Air pollution exacerbates nociceptor neuronal activity

(A–C) Male and female C57BL/6 mice (6–10 weeks old) were sensitized via intraperitoneal injection with an emulsion of ovalbumin (OVA; 200 µg/dose) and aluminum hydroxide (1 mg/dose) on days 0 and 7. On days 14–16, mice were challenged intranasally with OVA (50 µg/dose), either alone or in combination with fine particulate matter (FPM; 20 µg/dose). Bronchoalveolar lavage fluid was collected, and jugular-nodose complex neurons were cultured on day 17 for 24 hours before being loaded with the calcium indicator Fura-2AM. Cells were sequentially stimulated with the TRPA1 agonist AITC (successively to 10 µM at 60–90 seconds, 30 µM at 90–120 seconds, 100 µM at 120–150 seconds) and then with KCl (40 mM at 420–435 seconds). Calcium flux was continuously monitored throughout the experiment. The amplitude of AITC responses was measured by calculating the ratio of peak F340/F380 fluorescence after stimulation to the baseline F340/F380 fluorescence measured 30 seconds prior to stimulation. Data are plots as the per dish average of AITC and KCl responsive neurons and show that AITC (10 µM) responses was higher in JNC neurons from OVA-FPM exposed mice when compared to vehicle or OVA alone (C). Data in are presented as means ± SEM (B-C). N are as follows: B: n = 35 neurons (control group), 19 neurons (OVA group), and 38 neurons (OVA + FPM group), C: n = 5 dishes totalling 35 neurons (control group), 8 dishes totalling 42 neurons (OVA group), and 10 dishes totalling 76 neurons (OVA + FPM group). P-values were determined by nested one-way ANOVA with post-hoc Bonferroni’s. P-values are shown in the figure.

Air pollution reprograms the transcriptome of nociceptor neurons

(A–F) Naïve male and female TRPV1cre::tdTomatofl/wt mice (6–10 weeks old) underwent either a pollution-exacerbated asthma protocol, the classic OVA protocol, or remained naïve. On day 17 (peak inflammation), jugular-nodose-complex (JNC) neurons were harvested and dissociated, and TRPV1⁺ neurons (tdTomato⁺) were sorted via FACS to remove stromal cells and non-peptidergic neurons. RNA was then isolated for sequencing.

Volcano plots (A, C, E) and heatmaps (B, D, F) show differentially expressed genes (DEGs) for three comparisons: OVA + FPM vs. naïve (A-B), OVA + FPM vs. OVA alone (C-D), and OVA alone vs. naïve (E-F). Notable genes with increased expression include Lifr and Oprm3 in OVA + FPM vs. naïve, Oprm1, Nefh, P2ry1, Prkcb, Gabra1, and Kcnv1 in OVA + FPM vs. OVA, and Npy1r and Kcna1 in OVA alone vs. naïve.

Data are presented either as volcano plots (A, C, E), showing the log2 fold change of TPM between groups along with the corresponding –log10 p-values from DESeq2 analysis, or as heatmaps (B, D, F), showing the z-scores of rlog-transformed normalized counts. The experimental groups were naïve (n=2; A–B, E–F), OVA (n=3; C–F), and OVA-FPM (n=3; A–D). P-values were determined by DESeq2 (A, C, E) and are indicated in the figure.

Nociceptor neurons control pollution-exacerbated asthma

(A–B) Male and female C57BL/6 mice (6–10 weeks old) were sensitized intraperitoneally with ovalbumin (OVA; 200 µg/dose in 200 µL) and aluminum hydroxide (1 mg/dose in 200 µL) on days 0 and 7. On days 14–16, mice were challenged intranasally with OVA (50 µg/dose in 50 µL) alone or with fine particulate matter (FPM; 20 µg/dose in 50 µL). On day 16, 30 minutes after the final challenge, mice received intranasal QX-314 (5 nmol/dose in 50 µL). Bronchoalveolar lavage fluid (BALF) was collected on day 17 and analyzed by flow cytometry. Compared with naïve or OVA-exposed mice, those co-challenged with OVA + FPM showed increased BALF neutrophils (A). QX-314 treatment normalized these levels, while BALF eosinophil levels remained comparable (B).

