LPS from E. coli 0111:B4 (MilliporeSigma), 0.10 or 0.03 μL of a protease mix derived from Aspergillus oryzae (ASP) (MilliporeSigma) (19). Sensitized mice were challenged 1 wk after the second sensitization with an aerosol of 1% OVA (grade V, Millipore Sigma) in PBS for 1 h.

Flow cytometry.

Lungs were excised, minced, digested, and filtered to generate a single cell suspension of cells as described previously (20). After Histopaque 1119 centrifugation, banded cells were diluted to 1–2 x 10⁶/100 µL and incubated with a non-specific binding blocking reagent cocktail of anti-mouse CD16/CD32 Ab (2.4G2) (10% culture supernatant), 5% normal mouse and 5% rat serum (Jackson ImmunoResearch). Cells were then incubated with Fluorochrome-conjugated antibodies and analyzed as described in Supplementary Methods.

Migration assays

BALF cells were collected 2 d after OVA challenge of ASP/OVA-sensitized mice and subjected to assays of chemokinesis and chemotaxis, as described in Supplementary Methods.

Single cell RNA sequencing

Detailed experimental methods for single cell RNA sequencing are described in Supplementary Methods. The data have been submitted to the Gene Expression Omnibus (GEO) and can be found under GSE297229 (CD45 negative cells) and GSE297230 (CD45 positive cells).

Statistics

Data are shown as means ± SEM. To compare two groups, Paired Student’s t test (two-tailed, parametric) was used. For comparisons of three or more groups either a one-way, parametric ANOVA with Sidak’s post-hoc tests was performed or a two-way ANOVA with Sidak’s post hoc test when two factors varied. The qRT-PCR analysis was performed with the ΔΔCt– method and expressed as arbitrary units (AU) of the fold-induction of mRNA compared with the reference gene, Gapdh (1.0-fold induction). Analyses were conducted in GraphPad Prism and values of p < 0.05 were considered statistically significant.

Results

Induction of Plaur and related genes mouse models of asthma.

To investigate potential roles of uPAR in allergic asthma, we first tested if Plaur was induced in animal models of that disease. Such models typically involve two phases: an allergic sensitization phase that induces allergen-specific adaptive immune responses; and an allergen challenge phase that elicits pulmonary inflammation. OVA is often used as an experimental allergen. On its own, highly purified OVA is not allergenic, but mixing it with potent adjuvants prior to its instillation into the airways of mice elicits strong OVA-specific adaptive immune responses. The use of LPS as an adjuvant leads to relatively weak type 2 responses, but triggers strong, IL-17-dependent neutrophilia (3). By contrast, proteases from Aspergillus oryzae (ASP) drive highly polarized type 2 responses and eosinophilia (18). We therefore mixed OVA with either LPS (LPS/OVA) or ASP (ASP/OVA), and instilled these different mixtures into the airways of separate groups of mice to investigate the expression of genes relevant to uPAR signaling. Animals were harvested at either 2 h or 4 h post-sensitizations and their lungs were examined for expression of Plaur, Plau (encoding uPA), Vtn, and tPA (Figure E2A). Expression of each of these genes was significantly increased in lungs of mice receiving adjuvant alone or OVA plus adjuvant compared with untreated mice or animals receiving OVA only (Figure E2, B–C).

We next tested whether Plaur, Plau, Vtn, and tPA are also induced by allergen challenge of previously sensitized mice. To do this, animals were sensitized twice by instillations of OVA together with either LPS or ASP and then challenged one wk after the second sensitization by exposing them to aerosolized OVA alone (Figure E3A). In these experiments, OVA challenge increased expression of Plaur, Plau, Vtn, and tPA in mice that had been previously sensitized, but not in mice that had not been sensitized (Figure E3, B–C). The increased levels of Plaur during both the sensitization and challenge phases of these asthma models are consistent with the elevated levels of uPAR in asthmatic patients.

uPAR limits allergic airway inflammation.

To investigate the impact of uPAR on allergic airway disease, we compared the responses of male WT and Plaur^−/− mice to LPS/OVA sensitization followed by OVA challenge (Figure 1A). As expected, control male mice that had not been previously sensitized did not develop inflammation upon challenge with aerosolized OVA (Figure 1B). By contrast, animals that had been sensitized with LPS/OVA displayed increased numbers of inflammatory cells, particularly neutrophils, after OVA challenge. Of note, however, airway inflammation was even more robust in Plaur^−/− mice than in WT mice. This heightened response was not sex-specific, as increased airway inflammation was also seen in female Plaur^−/− mice (Figure E4). When we used a different model of asthma, the ASP/OVA model, Plaur^−/− mice once again displayed increased airway inflammation, especially eosinophils (Figure 1C). Taken together, these results demonstrate that uPAR functions to limit pulmonary inflammation in multiple mouse models of allergic asthma.

