To study the impact of nafamostat in allergic airway inflammation, we used a protocol based on sensitization with house dust mite (HDM) extract. The used model was shown to replicate major features of the clinical asthma, including an increased cell infiltration to the airway lumen ( Figure 1A ) and raised airway hyperreactivity (AHR) in response to methacholine ( Figure 1B ). In contrast, no effects of HDM sensitization on dynamic compliance (Cdyn) was seen ( Supplementary Figure 1 ). The increase in inflammatory cell infiltration to the airway lumen was primarily due to a profound increase of eosinophils ( Figure 1C ) but increased infiltration was also seen for macrophages, lymphocytes and neutrophils ( Figures 1D–F ) as compared to control mice. Treatment of HDM-sensitized mice with nafamostat significantly reduced the airway inflammation and lung resistance (R_L) ( Figures 1A, B ), which was accompanied by a substantial reduction in the infiltration of eosinophils ( Figure 1C ) and lymphocytes ( Figure 1E ) to the airway lumen, whereas a tendency of reduced neutrophil influx was seen ( Figure 1F ).

To further evaluate the effect of nafamostat on lung inflammation, we examined effects on the peribronchial and perivascular infiltration of inflammatory cells by H&E staining, and infiltration of eosinophils into the lung tissue was assessed by chromotrope 2R staining. Consistent with the findings above, lung sections from HDM-sensitized mice showed a significant increase in peribronchial and perivascular inflammation ( Figure 2A ), which was mainly attributed to an excessive infiltration of eosinophils around the airways when compared to control mice ( Figure 2B ). Furthermore, treatment of the HDM-sensitized mice with nafamostat significantly attenuated the peribronchial and perivascular inflammation and caused a reduced infiltration of tissue eosinophils ( Figures 2A, B ). Together, these findings suggest that nafamostat could be beneficial to mitigate the hallmark features of allergic airway responses.

Next, we examined the efficacy of nafamostat as an anti-inflammatory agent in allergic airway responses by measuring effects on inflammatory mediators recovered from the BAL fluid, by adopting a cytokine array approach. This assessment revealed increased levels of various cytokines/chemokines in BAL fluid from HDM-sensitized vs. control mice, including IL12p40/p70, MIP-1γ/CCL9, TARC/CCL17, TIMP-1, VCAM-1/CD106 and KC/CXCL1 ( Figure 3 ). Notably, nafamostat treatment caused a profound reduction in the levels of each of these compounds in the BAL fluid, in most cases down to the baseline level ( Figure 3 ). Of note, the HDM sensitization did not result in increased levels of Th2 cytokines (IL-4, IL-5, IL-13) in the BAL fluid.

Airway remodelling is a prominent feature of asthma, as characterized by increased smooth muscle cell proliferation and thickening of the smooth muscle layer, contributing to the airway narrowing and increased AHR. Considering our observed beneficial effects of nafamostat on HDM-induced airway inflammation and AHR, we next investigated the effect of nafamostat treatment on airway remodelling as manifested by goblet cell hyperplasia and smooth muscle layer thickening. Histological analysis demonstrated a significant rise in goblet cell density (assessed by PAS staining; Figure 4A ) and an increased smooth muscle layer thickness ( Figure 4B ) in HDM-sensitized vs. control mice. Further, and consistent with its beneficial effect on other features of allergic airway inflammation, nafamostat treatment caused a modest, yet significant attenuation of the HDM-induced goblet cell hyperplasia and smooth muscle layer thickening ( Figures 4A, B ).

To provide a more extended insight into the mechanism by which nafamostat ameliorates allergic airway inflammation, we next performed a transcriptomic analysis of how nafamostat affects the gene expression profiles in the lung, by adopting the Ampliseq platform. Transcripts with less than 30 counts were considered to be poorly expressed and were disregarded from the analysis. Visualization using principal component analysis (PCA) of the total lung transcriptome revealed a well-defined separation between control- (PBS) vs. HDM-sensitized mice (PC1 93% of variance between the groups and PC2 2% of variance) ( Figure 5A ) and also between HDM-sensitized vs. HDM-sensitized/nafamostat-treated mice (PC1 66% of variance between the groups and PC2 14% of variance) ( Figure 5B ). These findings suggest that the HDM-sensitization causes major effects on the lung transcriptome, and also that nafamostat has the capacity to modulate the expression of genes that are induced by HDM sensitization.

