Airway administration of OM-85, a bacterial lysate, blocks experimental asthma by targeting dendritic cells and the epithelium/IL-33/ILC2 axis

Abstract

Background

Microbial interventions against allergic asthma have robust epidemiologic underpinnings and the potential to recalibrate disease-inducing immune responses. Oral administration of OM-85, a standardized lysate of human airways bacteria, is widely used empirically to prevent respiratory infections and a clinical trial is testing its ability to prevent asthma in high-risk children. We previously showed that intra-nasal administration of microbial products from farm environments abrogates experimental allergic asthma.

Objectives

To investigate whether direct administration of OM-85 to the airway compartment protects against experimental allergic asthma, and to identify protective cellular and molecular mechanisms activated through this natural route.

Methods

Different strains of mice sensitized and challenged with Ovalbumin or Alternaria received OM-85 intra-nasally, and cardinal cellular and molecular asthma phenotypes were measured. Airway transfer experiments assessed whether OM-85-treated dendritic cells protect allergen-sensitized, OM-85-naive mice against asthma.

Results

Airway OM-85 administration suppressed allergic asthma in all models acting on multiple innate and adaptive immune targets: the airway epithelium/IL-33/ILC2 axis, lung allergen-induced type-2 responses, and dendritic cells whose Myd88/Trif-dependent tolerogenic reprogramming was sufficient to transfer OM-85-induced asthma protection.

Conclusion

We provide the first demonstration that administering a standardized bacterial lysate to the airway compartment protects from experimental allergic asthma by engaging multiple immune pathways. Because protection required a cumulative dose 27 to 46-fold lower than the one reportedly active through the oral route, the efficacy of intra-nasal OM-85 administration may reflect its direct access to the airway mucosal networks controlling the initiation and development of allergic asthma.

GraphicalAbstract

CAPSULE SUMMARY

Administration of the OM-85 bacterial lysate to the airway compartment suppresses multiple components of experimental allergic asthma. The efficacy of this treatment may reflect its direct access to the airway mucosal networks that regulate asthma pathogenesis.

INTRODUCTION

Asthma, the most common chronic disease of childhood¹, imposes a societal burden higher than the one due to tuberculosis and HIV/AIDS combined². Profound changes in lifestyle and hygiene, and the concomitant sharp increase of asthma prevalence in the western world over the last decades, suggest that asthma pathogenesis is critically influenced by alterations in the environmental and/or endogenous microbial exposures with which humans evolved^{3, 4}. The role of environmental microbiota in promoting calibrated, timely immune responses and their ability to protect from asthma are highlighted by epidemiologic studies in children raised on traditional dairy farms, where frequent contact with livestock provides intense microbial exposures and asthma prevalence is extremely low^5-8. Studies in humans and germ-free mice suggest that live gut microbiota may also protect against asthma^9-12. However, the manufacturing, standardization and regulation of live microbial interventions remain challenging¹³.

These considerations likely motivate current efforts to decrease asthma burden using standardized microbial products that modulate immune maturation, thereby enhancing resistance against childhood asthma and lower respiratory tract infections, common harbingers of the disease^{14, 15}. OM-85 (Broncho-Vaxom™), a standardized, low-endotoxin alkaline lysate of 21 bacterial strains from five genera (Moraxella, Hemophilus, Klebsiella, Staphylococcus and Streptococcus) found in the human airways^{16, 17}, has been used empirically by over 40 million individuals for the prophylaxis of early-life recurrent upper respiratory infections^{18, 19}. OM-85 reduced the rate and duration of wheezing attacks in pre-school children with acute respiratory infections²⁰ and increased the time to severe lower respiratory illnesses in at-risk infants²¹. The NIH-sponsored ORal Bacterial Extract (ORBEX) Trial (NCT02148796) is currently testing whether OM-85 (Broncho-Vaxom®) given to high-risk, 6-18 months old infants for 10 days, monthly, for two consecutive years can increase time to occurrence of the first episode of wheezing lower respiratory tract illness during a three-year observation period off therapy.

So far, OM-85 has been administered orally - a choice shaped by decades of clinical practice²² and the notion that immune cells programmed in the gut may modulate distant mucosal responses^{23, 24}. On the other hand, we previously demonstrated that microbial products extracted from traditional farm dust abrogate allergic asthma in mice when administered intra-nasally (i.n.)^{7, 25} - a route that grants direct access to the regulatory networks residing in the airway mucosa. Therefore, we investigated whether direct administration of OM-85 to the airway compartment protects against experimental allergic asthma, and we characterized the cellular and molecular mechanisms underpinning protection through this natural route.

METHODS

Allergic asthma models

Different allergens and mouse strains were used to explore the effects of OM-85 in distinct genetic backgrounds. Protocols were chosen because they consistently induced robust cardinal experimental asthma phenotypes [airway hyperresponsiveness (AHR), broncho-alveolar lavage (BAL) eosinophilia and type-2 cytokine production in the lung] in all strains and allowed to keep cumulative OM-85 dosage constant across the study, except when otherwise stated.