Figure 1.

Effect of uPAR on mouse models of allergic asthma. (A) Timeline for LPS/OVA- and ASP/OVA-mediated models of asthma. (B and C) Allergic airway inflammation following OVA challenge of (B) LPS/OVA- and (C) ASP/OVA-sensitized Plaur^+/+ (WT) and Plaur^−/− (KO) mice. Values shown represent mean cell numbers ± SEM. Data are compiled from (A) 3 experiments (n=16–18 ‘challenged only’ mice; n=15–24 ‘sensitized and challenged mice) or (B) 2 experiments (n = 6 ‘challenged only’ mice; n=14–15 ‘sensitized and challenged’ mice). (D and E) AHR in LPS/OVA-sensitized (D) and ASP/OVA-sensitized mice (E) following OVA challenge. Airway resistance (R) was calculated at baseline (B) and after exposure to increasing doses of methacholine (MCH). Dots represent individual mice with rectangles showing mean values ± SEM. Data shown are compiled from 2 experiments: (D) n=10–14 for ‘challenged only’ mice; n=15 for ‘sensitized and challenged’ mice; (E) n=6 for ‘challenged only’ mice; n=11–12 for ‘sensitized and challenged’ mice. Statistical significance is indicated: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (Two-way ANOVA with Sidak’s multiple comparison test).

uPAR limits allergen-induced AHR.

AHR is a salient feature of allergic asthma, and the increased inflammation in Plaur^−/− mice suggested that they might also have heightened physiologic responses to allergens. To test this, male WT and Plaur^−/− mice were sensitized and challenged as before, and at 48 h post-challenge anesthetized mice were analyzed by invasive measurements of AHR. Once again, two different adjuvants were separately mixed with OVA for the sensitizations: LPS and ASP, and in each of these experiments, Plaur^−/− mice displayed greater increased maximal airway responsiveness to methacholine (Figures 1, D and E) and increased sensitivity to methacholine (PC200) (data not shown). These results are consistent with the previously described association between Plaur polymorphisms and human asthma and show that in multiple mouse models of this disease, uPAR acts to limit both inflammatory and physiologic responses to allergen challenge.

uPA is nonessential for limiting allergic airway disease.

Given that Plau encodes the uPAR ligand, uPA, and is increased during both allergic sensitization and challenge, we studied Plau-deficient mice (lacking uPA). However, Plau-deficient mice had similar cellular inflammation and AHR as the wild-type mice in the LPS-OVA and ASP-OVA models of asthma (Figure E5A). This suggests that a ligand other than uPA likely acts together with uPAR attenuate allergic airway disease.

uPAR is dispensable for allergic sensitization.

During the allergic sensitization phase of asthma models, lung dendritic cells take up inhaled allergens and migrate to regional LNs to interact with and activate allergen-specific, naïve CD4⁺ T cells. This activation can be indirectly assessed by cytokine production in supernatants of cells prepared from regional LNs of sensitized mice and cultured in the presence of OVA (Figure E6A). Analysis of type 2 and type 17 cytokines in supernatants of LN cells did not reveal differences between WT and Plaur^−/− mice, regardless of whether LPS or ASP was used as the adjuvant during OVA sensitization (Figure E6, B and C). These findings suggest that uPAR is dispensable for allergic sensitization and instead acts at a later point in asthma models.

uPAR suppresses responses to allergen challenge.

We next focused on the role of uPAR during the challenge phase of the asthma models. To do this, we bypassed the sensitization phase by generating OVA-specific Th2 cells in vitro from naïve CD4⁺ T cells prepared from transgenic OT-II mice, which carry a T cell receptor specific for OVA. Equal numbers of these Th2 cells were adoptively transferred into either naïve WT or Plaur^−/− mice, and the recipient animals were then challenged with aerosolized OVA (Figure 2A). Analysis of inflammatory cells in the bronchoalveolar lavage fluid (BALF) of the challenged mice did not reveal genotype-specific differences in numbers of eosinophils (Figure 2B). However, Plaur^−/− mice that had received OVA-specific Th2 cells did develop significantly greater AHR than did similarly treated WT mice (Figure 2C). Thus, at least for Th2 cell-driven inflammation, uPAR acts during the challenge phase to attenuate AHR. The ability of uPAR to suppress AHR in this model without affecting inflammation also suggests that these two features of asthma might be controlled through distinct uPAR-dependent pathways.