Differential gene expression analysis revealed that a total of 1680 out of 14499 transcripts were differentially (p<0.05; log2FC ≥ 1) expressed in HDM-sensitized vs. control mice, out of which 1077 genes were upregulated and 603 genes were downregulated in response to HDM sensitization (see Supplementary Table 1 ). Further, a comparison between HDM-sensitized vs. HDM-sensitized/nafamostat-treated mice revealed that 139 genes were differentially expressed, out of which 19 genes were upregulated and 120 genes were downregulated in response to nafamostat treatment (see Supplementary Table 2 ). Notably, the nafamostat treatment caused a reversal of the expression, back to baseline levels, of several of the genes that were induced by the HDM sensitization. The latter supports the notion that nafamostat, at least partly, can block the effects of HDM sensitization on global gene expression patterns in the lung. Using the enhanced volcano plot the profound effects of HDM sensitization on the lung transcriptome ( Figure 5C ; see also Supplementary Table 1 ), and the marked effects of nafamostat on the lung transcriptome in HDM-sensitized mice ( Figure 5D ; see also Supplementary Table 2 ) were visualized. The gene expression patterns in the lungs of control- vs. HDM-sensitized mice and in HDM-sensitized vs. HDM-sensitized/nafamostat-treated mice were also hierarchically clustered and presented as a heat map ( Figure 6 ). Overall, these analyses reveal extensive effects of the HDM sensitization on the lung transcriptome, and also reveal profound effects of nafamostat treatment on the lung transcriptome after HDM sensitization.

To further address the effects of nafamostat on parameters of HDM-induced allergic airway inflammation, enriched GO and KEGG analysis was performed (the top 25 categories were visualized). The GO analysis revealed that the HDM sensitization, as expected, caused significant effects on the transcription of genes associated with immune responses, including the categories: leukocyte chemotaxis and migration, regulation of cytokine production and cytokine mediated signalling ( Figure 7A ). Further, the GO analysis showed that several of these pathways were suppressed by nafamostat ( Figure 7B ), providing further support for an anti-inflammatory impact of nafamostat on allergic airway responses. Similarly, enriched KEGG analysis provided further support for a strong association between the HDM sensitization and pathways related to inflammation and immunity, including categories such as: cytokine-cytokine receptor interaction, chemokine signalling, IL-17 signalling pathway, toll-like receptor signalling, asthma, c-type lectin receptor pathway and ECM-receptor interaction ( Figure 7C ). Moreover, the KEGG analysis supported an anti-inflammatory impact of nafamostat on such pathways, including categories such as: cytokine-cytokine receptor interaction, asthma, and IL-17 signalling pathway ( Figure 7D ). Altogether, these findings highlight the profound effects of HDM sensitization on the lung transcriptome, and also that nafamostat has the capacity to partly block the effects of HDM sensitization on pro-inflammatory gene expression in the lungs.

Further analysis by gene set enrichment analysis on the KEGG pathways revealed that the HDM sensitization induced a positive enrichment of 62 pathways and negative enrichment of 49 pathways. The top positively enriched pathways (activated) that were mapped in response to HDM were: asthma, IL-17 signalling, linolenic acid metabolism, cytokine-cytokine receptor interaction, toll-like receptor signalling, chemokine signalling and c-type lectin receptor signalling ( Figure 8A and Supplementary Table 3 ). The pathways that were negatively enriched (suppressed) in response to HDM sensitization included: Hippo signalling, Rap1 signalling, circadian entrainment, ECM-receptor interaction and Wnt signalling ( Figure 8A and Supplementary Table 3 ). Moreover, a total of 46 pathways were positively enriched in HDM-sensitized/nafamostat-treated vs. HDM-sensitized mice, including several categories related to cardiomyopathy, and also including several of those categories that were associated with the HDM sensitization: Hippo signalling pathway, circadian entrainment, ECM-receptor interaction, Rap1 signalling ( Figure 8B and Supplementary Table 4 ). Further, 8 pathways were shown to be suppressed by the nafamostat treatment. Importantly, “asthma” was one of the major pathways that was suppressed by the nafamostat treatment. Nafamostat also suppressed several other pathways, including the categories: linolenic acid metabolism, ribosome, oxidative phosphorylation, complement and coagulation cascades and cytokine-cytokine receptor interaction ( Figure 8B and Supplementary Table 4 ). Hence, these findings provide further support for an ameliorating effect of nafamostat on the pro-inflammatory gene expression patterns that are associated with allergic airway inflammation.

