Pulmonary neuroendocrine cell–derived exosomes regulate iron homeostasis and oxidative stress in lung neurons

Abstract

Nicotine, the principal addictive component of cigarettes, is linked to cognitive decline and neurodegenerative alterations, likely through oxidative stress and impaired iron regulation in neurons. Yet, underlying molecular pathways remain unclear. This study examined the role of pulmonary neuroendocrine cells (PNECs) in smoke-induced neural changes. Using human pluripotent stem cells, we generated induced PNECs (iPNECs) to overcome culture limitations and performed mechanistic analyses. We found that nicotine exposure stimulates iPNECs to secrete exosomes enriched with serotransferrin, an iron-binding glycoprotein. Neurons internalizing these exosomes displayed elevated levels of transferrin receptor 1 (TFR1), divalent metal transporter 1, and duodenal cytochrome b, associated with ferritin accumulation, oxidative stress, and adenosine triphosphate depletion. Inhibition of TFR1 alleviated these effects. Furthermore, nicotine-triggered exosomes increased α-synuclein expression in neurons in a manner consistent with stress- and vulnerability-associated signatures observed in human lungs and nicotine-exposed mice, highlighting PNEC-derived exosomal signaling that may contribute to neuronal dysfunction.

Nicotine triggers PNEC-derived exosomes that alter iron and redox balance associated with neuronal dysfunction.

INTRODUCTION

The association between cigarette smoking and the development of neuropathology, including cognitive decline and motor behavior deterioration, has been widely reported (1–6). Nicotine, one of the major components of tobacco smoke, has been found to trigger neurological complications including effects on cognition and memory (7–10) by increasing oxidative stress in brain cells (11, 12), which is also associated with iron dyshomeostasis (13–15). Intracellular accumulation of iron occurs before the development of amyloid-β and neurofibrillary tangles (16, 17), which damage neurons via ferroptosis, a type of programmed cell death relying on iron levels (18, 19), indicating an association of iron dyshomeostasis with neurodegenerative disease. However, the precise mechanism underlying iron dyshomeostasis–based oxidative stress in neuronal cells after smoking or nicotine exposure remains to be elucidated.

The brain is connected to various body organs including the lung via the vagus nerve, the 10th cranial nerve, which descends from the brain in the carotid sheath lateral to the carotid artery and then connects and communicates with the lungs (20). Among the various cell types in the lung, pulmonary neuroendocrine cells (PNECs) represent the most prominently structurally innervated epithelial populations. PNECs are rare (less than 1% of total lung cell population) airway epithelial cells that uniquely harbor both neuronal and endocrine characteristics (21–23). These cells function as an important airway sensor responding to chemical or mechanical stimuli by releasing neuropeptides and neurotransmitters (22, 24). At airway bifurcation sites, PNECs form small clusters called neuroepithelial bodies (NEBs), which interact with intraepithelial nerve fibers. The axons of nerve fibers enter the brain and transmit sensory information (25), and atrophy of this nerve has also been correlated with dementia (26–31), suggesting that PNECs in close proximity to neurons can establish intercellular communication.

Exosomes are nanovesicles with a diameter around 30 to 200 nm released by most cell types that circulate in biofluids including blood, cerebrospinal fluid, and pleural fluid (32–35). Exosomes play a crucial role in intercellular communication in disease progression including neuropathology via oxidative stress (36); however the role of PNEC-derived exosomes in iron dyshomeostasis has not yet been investigated. Currently, the study of exosome-based communication between PNECs and neuronal cells is difficult owing to the limited and rare availability of PNECs. Moreover, modeling PNECs of human origin is a difficult task as PNECs are rare and few methods have been established to grow and expand primary PNEC cells. To circumvent this barrier, we previously established the technology to efficiently generate PNECs via directed differentiation of human pluripotent stem cells (hPSCs), namely induced PNECs (iPNECs) (37). Cigarette smoking is known to lead to iron dysregulation and oxidative stress, which contribute to neuropathology, including neurodegenerative diseases (38, 39). Here, we report the use of iPNECs for the study of exosome-based intercellular communication with neuronal cells to examine the effect of nicotine on the release of exosomes from iPNECs, their altered cargoes, and their communication with neuronal cells.