Figure 2.

Effect of host uPAR on allergic responses to OVA challenge in mice receiving OVA-specific Th2 cells. (A) Timeline of Th2 cell transfers and OVA challenge. (B and C) Airway inflammation (B) and AHR (C) in mice receiving Th2 cells prior to OVA challenge. Data shown are compiled from 2 experiments giving similar results. (n=3–4, for ‘challenged only’ mice; and 12–13 for ‘sensitized and challenged’ mice. (D) Timeline for Plaur induction by LPS or ASP, followed by Th2 cell transfer and OVA challenge. (E and F) Airway inflammation as determined by cell number in BALF (E), and AHR (F) in mice treated with LPS prior to Th2 cell transfer. (G and H) Airway inflammation as assessed by cell number in BALF (G), and AHR (H) in mice treated with ASP prior to Th2 cell transfer. Dots represent individual mice, and rectangles show mean values ± SEM. Data shown were compiled from 2 experiments giving similar results. For G and H, n=9–12. Statistical significance is indicated: not significant (n.s.), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (B, C, F, H) Two-way ANOVA with Sidak’s multiple comparison test; (E-H) - Student’s t test.

In addition to bypassing the allergic sensitization phase, the above-described experiments also bypassed the induction of Plaur that occurs in WT mice following instillation of OVA together with an adjuvant (Figure E2, B–C). Thus, it remained unclear if this induction impacts Th2 cell-driven inflammation. We therefore performed additional adoptive transfer experiments, but this time we treated mice with adjuvant alone (without OVA) to induce Plaur in WT mice prior to the transfer of OVA-specific Th2 cells and OVA challenges (Figure 2D). Treatment with LPS prior to adoptive Th2 cell transfer led to greater eosinophilia and AHR in Plaur^−/− mice than in similarly-treated WT mice (Figures 2, E and F). Similarly, pretreatment with ASP prior to Th2 cell transfer also increased inflammation and AHR in Plaur^−/− recipients more than in similarly treated WT mice (Figure 2, G and H).

Given that Th17 cells are a major driver of airway neutrophilia, we generated OVA-specific Th17 cells and transferred them into WT and Plaur^−/− recipient animals prior to challenge with aerosolized OVA (Figure E7A). No genotype-specific differences in airway inflammation or AHR were seen upon transfer of these cells into naïve mice (Figures E7, B and C). However, Plaur^−/− mice that were given LPS prior to transfer of the Th17 cells had greater neutrophilic inflammation after OVA challenge than did similarly treated WT mice (Figures E7, D and E). By contrast, no differences in AHR were noted between these two strains (Figure E7F), in agreement with a previous report showing that on its own, IL-17 is insufficient to drive AHR (3). Taken together, these data show that uPAR functions during the challenge phase to suppress eosinophilic and neutrophilic inflammatory, as well as physiologic responses to inhaled allergen.

The increased inflammation of Plaur^−/− mice does not stem from increased inflammatory cytokines, immunoglobulins or mucus production.

Given that our experiments thus far had shown that uPAR acts during the challenge phase to suppress inflammation and AHR, we compared concentrations of pro-inflammatory cytokines in BALF of WT and Plaur^−/− mice. As expected, mice previously sensitized to LPS/OVA or ASP/OVA expressed higher amounts of pro-inflammatory cytokines upon OVA challenge than did unsensitized mice, with type 2 cytokines being highest in ASP/OVA-sensitized mice and IL-17 highest in LPS/OVA-sensitized mice (Figure E8A). Surprisingly, however, regardless of the adjuvant used for sensitization, there were no significant differences between OVA challenged WT and Plaur^−/− mice in levels of the pro-inflammatory cytokines, IL-4, IL-5, IL-13, or IL-17. There were also no genotype-specific differences in IL-6, TNF-α, nor in the regulatory cytokines, IL-10 and TGF-β. The differences in inflammation between WT and Plaur^−/− mice were also not due to differences in the eosinophil chemoattractants, CCL11 and CCL24, or the neutrophil chemoattractants, CXCL1 and CXCL5, as these molecules were similar in the two strains (Figure E8B). In keeping with these findings, WT and Plaur^−/− mice also had similar levels of serum IgE, IgG1 and IgG2a (Figure E8C), and had similar amounts of mucus-staining cells in the airway (Figure E9). Taken together, these data suggest that the differences in airway inflammation in WT and Plaur^−/− mice is not due to differences in concentrations of the classic soluble mediators of leukocyte recruitment and activation.

