Fibrinogen and lipopolysaccharide promote TLR4-dependent glucocorticoid resistance in airway mycosis-driven allergic airway disease

Abstract

Glucocorticoid-refractory asthma is a severe disease type linked to both T helper type 2 (T2) high and T2 low, neutrophil-high asthma endotypes, but the mechanisms governing glucocorticoid resistance are largely unknown. We show that glucocorticoid-resistant mouse allergic airway disease is linked to airway mycosis and lung-specific production of fibrinogen cleavage products (cryptokines). Mice challenged with the allergenic fungus Aspergillus niger and low-dose inhaled lipopolysaccharide (LPS) showed exaggerated airway hyperresponsiveness (AHR), developed T2 high and neutrophilic airway inflammation recapitulating severe asthma. AHR was not attenuated by glucocorticoid therapy, but required expression of Toll-like receptor 4 (TLR4). Low-dose intranasal LPS alone induced AHR in a manner previously shown for Aspergillus proteinase-derived cryptokines that required TLR4. We show further that cryptokines significantly enhance fibrinogen production in both the mouse lung and from primary human airway epithelial cells and that cryptokine-induced AHR is glucocorticoid resistant. Thus, glucocorticoid resistance in asthma may in part result from airway-specific production of the TLR4 ligands cryptokines and LPS. These findings suggest novel approaches to the management of glucocorticoid-resistant asthma.

Graphical Abstract

Introduction

Asthma is a non-communicable obstructive lung disease afflicting more than 300 million people worldwide. ^1,2 Although generally characterized by chronic airway inflammation, remodeling, and AHR, considerable phenotypic heterogeneity exists in asthma (e.g., allergic/atopic, late onset, non-atopic, obesity related, etc.) due to differences in underlying cause, disease progression, and severity, and sensitivity to current therapies. Asthma is typically well managed with inhaled bronchodilators and anti-inflammatory glucocorticoids, but can progress to a more severe, life-threatening glucocorticoid-resistant disease in up to 10% of patients. Understanding the mechanism of disease progression is an area of active investigation and necessitates a clear understanding of the immune responses driving asthma.

Asthma phenotypes are ascribed to two distinct pathophysiologic mechanisms or endotypes: T helper type 2 (T2)-high, eosinophilic, and T2-low, neutrophilic.^2–4 Mild to moderate atopic asthma is thought to reflect the T2-high eosinophilic endotype primarily and is associated with exposure to airborne allergens derived from animals (e.g., cat dander), arthropods (dust mites), molds (Aspergillus spp., Alternaria spp., etc.), and other fungi (Candida spp.) and is generally responsive to inhaled glucorticoids.⁵ However, T2-high asthma can progress to a more severe glucocorticoid-resistant disease and accounts for approximately 33% of all glucocorticoid resistant asthmatics.⁶ Glucocorticoid resistance is furthermore an intrinsic characteristic of the T2-low, neutrophilic asthma endotype, which has been associated with chronic fungal, bacterial (Chlamydia or Haemophilus) and viral (Influenza or Respiratory Syncytial Virus) infections, and prominent T1 and T17 responses.^3,7 Progression from a T2-high to a T2-low endotype may occur due to secondary infection of the airways with especially atypical bacteria³.

Multiple murine models of allergic asthma have been developed by sensitizing against ovalbumin, cat dander, house dust mite, fungal extracts, and other allergens to induce a T2-high disease characterized by T2 cytokine production including interleukin 4 (IL-4), IL-5, and IL-13, eosinophilia, and AHR. Adaptation of the allergic models to develop T2-low disease has been achieved by the addition of LPS during antigen challenge or secondary infection with gram-negative bacteria that express LPS, activating innate immune receptors (e.g., Toll-like receptor 4 (TLR4) and/or inflammasome) to promote T17 inflammation (IL-17), neutrophilia, and glucocorticoid-resistant AHR.^7–11 The dependency of innate immune signals in determining the disease state highlights the critical role of LPS/TLR4 in promoting disease.

Despite these robust observations, childhood LPS exposure is protective against atopic sensitization and asthma in adulthood.¹² LPS contamination suppresses AHR in ovalbumin murine models¹³, and farm-related LPS exposure can protect against asthma.¹⁴ In contrast to these findings, other studies have demonstrated that farm-related LPS exposure does not necessarily protect from asthma.¹⁵ Furthermore, LPS, presumably signaling through TLR4, can enhance expression of AHR¹⁶ and allergic airway inflammation in mice¹⁷. Thus, LPS exposure differentially affects asthma and allergic airway disease in complex ways through mechanisms that yet remain poorly understood.

Severe T2-low, neutrophilic disease is accompanied by – and largely due to – prominent airway obstruction and mucus plugs that contain fibrin. When fibrin deposition becomes excessive, the airway plugs acquire a rigid, clot-like consistency that can completely occlude major airways.¹⁸ Such physical obstruction of the airways by fibrin, termed plastic bronchitis^18–21, is accompanied by the production of fibrinogen cleavage products (FCPs) or cryptokines by endogenous (thrombin) and exogenous (fungal, dust mite, etc.) proteinases that promote AHR, goblet cell metaplasia, and mucus hypersecretion.^22,23 Fibrinogen cryptokines, like LPS, signal through TLR4 and can promote allergic airway disease.²³.

