A TRPA1-mediated substance P release assay as an in vitro screening tool for airway irritant potential

Environmental exposure to airway irritants can elicit sensory irritation and contribute to various disease outcomes. Although in vitro assays have been developed to measure skin and eye irritation, evaluation of respiratory sensory irritation relies predominately on in vivo testing using the RD 50 assay, or Alarie test. Although the RD 50 assay aligns well with human responses, various disadvantages, such as low throughput and interspecies differences in receptors responsible for detecting noxious stimuli, highlights a few of the many requirements for alternative, or complementary in vitro approaches. Recently, increased emphasis has been placed on methods that predict human responses at the mechanistic level. In this study, we have developed a human relevant in vitro model for airway irritant detection based around molecular events of airway irritation, including transient receptor potential channel activation and substance P release. Using alveolar- and bronchial-derived cell lines, we demonstrate endogenous TRPA1-driven Substance P release following exposure to prototypical irritants. Calcium response was confirmed by flow cytometry following exposure to irritants, with substance P release demonstrated at both the gene and protein levels by qPCR, immunocytochemistry and ELISA. Furthermore, our model demonstrates some concordance with in vivo RD 50 data.

Environmental exposure to airway irritants can trigger a range of acute respiratory symptoms, including pain, cough, dyspnea, and bronchospasm. Chronic or repeat exposure can lead to longer term conditions such as airway hyper-responsiveness, asthma, and chronic obstructive pulmonary disorder (COPD). These disorders contribute to reduced economic productivity and increased public health costs. The use of volatile organic chemicals (VOCs) in consumer products including those found in solvents, vapours and gases continues to increase. Furthermore, many of these VOC’s are released into the environment and remain poorly characterised with respect to their potential to cause airway irritation (Wallbanks et al., 2024; Jung et al., 2025; Dalton, 2003).

The human respiratory tract is composed of various cell types, such as epithelial cells, that express transient receptor potential (TRP) ion channels. Activation of these TRP ion channels has been described as a front-line mechanism for detecting environmental irritants (Gerhold and Bautista, 2008). TRPA1 and TRPV1 are two key ion channel receptors present in the airways that mediate irritation as part of a physiological response to make the subject aware of the presence of chemicals and initiate several defensive biological responses. The activation of TRPA1 and TRPV1 has been well characterised as a gatekeeper of inflammation and is an important process in regulating the airway response following exposure to VOCs (Bautista et al., 2013; Müller et al., 2022; Bessac and Jordt, 2008).

Chemical irritants can be identified based on physical and chemical properties. Traditional assays used to evaluate irritancy include the in vitro skin and eye corrosion/irritation assays (Lee et al., 2017; OECD, 2015; OECD, 2016). However, respiratory irritation is more complex to measure in vitro and currently used assays involve animal models. In these models, corrosive effects are identified through histopathological endpoints, such as cell necrosis and inflammatory cell infiltration. Whilst these observations are predictive for corrosive effects, in general, they are not predictive of sensory irritation (Bos et al., 2002). Determination of sensory irritation is limited to the in vivo sensory irritation test that measures the 50% reduction in respiration rate or RD 50, also known as the Alarie test (Alarie, 1981). Although this assay demonstrates correlation with irritancy in humans for certain compounds, limitations include inability to predict human mechanistic responses and lack of cost-effectiveness for high-throughput screening of potential irritants. Furthermore, limitations with currently used risk assessment strategies appear to be reflected in epidemiological data with healthcare and cleaning occupations demonstrating increased occurrences of irritant-induced asthma associated with irritant exposure levels below established occupation exposure limits (OELs) (Baur et al., 2012; Bos et al., 2002). The use of mode-of-action testing has been suggested for setting air quality guidelines. Furthermore, the ability to develop systematic approaches using in vitro techniques aligns with the changing landscape of risk assessment testing, with mechanistic testing, such as the use of adverse outcome pathways (AOP), becoming more important (Wallbanks et al., 2024; Martinez and Eling, 2019; Kim and Choi, 2026).

In this study, we developed an in vitro airway irritation model based on TRPA1 and TRPV1 activation, quantified using flow cytometry. We demonstrate the expression of TRPA1 and TRPV1 in the human alveolar cell line, A549, following which activation of the channels is confirmed using a combination of channel-specific agonists and inhibitors. Subsequently, we confirmed the effects of several known airway irritants on TRPA1 and TRPV1 activation using the developed model. Using previously described molecular pathways, we investigated the link between TRP channel activation and substance P expression–a downstream signalling mechanism known to drive the airway irritation response in humans (Martinez and Eling, 2019). The presence of TRPA1-gated calcium responses and substance P release following exposure to a panel of known airway irritants highlighted the potential use of our model as a novel in vitro tool for screening compounds for their irritant potential in the respiratory tract. Finally, we confirmed the presence of similar biomarkers in the human upper airway model, EpiAirway™.

In conclusion, we describe the development of promising human alveolar- and bronchial-derived in vitro models that could be used alongside current approaches, such as the Alarie test, during the assessment of compounds in their ability to induce respiratory irritation. Such an approach could help to improve cost-efficiency and reduce animal use during compound risk assessment.