Single cell RNA-seq analysis of Plaur expression.

To characterize cells in the mouse lung that express Plaur during allergic airway inflammation, we performed single cell RNA sequencing (scRNA-seq) on lungs of mice that had undergone LPS/OVA-induced airway disease. Flow cytometry was then used to separately sort hematopoietic CD45⁺ cells and non-hematopoietic CD45^– cells 3 h post challenge for analysis. Using specific marker genes, we identified 31 cell clusters of CD45⁺ cells, including T cells (CD4⁺, CD8⁺, and Tregs), ILC2s, B cells, NK cells, dendritic cells (DC1s and DC2s), monocytes, alveolar macrophages (AlvMacs), recruited macrophages (recMacs), interstitial macrophages (IMs), and granulocytes (Figures 3A; Figure E10, A–C). The latter included neutrophils, eosinophils, basophils and mast cells, although these cell types were not clearly resolved from one another, likely because of their fragility during digestion. Plaur was highly expressed in granulocytes, although other cell types expressed Plaur at lower levels, including T cells, B cells, monocytes and recMacs (Figure 3, B and C). Analysis of the CD45^– non-hematopoietic cells revealed clusters of epithelial cells (alveolar type I and II, ciliated, club and basal cells), endothelial cells, fibroblasts and smooth muscle cells (Figures 3D; Figure E11, A and B). In addition to CD45⁺ cells, some CD45^– cells also expressed Plaur, including basal epithelial cells, endothelial cells, fibroblasts and smooth muscle cells (Figure 3, E and F).

Figure 3.

Single cell RNA-seq analysis of lung cells. Expression of Plaur in mouse lung cells following allergic inflammation. (A-F) Mice were sensitized to LPS/OVA and then subsequently challenged with aerosolized OVA, as in Figure 1A. At 3 h post-challenge, lung cells were separately sorted into CD45⁺ and CD45^– cells, and each cell type analyzed by single cell RNA sequencing (scRNA-Seq). (A - C) UMAP plots showing annotation of CD45⁺ cells (A) and their expression of Plaur (B). A heat map of Plaur expression in the indicated cell types (C). (D - F) UMAPS of CD45⁻ cell annotation (D) and Plaur expression (E), and a heat map of Plaur expression in the indicated cell types (F).

Plaur functions in eosinophils to limit allergic inflammation.

Given our finding that both hematopoietic and non-hematopoietic cells express Plaur during allergic airway disease, we tested the functional requirement of Plaur expression in these general cell types by generating reciprocal bone marrow chimera mice using WT and Plaur-/- mice (Figure 4A). WT and Plaur^−/− recipient animals were irradiated, given bone marrow cells from either WT or Plaur^−/− mice, and allowed to reconstitute for 12 wk. These reciprocal chimeric mice then underwent the ASP/OVA model of asthma. As anticipated, Plaur^−/− mice reconstituted with Plaur^−/− bone marrow (KO donors → KO recipients) had more airway eosinophilia than WT mice that received WT bone marrow (WT → WT) (Figure 4B). WT mice that received Plaur^−/− bone marrow (KO → WT) also had more eosinophilic inflammation than WT → WT mice, but to a lesser degree than the KO → KO mice. However, Plaur ^−/− mice that received WT bone marrow (WT → KO) only had minor increases in cellular inflammation compared with WT → WT mice. These findings demonstrate that Plaur expression in hematopoietic cells is primarily responsible for limiting allergic eosinophilic inflammation.

Figure 4.

Effect of cell autonomous Plaur on eosinophil migration in vivo. (A) Generation of reciprocal bone marrow-chimeric mice using Plaur^+/+ (WT) and Plaur^−/− (KO) marrow. (B) Airway inflammation in chimeric mice. Mean cell numbers in BALF ± SEM are shown. The data are compiled from 2 experiments yielding similar results. n=5, control group; n=18–20, ‘sensitized and challenged’ group. (C and D) Effect of Plaur disruption in eosinophils (Epx^cre x Plaur^fx/fx mice) on airway inflammation in the (C) ASP/OVA and (D) LPS/OVA models of asthma. Dots represent individual mice, and rectangles show mean values ± SEM. Data shown were compiled from 2 experiments giving similar results. n=8, controls; and n=16–22, sensitized and challenged mice. Statistical significance is indicated: not significant (n.s.), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. (Two-way ANOVA with Sidak’s multiple comparison test).