Ovalbumin (OVA) model⁷:

7-8 week old Balb/c mice were sensitized intra-peritoneally (i.p.) with OVA (chicken egg white albumin, Sigma: 20 μg) + Alum (Imject Alum, Thermo Fisher: 50 μl) at day 0 and 14 and challenged i.n. with OVA (50 μg) at day 28 and 38 (Fig. E1A). Terminal assessments at day 39 included invasive lung function measurements, BAL cellularity with differentials, lung cytokine mRNA by reverse transcription (RT) quantitative (q)-polymerase chain reaction (PCR), lung and airway draining lymph nodes (ADLN) cell proportions and phenotypes by flow cytometry, and serum OVA-specific IgE by enzyme-linked immunoassay (ELISA). OM-85 (1 mg in 0.9% saline/treatment x 14 treatments) was administered i.n., 25 μl/nostril, under light anesthesia (1-5% isoflurane) every 2-3 days from day 0. The OM-85 dose was chosen because dose-response studies (0.001- 2.25 mg of OM-85/treatment) showed virtual abrogation of OVA-induced BAL eosinophilia at the 1 mg/treatment dose (Fig. E1B). Control mice received saline at the time of sensitization, challenge and treatment.

In selected experiments, Balb/c mice were immunized with OVA (50 μg) − Alum at day 0 and 7 and received OM-85 (1 mg/treatment) i.n. 8 times between day 0 and day 14. Mice were challenged with OVA (75 μg) i.n. at day 15 and 17 and assessed at day 19 (Fig. E2A).

In other experiments, OM-85 (1 mg in 50 μl 25 μl/nostril) was instilled i.n. every 2-3 days (14 times total) from day −5 into 7-week-old wild-type (WT) or MyD88^−/−Trif^−/− C57BL/ 6 mice that were sensitized i.p. with OVA (20 μg) + Alum (150 μl) at day 0 and 14 and challenged i.n. with OVA (75 μg) at day 26-28. Terminal assessments were performed at day 30 (Fig. E9).

Alternaria model:

7-8 week old Balb/c mice were sensitized i.n. with Alternaria alternata extracts (Greer Laboratories: 50 μg of dry weight/10 μg of protein in 50 μl of saline) at day 0 and 1, and challenged i.n. with Alternaria extracts (25 μg of dry weight/5 μg of protein in 50 μl at day 17, 18 and 19 ²⁶. OM-85 (1 mg in 50 μl, 25 μl/nostril) was administered every 2 days for 14 times from day −10. Terminal assessments were performed at day 20 (Fig. E11).

Statistical analyses

Because this study was designed to test the hypothesis that airway OM-85 administration affects allergen-induced asthma-related phenotypes, statistical analyses primarily compared allergen- and allergen+OM-85-treated groups, with other groups (saline, allergen only and OM-85 only) as experimental controls. Differences between allergen- and allergen+OM-85-treated mice were assessed by an unpaired, two-tailed t test or a Wilcoxon two-sample test after evaluating the normality of sample distribution with the Shapiro-Wilk normality test. When multiple groups were compared without an a priori hypothesis, inter-group differences were assessed using ANOVA on ranked data and Tukey’s post-hoc analyses for multiple testing adjustment. AHR dose-response curves for allergen- and allergen+OM-85-treated groups were compared using linear mixed-effects models, with treatment group and dosage of bronchoconstrictor as independent variables (fixed effects) and using each mouse identifier to account for repeated measures within each subject (random effects).

OM-85 preparations and all other experimental and analytical procedures are further detailed in the Online Repository.

RESULTS

Airway OM-85 administration suppresses OVA-induced experimental asthma

The impact of airway OM-85 administration on experimental allergic asthma was initially characterized in a classic OVA model⁷. Balb/c mice were sensitized with OVA+Alum intra-peritoneally (i.p.) at days 0 and 14 and challenged i.n. with OVA at days 28 and 38. Terminal assessments were performed at day 39 (Fig. E1A). OM-85 was administered i.n. 14 times from day 0 at 1 mg/treatment, a dose that maximally suppressed OVA-induced BAL eosinophilia in preliminary experiments (Fig. E1B). Fig. 1 shows that OM-85 robustly inhibited multiple features of OVA-dependent experimental asthma. Invasive lung function measurements revealed profound changes in both airway and tissue mechanics, with significant inhibition of total airway resistance (Rrs: P=9.5e-07, Fig. 1A), conducting airway (Newtonian) resistance (Rn: P=3.4e-07, Fig. 1B), and tissue damping (G), a parameter that reflects energy dissipation in the alveoli and is closely related to tissue resistance (P=1.3e-06, Fig. 1C). OVA-induced BAL eosinophilia was also essentially abrogated in OM-85-treated mice (P=0.005, Fig. 1D). These dramatic effects of i.n. OM-85 administration on allergic airway inflammation were accompanied by marked suppression of type-2 cytokine responses in the lung. Interleukin (Il) 13 and Il5 mRNA levels were significantly decreased (P=0.013 and P=0.043, respectively) in OVA+OM-85- compared to OVA-treated mice (Fig. 1E), and IL-13 protein levels in BAL were also markedly suppressed (P=0.012, Fig. 1F), whereas expression of the type-1 cytokine interferon-γ (Ifng) was unaffected. OM-85 also virtually abrogated OVA-induced lung goblet cell metaplasia, a signature of IL-13-induced responses²⁷ (P<0.001, Fig. 1G-H). Of note, in an abbreviated OVA model (Fig. E2A), as few as eight i.n. administrations of OM-85 (1 mg/treatment) were sufficient to significantly reduce BAL eosinophilia (P=0.03) and decrease lung Il13 and Il5 mRNA expression (P=0.02 and P=0.003: Fig. E2B-C). These data show that i.n. OM-85 administration suppresses cardinal cellular and molecular phenotypes of experimental allergic asthma. Interestingly, OVA-specific serum IgE levels were unaffected by i.n. OM-85 treatment (Fig. 1I), suggesting that the latter preferentially targets local rather than systemic immune responses.