We investigated whether serotransferrin-enriched exosomes from iPNECs influence iron homeostasis and oxidative stress in neurons. We examined the effect of iPNEC-derived exosomes on expressions of key iron transporters transferrin receptor 1 (TFR1), DMT1, and DCYTB in HD10.6 neuronal cells, which express sensory neuron–associated transcription factors and, as a human dorsal root ganglion (DRG) neuronal cell type, share some molecular features with vagal neurons (22, 40, 41). We used magnetic bead–based immunocapture of SYN⁺ exosomes derived from PNECs of ex vivo–cultured lungs from control and nicotine-treated mice, to study their effects on iron homeostasis, oxidative stress, and mitochondrial function on neuronal cells. To validate the role of exosome release and serotransferrin, we inhibited exosome secretion in iPNECs using GW4869 (an inhibitor for exosome release) and performed a loss-of-function study of transferrin (TF) using short hairpin RNA (shRNA). This study revealed a previously unknown mechanism of iron dyshomeostasis–dependent oxidative stress in neurons mediated by exosomes released from nicotine-exposed iPNECs, providing previously unidentified mechanistic insight into exosome-mediated neuronal stress pathways relevant to neurodegenerative disorders.

RESULTS

PNECs are rare airway sensory epithelial cells innervated by neurons

To identify the distribution of PNECs and their innervation by neurons in the lung, we used immunohistochemistry, staining for CGRP (a PNEC-associated marker), β-III-tubulin (pan-neuronal marker), and TRPV1 (a sensory neuron–associated marker–-positive cells in lung tissue samples of healthy humans and C57BL/6 wild-type (WT) adult mice. Solitary PNECs were found to be bean-shaped and more abundant than clusters of PNECs (often referred to as NEBs) in human lung bronchial epithelium stained for CGRP. The immunofluorescence staining of neurons with β-III-tubulin showed their distinct colocalization with CGRP⁺ PNECs, indicating the direct interaction of neurons with PNECs in human (Fig. 1, A and B) and mouse (Fig. 1, C and D) lung tissue. TRPV1- or β-III-tubulin–positive vagal sensory neurons colocalized with CGRP⁺ PNECs in both human (Fig. 1, A, B, and E) and mouse (Fig. 1, C to E) lung tissues. The protein expression of P2Y1 in neuronal cells innervating the CGRP⁺ PNECs is consistent with the identity of PNEC-innervating neuronal cells as vagal neurons in human and mouse lung tissue samples (fig. S1, A and B). To further understand the neuroinnervation of PNECs in three-dimensional (3D), we used 3D iDISCO tissue clearing, which enabled 3D visualization and immunolabeling in unsliced whole mouse lungs (42). This approach revealed the close proximity of PNEC and neurons (Fig. 1F), confirming the interaction of PNECs with neurons in both human and mouse lungs. This finding prompted us to study their communication.

Fig. 1. PNECs are innervated by neurons in the lungs.