Chemicals tested in the study

All chemicals tested in the study were purchased from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany). The chemicals tested include; Allyl Isothiocyanate (AITC) (CAS No. 57–06–7), Capsaicin (CAS No. 404–86–4), acetone (CAS. No. 67–64–1), Nicotine (CAS. No. 54–11–5), p-Xylene (CAS. No. 106–42–3), Methanol (CAS. No. 67–56–1), Toluene (CAS. No. 108–88–3), Heptanal (CAS. No. 111–71–7), 2-Ethoxy Ethyl Acetate (CAS. No. 111–15–9), formaldehyde (CAS. No. 50–00–0), Isopropanol (CAS. No. 67–63–0), acetic acid (CAS. No. 64–19–7), HC-030031 (CAS. No. 349085–38–7) and Capsazepine (CPZ) (CAS. No. 138977–28–3). These chemicals were tested in this study to provide a range of responses based on responses observed in animal models (Schaper, 1993).

Cell culture

A549 cells were purchased from Culture Collections (Porton Down, Wiltshire) and cultured in Roswell Park Memorial Institute (RPMI) 1,640 medium containing 2 mM glutamine (Cat. No. 11875093, Thermo Fisher Scientific, Waltham, MA, United States) supplemented with 10% fetal bovine serum (FBS) (Cat. No. F9665, Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) and 1% penicillin and streptomycin (Cat. No. P4333, Sigma-Aldrich, Merck KGaA, Darmstadt, Germany). Cells were provided by the vendor at passage 91 and used in experiments at passage 93 with routine mycoplasma testing performed by Mycoplasma Experience (Mycoplasma Experience Limited, Surrey, United Kingdom). One day before treatment, A549 cells were seeded at a cell density of 6 × 10 4 cells per well in 24-well plates. A549 cells were grown in 1.0 mL medium until treatment commenced. For viability measurement using WST-8, cells were treated with 1.0 mL of the appropriate compound at specified concentrations for a 24 h duration. Compound treatments for the calcium response assay are described in subsequent sections.

EpiAirway™ tissues were purchased in 12-well Transwell inserts from MatTek (Ashland, MA, United States) and cultured in proprietary EpiAirway™ Maintenance Medium (AIR-100-MM; MatTek) according to the manufacturer’s protocol. EpiAirway™ tissues were maintained at the air liquid interface (ALI) with 1 mL medium in the basal compartment with medium replaced every 24 h and experiments performed 7 days following tissue culture set-up. EpiAirway™ tissues were of bronchial origin from a healthy donor.

Viability measurements by WST-8 assay

Following the 24 h treatment duration, treatment medium was removed from cells and the contacted surface rinsed with calcium and magnesium-free PBS (CMF-PBS) (Cat. No. 10010023, Thermo Fisher Scientific, Waltham, MA, United States). Viability was measured using the WST-8 assay. Briefly, 0.4 mL of WST-8 solution (Cat. No. 96992, Merck KGaA, Darmstadt, Germany), diluted 1 in 10 using appropriate medium, was applied to A549 cells, or basolateral surface of EpiAirway™ tissues, for 1 h. Subsequently, 0.1 mL aliquots were plated in triplicate and absorbance measured at 450 nm. To determine IC 20 and IC 50 values, viability was expressed as relative viability in comparison to the vehicle control and cell viability was interpolated to assign the concentrations required to reach 80% and 50% viability respectively. Cells were rinsed with CMF-PBS and stored at < -50 °C for future qPCR analysis.

RNA analysis

Total RNA was isolated from A549 cells and EpiAirway™ tissues following treatment using the RNAqueous™ Total RNA Isolation Kit (Cat. No. AM 1912, Thermo Fisher Scientific, Waltham, MA, United SA) according to the manufacturer’s instructions. A single EpiAirway™ tissue was used for each treatment condition per experiment, in three independent experiments, resulting in three tissues per treatment condition in total (n = 3). A549 measurements were performed in duplicate, per experiment, resulting in six individual replicates analysed in total. RNA concentration and purity was measured at 260 nm/280 nm by spectrophotometry. RNA concentrations were normalized to 100 ng of RNA per reaction and converted to complementary DNA (cDNA) using QuantiTect® Reverse Transcription Kit (Cat. No. 205311, Qiagen, Hilden, Germany). TaqMan® Gene Expression primer/probes (Thermo Fisher Scientific, Waltham, MA, United States) were purchased for the following genes: GAPDH, TRPA1, TRPV1, TAC1. Quantitative PCR (qPCR) was performed using 1 ng cDNA per reaction with each primer/probe added to separate wells with the PCR performed as single plex reactions. Subsequently, qPCR was performed using a QuantStudio™ 5 Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, United States) and data was analysed by relative quantification using the ΔΔCt method and quantification normalized to the endogenous control, GAPDH. Throughout the study, GAPDH Ct values measured 17.898 ± 0.561 SD. Normalised values were subsequently compared to the normalized vehicle control values to calculate the RQ, or fold-change, values. All data calculations were performed using QuantStudio™ 5 Design and Analysis Software version 1.5.2.