Our experiments thus far had shown that Plaur was most highly expressed in hematopoietic cells and that this expression was important for the attenuation of allergic airway inflammation. However, it remained unclear in which specific hematopoietic cell type(s) Plaur expression was required for the observed suppression of airway inflammation. To address this, we generated conditionally mutant Plaur mice (Plaur^fx/fx) (Figure E1) with the goal of crossing them to mice that selectively express Cre recombinase (Cre) in specific cell types. We first tested the functionality of the Plaur^fx/fx animals by crossing them to mice with a human cytomegalovirus (CMV) promoter-driven Cre, which deletes Plaur in all cells, including germ cells. As expected, these CMV-cre Plaur^fx/fx mice behaved similarly to the original Plaur^−/− mice in that they also displayed increased airway eosinophilia in both the ASP/OVA and LPS/OVS models of asthma compared with WT mice (Figure E12, A and B). By contrast, mice carrying the Plaur^fx/fx gene without Cre responded similarly to WT mice in the asthma models. Having thus established the functionality of the Plaur^fx/fx mice, we crossed them to animals that express Cre in various cell types, including antigen-presenting cells through an integrin alpha X (Itgax) (Cd11c)-cre transgene; myeloid cells using a Cre inserted into the lysozyme (Lyz)2 locus (Lyz2^Cre); and eosinophils through expression of Cre at the eosinophil peroxidase locus (Epx^cre) (21). Plaur was also deleted in epithelial cells using a Surfactant protein c (Sftpc)-cre (SPC-cre) transgene; and in endothelial cells using the TEK receptor tyrosine kinase (Tie2)-Cre transgene. Each of these novel strains was tested in the LPS- and protease-induced models of asthma. We found that Epx^cre Plaur^fx/fx mice were the only ones in which airway eosinophilia was greater than that of WT mice (Figure 4, C and D; Figures E13, A–D and E14, A–D). Collectively, these data suggest that during eosinophilic inflammation, uPAR functions in a cell autonomous manner to suppress eosinophil migration into the airway, and that loss of this protein leads to increased airway eosinophilia.

uPAR suppresses eosinophil recruitment to the airway.

The increased airway eosinophilia in Plaur^−/− mice might have resulted from increased migration of these cells to the airways, or to enhanced eosinophil survival. To study that, we assessed inflammation at various times post-allergen challenge. We found that the differences in inflammation between WT and Plaur^−/− mice occurred very soon after allergen challenge, and that the magnitude of those differences did not increase over time (Figure E15, A and B). This suggests that the difference in inflammation between the two strains stems from a difference in the recruitment of eosinophils to the airways, rather than from an enhanced survival of those cells.

uPAR suppresses eosinophil chemotaxis ex vivo.

To gain greater insight into the mechanisms by which uPAR affects eosinophil migration, we conducted ex vivo migration experiments using cells derived from the airways of WT or Plaur^−/− mice undergoing allergic airway inflammation. The protease-mediated model of asthma was used because it generates a large number of airway eosinophils. Equal numbers of viable eosinophils from the BALF were placed in the top well of a Boyden chamber, and migratory activity was assessed by counting cells that migrated from the top to the bottom chamber (Figure 5A). Chemokinetic, or non-directional, migratory activity was assessed by either adding no chemokine to the top or bottom chamber, or by adding the same amount of the eosinophil chemoattractant, CCL11, to both the top and bottom chambers. By contrast, chemotactic activity was assessed by adding the eosinophil chemoattractant, CCL11 (eotaxin1) to only the bottom chamber. No genotype-specific differences were noted for chemokinetic activity (Figure 5, B and C). However, after adding CCL11 to the bottom chamber only, eosinophils from Plaur^−/− mice migrated more strongly than did their counterparts from WT mice (Figure 5D). By contrast, no differences in cell migration were seen for eosinophils from Itgax^cre Plaur^fx/fx mice or Lys2^cre Plaur^fx/fx mice compared to their counterparts from Plaur^fx/fx mice (Figure 5E), showing that Plaur expression in antigen presenting cells does not affect eosinophil migration. However, eosinophils from Epx^cre Plaur^fx/fx mice also displayed enhanced migration compared to eosinophils from control Plaur^fx/fx mice lacking Cre. These results are consistent with the increased eosinophilic inflammation in Epx^cre Plaur^fx/fx mice and provides additional evidence that uPAR acts in a cell autonomous manner to suppress eosinophil migration.