(A and B) Representative immunofluorescence staining images showing expression of (A) β-III-tubulin (pan-neuronal marker) and CGRP (PNEC-associated marker), (B) TRPV1 (sensory neuron–associated marker) and CGRP, in the lung tissue sample of a healthy donor. (C and D) Representative immunofluorescence staining images showing expression of (C) β-III-tubulin (a pan-neuronal marker) and CGRP, (D) TRPV1 and CGRP, in the lung tissue sample of C57BL/6 adult mice. (E) The quantitative bar graph, depicting the percentage of the PNEC population linked to TUBB3⁺ or TRPV1⁺ neurons, in human and mouse lungs. (F) Whole-tissue 3D assessment of local neural innervations of PNEC in the lung tissues using iDISCO tissue clearing. Zoom-in: Representative unsectioned lung of the C57BL/6 adult mice for the whole-tissue immunolabeling of anti–β-III-tubulin and anti-CGRP, showing the PNEC with neuroinnervation, depicted by the coexpression of CGRP and β-III-tubulin. In (A) to (D) and (F), yellow-colored arrowheads indicate the neuroinnervation of PNEC in lung samples of human or mouse. Scale bars, (A) to (D) 20 μm; (F) 100 μm; (zoom-in) 50 μm. In bar graph shown in (E), data are presented as means ± SEM; n = 4 human donors and n = 6 mice. Two fields per donor per mouse were quantified as technical replicates and averaged to obtain one value per donor per mouse. Each data point represents one biological sample (donor or mouse). Significance level: *P < 0.05, TUBB3⁺ or TRPV1⁺ neuron–linked PNEC population in human lung versus mouse lung.

hPSC-derived iPNECs have iron homeostasis machinery like native PNECs

In light of the unavailability of any representative human cell line as a model of PNEC and the challenge in using primary culture of PNEC from mouse due to their rare numbers, we differentiated hPSC into iPNECs. To facilitate the purification of PNECs for fractionated culture followed by functional studies, we engineered a reporter hPSC line with lentiviral vectors carrying the fluorescent protein enhanced yellow fluorescent protein (eYFP) (pseudo-green color was used for better visibility) controlled by the CGRP gene promoter. By replicating the in vitro development of lung environment in stages, we induced the hPSCs into definitive endoderm (DE), followed by anterior foregut endoderm (AFE), and eventually to lung progenitor (LP) cells (Fig. 2A and fig. S2, A to C). By changing NOTCH signaling, which directs the differentiation of neuronal and secretory cells, we achieved considerable proportions of PNECs (8 to 10% of the total cell population) from the dedicated lung precursors (Fig. 2, B and C, and fig. S2, A to C). Further, we validated the expression of CGRP in YFP reporter cell line (fig. S2D). We analyzed the iLung (batches 1 and 2) using single-cell RNA-sequencing (scRNA-seq). Figure 2D shows a t-distributed stochastic neighbor embedding (t-SNE) plot depicting the clusters of major cell types in the mixed culture, enriched with iPNEC in cluster 7 as evidenced by the predominant expression levels of CALCA and ASCL1, canonical PNEC markers (Fig. 2, E to H). The evidence indicates that the expression of key maturity markers in hPSC-derived iPNECs suggests a mature-like state with transcriptional features resembling native human PNECs. A study reporting that PNECs are among the first specialized epithelial cell types to appear during lung development (37, 43) supports a short and feasible differentiation trajectory between these cells and committed progenitor cells. scRNA-seq data pointed to high similarity between the profiles of the iPNECs and those of native PNECs in humans (fig. S3, A and B) (37). The iPNECs were also shown to have neuroendocrine (NE) characteristics and coexpress NE markers, transcription factors, and signaling molecules that have been associated with native human PNECs [e.g., CGRP, GRP, SYP, UCHL1, substance P, chromogranin A (CgA), NCAM1, NE transcription factor ASCL1, and lung lineage marker NKX2.1]. In addition, they express the main neurotransmitter transporters or enzymes (e.g., GAD1, ACHE, and SLC6A4) and are capable of secreting major PNEC-related neurotransmitters [e.g., serotonin, γ-aminobutyric acid (GABA), adenosine triphosphate (ATP)] (Fig. 2I and fig. S4, A and B) (37). Furthermore, we confirmed the presence of neurotransmitters including GABA, acetylcholine, and serotonin in lysed iPNECs through metabolomic profiling and in the conditioned medium of iPNECs using surface plasmon resonance analysis (fig. S4C). Notably, iPNECs also express ACO1, SLC40A1, and HFE (fig. S4D), which function during iron metabolism. For example, ACO1 is a major source of cellular iron-sulfur (Fe-S) clusters, which are essential cofactors for many enzymes involved in cellular metabolism and redox homeostasis (44). SLC40A1 is regulated by the iron-regulatory hormone hepcidin (45), and HFE plays a role in iron regulation (46). Similar validation of PNEC associated markers was performed using publicly available mouse scRNA-seq data (fig. S5, A to E). This evidence implies that iPNECs are physiologically relevant and suitable to serve as a novel cell resource for studying cross-talk with neuronal cells under smoke or nicotine conditions in the context of iron homeostasis–dependent oxidative stress.