Calcium influx using flow cytometry

Intracellular calcium influx was assessed in A549 and EpiAirway™ cells using the calcium-sensitive fluorescent dye Fluo-4, AM. A549 or EpiAirway™ cells were incubated with 1 μM Fluo-4, AM (Cat. No. F14201, Invitrogen, Thermo Fisher Scientific, Waltham, MA, United States) for 1 hour at 37 °C ± 1 °C in an atmosphere of 5% ± 1% (v/v) CO 2 in air. Fluo-4, AM was reconstituted in DMSO to a stock concentration of 1 mM and diluted in appropriate medium to achieve the final working concentration (1 μM).

Cells were washed three times with Hank’s Buffered Saline Solution (HBSS) containing Ca 2+ and Mg 2+ and incubated in HBSS containing Ca 2+ and Mg 2+ for 30 min at room temperature protected from light to allow complete de-esterification of intracellular AM esters. Cells were trypsinised using 0.25% Trypsin-EDTA until detached and cells were then centrifuged and resuspended in HBSS (containing Ca 2+ /Mg 2+). HC-030031 or capsazepine (CPZ) were added to cells for TRPA1 or TRPV1 inhibition, respectively, for 30 min on ice. In addition, cells were supplemented with 1 μg/mL propidium iodide as a cytotoxicity marker. Cells remained on ice until agonist addition and measurement of intracellular calcium movement by flow cytometry.

Flow cytometry measurements were performed using a BD FACSCanto™ II (3 lasers, 4-2–2 configuration, BD Biosciences, Franklin Lakes, NJ, United States). During measurement of calcium response, 0.1 mL of sample supplemented with 0.1 mL HBSS (containing Ca 2+ /Mg 2+) (baseline state) was recorded for 3,000 events prior to immediate measurement of an additional 3,000 events following addition of agonist compound (activation state). Activation by agonist was measured immediately on addition until 3,000 events were acquired, with the mean FITC measurement used to assess level of calcium influx, and binding to Fluo-4, AM. The mean FITC measurement from the activated state was subtracted from the mean FITC measurement from the baseline state. Normalisation to the maximum response induced by 1 μg/mL ionomycin and the vehicle control only response was subsequently performed to calculate a percentage maximum calcium response. In addition, to control for potential fluorescence drift during the assay, for example, from measurement of the first concentration of compound to measurement of the final concentration of compound, baseline fluorescence measurements were taken before each concentration of compound.

Detection of hTRPA1 and hTRPV1 using flow cytometry

A549 cells were cultured as previously described following which cells were trypsinised, pelleted and fixed using 3% paraformaldehyde for 1 h at room temperature. Cell pellet was washed and cells permeabilised for 1 h using 1% (v/v) Triton X-100 in PBS. Fixed and permeabilised cells were incubated with anti-TRPA1 (Cat. No. PA1-46159, Thermo Fisher Scientific, Waltham, MA, United States) and anti-TRPV1 (Cat. No. PA1-748, Thermo Fisher Scientific, Waltham, MA, United States) specific antibodies overnight at 4 °C. Cells were washed and incubated overnight using Alexa Fluor™ 488 conjugated anti-Rabbit secondary antibodies (Cat. No. A-11008, Thermo Fisher Scientific, Waltham, MA, United States). Detection of bound fluorophore-tagged secondary antibodies was performed using a BD FACSCanto™ II (3 lasers, 4-2–2 configuration, BD Biosciences, Franklin Lakes, NJ, United States). In addition, cells incubated with buffer only and secondary antibody only were analysed for to confirm background fluorescence levels.

Detection of substance P in A549 cells using immunocytochemistry

A549 cells that had been treated for a 24-h duration were washed and subsequently fixed using 3% paraformaldehyde for 1 h at room temperature and permeabilised for a further 1 h using 1% (v/v) Triton X-100 in PBS. Fixed and permeabilised cells were incubated with anti-substance P antibody (Cat. No. Ab14184, Abcam, Cambridge, United Kingdom) overnight at 4 °C. Cells were washed and incubated with Alexa Fluor™ 647-conjugated anti-mouse antibody (Cat. No. Ab150115, Abcam, Cambridge, United Kingdom) and Hoechst 33,258 solution (Cat. No. 94403, Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) overnight at 4 °C. Detection of fluorophore-tagged secondary antibodies was performed using an EVOS™ M3000 Imaging System (Cat. No. AMF3000, Thermo Fisher Scientific, Waltham, MA, United States). ImageJ v1.47 was subsequently used to calculate ratios between the blue (nuclei-staining) and red (substance P-staining) channels to determine overall substance P levels.

Detection of secreted substance P in EpiAirway™ by ELISA

Following treatment with compound for a 24 h duration, medium was recovered from EpiAirway™ tissues and used to measure secreted substance P by ELISA (Cat. No. EEL013, Thermo Fisher Scientific, Waltham, MA, United States) according to manufacturer’s instructions. Briefly, 50 μL of sample was added to the ELISA plate for 45 min at 37 °C and then washed three times. 100 μL HRP-Conjugate was added to each well for 30 min at 37 °C and washed five times. Subsequently, 90 μL Substrate Reagent was added to each well for 15 min at 37 °C followed by addition of 50 μL Stop Solution. The plate was measured at 450 nm using a spectrophotometer and data analysed by subtracting the medium only values from each sample value. Fold-change compared to vehicle control was calculated. In addition, a standard curve was used to interpolate substance P concentration in each sample (pg/mL) using a four-parameter logistic curve.