Fig. 1. Airway administration of OM-85 suppresses OVA-induced experimental asthma.

Balb/c mice were treated with OVA and/or OM-85 as in Fig. E1A and phenotyped at day 39. A-C: Invasive lung function measurements in n=9 (A) and 4-6 (B, C) mice/group from two independent experiments. Symbols and bars denote the mean and standard error of the mean (SEM), respectively. Statistical differences between lung function dose-response curves of OVA- and OVA+OM-85-treated mice were assessed by linear mixed-effects models. D: Total BAL cellularity with differentials. E: Lung cytokine mRNA levels. F: BAL IL-5 and IL-13 protein levels. G-H: PAS lung section staining (20x magnification) and percentages of PAS⁺ airway cells. I: OVA-specific serum IgE levels. Data in D-I were obtained in 9 mice/group (except for H: n=4-6 mice/group) from two independent experiments. In D-F and H-I, filled symbols denote individual mice and bars denote mean ± SEM. Statistical differences between OVA- and OVA+OM-85-treated groups were assessed by an unpaired, two-tailed t test (D, H, I) or a Wilcoxon two-sample test (E, F).

Whole lung transcriptome analysis identifies genes and pathways targeted by OM-85

The lung transcriptome of mice treated i.n. with OM-85 in the presence or absence of OVA was then profiled by RNA-Seq to gain unbiased clues about the genetic networks that mediate OM-85-induced asthma protection. Principal component analysis demonstrated that the lung transcriptomes of mice treated with OVA+OM-85 tended to cluster with those of mice treated with saline or OM-85 alone but were distinct from those of OVA-treated mice (Fig. 2A). A total of 1,027 differentially expressed genes (∣log₂ fold change∣ > 0.5, FDR-adjusted P-value < 0.05) were identified comparing mice that had been treated with OVA and exhibited experimental asthma phenotypes (AHR and BAL eosinophilia) with mice that had received OVA+OM-85 and were protected from asthma (Table E1). Weighted gene correlation network analysis (WGCNA)²⁸ was then performed to identify modules of highly correlated genes within the latter data set and relate these modules to cardinal asthma traits (AHR and BAL eosinophilia). Four gene modules (blue, brown, yellow and turquoise) were identified (Fig. 2B), three of which were strongly and significantly associated with BAL eosinophilia and/or AHR (∣Pearson’s r∣ > 0.5 and P ≤ 0.004: Table E2). When the genes in the turquoise, blue and brown modules were further clustered and ranked by their differential expression in lungs of OVA+OM-85- vs. OVA-treated mice (Fig. 2C), it became evident that the genes most downregulated in OVA+OM-85-treated mice belonged to the large turquoise module, which was positively associated with asthma traits (Table E2 and Fig. E3). Among them were classical type-2 pathway members that promote allergic inflammation and remodeling (e.g., Il13, Retnlb), eosinophil activation and recruitment (e.g., Epx, Ear6, Ear7, Ccl24) and alternative macrophage activation (e.g., Arg1, Cd163) as well as pro-inflammatory cytokines and chemokines (e.g. Il6, Ccl4, Cxcl2, Cxcl3) and innate type-2 lymphoid cell (ILC2) signature genes (Il1rl1/ST2, Il33). In contrast, the smaller brown module, which was negatively associated with AHR and/or BAL eosinophilia (Table E2), included genes upregulated in mice treated with OVA+OM-85 compared to mice treated with OVA alone. Interestingly, the most upregulated gene in this module was CD207/langerin, which is expressed primarily on a subset of tolerogenic CD103⁺ DCs. These cells form an extensive network in the airway epithelial layer²⁹ and are uniquely capable of inducing FOXP3⁺ T regulatory (Treg) cells in vivo³⁰. Cldn1, a tight junction gene critical for epithelial barrier function³¹, was also upregulated in OVA+OM-85- vs. OVA-treated mice, pointing to direct barrier-enhancing effects of OM-85.

Fig. 2. Whole lung RNA-Seq analysis identifies genes and pathways targeted by OM-85 and allergen.

A: Clustering of mouse lung transcriptome profiles by principal component analysis. Each dot represents the transcriptome of a single mouse. The distance between two dots indicates how similar they were to one another. Ellipses represent the 95% confidence intervals for each group. (saline, n=6 mice; OVA, n=7 mice; OM-85, n=7 mice; OVA+OM-85, n=7 mice). B: Hierarchical clustering of genes differentially expressed in OVA+OM-85- vs. OVA-treated mice, as implemented by WGCNA. Four distinct modules were identified and are labeled with colors below the dendrogram (turquoise, blue, brown and yellow). C: Sample clustering of genes differentially expressed in OVA+OM-85- vs. OVA-treated mice. The heatmap shows clustering based on expression levels of the genes (rows, with gene symbols on the right) within each mouse (columns). Shown are the 100 most differentially regulated genes, ranked by ∣log₂(fold change)∣. The bar at the top is color-coded by treatment as in panel A.