These observations indicate that T2 low, neutrophilic asthma reflects a more complex pathophysiology that could result from secondary exposure to LPS, significant innate immune activation involving fibrinogen cleavage, and airway infection. We have demonstrated that high dose (> 0.1 μg) airway exposure to LPS abrogates airway mycosis-induced T2-high, eosinophilic disease and instead induces a hypersensitivity pneumonitis (HP)-like syndrome.²⁴ In the current study, we examined the effects of LPS on the evolution of asthma-like endotypes using both an A. niger-induced model of allergic lung inflammation and an OVA immunization model. We show that exposure to fungus and LPS results in distinct asthma-like endotypes, including glucocorticoid-resistant eosinophilic and neutrophilic lung inflammation. Using relevant transgenic mice, we provide a unifying signaling mechanism that sheds insight into how allergic mediators (e.g., T2 cytokines, cryptokines), coordinate with LPS signaling in the airways to culminate in glucocorticoid resistant disease.

Results

Low-dose LPS exacerbates fungal-induced allergic airway disease

Previous studies have shown that LPS signals through TLR4 and that TLR4 is required for allergic airway disease in mice immunized and challenged with ovalbumin^22,25. However, LPS can also inhibit allergic airway disease and either promotes or inhibits asthma, suggesting that the relationship between LPS and asthma is complex and incompletely resolved. To address this further, we determined the effect of LPS (10–100 ng) exposure on allergic airway inflammation using an A. niger (AN) model of asthma (Fig. 1A). As expected, mice challenged with live AN spores alone developed significant AHR as assessed by acetylcholine-dependent increases in respiratory system resistance (R_RS) when compared to control (PBS treated) mice (Fig. 1B).²⁶ However, exposure to LPS in AN-treated mice resulted in markedly divergent dose-dependent responses, with lower doses (10 and 30 ng) significantly enhancing AHR, and higher LPS doses (100 ng and above²⁴) abrogating AHR (Fig. 1B).

Fig. 1.

Dose-dependent divergent effects of LPS on airway mycosis-induced allergic airway disease. (A) Mice were intranasally (i.n.) challenged with 4 × 10⁵ A. niger (AN) conidia without and with indicated doses of LPS over 17 days, and (B) respiratory system resistance (R_RS) in response to acetylcholine challenge, (C) bronchial airway lavage fluid (BALF) differential cell counts (eos: eosinophils, neutro: neutrophils, macro: macrophages, lymph: lymphocytes), and (D) lung CD3⁺CD4⁺ T cell cytokine production were measured. (E) Total serum IgE and fungus-specific IgG1 were measured by ELISA and expressed as optical density (O.D.) units. n ≥ 3 as indicated, *=p < 0.05. Two-way ANOVA was used to analyze data in (B), and other data were analyzed by one-way ANOVA. Data are from one of at least two separate and representative experiments.

Total inflammatory cells recruited to the airway were significantly increased in all AN-challenged mice, with fungus alone demonstrating the greatest increase (Fig. 1C). AN challenge induced predominant airway eosinophilia that was significantly suppressed when mice received LPS, with the highest LPS dose abrogating this response. In contrast, the addition of LPS induced airway neutrophilia, which became the predominant granulocytic response at all doses examined. Although no significant differences were found in airway macrophage abundance, a significant increase in airway lymphocytes was observed in mice challenged with fungus alone (Fig. 1C).

We further determined the cytokine profiles of T helper cells recruited to the lungs of fungus and LPS-challenged mice using flow cytometry. Fungal challenge significantly increased recruitment of T2-like cells expressing the cytokines IL-5, IL-13, and T17 cells expressing IL-17A, but not T1 cells producing IFN-γ, when compared to control mice (Fig. 1D). At all doses, LPS significantly reduced lung T2 and T17 cells relative to fungal treatment alone, particularly reducing the number of IL-13⁺ T helper cells in a dose dependent manner. Although lung T1 cells were present in very low abundance in all groups, their numbers were significantly increased in AN-treated mice receiving LPS.

Serum total IgE and fungus-specific IgG₁ antibody levels were significantly increased in all AN-treated groups, which were not suppressed by LPS (Fig. 1E). Although the addition of LPS to AN-treated groups did not affect IgG₁ responses, low-dose LPS significantly enhanced total IgE responses. Additionally, AN enhanced the infiltration of inflammatory cells into peribronchiolar and perivascular areas (Suppl. Fig. 2A), increased goblet cell metaplasia, and mucus production in conducting airways (Suppl. Fig. 2B), regardless of LPS dose. Thus, in this fungal-induced model of asthma, low doses of LPS skew airway mycosis-induced allergic airway disease towards a T2-low, neutrophilic inflammatory phenotype with enhanced AHR. Notably, in the same model, at higher doses (e.g., 100 ng), LPS abrogates AHR, eosinophilia, and IL-13, while other markers (e.g., IL-5, IL-17, and IgE) remain unaltered in comparison to the lower LPS doses.