Statistical analysis

Data are presented as the mean ± SEM (unless otherwise stated) of three independent experiments. All statistical analyses were performed using GraphPad Prism ®Software 10.4.2 (GraphPad Software Inc., La Jolla, CA, United States). All data was tested for normality using the Shapiro-Wilk normality test. Subsequently, comparisons between groups were performed by one-way analysis of variance (ANOVA). Values of p < 0.05 were considered to be statistically significant. Dose-response data were fitted using a four-parameter logistic (4 PL) model to accommodate flexible estimation of minimum and maximum responses and provide an appropriate sigmoidal fit.

Assessment of TRPA1 and TRPV1 expression in A549 cells

To confirm that A549 cells were an appropriate model for testing compound risk for airway sensory irritation induction, the presence of TRPA1 and TRPV1 ion channels was initially confirmed by qPCR (Figure 1A). Although expression was relatively low in comparison to the endogenous gene control, GAPDH, Ct values of 30.872 and 32.047 for TRPV1 and TRPA1, respectively, confirmed their presence. Despite this, gene expression does not necessarily confirm functional protein presence.

To confirm the presence of TRPA1 and TRPV1 protein in A549 cells, cells were labelled with anti-TRPA1 and anti-TRPV1 antibodies respectively, with detection performed by flow cytometry (Figure 1B). Increased fluorescence intensity of anti-TRPA1 and anti-TRPV1-labelled cells compared to cells only demonstrated increased expression of both channels. Furthermore, secondary-antibody only labelled cells demonstrated lower fluorescence intensity than the primary antibody-labelled cells, demonstrating that the fluorescence intensity was not a result of the secondary antibody binding to non-specific targets in A549 cells. It should be noted however, that due to the primary antibody epitopes being present on the C-terminus of the TRPA1 and TRPV1 channels, the cells had to be permeabilised to enable antibody access to their targets. Consequently, the increased signal does not necessarily confirm channel presence at the cell membrane, or channel functionality. Furthermore, despite the higher fluorescent signals observed following use of TRPA1 and TRPV1-specific antibodies, the high concentration of Triton X-100 (1%) used to permeabilise the fixed cells may have resulted in potential extraction of the ion channels from the membrane, further supporting the possibility that antibodies were binding to intracellular targets.

To confirm function of the expressed ion channels, treatment with TRPA1 and TRPV1-specific agonists and antagonists was performed and calcium mobilisation confirmed using the calcium probe, Fluo-4, AM, with measurement performed by flow cytometry using previously described methods, with minor adjustments (Bajnok et al., 2023). During calcium measurement, the calcium ionophore, ionomycin, was used to determine a maximum calcium response (100%) and the vehicle control, DMSO, used to determine the minimum calcium response (0%). The TRPA1 channel agonist, allyl isothiocyanate (AITC) was added to cells at a final concentration of 250 μM with a mean calcium response measured as 67.72% (±14.66%) activation (compared to the 100% ionomycin response). In contrast, treatment with AITC after a 30-min treatment with the TRPA1 antagonist, HC-030031, at a concentration of 100 μM resulted in a mean calcium response of 30.10% (±8.92%), with responses in the absence and presence of antagonist demonstrating a significant difference (p < 0.05) (Figure 1C). To confirm TRPV1 channel function, the TRPV1 channel agonist, capsaicin, was added to cells at a final concentration of 250 μM with a mean calcium response measured as 48.47% (±13.71%) activation. In contrast, treatment with capsaicin after a 30-min treatment with the TRPV1 antagonist, capsazepine (CPZ), at a concentration of 1 μM, resulted in a mean calcium response of 6.49% (±22.90%), with responses in the absence and presence of antagonist demonstrating a significant difference (p < 0.01) (Figure 1D). To confirm that the reduced responses observed in the presence of antagonist were not driven by toxicity, A549 viability was measured following treatment with a range of concentrations of HC-030031 (Figure 1E) and CPZ (Figure 1F), with no, or limited, loss in viability observed at the selected concentrations of HC-030031 (100 μM) or CPZ (1 μM), suggesting that loss of channel activation was a direct result of inhibition, rather than an artefact of toxicity. Viability following treatment at the selected concentrations of HC-030031 (98.1% v vehicle control) and CPZ (93.5% v vehicle control) was also confirmed during the calcium response assay by measuring propidium iodide uptake (Figures 1E,F), again, with minimal decrease in viability measured.