Five hundred and eighteen genes were differentially expressed in mice treated with OM-85 alone compared to saline-treated, allergen-naive mice ∣log₂ fold change∣ > 0.5, FDR-adjusted P-value < 0.05, Table E3). Some of those genes are involved in immunoregulation (Tigit, Ctla4, Icos, Irf4) and recognition of microbial metabolites (Gpr55, Gpr65, Gpr82 and Gpr25), but – as in the OVA+OM-85 vs. OVA treatment comparison – one of the genes most upregulated by OM-85 in allergen-naïve mice was tolerance-associated CD207/langerin (Fig. E4). In combination, these unbiased analyses suggest that the asthma protection induced by i.n. OM-85 treatment involves a rewiring of lung transcriptional networks in which upregulation of DC-based tolerogenic pathways promotes downregulation of type-2 and pro-inflammatory responses.

i.n. OM-85 treatment induces a tolerance-promoting landscape in the lung

To develop the leads provided by transcriptome analyses and further dissect the mechanisms underlying OM-85-dependent suppression of experimental asthma, lung and airway draining lymph node (ADLN) cells were isolated from mice treated with saline, OVA or OVA+OM-85 as in Fig. E1A. DCs and Treg cells were phenotyped by flow cytometry using the gating strategies shown in Fig. E5-E6. Fig. 3A shows that overall, conventional CD45⁺MHC class II⁺F4/80⁻CD11c⁺ DCs (cDCs) were similarly expanded in the lungs of mice treated with OVA or OVA+OM-85 compared to saline-treated mice. However, when cDCs were further subdivided into CD11b⁺CD103⁻ and CD11b⁻CD103⁺ populations representing dominant airway cDC subsets with specialized Th2-promoting or tolerogenic roles, respectively^32-34, it became clear that numbers of OVA-induced Th2-promoting lung CD11b⁺CD103⁻ cDCs were significantly (P=0.01) reduced in OVA+OM-85-treated mice. Conversely, tolerogenic CD11b⁻CD103⁺ cDCs were significantly expanded (P=0.03) in OVA+OM-85-treated compared to OVA-treated animals. No significant differences in these cDC subpopulations were detected in ADLNs even though the OVA+OM-85-induced increase in CD11b⁻CD103⁺ cDCs approached statistical significance (P=0.06) (Fig. 3B). Notably, as suggested by our transcriptomic analyses, proportions of CD207⁺CD103⁺ lung cDCs were significantly (P=0.001) increased in OVA+OM-85- vs. OVA-treated mice (Fig. 3C). Because of their unique capacity to induce and activate FOXP3⁺ Treg cells³⁰, these cells likely promote a tolerogenic landscape following OM-85 treatment. Indeed, total numbers and proportions of lung FOXP3⁺CD3⁺CD4⁺ Treg cells were robustly increased in OVA+OM-85- compared with OVA-treated mice (P=0.015 and P<0.001, respectively; Fig. 3D-E). Moreover, CTLA-4⁺ Treg cell proportions and CD69 expression, which are associated with Treg function and activation³⁴, were significantly upregulated in the lungs of OVA+OM-85-treated mice relative to animals that had received only OVA (P=0.01 and P=0.003, respectively; Fig. E7). Finally, a striking, highly significant (P<0.001) increase in the ratio between FOXP3⁺CD3⁺CD4⁺ Treg and IL-13⁺CD3⁺CD4⁺ Th2 cells was found in the lungs of OVA+OM-85- vs. OVA-treated mice (Fig. 3F and Fig. E8). These data converge to demonstrate that i.n. OM-85 administration induces tolerance-promoting pathways in the lungs of allergen-treated mice.

Fig. 3. i.n. OM-85 treatment induces a tolerance-promoting landscape in the lung.

Lung and ADLN cells were isolated at day 39 from Balb/c mice treated with saline, OVA or OVA+OM-85 as in Fig. E1A. cDCs (CD45⁺MHC class II⁺F4/80⁻CD11c⁺: Fig. E5), Treg cells (FOXP3⁺CD3⁺CD4⁺: Fig. E6) and IL-13⁺CD3⁺CD4⁺ Th2 cells (Fig. E8) were analyzed by flow cytometry. A, B: Total numbers of lung (A) and ADLN (B) CD11c⁺, CD11b⁺CD103⁻ and CD11b⁻CD103⁺ cDCs from one experiment (n=4-6 mice/group) representative of three independent experiments. C: Proportions of CD207⁺CD103⁺ DCs among CD11c⁺ cDCs (n=4 mice/group from one experiment representative of two independent experiments). D-E: Total numbers/lung (D) and proportions (E) of lung FOXP3⁺CD3⁺CD4⁺ Treg cells. Data are from 7-8 mice/group pooled from three independent experiments. F. FOXP3⁺CD3⁺CD4⁺ Treg/CD3⁺CD4⁺IL-13⁺Th2 cell ratios in the lungs. Cells were analyzed by flow cytometry after a 5-hour in vitro restimulation with PMA + ionomycin. Data are from 5-6 mice/group pooled from two independent experiments. In all panels, filled symbols denote individual mice and bars denote mean ± SEM. An unpaired, two-tailed t test (A, B, C, F) or a Wilcoxon two-sample test (D, E) were used for statistical analysis.