Glucocorticoids alleviate OVA-induced allergic airway disease

We next determined the minimum efficacious dose of the common inhaled glucorticoid, fluticasone propionate (FP), using an ovalbumin (OVA) model of allergic airway disease that allows the tracking of antigen-specific endpoints (Fig. 2A). As expected, treatment with FP during intranasal OVA exposure resulted in a marked, dose-dependent reduction in AHR (Fig. 2B)²⁷. Notably, higher doses of FP (2 and 5 μg) reduced total airway inflammatory cells including eosinophils (Fig. 2C) and all FP doses reduced IL-5 and IL-13-producing T helper cells, but had no significant impact on T1 or T17 responses (Fig. 2D). The lowest dose of FP tested (0.5 μg) significantly enhanced airway neutrophils and lymphocytes with a trend toward enhancing lung T17 responses. The highest dose of FP used (5 μg) also reduced circulating OVA-specific IgE (Fig. 2E). Thus, FP suppresses AHR and key allergic inflammatory features of OVA-induced allergic airway disease.

Fig. 2.

Fluticasone attenuates allergic airway disease in ovalbumin (OVA)-challenged mice. (A) Protocol for challenging mice intraperitoneally (i.p.) with OVA and alum and intranasally (i.n.) with OVA and fluticasone propionate (FP) or dilauroylphosphatidylcholine (DLPC) liposome vehicle. (B) R_RS, (C) BALF inflammatory cells, (D) lung CD3⁺ CD4⁺ T cells expressing IFN-γ, IL-5, IL-13, and IL-17A, and (E) OVA-specific serum IgE were quantitated. n ≥ 3 in each group, *=p < 0.05. Two-way ANOVA was used to analyze data in (B) and one-way ANOVA in all else. Data are from one of at least two separate and representative experiments.

Fungal-induced allergic airway disease is refractory to glucocorticoids

The OVA model of asthma is advantageous given its simplicity and antigen specificity, but it fails to include critical biological properties of environmental allergens such as proteinase or live fungal activity that influence the evolution of immunological responses to airway allergens^26,28–30. While clearly anti-inflammatory in the OVA model, the effect of glucocorticoid treatment in a model of fungal-induced asthma that recapitulates many aspects of human disease,²⁶ is not known. Therefore, we next examined the effect of glucocorticoids in the fungal (A. niger) model of asthma. Mice were challenged with A. niger (Fig. 3A) for over 2 weeks while receiving daily i.n. FP or vehicle. All AN-treated mice developed AHR as compared to vehicle controls (Fig. 3B). Notably, treatment with FP did not attenuate AHR, with those receiving the low-dose (0.5 μg/day) of FP developing significantly enhanced AHR as compared to the group treated with higher FP doses (2 or 5 μg/day) or vehicle (Fig. 3B). Therefore, in contrast to OVA-sensitized mice, glucocorticoid treatment failed to suppress fungus-induced AHR at any dose. In contrast, as expected from the general anti-inflammatory effect of glucocorticoids, total inflammatory cells and eosinophils in AN-treated mice were significantly reduced by FP treatment at higher doses (Fig. 3C). The predominant airway neutrophilia observed in AN-treated mice was also significantly reduced in mice treated with FP, indicating enhanced susceptibility of neutrophilic responses to FP as compared to eosinophilic responses. High-dose FP treatment also significantly suppressed the recruitment of AN-dependent lung IL-13- and IL-17-cytokine-producing T cells (Fig. 3D). FP treatment did not reduce total IgE responses in all fungal-challenged groups; however, fungal-specific serum IgG₁ responses were significantly reduced in a dose-dependent manner (Fig. 3E).

Fig. 3.

Fluticasone attenuates inflammation but not AHR in mice with airway mycosis-dependent allergic airway disease. (A) Mice were challenged i.n. with AN and either FP or DLPC liposome vehicle over 17 days and then assessed for changes in (B) R_RS, (C) BALF inflammatory cells, and (D) total lung IFN-γ, IL-5, IL-13, and IL-17A-producing CD3 + CD4 + T cells. (E) Total serum IgE and fungal-specific IgG1 were quantified by ELISA. n ≥ 3 in each group, *=p < 0.05. Two-way ANOVA was used to analyze data in (B); for all other data, one-way ANOVA was used. Data are from one of at least two separate and representative experiments.