TRP channel activation assessed by a panel of known airway irritant compounds

Following successful use of the flow cytometry-based TRP channel activation assay using established agonists and antagonists, we next assessed its suitability for screening airway irritant potential by testing a panel of compounds with known in vivo irritancy profiles. A range of calcium responses were observed across the panel (Figure 2A) with statistical significance outlined (Table 1). For example, nicotine induced a marked increase in intracellular calcium that was unaffected by TRPA1 or TRPV1 inhibition, an expected response considering the role of nicotinic acetylcholine receptors (nAChRs) in nicotine binding and the high expression levels observed in A549 cells (Sun et al., 2017; Mucchietto et al., 2018). Interestingly, p -xylene and toluene have previously demonstrated TRPA1 and TRPV1mediated responses in mouse models (Sándor et al., 2009; Norões et al., 2019). Although a significant difference was not observed when comparing non-inhibited and TRPA1/TRPV1-inhibited responses in our model, non-inhibited responses did demonstrate a slightly greater response compared to TRPA1 and TRPV1 inhibited responses. Overall, non-inhibited percentage maximum calcium response to p -Xylene treatment was 43% (p < 0.01 v vehicle control) while responses following TRPA1 inhibition and TRPV1 inhibition were 27% and 29%, respectively, neither of which were significantly different compared to vehicle control response. Following toluene treatment, non-inhibited percentage maximum calcium response was 40% (p < 0.0001) compared to TRPA1 (p < 0.001) and TRPV1 (p < 0.01) inhibited responses of 32% and 27%, respectively. In contrast, 2-ethoxyethyl acetate and isopropanol did not induce statistically significant calcium responses. It should be noted however, that formaldehyde responses were inconsistent. Despite this, in comparison with previous studies, a minor response was observed (p < 0.05) that was not observed following TRPA1-and TRPV1 inhibition (Macpherson et al., 2007; Song et al., 2017). Notably, acetic acid produced a pronounced, concentration-dependent decrease in calcium signal, highlighting the challenges of measuring calcium dynamics for certain compounds. For example, even low concentrations of acetic acid may result in disrupted membrane integrity and toxicity, even during short-term application, thereby promoting leakage of calcium sensitive dyes from cells. Indeed, propidium iodide data suggests greater toxicity induction following acetic acid application, in comparison to other compounds. Alternatively, the reduced pH following acetic acid addition may result in reduced fluorescence intensity of calcium probes such as Fluo-4, AM via a quenching mechanism (Maruta and Yamashita, 2020; Baker et al., 1994; Rohrbach et al., 2005).

To further evaluate assay performance, we monitored baseline fluorescence stability and cell viability throughout the experimental window (Figure 2B). As the assay relies on repeated baseline measurements between compound additions, minimal variability in these readings, as well as sustained cell health, is essential. Substantial declines in baseline fluorescence, for instance, could indicate dye leakage and compromise interpretation of compound-induced responses. Overall, baseline fluorescence remained stable across experiments for most compounds, although measurements during toluene testing were more variable. Although there is the possibility that this is a result of the high volatility of toluene, or autofluorescence of residual toluene remaining in the flow cytometer during baseline measurements, this is unlikely as stable baseline fluorescence was observed for the similar aromatic hydrocarbon compound, p -xylene. Nevertheless, further investigations would be required to further understand the fluorescent drift observed during toluene testing. Cell viability remained high throughout for all compounds tested, with toxicity consistently below 10%, as measured by propidium iodide uptake, supporting the robustness of the assay conditions.

Activation of TRPA1 in A549 cells promotes substance P signalling

Paracrine communication between neuronal and epithelial cells has been identified as a key driver of airway sensory irritation and subsequent inflammatory responses. A key peptide in this signaling cascade is substance P, expressed from the preprotachykinin-1 gene, Tac1. Although previous studies have demonstrated oxidative stress and inflammatory induction following substance P expression in neuronal cells, or in A549 cells following treatment with exogenous substance P, we sought to determine the ability of A549 cells to express endogenous substance P following activation of TRPA1 and TRPV1 (Xu et al., 2018; Springer et al., 2005; Nguyen et al., 2023).

Initially, A549 cells were treated with AITC at a range of concentrations for a 24 h duration and viability measured, with IC 20 and IC 50 values of 0.6 μM and 2.8 μM determined, respectively. Subsequently, using the flow cytometric-based calcium mobilization assay described in Figures 1, 2, the ability of AITC to evoke a concentration-related response was investigated in the absence (EC 50 22 μM) and presence (EC 50 385 μM) of 100 μM HC-030031, with a significant difference observed. Furthermore, a significant increase in endogenously expressed Tac1 was confirmed by qPCR following treatment with 2.8 μM AITC for 24 h (p < 0.01), an increase that was significantly diminished in the presence of 1 μM HC-030031 (p < 0.05) (Figure 3A). It should be noted that although only a small increase in Tac1 expression (not statistically significant) was observed following AITC treatment at 0.6 μM, this may have been a result of variability in viability, with viability being as high as 90% during the Tac1 expression experiments presented in Figure 3.

Following treatment with capsaicin, although decreased cell viability and increased calcium mobilization was observed, increased Tac1 expression was not observed at the capsaicin concentrations tested (Figure 3B). However, it should be noted that higher concentrations of capsaicin may have induced higher levels of Tac1 expression. Alternatively, the lack of Tac1 expression following capsaicin-induced TRPV1 activation, may be explained by expression of the ion channel receptor on an intracellular organelle, rather than at the plasma membrane. Both possibilities should be further explored in future studies. Treatment with acetone, a compound that has demonstrated low respiratory irritation when tested using in vivo models, demonstrated negligible toxicity, calcium response or Tac1 expression (Figure 3C). To confirm that increased Tac1 expression following treatment with 2.8 μM AITC translated to enhanced substance P protein expression, immunocytochemistry was performed and substance P-specific staining intensity normalized to nuclei intensity (Figure 3D). Results demonstrated a significant increase in substance P staining following treatment with AITC compared to vehicle treatment (p < 0.01). A statistically significant increase was not observed following treatment with capsaicin (500 μM) or acetone (3,522 μM). The endogenous gene control, GAPDH, response remained consistent during compound treatments, with an average Ct value of 17.571 ± 0.284 SD observed.