Airway transfer of OM-85-reprogrammed DCs is sufficient to protect mice from allergic asthma

To directly assess the functional nexus between the tolerogenic effects of OM-85 and suppression of allergic asthma, bone marrow-derived DCs (BMDCs) were generated and pulsed with OVA in vitro for 2 days after a 2 day-pre-incubation with saline or OM-85 (Fig. 4A). Flow cytometry analysis using the gating strategy developed to phenotype lung DCs in vivo (Fig. E5) revealed that total numbers of Th2-promoting CD11c⁺CD11b⁺ BMDCs decreased significantly in OVA-stimulated samples pretreated with OM-85 (Fig. 4B). In parallel, mean fluorescence intensity (MFI) for CD103, a marker of tolerogenic activity, increased substantially in OM-85-pretreated BMDCs (Fig. 4C), highlighting the overall phenotypic similarity between BMDCs and lung cDCs exposed to OVA and/or OM-85 (Fig. 3).

Fig. 4. Airway transfer of OM-85-reprogrammed BMDCs is sufficient to protect mice from allergic asthma.

A: BMDCs were pre-treated with saline or OM-85 and pulsed with saline or OVA, and then transferred i.n. into naive mice that were challenged i.n. with OVA (100 μg) at day 20-22 and characterized at day 24. B: Total numbers of CD11c⁺CD11b⁺ BMDCs. C: CD103 expression in saline/saline-, saline/OVA- and OM-85/OVA-treated BMDCs. Data are from 4 mice/group pooled from two independent experiments (gating strategies in Fig. E5 and Fig. 3). D: FOXP3⁺ proportions (left) and expression (right) among CD3⁺CD4⁺ T cells co-cultured with saline/saline-, saline/OVA- or OM-85/OVA-treated BMDCs (n=6 samples from two independent experiments). E-J: Asthma phenotypes in mice transferred i.n. with saline/saline-, saline/OVA- or OM-85/OVA-treated BMDCs (n=8 mice/group from one experiment representative of three independent experiments). E-G: Invasive lung function measurements. Symbols denote mean and bars denote SEM. Statistical differences were assessed by repeated measures mixed models. H: Total BAL cellularity with differentials. I-J: Cytokine mRNA levels in lungs (I) and ADLNs (J). Filled symbols denote individual mice and bars denote mean ± SEM. An unpaired, two-tailed t test (B, D, H-I) or a Wilcoxon two-sample test (J) were used for statistical analysis.

To determine whether functional reprogramming paralleled the phenotypic changes detected in BMDCs treated with OM-85 in vitro, saline- or OM-85-pretreated BMDCs derived from naïve mice were pulsed with OVA and co-cultured for four days with splenic CD4⁺ T cells from OVA-immunized mice. Flow cytometry analysis revealed significantly (P<0.001) increased FOXP3⁺CD3⁺CD4⁺ T cell proportions in co-cultures containing OM-85-pretreated, OVA-pulsed BMDCs compared to saline/OVA BMDC-CD4⁺T cell co-cultures (Fig. 4D, left). FOXP3 MFI was also significantly (P<0.001) enhanced in OM-85/OVA BMDC-CD4⁺T cell co-cultures (Fig. 4D, right), demonstrating that the Treg cell-inducing capacity of BMDCs was decisively amplified by in vitro reprogramming with OM-85.

Next, these in vitro-generated, saline- or OM-85-pretreated BMDCs pulsed with OVA were transferred i.n. into naïive Balb/c mice that received three consecutive OVA challenges 10 days later. As expected, saline-pretreated, OVA-pulsed BMDCs elicited brisk experimental asthma, as revealed by exacerbated responses of both the conducting airways and peripheral lung tissue (Fig. 4E-G), increased BAL eosinophilia (Fig. 4H), and enhanced type-2 cytokine expression in the lungs and ADLNs (Fig. 4I-J). Strikingly, all these allergen-driven responses were significantly and strongly suppressed in mice that had received i.n. transfers of OVA-pulsed BMDC pre-treated with OM-85 (Fig. 4E-J). These results demonstrate that adoptively transferred, OM-85-reprogrammed DCs were sufficient to mediate the asthma protection induced by airway administration of OM-85.

OM-85-induced inhibition of allergic asthma is mediated by Myd88/Trif-expressing DCs

To assess whether i.n. OM-85-dependent suppression of experimental asthma is strain-specific, the OVA model used in Balb/c mice was adapted to C57BL/6 mice⁷, keeping the cumulative dose of OM-85 constant (Fig. E9). The asthma-suppressive effects of OM-85 were found to be strain-independent because the cardinal phenotypes detected in OVA+OM-85-treated Balb/c mice (inhibition of OVA-induced AHR, BAL eosinophilia and type-2 cytokine gene expression in the lung) were also observed in similarly treated C57BL/6 mice (Fig. 5A-C and Fig. E10). Of note, OM-85-induced asthma protection required innate immune signaling, because OM-85-treated Myd88^−/−Trif^−/− C57BL/6 mice failed to show significant suppression of cardinal experimental asthma phenotypes (Fig. 5D-F).

Fig. 5. The Myd88/Trif pathway is required for OM-85-induced inhibition of allergic asthma.

WT (A-C) and Myd88^−/−Trif^−/− (D-F) C57BL/6 mice were sensitized with OVA and treated i.n. with OM-85 (1 mg/treatment x 14 treatments: Fig. E9), and phenotyped at day 30. A, D: Respiratory system resistance in n=7-8 (A) and 4-9 mice/group (D) from two independent experiments. Symbols denote mean and bars denote SEM. Statistical differences between lung function dose-response curves of OVA- and OVA+OM-85-treated mice (A, D) or saline- and OVA-treated mice (D only) were assessed by repeated measures mixed models. B, E: Total BAL cellularity with differentials. C, F: Lung cytokine mRNA levels. Data were from n=7-8 (B, C) and 4-9 mice/group (E, F) from two independent experiments. Filled symbols denote individual mice, and bars denote mean ± SEM. A Wilcoxon two-sample test was used for statistical analysis.