(C–E) Male and female littermate control (TRPV1WT) and nociceptor-ablated (TRPV1DTA) mice (6–10 weeks old) were sensitized and challenged under the same OVA ± FPM protocol (days 0, 7, and 14–16). BALF or lungs were collected on day 17 and assessed by flow cytometry. Compared with naïve or OVA-exposed mice, OVA + FPM co-challenged mice exhibited higher BALF neutrophils (C) and lung γδ T cells (E). Nociceptor ablation protected against these increases (C, E), while BALF eosinophil levels remained comparable (D).

Data are shown as mean ± SEM (A–E). Experiments were replicated twice, and animals pooled (A-E). N are as follows: A-B: control (n=6), OVA (n=7), OVA-FPM (n=12), OVA-FPM+QX-314 (n=10), C-D: TRPV1WT + control (n=9), TRPV1WT + OVA (n=13), TRPV1WT + OVA-FPM (n=18), TRPV1DTA + OVA-FPM (n=19), E: TRPV1WT + control (n=3), TRPV1WT + OVA (n=3), TRPV1WT + OVA-FPM (n=4), TRPV1DTA + OVA-FPM (n=5). P-values were determined by a one-way ANOVA with post-hoc Tukey’s (A-E). P-values are shown in the figure.

Vagal sensory neurons gatekeep alveolar macrophage motility and neutrophil numbers

(A–D) Male and female littermate control (NaV1.8wt::DTAfl/wt denoted as NaV1.8WT) and nociceptor-ablated (Nav1.8cre::DTAfl/wt denoted as NaV1.8DTA) mice (6–10 weeks old) were sensitized via intraperitoneal injection with an emulsion of ovalbumin (OVA; 200 µg/dose) and aluminum hydroxide (1 mg/dose) on days 0 and 7. Phagocytes were labeled by intranasal injection of PKH26 Red Fluorescent Cell Linker Kit (25 pmol/dose) on day 10. Mice were then challenged intranasally with OVA (50 µg/dose) alone or in combination with fine particulate matter (FPM; 20 µg/dose) on days 14–16, and images were acquired on day 17. (A) Representative maximum-intensity projection of alveolar macrophages (AMs, red). Scale bar: 20 µm. (B) Quantification of AM numbers per field of view (FOV). (C) Net displacement of AMs over 1 hour. (D-F) Representative tracks of individual AMs (each color represents a single AM) over 1 hour. While the AM numbers (A-B) were not impacted, their net displacement (C-F) was reduced in OVA-FPM-exposed NaV1.8DTA mice.

(G-L) Male and female littermate control (NaV1.8wt::DTAfl/wt denoted as NaV1.8WT) and nociceptor-ablated (NaV1.8cre::DTAfl/wt denoted as NaV1.8DTA) mice (6–10 weeks old) were sensitized via intraperitoneal injection with an emulsion of ovalbumin (OVA; 200 µg/dose) and aluminum hydroxide (1 mg/dose) on days 0 and 7. Mice were then challenged intranasally with OVA (50 µg/dose) alone or in combination with fine particulate matter (FPM; 20 µg/dose) on days 14–16, and images were acquired on day 17. Immediately prior to perform the intravital imaging, we administered an intravenous Ly6G antibody to label neutrophils. (G) Representative maximum-intensity projection of neutrophils (black). Scale bar: 20 µm. (H) Quantification of neutrophil numbers per FOV. (I) The total displacement of neutrophils over 20 minutes. (J-K) Representative tracks of individual neutrophils (each color represents a single neutrophil) over 20 minutes. (L) Frequency of neutrophil behaviors per FOV (adherent, crawling, patrolling, or tethering). Data show that OVA-FPM-exposed littermate control mice present an increase in neutrophil numbers per field of view (G-H), an effect absent in NaV1.8DTA. Interestingly, OVA-FPM-exposed NaV1.8DTA mice show an increase in neutrophil net displacement (I-K), an effect irrespective to one of the specific behaviors tested (L).