To further validate the effects of HDM sensitization and nafamostat on the gene expression patterns in the lung, we performed qPCR analysis. To this end, we focused on genes of particular relevance in asthmatic settings, including genes involved in eosinophil chemotaxis. These assessments thus included an analysis of the expression of Ear2, Ear6, Clec7a, Slc26a4, Fcer1g, CCL8, CCL9 and CCL11. In concordance with the transcriptome analysis, qPCR analysis supported that HDM-sensitized mice display a significant increase in the expression of all these genes in lung, when compared with control mice ( Figures 9 ). Further, it was demonstrated that treatment of HDM-sensitized mice with nafamostat caused a significant attenuation of the expression of clec7a (Dectin-1), CCL8, CCL9 and Ear6 ( Figure 9 ), and a trend of reduced Slc26a4 and CCL11 expression ( Figure 9 ), whereas no effects of nafamostat on Fcerg1 or Ear2 expression were seen.

Female BALB/c mice (8 – 9 weeks of age) were purchased from Taconic Biosciences (Lille Skensved, Denmark). All procedures were performed at the Swedish University of Agricultural Sciences animal facility under protocols compliant with the EU Directive 2010/63/EU for animal experiments and approved by the local ethical committee (Uppsala djurförsöksetiska nämnd; Dnr 5.8.18-12873/2019). Mice were acclimatized for one week in the experimental room prior to the start of the experiment. Mice were lightly anaesthetized with isoflurane using a portable isoflurane vaporizer. Anaesthetized mice were instilled with 10 μg of house dust mite (HDM) extract reconstituted in 30 μl PBS (Dermatophagoides pteronyssinus, CiteQ BV, Groningen, Netherlands) intranasally twice a week for three weeks. Control mice received 30 μl PBS via the intranasal route. To investigate the efficacy of nafamostat on asthmatic features, nafamostat (20 mg/kg) (Sigma-Aldrich, St Louis, MO) or vehicle (3% acetic acid in PBS) were administered thirty minutes before each PBS and/or HDM instillation via the intraperitoneal route. A total of 48 mice were used in two individual experiments and the mice were allocated to the following experimental groups: PBS (n = 8), HDM (n = 9), HDM + vehicle (n = 9), nafamostat (20 mg/kg) (n = 8) and HDM + nafamostat (20 mg/kg) (n= 9).

Twenty-four hours after the last HDM instillation, airway hyperreactivity (AHR) was measured using a Buxco small ventilator (Buxco^® FinePointe Resistance and Compliance, Winchester, UK). Mice were anaesthetized with pentobarbital sodium (50 mg/kg; Sigma-Aldrich), and a tracheostomy was performed to insert a 20-gauge needle. Mice were kept under ventilation at 160 breaths/minute with a tidal volume of 0.25 mL. Mice were acclimatized to the ventilator before measuring the baseline. Lung resistance (R_L) and dynamic compliance (C_dyn) were measured using a dose-response curve for methacholine by administering increasing doses of nebulized methacholine (0 – 50 mg/mL; Sigma-Aldrich).

Immediately after measuring lung function parameters, blood was collected and centrifuged at 2300 x g for 15 minutes at 4°C and serum was separated to measure HDM-specific IgE using ELISA (Chondrex, Woodinville, WA) following the manufacturer’s instructions. For bronchoalveolar lavage (BAL), the cannulated lungs were lavaged twice with 0.5 mL of sterile Hanks Balanced Salt Solution (HBSS). The collected BAL fluid was centrifuged at 600 x g for 10 min at 4°C. Supernatants were collected for analysis of cytokines and chemokines, while the cell pellet was resuspended in 1 mL sterile HBSS for enumeration of total and differential cell counts. For total cell counts, a hemocytometer was used. For differential leukocyte counts, cytospins were performed by spinning 100 µl cell suspensions onto glass slides at 26 x g for 5 min using a cytospin centrifuge. Cells were stained with May Grünwald/Giemsa, and a minimum of 200 cells per slide cells were counted. Cytokines levels in the BAL fluid were measured using the C3 Mouse Cytokine Array (RayBiotech, Norcross, GO) according to the manufacturer’s instructions by pooling the individual samples within the group. Signal pixel density assessment was conducted by using the ImageJ software (https://imagej.nih.gov/ij/).