Figure 5.

Effect of Plaur on the ex vivo migration of airway eosinophils. (A) Schematic representation of in vitro eosinophil (Eos) migration assays. Eos were derived from BALF of OVA-challenged mice undergoing the ASP/OVA model of asthma. (B - E) Percentages (mean ± SEM) of eosinophils that migrated from the top to bottom chamber of transwell plates. (B and C) Chemokinesis assays having CCL11 in neither (B) chamber or both (C) chambers. (D,E) Chemotaxis assay with CCL11 in the bottom chamber only. (D) Comparison of eosinophils from WT and Plaur^–/– mice. (E) Comparison of eosinophils from control Plaur^fxfx mice and eosinophils in Itgax^cre Plaur^fx/fx mice, Lyz2^cre Plaur^fx/fx mice, and Epx^cre Plaur^fx/fx mice. (B - E) Data are from a single experiment, representative of two (n=6). Statistical significance is indicated: Not significant (n.s.), *P < 0.05, **P < 0.01, ***P < 0.001 (Student’s t test).

Effect of uPAR on CCR3 and Siglec F.

Given that eosinophils were significantly increased in the BALF of Plaur^−/− mice and that these cells had increased migration ex vivo, we used flow cytometry to further study the effect of uPAR on eosinophils. To do this, we gated on single, live cells that were negative for CD88 and positive for CD11b, Siglec F, and CCR3 (Figure 6A). A comparison of eosinophils in BALF and lungs revealed an increase in cells of Plaur^−/− mice compared with WT mice in both compartments, although this difference was greater for BALF cells than for lung-derived cells (Figure 6B). Given the enhanced ability of eosinophils from uPAR-deficient mice to migrate towards CCL11, we investigated whether this was due to increased display levels of its receptor, CCR3, on Plaur-deficient eosinophils. However, the mean fluorescence intensity (MFI) of CCR3 was not different between WT and Plaur^−/− mice, regardless of whether the eosinophils were from the BALF or lungs (Figure 6C).

Figure 6.

Flow cytometric analysis of eosinophils in WT and Plaur^fx/fx mice. (A) Gating strategy for single, live, CD88^– Cd11b⁺ F4/80⁺ eosinophils from lungs of mice undergoing the ASP/OVA model. Also shown is a representative flow plot showing staining for CCR3 and Siglec F. (B) Eosinophils in BALF (left) and lungs (right). Data shown represent eosinophil numbers normalized to mean eosinophil number of WT mice. (C, D) Compiled data showing geometric mean of staining for CCR3 (C) and Siglec F (D) on eosinophils from the BALF (left) and from whole lung (right). Values shown represent the geometric mean ± SEM normalized to the geometric mean of WT eosinophils. Data shown were compiled from 2 experiments giving similar results (n=10–12). (E) Heat map showing expression of the indicated genes in granulocytes from WT and Plaur^−/− mice following of challenge of mice undergoing the OVA/LPS model. Statistical significance is indicated: not significant (n.s.), *P < 0.05, **P < 0.01 (Student’s t test).

Sialic acid-binding immunoglobulin-like lectin f (Siglec F) is a well-studied protein found on the surface of eosinophils and macrophages. Indeed, its display is routinely used, along with other cell surface proteins, to identify eosinophils in flow cytometry. In addition to this useful analytic feature, Siglec F has also been described a negative regulator of allergen-induced eosinophilia (22, 23). We therefore considered the possibility that uPAR deficiency might reduce display of Siglec-F. To test that, we used flow cytometry to compare display levels of Siglec F in WT and Plaur^−/− mice. Interestingly, we found that Siglec-F was indeed reduced on eosinophils from Plaur^−/− mice (Figure 6D), raising the possibility that uPAR deficiency leads to enhanced eosinophil chemotaxis at least in part by suppressing the expression of Siglecf or inhibiting the trafficking of Siglec F to the cell surface.