Glucocorticoids fail to reduce LPS-fungal-induced allergic airway disease

In the presence of low-dose LPS, fungal-induced allergic airway inflammation becomes neutrophil-predominant, and despite a reduction in T2 cytokines, AHR remained intact (Fig. 1). We next examined whether treatment with glucocorticoids can inhibit AHR in the LPS-AN-induced model of neutrophilic asthma (Fig. 4A). We found that mice receiving LPS-AN, regardless of glucocorticoid treatment, show significantly greater AHR as compared to fungal challenge alone (Fig. 4B). As expected, treatment with FP resulted in a significant reduction in airway neutrophilic and eosinophilic inflammation (Fig. 4C). Consistently, FP treatment significantly suppressed the recruitment of AN-mediated lung IL-5, IL-13- and IL-17-cytokine-producing lung T cells (Fig. 4D). Conversely, LPS markedly enhanced the lung and airway secretion of many cytokines and chemokines that were reduced in response to FP treatment (Suppl. Fig. 3). Low-dose LPS significantly enhanced total serum IgE, but not fungal-specific IgG1 responses. In contrast, FP had no effect on total serum IgE, but markedly suppressed fungus-specific IgG1, an effect that was reversed by LPS (Suppl. Fig. 4).

Fig. 4.

Airway mycosis and LPS-induced airway disease is resistant to intranasal glucocorticoids. (A) Mice intranasally (i.n.) challenged with A. niger (AN) with or without 10 ng of LPS and/or 5 μg of FP or DLPC were assessed for changes in (B) airway hyperreactivity measured by respiratory system resistance (R_RS) (C) BALF inflammatory cells, and (D) lung CD3⁺ CD4⁺ T cells producing IFN-γ, IL-5, IL-13, and IL-17A. n ≥ 3, * = p < 0.05. Two-way ANOVA was used to analyze data in (B) or by one-way ANOVA with post-tests otherwise. All data are from one of two separate and representative experiments.

We next assessed whole lung to determine the effect of FP and LPS on fungal growth and lung histology. We found that the addition of FP significantly impaired lung clearance of AN in mice challenged with and without LPS. In contrast, LPS alone markedly enhanced clearance of AN (Fig. 5A). Histologic analysis of the lungs revealed that neither FP nor FP and LPS challenges substantially reduced AN-induced peribronchovascular inflammation, nor did these interventions grossly reduce goblet cell metaplasia as assessed by PAS staining (Fig. 5B).

Fig. 5.

Glucocorticoids impair clearance of airway mycosis without inhibiting fungus-induced AHR and other asthma indices regardless of route of administration. (A) Fungal colony forming units (CFU) recovered from whole lungs normalized to gram (g) lung weight from Fig. 4 experiment. (B) H&E and PAS staining of lung tissue from indicated challenge groups. Scale bars in the low and high magnification are 750 μm and 150 μm, respectively. n > 3, *: p < 0.05. Two-way ANOVA was used to analyze data in (B) or by one-way ANOVA. (C) Mice were challenged every other day with A. niger (AN) or vehicle (PBS) for 4 wks (Phase 1 and Phase 2) as indicated. Dexamethasone (Dex) was given i.p. during Phase 2, with mice receiving an acute high dose (aDex) in the final 3 days of fungal challenges or a chronic immune suppressive dose (cDex) during Phase 2, after which R_RS was assessed (D). (E) Mice received A. niger and low dose (30 ng) LPS intranasally without and with dexamethasone (2 mg/kg i.p.) on the final three days, after which R_RS was assessed (F). n = 3–5, * = p < 0.05.

To confirm the glucocorticoid-resistant nature of the AN-LPS model of neutrophilic asthma, we assessed the effect of systemic dexamethasone (Dex) using acute and chronic protocols (Fig. 5C). Like the effect of FP on AHR, neither acute nor chronic Dex treatment reduced AHR (Fig. 5D). Similar results were observed in mice with severe neutrophilic disease that received both AN and LPS without and with Dex (Fig. 5E, F). Taken together, these data indicate that AHR as assessed in the AN-LPS asthma model is glucocorticoid-resistant regardless of the glucocorticoid used or the route of administration.

Glucocorticoid resistance is mediated by cryptokines and LPS

The previous findings paradoxically show that glucocorticoids fail to inhibit AHR induced by AN despite reducing lung T helper cells that promote AHR, i.e., IL-4, IL-13, and IL-17-secreting cells. In addition to these cytokines, our previous studies demonstrated that AHR in mice is also dependent on the presence of airway cryptokines (fibrinogen cleavage products) generated by fungal proteinases acting on fibrinogen and which signal through TLR4.^22,23 This suggests that cryptokines and LPS, both of which are TLR4 ligands, can independently induce AHR even when T cell-derived IL-4, IL-13, and IL-17 are suppressed, as we showed in response to glucocorticoid therapy.

To further test this possibility, we determined if low-dose LPS directly enhances AHR when delivered to the lungs (Fig. 6A). Intranasal challenge with LPS for 3 consecutive days induced minimal, but significant, AHR as compared to vehicle exposed control mice (Fig. 6B). We further exposed wild type and TLR4-deficient mice to AN with and without low-dose LPS (Fig. 6C). As expected, wild type mice treated with AN developed significant AHR, with LPS treated mice developing significantly stronger responses over AN alone (Fig. 6C). However, AN-challenged TLR4-deficient mice failed to develop significant AHR compared to vehicle-treated wild type or littermate TLR4^−/− controls (Fig. 6C). Consistent with our prior observations of the importance of TLR4 for expression of AHR in fungus-challenged mice, low-dose LPS enhanced expression of AHR through TLR4.^22,23.