Adoptive transfer experiments were then undertaken to identify cell type(s) sufficient for OM-85-induced, Myd88/Trif-dependent asthma protection. I.n. transfer of saline-pretreated, OVA-pulsed, WT BMDCs to naïive WT C57BL/6 mice induced AHR, BAL eosinophilia, and Il5 and Il13 expression in the lung. All these responses were significantly suppressed in mice receiving OVA-pulsed WT BMDCs pre-treated with OM-85 in vitro for 2 days (Fig. 6A-C). Transfer of saline-pretreated, OVA-pulsed Myd88/Trif-deficient BMDCs also promoted lung type-2 responses in WT mice, albeit at attenuated levels, but transfer of OM-85-pretreated, OVA-pulsed Myd88/Trif-deficient BMDCs failed to suppress OVA-induced AHR, BAL eosinophilia, and lung type-2 cytokine expression (Fig. 6D-F). These results conclusively demonstrate that Myd88/Trif expression in DCs is necessary and sufficient for OM-85-dependent DC reprogramming and OM-85-induced asthma protection in the OVA model.

Fig. 6. OM-85-induced inhibition of allergic asthma is mediated by Myd88/Trif-expressing DCs.

BMDCs from WT (A-C) and Myd88^−/−Trif^−/− (D-F) C57BL/6 mice were generated in vitro, pre-treated with saline or OM-85 (0.3 mg/ml: A-C, or 1 mg/ml: D-F) for 2 days, and pulsed with saline or OVA (100 μg/ml) for 2 additional days. On day 10, saline/saline-, saline/OVA- or OM-85/OVA-treated BMDCs were transferred i.n. into untreated WT C57BL/6 mice (1x10⁶ cells/mouse). Mice were challenged i.n. with OVA (100 μg) at day 20-22 and phenotyped at day 24. A, D: Respiratory system resistance in 6-7 mice/group from two independent experiments. Symbols denote mean and bars denote mean ± SEM. Statistical differences between lung function curves in mice that had received saline/OVA- or OM-85/OVA-treated BMDC (A, D) were assessed by repeated measures mixed models. B, E: Total BAL cellularity with differentials. C, F: Lung cytokine levels. B, C, E and F: data from 6-7 mice/group from two independent experiments. Filled symbols denote individual mice and bars denote mean ± SEM. An unpaired, two-tailed t test (B) or a Wilcoxon two-sample test (C) were used for statistical analysis.

Airway administration of OM-85 suppresses Alternaria-induced experimental asthma

In the final set of experiments, we asked whether airway OM-85 administration affects experimental asthma driven by Alternaria alternata. Exposure to Alternaria, a protease-rich fungus, is strongly associated with the development of allergic asthma in children and adults³⁵, especially in semi-arid environments³⁶, and these effects reflect its ability to damage the airway epithelial barrier. In mice, airway exposure to Alternaria in the absence of adjuvants induces allergic lung inflammation that closely recapitulates the human phenotype and relies on the activation of innate pathways involving IL-33 and ILC2s^{26, 37-40}. We sensitized Balb/c mice i.n. with Alternaria at day 0 and 1, challenged these animals i.n. at days 17-19, and phenotyped them at day 20²⁶. I.n. administration of OM-85 (1 mg/treatment x 14 treatments) began 10 days before Alternaria sensitization and ended 1 day before challenge (Fig. E11). This prophylactic/concurrent regimen enhanced the translational significance of our studies because it kept the cumulative dose of OM-85 constant across models and was partially akin to the design of the ORBEX trial (NCT02148796).

Fig. 7A-C shows that i.n. sensitization and challenge with Alternaria profoundly altered airway and tissue mechanics, with strong increases in total and conducting airways resistance and tissue damping, all of which were virtually abrogated by i.n. administration of OM-85. Moreover, Alternaria-dependent BAL eosinophilia and lung Il5 mRNA levels were strongly suppressed (P=0.001 and P=0.004, respectively: Fig. 7D-E). Expression of Il13 but not Il10 or Ifng also appeared to decrease. As in the OVA model, lung immune cell phenotyping revealed a significant increase of tolerogenic CD103⁺ cDCs and FOXP3⁺CD3⁺CD4⁺ Treg cells in mice treated with Alternaria+OM-85 compared to mice treated with Alternaria alone (Fig. 7F and H). However, CD207⁺CD103⁺ cDC populations were not significantly expanded in Alternaria+OM-85-treated mice (Fig. 7G). Interestingly, as few as nine i.n. OM-85 administrations beginning one day before Alternaria sensitization (Fig. E12A) significantly inhibited BAL eosinophilia (P=0.02) and expression of Il5 and Il13 mRNA in the lung (P<0.001 and P=0.03, respectively: Fig. E12B-C).

Fig. 7. Airway administration of OM-85 suppresses Alternaria-induced experimental asthma.