We next examined the scRNA-seq data to study differentially expressed genes in granulocytes (clusters 0, 1 and 12) of WT and Plaur^−/− mice. We found that several genes associated with eosinophil activation (24), including Cd69, Cd80, Cd274, Ptgs2, IL1b, Ccl3l3, and Nfkb1, were more highly expressed in cells of Plaur^−/− mice than in their WT counterparts (Figure 6E). Therefore, uPAR may limit eosinophilic chemotaxis in part by suppressing genes that enhance granulocyte activity or migration.

Increased AHR in Sftpc-cre Plaur^fx/fx mice.

We next followed up on our previous observation that naïve Plaur^−/− mice receiving OVA-specific Th2 cells developed significantly greater AHR than did similarly treated WT mice, even though the two strains had similar levels of eosinophilic inflammation. Given that Epx^cre Plaur^fx/fx mice, which selectively lack Plaur in eosinophils, displayed increased inflammation compared with WT mice, we compared AHR in these two strains. We found that despite their enhanced eosinophilic inflammation, Epx^cre Plaur^fx/fx mice did not develop greater AHR than did WT animals (Figure 7A). Together, these results show that inflammation and AHR are separable, at least to some extent. Given our finding that Plaur is expressed in some endothelial cells, we studied AHR in Tie2-Cre Plaur^fx/fx mice, but their responses did not differ from those of WT mice (Figure 7B). Finally, given the growing body of literature showing that changes in the lung epithelium can contribute to AHR (25), we studied this response in Sftpc-cre Plaur^fx/fx mice. It is well-known that Sftpc is expressed in type 2 epithelial cells (EC), but it is also expressed in the progenitors of most mature ECs. Thus, genetic loci that undergo Cre-mediated recombination at an early stage of development maintain that recombination in mature ECs. Accordingly, Sftpc-cre Plaur^fx/fx mice lack Plaur in the entire lung epithelium. Remarkably, Sftpc-cre Plaur^fx/fx mice displayed much greater increased maximal airway responsiveness to methacholine (Figure 7C) and sensitivity, as seen with increased PC200 (data not shown), than did WT mice (data not shown), even though the two strains had similar levels of inflammation. This was true both in mice undergoing the ASP/OVA model of asthma and those undergoing the LPS/OVA model. Taken together with our previous experiments, this result shows that uPAR acts in eosinophils to suppress inflammation and acts in epithelial cells to suppress AHR.

Figure 7.

Epithelial cell expression of Plaur limits airway hyperresponsiveness. (A-C) AHR in Plaur^fx/fx mice with no Cre, as well as in Plaur^fx/fx mice crossed to (A) Epx^cre mice (eosinophils), (B) Tie2^cre mice (endothelial cells), and (C) Sftpc^cre mice (epithelial cells). Results are shown for OVA-challenged animals previously sensitized to OVA using LPS (left) or ASP (right) as the adjuvant. Mean resistance values ± SEM are shown at baseline (B) and following exposure to increasing doses of methacholine (MCH). Data are compiled from (A) 2 experiments in the LPS/OVA model (n=10–11) and ASP/OVA model (n=12) mice; (B) 3 experiments LPS/OVA model (n=16–18) and 2 experiments in ASP/OVA model (n=13); and (C) 1 experiment each for LPS/OVA model (n=5–8) and ASP/OVA model (n=6). Statistical significance is indicated: not significant (n.s.), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001 (Two-way ANOVA with Sidak’s multiple comparison test).

Discussion

While the role of uPAR in hemostasis and clot resolution is well-described, the function of this molecule in inflammation is more complex and situation dependent. Early studies with Plaur-deficient mice showed that they have attenuated neutrophilic responses to infection with Pseudomonas aeruginosa with consequent increased spread of that pathogen (26). However, the role of uPAR during infection might depend on the nature of the infecting organism because uPAR is dispensable for inflammatory responses to respiratory viruses (27). In the current study, we found that Plaur is strongly induced in asthma models, but rather than enhancing inflammation, uPAR acted to suppress both allergic airway inflammation and AHR. Although perhaps surprising, our current findings are nonetheless consistent with a previous study of uPAR in a murine model of colitis. That study found that Plaur expression increases during colitis progression and that uPAR-deficient mice have an increased severity of inflammation and tissue damage compared with their WT littermates (28). That study also used bone marrow chimeric mice, and in agreement with our findings, found that loss of Plaur in hematopoietic cells drove the hyperinflammatory phenotype. Furthermore, analysis of macrophages from human patients with inflammatory bowel disease showed that they express lower levels of uPAR than macrophages of healthy individuals. Together, these various studies show that the tendency of uPAR to exacerbate inflammation or protect against it is highly dependent on the nature of the provoking substance or pathogen. It is also possible that the qualitative effect of uPAR on inflammation depends in part on the anatomic region, as loss of uPAR can lead to increased inflammation in both the gut and airway, which are both mucosal tissues with an active epithelial lining.