Fig. 6.

Fungus-induced, glucocorticoid-resistant AHR is LPS, cryptokine, and TLR4 dependent. (A) Mice were challenged daily for 3 days i.n. with LPS (30 ng) and assessed for changes in (B) R_RS. Additional groups of wild-type and syngeneic mice deficient in TLR4 (TLR4^−/−, C) were assessed for changes in R_RS following i.n. challenge with AN with and without 10 ng LPS as in Fig. 1. D) After 14 days of i.n.challenge with either PBS or A. niger conidia, the left lung was obtained, mRNA extracted and RT-qPCR performed for the indicated fibrinogen (FBG) chains and prothrombin. n ≥ 3 in each group, *=p < 0.05 using two-way ANOVA or Student’s t-test for direct comparisons; ***=p < 0.001 by Student’s t-test. Data are from one of two separate and representative experiments.

To refine the mechanism by which airway mycosis leads to AHR, we next determined how fibrinogen in the airway becomes a substrate for AN proteinases.²⁶ Fibrinogen and its cleaved products (cryptokines) may be generated in the airway through inflammation-related extravasation (e.g., airway extrinsic fibrinogen)^31,32. However, we determined if transcripts for the three fibrinogen chains (α, β, and γ) as well as prothrombin could be expressed in the mouse lung (e.g., airway intrinsic fibrinogen). We found that expression of all four-coagulation factor-related genes occurs in the naïve mouse lung and that intranasal challenge of mice with A. niger significantly upregulates most transcripts (Fig. 6D).

Cryptokines induce coagulation factor secretion from human airway epithelia and contribute to glucocorticoid-resistant AHR

We could not confirm the presence of fibrinogen and prothrombin in the lungs of mice at the protein level due to the lack of suitable reagents. Therefore, we next cultured human primary sinonasal epithelial cells (SNECs) and found that they secrete fibrinogen (Fig. 7A), prothrombin (Fig. 7B), as well as antithrombin and plasminogen at baseline (Suppl. Fig. 5). Notably, we found that cryptokines consistently and significantly induced their secretion (up to 10,000-fold for fibrinogen) in vitro (Fig. 7A), with whole fibrinogen variably enhancing secretion of these factors (Fig. 7A, B; Suppl. Fig. 5). These findings demonstrate that airway epithelia participate in a feedforward loop involving a network of secreted coagulation factors that is strikingly augmented by airway cryptokines that are produced in situ in response to airway mycosis^22,23.

Fig. 7.

Cryptokines enhance secretion of fibrinogen and prothrombin from asthmatic airway epithelial cells and promote AHR independent of topical glucocorticoid. (A, B) Human sinonasal epithelial cells (SNEC) were cultured from healthy control, CRSwNP without asthma (CRSwNP) and CRSwNP with asthma (CRSwNP + asthma) groups and stimulated with either standard serum-free media or media supplemented with cryptokines (Crypto) for 3 days, after which supernatants were obtained and (A) fibrinogen and (B) prothrombin were quantified. (C) Mice were challenged intranasally with proteinase control (pPBS), whole fibrinogen (FBG), or cryptokines. Cryptokine treated mice received vehicle, FP i.n. or Dex i.p. in the final 3 days of challenge and (D) airway resistance across all Ach doses and (E) peak resistance are shown. n ≥ 3 in each group, *=p < 0.05 using two-way ANOVA or student t-tests for direct comparisons.

To determine if cryptokine-induced allergic inflammation responds to glucocorticoids, we assessed AHR after treatment with intranasal FP or systemic dexamethasone (Fig. 7C). Relative to vehicle-challenged controls containing residual fungal proteinase as used to produce cryptokines in vitro (pPBS), cryptokine-challenged mice had significantly higher AHR relative to vehicle control and fibrinogen (Fig. 7D, E). In contrast, mice receiving unproteolyzed whole fibrinogen were not significantly different from vehicle controls, highlighting the importance of cryptokine formation and proteinases for induction of AHR (Fig. 7E).²⁶ Mice challenged with cryptokines and either fluticasone or dexamethasone were not significantly different from cryptokine alone-treated mice. These data show that cryptokines derived from the proteolysis of fibrinogen promote topical glucocorticoid-resistant AHR.