Viability, calcium responses and Tac1 expression following treatment with known airway irritants correlate with in vivo responses

To further evaluate the ability of our in vitro model to identify airway irritants, we compared 24 h viability outcomes (Figure 4A) with calcium responses measured at a 1% (v/v) concentration to enable comparison between compounds when tested at the same concentrations (Figure 4B). Overall, viability and calcium signaling showed strong concordance. For example, nicotine, p -xylene, and heptanal each produced low IC 20 values and elicited significant calcium responses under non-inhibited (TRPA1/TRPV1-active) conditions. AITC demonstrated a similar alignment between viability and calcium response (Figure 3A). In contrast, acetone (Figure 3C) and isopropanol (Figure 4A) - both associated with minimal airway irritation in vivo - showed low toxicity (IC 20 values of 2,269 and 1771 mM, respectively) and did not induce calcium responses (Figure 4B).

Despite these general trends, certain compounds produced divergent outcomes. Methanol induced a calcium response without causing toxicity, whereas formaldehyde caused substantial toxicity but triggered a variable calcium response. These discrepancies underline the importance of integrating data from both assays when evaluating potential irritant risk.

To further leverage mechanistic pathways associated with airway irritation, Tac1 expression was quantified following treatment with the full panel of test compounds (Figure 4C). In general, Tac1 induction was greater for compounds that produced significant calcium responses. As previously demonstrated, the endogenous control gene, GAPDH, response remained consistent during compound treatments, with an average Ct value of 17.953 ± 0.578 SD observed.

Results from the in vitro viability, calcium response, and Tac1 assays were then compared with historical in vivo irritancy data (Table 2). Acetone and isopropanol showed strong concordance with their high in vivo RD 50 values (23,480 and 5,000 ppm, respectively), exhibiting low toxicity, negligible calcium responses, and minimal Tac1 expression using our in vitro model. Several compounds known to cause pronounced irritation in vivo - including AITC, nicotine, p -xylene, and toluene - also produced high toxicity, robust calcium responses, and elevated Tac1 expression in our model. As noted above, interpretation of compound risk requires consideration of all endpoints collectively. For instance, although formaldehyde induced a variable calcium response and did not induce Tac1 expression, its high toxicity aligned with in vivo observations. Conversely, methanol elicited calcium and Tac1 responses but only induced toxicity at high concentrations both in vitro and in vivo. Interestingly, methanol has previously been shown to increase intracellular calcium in mammalian cells, however, to our knowledge, its correlation with Tac1 or substance P expression has not been investigated (Li et al., 1999). Furthermore, the discrepancy between Tac1 expression and limited toxicity would require further investigation to highlight the mechanism involved.

The organotypic, bronchial epithelial-derived model, EpiAirway™ demonstrates TRPA1-driven substance P responses

Compared with the lower airway, epithelial cells of the upper airway serve as an early defensive barrier, detecting inhaled environmental stimuli and shaping the body’s subsequent response. The bronchial-derived EpiAirway™ model offers several advantages over traditional submerged cell culture systems, including the ability to respond to compounds at physiologically relevant concentrations due to features such as ciliation and mucus production. Building on these attributes, we evaluated whether the EpiAirway™ model could also generate biomarkers associated with the onset of airway irritation, including TRPA1-mediated calcium signaling and substance P release. Integrating these mechanistic endpoints alongside phenotypically protective features of ciliation and mucus production, within a single model, may help to reduce false-positive classifications during compound risk assessment.

To determine the suitability of the EpiAirway™ model for producing irritation-related biomarkers, we first confirmed the presence of TRPA1 and Tac1 transcripts, with Ct values of 33.271 (±0.67) and 32.954 (±1.08), respectively (Figure 5A). Tissues were then exposed for 24 h to a concentration range of AITC or acetone, with or without 1 μM HC-030031, and viability was assessed by RNA yield (Figure 5B). AITC produced a significant calcium response (p < 0.05), which was inhibited following treatment with 100 μM HC-030031 for 30 min (Figure 5C). To further confirm TRPA1-dependent substance P signaling, Tac1 expression was measured following AITC or acetone treatment. A significant increase in Tac1 expression was observed at 0.01 μM AITC treatment (p < 0.05), followed by a decrease at 0.1 μM AITC, likely reflecting high cytotoxicity at this concentration. When 0.01 μM AITC was tested in the presence of TRPA1 inhibition, Tac1 expression was significantly reduced compared with non-inhibited conditions (p < 0.01), despite similar viability observed (∼50%), confirming TRPA1 involvement (Figure 5D).