Data are presented as representative image (A, G; scale bar: 20µm), mean ± SEM (B, C, H), spider plot (D-F, J-K), violin plot showing median (I), and staked bar graph showing mean ± SEM (L). N are as follows: B: NaV1.8WT + control (n=6), NaV1.8WT + OVA-FPM (n=6), NaV1.8DTA + OVA-FPM (n=6), C: NaV1.8WT + control (n=42), NaV1.8WT + OVA-FPM (n=42), NaV1.8DTA + OVA-FPM (n=42), H: NaV1.8WT + control (n=6), NaV1.8WT + OVA-FPM (n=6), NaV1.8DTA + OVA-FPM (n=6), I: NaV1.8WT + OVA-FPM (n=385), NaV1.8DTA + OVA-FPM (n=469), J: NaV1.8WT + OVA-FPM (n=10), NaV1.8DTA + OVA-FPM (n=11). P-values were determined by a one-way ANOVA with post-hoc Tukey’s (B, C, H) or unpaired Student T-test (I, L). P-values are shown in the figure.

Artemin sensitizes TRPA1 activity in vagal sensory neurons

(A-B) Male and female littermate control (TRPV1WT) and nociceptor-ablated (TRPV1DTA) mice (6–10 weeks old) were sensitized and challenged under the same OVA ± FPM protocol (days 0, 7, and 14–16). BALF was collected on day 17 and assessed by multiplex array and ELISA. Compared with naïve or OVA-alone groups, OVA + FPM co-challenged mice exhibited levels of TNFα, and artemin. Notably, ablating nociceptors prevented these increases.

(C) In-silico analysis of the GSE124312 dataset51. The heatmap displays transcript expression levels for the pan neural-crest lineage transcription factor (Prdm12), voltage-gated sodium channels (Scn9a, Scn10a), jugular subset markers (Wfdc2, Mrgprd, Osmr, Sstr2, Nefh, Trpm8), peptidergic neuron markers (Trpa1, Trpv1, Calca, Tac1, Gfra3), and the pan placodal lineage marker (Phox2b). Gfra3 expression is enriched in the peptidergic neuron cluster labeled JG4. Experimental details and cell clustering are described by Kupari et al51.

(D) In-silico analysis of GSE19298752 showing co-expression of Gfra3 with Trpa1 and other inflammatory markers. Data are visualized as row z-scores in a heatmap or via UMAPs (TPTT > 1). Experimental details and cell clustering are described by Zhao et al52.

(E-G) Alveolar macrophages (3 × 10⁵ cells/well) from naïve male and female C57BL/6 mice were cultured overnight and then stimulated with vehicle (DMSO) or FPM (100 µg/mL). RNA was extracted 1- and 4-hours post-stimulation and Artn expression was assessed using qPCR. FPM exposure increased Artn transcript levels at both 1 and 4 hours (F, G).

(H-J) Naïve mice jugular-nodose-complex neurons were harvested, pooled, and cultured overnight with either vehicle or artemin (100 ng/mL). Cells were sequentially stimulated with AITC (TRPA1 agonist; 300 µM at 240–270 seconds), capsaicin (TRPV1 agonist; 300 nM at 320–335 seconds), and KCl (40 mM at 720–735 seconds). The percentage of AITC-responsive neurons (among all KCl-responsive cells) was normalized to vehicle-treated controls for each batch of experiments. Artemin-treated neurons showed increased responsiveness to AITC, while responses to capsaicin and KCl were unchanged (I-J).