Discussion

In contrast to asthma clinical phenotypes, the asthma endotypes of T2 high/eosinophil predominant and T2 low/neutrophil predominant disease are clinically significant because they predict clinical responses to glucocorticoids and biologics. To improve the management of glucocorticoid-resistant asthma, it is therefore essential to understand the immunologic, molecular, and environmental factors that contribute to these endotypes. We show that, in contrast to OVA challenge, airway mycosis, a powerful and common driver of asthma and chronic fungal rhinosinusitis (CFRS),^33,34 elicits glucocorticoid-resistant eosinophilic allergic airway disease in mice. Moreover, the addition of low-dose LPS- a commonly encountered immune modifier-to this model results in enhanced, T2 low/neutrophil-predominant disease that remains glucocorticoid-resistant. In part, the glucocorticoid-resistant nature of this model is due to the ability of LPS to directly promote AHR through TLR4. Notably, our studies reveal the existence of a fungus-inducible, airway epithelial-intrinsic coagulation factor network that enhances cryptokine production locally and also promotes glucocorticoid-resistant AHR. Cryptokines generated directly within the airways and LPS are therefore important mechanistic elements that explain glucocorticoid-resistant asthma with and without atopy, as well as glucocorticoid susceptibility of OVA-induced allergic airway disease where these factors are presumably reduced or absent. Our findings provide key insights into the management of glucocorticoid-resistant asthma and future research in this area.

Exaggerated airway contractility is ultimately induced immunologically, in part by the cytokines IL-4 and IL-13 signaling through their shared receptor chain, IL-4Rα, which activates STAT6 to drive disease^35–37. The dominant mechanism driving AHR in the OVA challenge model, an atopy-predominant asthma model, is overwhelmingly dependent on Th2 and Th17 cytokines. However, in the more complex and relevant airway mycosis model, a second immunological mechanism beyond Th2 and Th17 responses involving fungal proteinase-derived fibrinogen cryptokines is also essential to the expression of AHR²². Like LPS, cryptokines induce AHR through TLR4²². In contrast to LPS, cryptokine activation of TLR4 requires an intermediate binding partner, CD11b²³. Whether cryptokines can also activate STAT6 directly or indirectly through TLR4 and CD11b is the subject of ongoing studies in our laboratory.

We propose that expression of AHR and other aspects of mild/moderate atopic asthma with minimal concomitant airway mycosis may be due primarily to the airway production of IL-4 and IL-13 by allergen-activated Th2 cells that can be accelerated by Th17 cells (see Graphical Abstract)³⁸. We further propose that more complex asthma is driven by both antigen-activated T helper cells and fungal proteinase-dependent cryptokines, the latter becoming increasingly dominant as the burden of fungi and proteinases increase in the airways over time. Even more complex diseases may involve LPS or concomitant airway bacterial, viral, or parasitic infections that further drive STAT6-dependent AHR and inflammation. Because glucocorticoids promote the overgrowth of LPS-producing airway bacteria and fungi³⁹, this model provides a means for understanding why asthma becomes increasingly glucocorticoid resistant as airway mycosis and bacterial burdens advance due to chronic glucocorticoid-dependent immune suppression.

The distinct mechanisms through which LPS and possibly other pathogen-associated molecular patterns (PAMPs) signal to induce diverse inflammation in the lungs, including glucocorticoid-resistant allergic airway disease, remain poorly understood. Low dose LPS-induced exaggerated lung Th2 responses were previously traced to effects on dendritic cells¹⁷, but TLR4 downregulation, immune exhaustion, or induction of distinct inflammatory endotypes are additional possibilities deserving further exploration.

The crosstalk between Th2 and Th17 cells is also complex as they can negatively regulate each other.^40,41 Th17-mediated neutrophilic inflammation is glucocorticoid-insensitive as assessed in experimental models^42,43, due to impaired nuclear recruitment of histone deacetylase 2 (HDAC2), leading to continued enhanced expression of neutrophil chemoattractant and survival factors⁴³ or defective glucocorticoid receptor binding and translocation^44,45. These findings contrast with the enhanced lung (IL-4, IFN-γ, IL-17A) and BALF (TNF) cytokines secreted in response to low-dose LPS and AN challenge, as we show here, which potentially contribute to the enhanced AHR observed. Despite the uncertainties of LPS signaling in asthma and allergic airway disease models, our studies demonstrate that LPS-induced phenotypes are dose-dependent and require TLR4. We therefore propose that the paradoxical ability of LPS to be both pro- and anti-inflammatory in human and experimental allergic diseases may be related to the dose of LPS in the airway, a quantity that is difficult to ascertain in human populations.

Our findings are consistent with the previously demonstrated ability of inhaled fluticasone to improve asthma control and AHR^46,47. They also confirm the propensity of glucocorticoids to promote the overgrowth of airway fungi^48,49, but overgrowth of bacteria such as Chlamydia has also been demonstrated⁵⁰. Surprisingly, low-dose FP (0.5 μg) worsened disease control in the context of airway mycosis as assessed by enhanced AHR and airway eosinophilia and neutrophilia. In part, the mechanism that accounts for this worsening is likely related to the enhanced lung production of IL-13, but enhanced fungal growth with attendant increases in airway cryptokine generation could also be contributing. These new findings are important because they reveal multiple mechanisms by which inhaled glucocorticoids could potentially contribute to the gradual emergence of glucocorticoid-resistant severe asthma in humans.