Finally, to verify that Tac1 transcription translated into substance P protein release, substance P levels were quantified by ELISA (Figure 5E). A concentration-dependent increase in substance P release was observed following AITC exposure, with significant elevations at 0.1 μM (p < 0.001) and 1 μM (p < 0.0001). TRPA1 inhibition significantly reduced substance P release at both concentrations (p < 0.05 and p < 0.01, respectively), further supporting TRPA1-driven substance P signaling in this model.

Respiratory exposure to environmental irritants contributes to a range of acute and chronic airway disorders, and the prevalence of these conditions continues to increase despite the implementation of occupational exposure limits (OELs). Sensory irritation from inhaled chemicals is still assessed primarily using non-human animal models, most notably the RD 50 assay (Alarie, 1973). Although RD 50 values are informative for setting exposure thresholds and correlate well with lowest observed adverse effect levels (LOAELs) in humans (Kuwabara et al., 2007), a major limitation of this model is the substantial interspecies variability in ion channels, particularly TRPA1, that are central to chemosensory detection.

This variability in ion channel homology between species can lead to striking differences in channel pharmacology. For example, the human TRPA1 antagonists AMG5445 and AMG9090 act as agonists at the rat TRPA1 ortholog (Klionsky et al., 2007). In many cases, these species-specific differences arise from only a small number of amino acid substitutions. CMP1, which activates rat TRPA1 but inhibits human TRPA1, is functionally distinguished by Ala-946 and Met-949 in rat versus Ser-943 and Ile-946 in human (Chen et al., 2008). Similarly, the bimodal effects of menthol on mouse TRPA1, contrasted with its activation of human TRPA1, are determined by two residues in the pore region, Ser-876 and Thr-877 (Xiao et al., 2008). Even a single amino acid change can be decisive: substitution of Met268Pro in mouse TRPA1 converts caffeine from an agonist to an antagonist, and human TRPA1, activated by caffeine, naturally contains a proline at this position (Nagatomo et al., 2010).

Collectively, these examples illustrate that despite high overall conservation, subtle sequence differences between species can profoundly alter ion channel kinetics. Consequently, a compound’s ability to initiate downstream protective signaling pathways, such as the release of the neuro-inflammatory peptide substance P, may differ markedly between humans and commonly used experimental animal models, highlighting the need for species-relevant approaches to evaluating airway irritants.

Although substance P is classically released from neuronal populations to initiate inflammatory signaling, in this study we examined whether the human alveolar epithelial cell line A549 is capable of endogenous TRPA1–substance P signaling. TRPA1 expression has previously been reported in A549 cells, and its activation has been linked to cellular responses to wood smoke particulate matter; however, downstream biological consequences of TRPA1 activation in this context were not characterized (Shapiro et al., 2013). Additional studies in other cell types including rat dorsal root ganglia and human osteoarthritic chondrocytes have demonstrated that TRPA1 activation can drive inflammatory mediator release, including substance P, IL-6, and IL-1β (Nakamura et al., 2012; Nummenmaa et al., 2020; Lee et al., 2016; Nummenmaa et al., 2016). Moreover, the substance P receptor neurokinin-1 has been detected in A549 cells (Nguyen et al., 2023). Together, these independent findings suggest that A549 cells possess the molecular components necessary to support TRPA1-dependent neuro-inflammatory signaling. However, to our knowledge, endogenous TRPA1-driven substance P release has not been investigated.

To validate this potential, we first confirmed functional expression of both TRPA1 and TRPV1 in A549 cells using a flow cytometry - based calcium response assay (Figure 1). Although this approach has been applied previously in human B lymphocytes and macrophages, we sought to establish its suitability for calcium imaging in A549 cells (Bajnok et al., 2023; Schlautmann and Burgdorf, 2026). We therefore assessed calcium responses following exposure to a panel of chemicals known to elicit varying degrees of airway irritation in animal models (Figure 2A). Baseline fluorescence and cell viability, assessed using propidium iodide staining, remained stable across treatments, supporting the robustness of the assay for this application (Figure 2B). As expected, known irritants, including nicotine, p -xylene, and toluene, elicited clear calcium responses, whereas isopropanol, a compound associated with minimal irritation in humans, did not induce detectable calcium flux. These results indicate that the assay can discriminate between chemicals with differing irritancy profiles.

We next assessed concordance between calcium response and Tac1 gene induction - a necessary precursor for substance P production, following TRPV1 and TRPA1 activation (Figure 3). Results indicated that Tac1 expression was driven to a greater extent by TRPA1 activation, with TRPV1 activation inducing negligible Tac1 expression, although a small (∼1.5-fold) increase was observed when capsaicin was added at 500 μM. Treatment with acetone did not induce Tac1 expression. Furthermore, TRPA1-driven Tac1 gene expression translated to substance P protein expression as measured by immunocytochemistry. In contrast to our results, TRPV1 activation has been demonstrated to induce substance P release in stomach and pancreatic sensory neurons (Gazzieri et al., 2007; Wick et al., 2006). However, it should be noted that in corneal sensory neurons, TRPV1 activation has been shown to facilitate, rather than directly induce, substance P release in TRPM8+ neurons (Li et al., 2019). As a result, further investigative work would be required to confirm the presence and effect of additional molecular markers that may be required for TRPV1 driven substance P release in A549 cells. Furthermore, it is possible that our model would demonstrate TRPV1 mediated substance P release when tested at higher concentrations. In our model, when testing TRPV1-mediated responses, although a concentration-related calcium response was observed on immediate application of capsaicin at the selected concentrations, the concentrations selected for 24 h treatments prior to viability and Tac1 expression measurements, were not sufficient to induce an IC 50 viability response. Therefore, greater concentrations of capsaicin may have induced higher levels of Tac1 expression. Alternatively, TRPV1 may not have been localized to the plasma membrane, resulting in higher concentrations of capsaicin being required to evoke a calcium response than commonly used in neuronal cells (Koplas et al., 1997). Furthermore, this may explain the limited substance P expression observed following capsaicin treatment.