After BAL fluid collection, the right lung lobes were ligated, dissected and snap-frozen in liquid nitrogen and stored at -80°C for gene and protein analysis. The left lung lobes were inflated with 0.5 mL buffered formalin via the trachea cannula and trachea were sealed and fixed in buffered formalin. Formalin-fixed lobes were embedded in paraffin and 6 μm sections were prepared using a microtome. Sections were dewaxed and dehydrated by treatment with xylene, followed by a series of graded alcohol. Dewaxed lung sections were stained with haematoxylin and eosin (H&E) (30). Airway tissue inflammation was quantified by scoring the inflammatory cell infiltrate surrounding the airway using a semi-quantified score method ranging from 0-4 (0: no inflammatory cell infiltrates around the airway, 1: low-level cell infiltrates around part of the airway, 2: moderate cell infiltrates around part of or entire airway, 3: significant inflammatory cell infiltrate around part of or entire airway, 4: airway surrounded by inflammatory cell infiltrates). To study eosinophil infiltration, lung sections were stained with the chromotrope 2R (47). Bright red nuclei-stained cells around the airways were counted with a Nikon Microphot-FXA microscope using a 40 x objective lens and the Eclipse Net software.

To determine goblet cell hyperplasia and mucus production, airway sections were stained with periodic acid-Schiff’s (PAS). The number of goblet cells per μm of the airway epithelial layer was counted using a Nikon Microphot-FXA microscope with a 40 x objective lens and the Eclipse Net software (version 1.20, Developed by Laboratory Imaging, Prague, Czech Republic). The smooth muscle layer thickness around the airways was measured on the PAS-stained sections by measuring the thickness of every 30 μm basement membrane around the airways with a Nikon Microphot-FXA microscope and the Eclipse Net software; the average thickness was calculated. All quantitative histological assessments were performed and quantified by trained investigators in a blinded manner.

Snap-frozen upper portions of the left lung lobes from separate experiments were grounded in liquid nitrogen, and RNA was extracted using the Qiagen RNeasy^® plus mini kit (Hilden, Germany) following the manufacturer’s instructions. cDNA libraries were created and amplified using the Ion AmpliSeq™ Transcriptome Mouse Gene Expression Kit (Life Technologies, Carlsbad, CA) following the protocol of the manufacturer; the sequencing was performed on an Ion S5™ XL Sequencer (Thermo Fisher Scientific, Waltham, MA). Generated data had read lengths with an average of 98-110 bp and high mapping (95% of aligned bases). Normalized expression values generated by the Torrent Suite™ Software (version 5.10.1) were used for downstream differential gene expression analyses. The differential gene expression analysis was performed using the R package DESeq2 and genes with a log2FC ≥ 1 and an FDR-adjusted p-value <0.05 were considered as differentially expressed genes. To perform the gene enrichment analysis, we used the clusterProfiler package on the R program; the gene ontology with the biological process (BP) and KEGG pathway analyses were performed using the enrichGO and enrichKEGG functions in the clusterProfiler package (48). Functional enrichment analysis was performed using the Gene set enrichment analysis (GSEA) and the enrichment terms were considered significant with adjusted p < 0.05.

Quantitative RT-PCR (qPCR) analysis was used for further validation of the expression of selected genes. RNA concentration and purity was measured using a NanoDrop device (ND-1000; Nano Drop Technologies, Wilmington, Delaware). cDNA was prepared using 1 µg of total RNA and the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA) following the manufacturer’s instructions. Validated primers for selected genes were purchased from BioRad (Hercules, CA) and qPCR analysis was performed on a C1000 Touch Thermal Cycler instrument (Bio-Rad, Hercules, CA) using the iTaq Universal SYBR Green Supermix (Bio-Rad). Each sample was run in duplicates and qPCR data analysis was performed using the Bio-Rad CFX Maestro program. Gene expression levels were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and the levels of gene expression were quantified using the 2^−ΔCT method

Data are presented as mean values ± SEM, as analysed using GraphPad Prism 8.0. Two-way ANOVA statistical analysis was conducted to compare the in vivo lung function measurements between the experimental groups in response to the methacholine dose-response curve using the Bonferroni post-hoc analysis. For the rest of the experimental endpoints, one-way ANOVA was performed for statistical comparison between the treatment groups using Tukey post-hoc multi-comparison analysis. Differences between the experimental groups were regarded statistically significant when the p-value reached <0.05.