Paradoxically, fluticasone was effective at reducing airway neutrophilia, but neutrophilic asthma is typically characterized as steroid-resistant. Glucocorticoids usually inhibit the recruitment of most inflammatory cells as documented in this report and elsewhere. They do so by signaling through the cytoplasmic glucocorticoid receptor to block pro-inflammatory transcription factors including NF-κB and AP-1⁵¹. Such blockade represses the expression of adhesion molecules, airway epithelial chemokines, and pro-inflammatory cytokines that promote cellular recruitment⁵². Glucocorticoids further act directly on inflammatory cells to repress responsiveness to pro-recruitment signals and induce apoptosis⁵². Here we propose that in some contexts in which asthma-related AHR is driven primarily by cytokines alone (exemplified by the ovalbumin asthma model (Fig. 2)), the steroid effect on immune cells may be sufficient to inhibit AHR. In contrast, in more complex clinical contexts in which asthma-related AHR is driven by a combination of immune cells, cryptokines and low dose LPS, glucocorticoid-related effects are insufficient to counter the AHR-promoting effects of these additional elements, which operate independently of leukocytes (see Graphical Abstract).

A strength of this study is the inclusion of complementary data derived from both human primary sinonasal epithelial cells and mice. However, this also illustrates a potential limitation as sinonasal epithelial cells may not behave similarly to bronchial epithelial cells that are more relevant in the asthma models. Although interspecies comparisons are lacking, multiple studies have compared the fidelity of nasal and bronchial epithelial cells with respect to inflammatory and other physiological outcomes and have found excellent agreement, sufficient to validate the use of upper airway epithelial cells to model the behavior of lower airway epithelial cells^53–56. An additional limitation is the acute nature of the mouse studies that may fail to reflect the chronic nature of asthma, including some features of glucocorticoid-resistant asthma. Future studies involving truly chronic asthma models, i.e., involving A. niger inhalational challenges for a minimum of 6 weeks or longer, are needed to verify the findings presented here.

In summary, our findings link glucocorticoid resistance observed in experimental asthma to the activation of multiple distinct pathways activated by the ligands IL-4, IL-13, cryptokines, and LPS. Of note, the IL-4Rα antagonist dupilumab is, like glucocorticoids, less effective against T2 low/neutrophilic asthma⁵⁷, likely due to the IL-4/IL-13-independent mechanisms leading to AHR operating in complex, fungal-driven disease as shown here. Our findings therefore suggest that improved management of glucocorticoid-resistant asthma could be gained through the use of direct inhibitors of TLR4 and airway coagulation factors combined with appropriate antifungal³³ and perhaps antibacterial⁵⁸ therapy, where bacterial bronchitis or pneumonia contribute LPS that potentially exacerbates AHR. As these therapies remain either experimental or under-investigated⁵⁹ in the context of asthma, our findings highlight the need for additional studies that clarify the relationships between airway fungi, bacteria, and inhaled PAMPs and glucocorticoids in the pathogenesis of severe, treatment-refractory asthma and related disorders.

Methods

Mice

C57BL/6J and TLR4^−/− (B6.B10ScN-Tlr4^lps-del/JthJ) mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and maintained under specific pathogen-free (SPF) conditions. Age and sex matched 4–8 wks old females were used and all procedures approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine.

Reagents

Fluticasone propionate (FP, Sigma-Aldrich) was prepared in 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC, Avanti) and reconstituted in sterile phosphate buffered saline (PBS, Sigma-Aldrich) as previously described⁶⁰. Dexamethasone-Water Soluble (Dex, Sigma-Aldrich) was dissolved in sterile PBS. Lipopolysaccharide (LPS, Sigma-Aldrich) was dissolved in tissue grade DMSO (Sigma-Aldrich) at 1000 × and diluted according to specified doses. Ovalbumin (OVA, Sigma-Aldrich) was precipitated on Al(OH)₃ (Alum, Thermo Scientific)⁹. Cryptokines (Fibrinogen Cleavage Products: FCP) were prepared as previously described from human whole fibrinogen and administered to mice as 60 mg/dose²².

Asthma models

Airway mycosis-induced allergic airway disease has previously been described.²⁶ In brief, mice were intranasally (i.n.) challenged with 4 × 10⁵ Aspergillus niger (AN) conidia, in the presence or absence of LPS as presented in Fig. 1A. In fungal-challenged steroid experiments, mice received daily i.n. FP therapy or intraperitoneal (i.p.) Dex for the final 3 days of fungal challenges or daily Dex for 2 wks. The ovalbumin-induced asthma model was adapted from a previously described method (Fig. 2A).⁹ Acute LPS challenged mice were i.n. treated with 30 ng of LPS for 3 consecutive days and then assessed for AHR (Fig. 5A).

Airway Mechanics

Airway physiologic analysis was assessed as previously described.²⁶ We assessed changes in respiratory system resistance (R_RS) after intravenous challenge of mice with increasing doses of acetylcholine chloride by whole-body, semi-invasive plethysmography.