Data are presented as means ± SEM (A-B, F-G, J), heatmap displaying the z-score of DESeq2 normalized counts (C), tSNE plots (D), schematics (E, H), means ± 95% CI of maximum Fura-2AM (F/F₀) fluorescence (I). N are as follows: A: TRPV1WT + control (n=2), TRPV1WT + OVA (n=3) TRPV1WT + OVA-FPM (n=3), TRPV1DTA + OVA-FPM (n=8), B: TRPV1WT + OVA (n=6) TRPV1WT + OVA-FPM (n=8), TRPV1DTA + OVA-FPM (n=14), F: n=2/time point, G: n=8/group, I: vehicle (n=107 neurons), Artemin (n=122 neurons); J: n=4/group. P-values were determined by a one-way ANOVA with post-hoc Tukey’s (A, B) or unpaired Student T-test (G, J). P-values are shown in the figure.

In silico analysis of Artn expression in mouse immune cells

(A) In-silico re-analysis of Artn expression in mouse immune cells using the ImmGen database53. Artn and Ahr are expressed in Itgam+ macrophages. Data are presented as per-gene z-scores of normalized gene expression, calculated by the median of ratios method.

(B) In-silico re-analysis of the single-cell RNA-seq dataset from Tavares-Ferreira et al., 105 (Sensoryomics; dbGaP accession phs001158) shows that Gfra3 is expressed in Trpa1-positive nociceptors, C-LTMRs, and silent nociceptors within the human dorsal root ganglion. Expression levels are reported as per-gene z-scores calculated with the median-of-ratios normalization method.

In-silico re-analysis of Ahr and Artn expression in human tissues

(A) In-silico re-analysis of data from Karlsson et al., 106 (Human Protein Atlas, proteinatlas.org) indicates that Ahr and Artn are expressed in human lung. Expression values are reported as per-gene z-scores using the median-of-ratios normalization method. Experimental details and cell clustering are described by Karlsson et al., 106

(B) In-silico re-analysis of data from Uhlén et al. 55 (Human Protein Atlas, proteinatlas.org) confirms Ahr protein expression in human lung macrophages, as shown by immunohistochemistry. These data are presented as protein-normalized expression levels. Experimental details and cell clustering are described by Uhlén et al. 55.

(C) In-silico re-analysis of data from Abdulla et al. 56 (CELLxGENE, CZI Single-Cell Biology) reveals co-expression of Artn and Ahr in lung and alveolar macrophages from patients. Expression values are provided as per-gene z-scores, calculated by the median-of-ratios normalization method. Experimental details and cell clustering are described by Abdulla et al. 56.

Schematic of nociceptor involvement in pollution-exacerbated allergic asthma

In our study, mice were exposed to PM₂₅ particles and ovalbumin (OVA) to model pollution-exacerbated asthma. Compared to mice exposed to OVA alone, co-exposure to PM₂₅ and OVA significantly increased bronchoalveolar lavage fluid (BALF) neutrophils and lung γδ T cells levels. To counteract this heightened airway inflammation, we administered intranasal QX-314—a charged lidocaine derivative—at the peak of inflammation, effectively normalizing BALF neutrophil levels. Ablation of TRPV1⁺ nociceptor neurons produced a similar effect.

Further analysis with calcium imaging revealed that neurons from the jugular-nodose complex in pollution-exposed asthmatic mice were more sensitive via their TRPA1 channels. Levels of TNFα and the growth factor artemin were also elevated in the BALF of these mice, returning to normal following nociceptor ablation.

We identified alveolar macrophages as the source of artemin, which they secrete upon sensing fine particulate matter (FPM) through aryl hydrocarbon receptors. Artemin, in turn, heightened TRPA1 responsiveness to its agonist (mustard oil), thereby exacerbating airway inflammation. Our findings suggest that silencing nociceptor neurons can disrupt this pathway, offering a novel therapeutic approach to mitigate neutrophilic airway inflammation driven by pollution.