We next used the validated approach to test viability, calcium response and Tac1 expression, following treatment with a panel of compounds known to induce varying levels of respiratory irritation using in vivo models (Figure 4). Overall, we observed strong concordance among viability, calcium responses, and Tac1 expression, with a few notable exceptions. For instance, AITC, nicotine, p -xylene, and toluene all produced high toxicity, robust calcium influx, and elevated Tac1 expression, aligning well with in vivo irritancy data. Methanol, however, induced a calcium response and Tac1 expression without causing toxicity, underlining the importance of evaluating multiple biological endpoints when assessing irritant potential. Acetic acid further illustrated this point: despite an unusual concentration-dependent decrease in calcium response, it produced substantial toxicity and strong Tac1 induction, consistent with its known irritant properties. This again highlights the need to interpret all mechanistic readouts collectively rather than relying on any single measure.

In contrast to the positive responses observed with known irritants, acetone and isopropanol - both associated with low irritancy in vivo (23,480 and 5,000 ppm, respectively), did not induce toxicity, calcium influx, or Tac1 expression in our system. Beyond aligning well with established in vivo data (Table 2), the model also maintained stable baseline fluorescence and cell viability measurements throughout testing, as previously mentioned, confirming that the assay conditions were robust and suitable for reliably assessing calcium responses over the required duration.

The alveolar epithelium serves as an important barrier in the human lower airway; however, the nasal and bronchial epithelium constitute the primary interface with inhaled irritants. Therefore, we extended our findings from the alveolar A549 model to evaluate whether the bronchial-derived EpiAirway™ tissue exhibits comparable responses. The presence of structural features such as ciliation and mucus production enables the EpiAirway™ model to more closely recapitulate human airway physiology, making it a preferred platform for toxicity testing relative to monolayer systems like A549 cells. Despite this advantage, the expression of TRPA1 and Substance P within the EpiAirway™ model has not, to our knowledge, been previously characterized.

In this study, we first confirmed the presence of both TRPA1 and Substance P in EpiAirway™ tissues. We then demonstrated that TRPA1 activation elicited robust calcium flux and induced substance P expression, mirroring the responses observed in our A549-based system. These findings highlight the potential of the EpiAirway™ model as a physiologically relevant screening platform for identifying airway irritants (Figure 5). Despite these findings, we unexpectedly found that the presence of HC-030031 resulted in higher levels of toxicity in the EpiAirway™ model. We suggest that this may be a result of loss of protective mechanisms provided by TRPA1 activation, such as responses linked to oxidative stress and inflammatory mediators. Indeed, the presence of HC-030031 also reduced the secretion of substance P, which has demonstrated protective mechanisms in neuronal cells. Furthermore, substance P has demonstrated protective mechanisms in lung cells following exposure to cigarette smoke and respiratory viruses (Ahn et al., 2023; Hasenöhrl et al., 2000; Harris and Witten, 2003; Mehboob et al., 2024). Furthermore, the EpiAirway™ model demonstrated higher levels of toxicity when treated with AITC only, compared to the A549 cells. We suggest that this may result from the lower overall expression of substance P, as measured by Tac1 expression, observed in the EpiAirway™ model, however this requires further investigation, for example, measurement of A549 substance P release by ELISA, as was performed for the EpiAirway™ model, may provide additional insights.

In conclusion, we have verified that A549 cells and the EpiAirway™ model express paramount molecular biomarkers implicated in sensory irritation of the human airway (Martinez and Eling, 2019). Moreover, to our knowledge, this study provides the first evidence that these models can autonomously regulate Substance P production following activation of endogenously expressed TRPA1. We subsequently leveraged this capability to establish an innovative in vitro assay capable of screening compounds for their modulation of these molecular biomarkers. This assay could also be utilized alongside other in vitro models used in respiratory hazard assessment, such as those that focus on respiratory sensitization, for example, ALIsens® and GARD® air (Gutleb et al., 2026; Forreryd et al., 2015). Despite these findings, future work should aim to expand the number of compounds tested using this model to further refine its potential as a screening tool for respiratory irritation risk. These additional data would help to provide criteria by which compounds can be categorized based on their irritation potential. Furthermore, our model does not account for compound effects in the air phase. The immediate calcium responses induced on ion channel activation makes their measurement during concurrent compound exposure at the air liquid interface difficult. Despite this, future studies may investigate downstream signaling responses, such as substance P release, following exposure to compounds in their gaseous forms (Nielsen and Wolkoff, 2017).