Sample collection and single cell isolation

Serum was collected from blood following centrifugation. Bronchoalveolar lavage fluid (BALF) cells were collected from mice viaintubation with a 20½G catheter (Smith Medical) and lavage with 1.6 mL PBS. Lung homogenates were prepared from PBS-perfused lungs by incubating mechanically disassociating tissue in digestion buffer (HBSS (Gibco), 20% FBS (Gibco), 2 mg/mL collagenase type II (Worthington), 0.04 mg/mL DNase I (Sigma-Aldrich)) for 30 min at 37°C and subsequently passed through a 70 μM mesh basket (Fisher Scientific). Cells were washed and resuspended in PBS for immediate flow cytometric analysis, or complete RPMI (GenClone) for overnight culture at 37°C with 5% CO₂.

Flow cytometry

Single cell suspensions were stained with differential inflammatory cell panel (Suppl Table 1) and Live/Dead Fixable Blue (ThermoFisher) with FACS Lysing solution (BD Biosciences). For Intracellular cytokine staining (Suppl Table 2), lung single cell suspensions were incubated with 50 ng/ml phorbol 12-myristate 13-acetate (PMA, Biogems), 500 ng/mL ionomycin (Biogems), and Brefeldin A solution (BioLegend) for 4 h at 37°C with 5% CO₂, prior to Live/Dead and surface staining, permeabilization with Cytofix/Cytoperm Plus Kit (BD Biosciences), and intracellular staining. Samples were analyzed using a BD LSRFortessa cytometry (BD Biosciences) and FlowJo 10.4 software (BD Biosciences). Cells were quantified with CountBright beads (ThermoFisher) and differentiated by expression of surface and intracellular markers (Suppl. Fig. 1).

Histology

Lungs were inflated with 10% formalin (Sigma-Aldrich) at 25 cm H₂O pressure, embedded in paraffin, and serially sectioned (5 μm thickness) for hematoxylin and eosin (H&E, Thermo Scientific) or periodic acid Schiff (PAS, Thermo Scientific) staining in accordance to the manufacturer’s protocol.

Fungal burden

Lungs were weighed, and fungal burden was determined with serial dilutions of lung homogenate on Sabouraud’s agar plates (BD Biosciences) containing chloramphenicol (Sigma-Aldrich), incubated at 37°C overnight. Colony-forming units (CFU) were enumerated and normalized by the tissue weight.

Fungal antigen preparation

Fungal antigen was prepared from hyphal mats of A. niger that were ground in a planetary ball mill grinder (Retsch/Verder Scientific, Newtown, Pa) with 2-mm zirconium oxide balls in PBS as described³⁴.

Antibody and cytokine analysis

Sandwich ELISA was performed to detect antibodies. Sera were incubated in OVA, anti-IgE (BD Pharmingen), or fungal antigen coated High Protein Binding plates (Corning), detected with biotinylated IgE or IgG1 (BD Pharmingen) by Sav-HRP (BD Biosciences) activity in TMB substrates (Thermo Scientific) and read by spectrophotometer absorbance. Multiplex cytokine analysis from lung supernatant or BALF were quantified with mouse cytokine/chemokine magnetic bead and mouse T_H17 magnetic bead panel kits (Millipore) on a Luminex-based multi-analyte platform (Bio-Plex; Bio-Rad, Hercules, CA) in accordance to the manufacture’s protocol.²².

Sino-nasal epithelial culture

Patient-derived epithelial cells were collected from healthy and disease patients, and cultured in collagen-coated flasks as previously described.⁶¹ Adherent epithelial cells were harvested and re-seeded in 96-well plate, incubated overnight, prior to stimulation. Cells were stimulated with 50 μg/mL of Fibrinogen (FGN) Cleavage Products (FCP), FGN, or vehicle (media) for 6 hrs, changed media and incubated for 24 hrs. Secreted clotting factors were quantified with the LEGENDplex Human Fibrinolysis Panel (Biolegend).

Real time quantitative PCR

Total lung mRNA was isolated by Trizol (Thermo) and cDNA was synthesized with SuperScript III kit (Thermo). Taqman gene expression master mix and probes (Applied Biosystems) for identified genes were normalized to 18 s rRNA.

Statistical analysis

The results are expressed as the mean ± SEM. Statistical analyses were performed via one-way ANOVA with Fisher’s LSD test or two-way ANOVA using GraphPad Prism 8 software. The Kruskal-Wallis test was used for analysis of non-normally distributed data and Welch’s correction was used to assess heterogeneity of variance. For normally distributed data involving two comparisons, Student’s t test was conducted. Values of p < 0.05 were considered statistically significant.

Abbreviations:

AHR

airway hyperresponsiveness

AN

Aspergillus niger

BALF

bronchoalveolar lavage fluid

FCP

fibrinogen cleavage products

FP

fluticasone propionate

Dex

dexamethasone

HDAC2

histone deacetylase 2

HP

hypersensitivity pneumonitis

LPS

lipopolysaccharide

OVA

ovalbumin

PAMP

pathogen-associated molecular pattern

STAT6

signal transducer and activator of transcription 6

TLR4

Toll like receptor 4

T17

T helper cell type 17

T2

T helper cell type 2

SNEC

sinonasal epithelial cell

CRSwNP

chronic rhinosinusitis with nasal polyps

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.mucimm.2026.100354.