Balb/c mice were treated with Alternaria alternata extracts and/or OM-85 (1 mg/treatment x 14 treatments) as in Fig. E11. Terminal assessments were performed at day 20. A-C: Invasive lung function measurements from one experiment (n=6 mice/group) representative of three independent experiments. Symbols denote mean and bars denote mean ± SEM. Statistical differences between lung function dose-response curves in Alternaria- and Alternaria+OM-85-treated mice were assessed by repeated measures mixed models. D: Total BAL cellularity with differentials. E: Lung cytokine mRNA levels. Data are from one experiment (n=6 mice/group) representative of four (D) and two (E) independent experiments. F-G: Total numbers of lung CD103⁺CD11c⁺ cDCs and percentages of lung CD207⁺CD103⁺ cDCs in one experiment (n=4 mice/group) representative of two independent experiments. H: Total numbers of lung FOXP3⁺CD3⁺CD4⁺ Treg cells in 6-7 mice/group pooled from two independent experiments. Filled symbols denote individual mice and bars denote mean ± SEM. An unpaired, two-tailed t test (D, F-H) or a Wilcoxon two-sample test (E) were used for statistical analysis.

Effects of airway OM-85 administration on early Alternaria-dependent innate immune activation in the lung

We then took a multi-pronged approach to assess whether OM-85 affects innate events that initiate Alternaria-induced mucosal type-2 inflammation. A 24-hour in vitro pre-treatment with OM-85 not only preserved trans-epithelial electrical resistance (TEER) in serum-starved human airway epithelial cells, but also markedly mitigated the decrease in TEER induced by a 3- or 6-hour exposure of these cells to Alternaria extracts (P<0.001 and P=0.004, respectively: Fig. 8A), thereby demonstrating that OM-85 supports airway epithelial barrier function. This notion was reinforced by RNA-seq analysis of the same human airway epithelial cells, which revealed upregulation of multiple signature genes related to tight junctions and epithelial barrier function³¹ following a 24-hour stimulation with OM-85 (Table E4). Significant boosting of TEER was also observed in serum-starved, OM-85-treated murine tracheal epithelial cells cultured at the air-liquid interface (Fig. E13).

Fig. 8. Effects of airway OM-85 administration on early Alternaria-dependent innate immune activation in the lung.

A: 16HBE14o- human airway epithelial cells were incubated in serum-containing (FCS⁺) or serum-free (FCS⁻) medium for 24 hours, with or without OM-85. Selected wells then received Alternaria (50 μg/well). TEER was measured 3 and 6 hours later. Filled symbols and bars denote mean TEER ± SEM from 4 independent experiments. B: Mice received one i.n. Alternaria treatment (50 μg) concurrently with OM-85 (1 mg) (Fig. E14, top) or after four OM-85 pre-treatments (Fig. E14, bottom). BAL IL-33 was measured 1 hour after Alternaria exposure (n=7-10 mice/group from two independent experiments). C-D: Induction of lung ILC2s by Alternaria and OM-85. Mice were treated i.n. with Alternaria (50 μg) for three consecutive days. OM-85 (1 mg) was administered i.n. three times, concurrently (C; Fig. E15, top) or before Alternaria (D; Fig. E15, bottom). Lung ILC2s were harvested for flow cytometry. Data are from 6 (C) or 10 (D) mice/group pooled from two independent experiments. Filled symbols denote individual mice, bars denote mean ± SEM. ANOVA on ranked data with Tukey’s post-hoc analysis (A, B) or an unpaired, two-tailed t test (C) were used for statistical analysis.

To complement these findings, we then tested the impact of i.n. OM-85 administration on Alternaria-elicited release of epithelial IL-33 in BAL and IL-33-dependent induction and activation of ILC2s in the lung. We focused on these innate asthma-promoting responses^{41, 42} because in their early stages they occur without significant adaptive, Th2 cell-mediated positive feedback⁴³. As reported by others⁴⁴, IL-33 became readily detectable in BAL 1 hour after a single i.n. exposure to Alternaria (Fig. E14, top and Fig. 8B). A single, concurrent i.n. treatment with OM-85 (1 mg) attenuated but could not significantly suppress (P=0.08) this acute response. In contrast, acute IL-33 release was strongly and significantly (P=0.006) prevented in mice that had been pre-treated i.n. four times with OM-85 (1 mg) (Fig. E14, bottom, and Fig. 8B).

Finally, three i.n. pre-treatments with OM-85 could significantly suppress total numbers (P=0.012), proportions (P=0.0013), and surface ST2 expression (P<0.001) of Lin⁻ Thy1.2⁺IL-33Rα/ST2⁺KLRG⁺ lung ILC2 cells⁴⁵ in mice exposed i.n. to Alternaria for three consecutive days⁴⁶ (Fig. E15, bottom and Fig. 8D). However, no inhibition was detected when the same number of OM-85 treatments was administered concurrently with Alternaria (Fig. E15, top and Fig. 8C). Thus, at least in the lysate dose range used for this work, prophylactic but not concurrent i.n. administration of OM-85 effectively blocked Alternaria-driven innate activation in acute and short-term models.

DISCUSSION

Robust epidemiologic underpinnings and a potential to recalibrate disease-promoting immune responses are motivating intense interest in microbial interventions against asthma and allergies³. Different approaches and administration routes are under evaluation in both human and animal models¹³. Oral administration of the OM-85 bacterial lysate is widely used empirically in Europe and Asia to prevent recurrent upper respiratory infections¹⁸ and is currently tested for its ability to prevent asthma symptoms in young high-risk children. Mechanistic studies in allergen-sensitized rodents treated orally with OM-85 showed inhibition of airway inflammation accompanied by increased proportions of activated Treg cells and attenuated DC responses in the airways, especially the trachea, upon allergen challenge^{17, 47}. Adoptive transfer of OM-85-treated unfractionated tracheal CD4⁺ T cells protected allergen-sensitized animals from airway inflammation, and these effects were reduced upon depletion of T cells bearing gut-homing receptors⁴⁷. A trans-placental immunomodulation study later described dampened allergic airway inflammation in the offspring of mothers treated orally with OM-85 during pregnancy, which was paralleled by altered numbers and phenotypes of cDCs in lungs and ADLNs³⁴ and transcriptomic signatures of accelerated bone marrow myelopoiesis⁴⁸.