The signaling pathways that lead from uPAR to the suppression of inflammation and tissue damage in the lung and gut remain poorly understood. It is possible that some of the apparent discrepancies regarding the pro-inflammatory or anti-inflammatory functions of uPAR stem from the potential involvement of multiple ligands and co-receptors for uPAR. Although uPA is the best described uPAR ligand, there are at least eight others, including several integrins and vitronectin (10, 33). Whether one or more of ligands selectively contributes to inflammation, or suppression of it, should be a topic for subsequent studies. There are several possible signaling pathways that might account uPAR function in inflammation. One such possibility could include the EGF receptor, which has been shown to be downstream of uPAR (29, 30). This might be relevant because one of the ligands for EGFR is amphiregulin (AREG) (31), a cytokine that inhibits inflammatory cell infiltration and reduces lung injury in mice (32). Thus, impaired EGFR signaling from AREG due to loss of uPAR might contribute to the increased inflammation seen in uPAR-deficient mice. Another possible explanation for the suppressive capacity of uPAR in our experiments involves the chymotrypsin-generated version of uPAR (D2D3_88–274). This soluble protein can promote chemotaxis, but it can also suppress chemokine-directed migration, probably by interfering with chemokine-induced integrin activation (34). D2D3_88–274 could thus act in an autocrine or paracrine manner to suppress CCL11-induced eosinophil chemotaxis, both in vitro and in vivo. Alternatively, uPAR might suppress activation of genes that act in eosinophils to promote activation and migration.

An additional possibility to explain the increased migration of Plaur^−/− eosinophils is their reduced levels of Siglec F. This finding is consistent with previous studies showing that mice lacking Siglec F have enhanced inflammation compared to their WT littermates. This is likely a direct effect on eosinophils because Siglec F is not expressed in lung structural cells. The mechanism by which Siglec F exerts its negative regulatory effect remains uncertain, but one possibility is that it interferes with the signaling of CCR3, which is the main chemokine receptor for eosinophil chemotaxis. Whether uPAR interferes directly with CCR3 signaling, or indirectly by suppressing Siglec F, is a question that warrants further study.

One surprising finding of the current work was that although selective deletion of Plaur in eosinophils increased airway eosinophilia, AHR was unaffected in those animals. Conversely, selective deletion of Plaur in lung epithelium led to dramatic increases in AHR, but did not affect inflammation. Taken together, these findings suggest that airway inflammation and AHR are separable and likely arise from distinct uPAR-dependent mechanisms. The mechanism by which Plaur expression in epithelial cells suppresses AHR is unclear. These cells are reported to contribute to allergic sensitization, in part through their production of cytokines, such as TSLP, IL-25, IL-33, and GM-CSF (35). However, we did not find differences between WT and Plaur^−/− mice in their extent of sensitization, as measured by cytokine production in regional lymph nodes, OVA-specific IgE, or even type 2 cytokine production in the lung. Given that Plaur is induced during allergen challenge of previously sensitized mice, it is possible that uPAR suppresses AHR during the challenge phase in a manner independent of its effects on immune cells. This would be consistent with a previous study showing that glucocorticoid receptor expression in epithelial cells, but not in hematopoietic cells, is critical for suppression of AHR (36). Another potential explanation for our findings relates to the effect of epithelial cells on neurons that control smooth muscle contraction. Neuroendocrine cells are specialized epithelial cells that were recently shown to control airway contraction by releasing purinergic transmitters that act on afferent neurons to trigger protective reflexes (37). Whether uPAR suppresses epithelial cell-derived molecules that act directly on smooth muscle cells or acts on neuroendocrine cells that in turn promote contraction, warrants further study. Finally, although we generated several new strains of mice that selectively delete uPAR in multiple individual cell types, we have not specifically studied the role of this protein in airway smooth muscle cells. It will be interesting to see whether uPAR also functions in these cells, given their role in AHR. Regardless of the mechanism involved, the ability of uPAR to suppress AHR in mouse models of asthma suggests that molecules that mimic this activity might have potential as a novel therapeutic approach to treat allergic asthma.