Our novel results significantly enrich this body of knowledge. After demonstrating that i.n. exposure to farm-derived microbial products virtually abrogates asthma in humans and mice^{7, 25, 49}, here we investigated whether and how direct administration of OM-85 to the airways protects from experimental allergic asthma. Although our combined hypothesis-generating and hypothesis-driven analyses relied on distinct allergens, different mouse strains, and a variety of models in which OM-85 was administered concurrently with, or before and during, allergen sensitization, our results consistently and conclusively showed that OM-85 administration to the airway compartment strongly protected against allergic asthma.

Multiple components of the innate and adaptive immune response were targeted by OM-85. Pre-eminent among them were the epithelium//IL-33/ILC2 axis that initiates type-2 inflammation in the airway mucosa, allergen-induced type-2 responses that promote asthma-related respiratory phenotypes, and Myd88/Trif-expressing DCs whose OM-85-induced reprogramming was sufficient to transfer protection from allergic asthma. It is noteworthy that i.n. OM-85 administration dampened allergen-induced type-2 airway responses rather than promote type-1 immune deviation, thus inducing immune profiles reminiscent of those we found in both asthma-protected farm children and asthma-protected mice exposed i.n. to farm dust^{7, 49}.

To our knowledge, this is the first demonstration that a defined cell type (OM-85-reprogrammed Myd88/Trif-expressing DCs with enhanced tolerogenic properties) is sufficient to transfer OM-85-dependent asthma protection to allergen-sensitized mice. The critical protective role played by these cells was highlighted not only by adoptive transfer experiments, but also by studies that linked OM-85-dependent phenotypic and functional changes in DC populations to enhanced induction of Treg cells and increased Treg/Th2 cell ratios in the lung.

On the other hand, our extensive phenotyping and transcriptomic analyses showed that the effects of OM-85 likely involved additional immune and structural lung cells and pathways. Especially informative was the capacity of airway OM-85 administration to fully suppress cardinal asthma phenotypes induced by Alternaria in a model that is driven exclusively by i.n. allergen exposure and thus carries high physiological relevance. Interestingly, OM-85 was also able to block early Alternaria-induced innate events (IL-33 secretion in BAL and ILC2 recruitment to the lung), and this ability appeared to reflect airway epithelial barrier stabilization⁵⁰ through profound, time-dependent effects on the epithelial transcriptome. In line with this notion, prophylactic but not concurrent i.n. treatment with OM-85 protected against innate responses acutely elicited by Alternaria-induced mucosal damage.

Notably, OM-85 administration through the nasal route achieved significant, often complete asthma suppression at a cumulative dose 27 to 46-fold lower than the one reportedly used through the oral route (8-14 vs. 375 mg/mouse, respectively)⁴⁷. We surmise that the high efficacy of airway-administered OM-85 may reflect both the diverse microbial stimulation this lysate provides and its direct access to the cellular and molecular networks that reside in the airway mucosa and regulate its response to allergens. Because previous evidence from oral treatment models could not identify cell-intrinsic effects sufficient for OM-85-dependent suppression of allergic asthma, at this time it is unclear whether oral and i.n. OM-85 administration trigger similar or distinct protective mechanisms. Regardless, the intense, multi-pronged effects of i.n. OM-85 on allergen-induced innate and adaptive murine airway responses warrant further studies to determine whether administration of this standardized, safe bacterial lysate to the airway compartment can protect humans from allergic asthma.

KEY MESSAGES.

• Direct administration of the OM-85 bacterial lysate to the airway compartment strongly suppresses experimental allergic asthma.

• OM-85 acts on multiple innate and adaptive immune targets: the airway epithelium/IL-33/ILC2 axis, lung allergen-induced type-2 responses, and dendritic cells whose Myd88/Trif-dependent tolerogenic reprogramming is sufficient to transfer OM-85-induced asthma protection.

• The high efficacy of intra-nasal OM-85 administration may reflect its direct access to the airway mucosal networks that regulate the initiation and development of allergic asthma.

ABBREVIATIONS

ADLN

airway draining lymph nodes

AHR

wairway hyperresponsiveness

BAL

broncho-alveolar lavage

BMDC

bone marrow-derived dendritic cells

cDC

conventional DCs

DC

dendritic cell

ELISA

enzyme-linked immunosorbent assay

FMO

fluorescence minus one

G

tissue damping

IFN

interferon

IL

interleukin

ILC2

innate type-2 lymphoid cells

i.n.

intra-nasal

MFI

mean fluorescence intensity

ORBEX

ORal Bacterial EXtract

OVA

ovalbumin

RT-qPCR

reverse transcription quantitative polymerase chain reaction

Rn

Newtonian resistance

Rrs

total airway resistance

TEER

trans-epithelial electrical resistance

Treg

T regulatory

WGCNA

weighted gene correlation network analysis

WT

wild-type