Fig. 2. Characterization of hPSC-derived iPNEC.

(A) Representative flow diagram depicting the strategy for the stepwise directed differentiation of hPSCs into day-3 DE, day-6 AFE, and increasing numbers of day-15 to -25 LP cells using differentiation mixtures I to V. The LP cells were further differentiated in mixture VI along with DAPT to induce the formation of PNECs with a mixed population of LP cells. The mixtures I to VI are defined in Materials and Methods. (B) Representative bright-field and fluorescence images of CGRP-YFP reporter cells representing iPNEC. Pseudo green color was used for depiction (Scale bar, 100 μm). (C) Flow cytometry analysis showing the CGRP-YFP reporter cells representing iPNEC. (D) t-SNE map showing clustering of various cell type populations including PNEC in mixed lung cell population. (E and F) t-SNE maps showing the major expression of (E) CGRP (or CALCA) and (F) ASCL1 in cluster 7 depicting iPNEC, and (G and H) the corresponding violin plots showing the expression of CGRP and ASCL1 in iPNEC. (I) Representative violin plots showing the expression of major markers of native PNEC in iPNEC.

Nicotine increases iPNEC exosome release and enriches serotransferrin cargo

iPNECs were purified using the CGRP-YFP reporter–based fluorescence-activated cell sorting (FACS) followed by fractionated culture and nicotine treatment. Exosome release was 2.8-fold higher in nicotine-treated iPNECs than in untreated controls. The diameters of the exosomes released from control iPNECs and nicotine-treated iPNECs ranged from 92.5 to 268.5 nm and 63.5 to 295.5 nm, respectively (Fig. 3, A and B). This indicates that the augmented release of exosomes could be a crucial biophysical phenotype because of nicotine’s effect on the pulmonary system, particularly on PNECs. Transmission electron microscopy (TEM)–based and atomic force microscopy (AFM)–based analyses showed the heterogeneous morphology as well as physical size of the iPNEC-derived exosomes (Fig. 3, C to F). These exosomes were found to be circular and cup-shaped with size around 100 nm from the representative TEM image (Fig. 3C), which was also supported by the height profiles from the AFM-based topographical analysis (Fig. 3, D to F). The expression of CD63 (a prevalent exosome marker) further validated the presence of exosomes (Fig. 3G). Exosomes from both control- and nicotine-treated iPNECs were found to be internalized by the neuronal cells (Fig. 3H), and exosomes from nicotine-treated iPNECs were associated with increased intracellular calcium and membrane potential in neuronal cells (fig. S6, A to D). In addition, we also cocultured the iPNECs and neuronal cells together, followed by immunostaining with FM4-64 and AFM-based topographical analysis, which revealed the interaction between single PNEC and neuronal cells (fig. S7, A to G), suggesting that PNECs may interact with neurons either through direct contact or indirectly via paracrine effects when in close proximity.

Fig. 3. iPNEC-derived exosomes, validated to contain serotransferrin, are associated with increased calcium and membrane-potential responses in neuronal cells.

(A) Size-distribution curves showing enhanced release of exosomes from iPNECs treated with nicotine compared with untreated controls and (B) their quantitative bar graph. (C) Representative TEM micrograph of iPNEC-derived exosomes. (D to F) AFM-based images: (D) 2D, (E) 3D, and (F) height profiles of iPNEC-derived exosomes. (G) Immunogold-EM micrograph showing CD63-positive 10-nm gold particles in exosomes isolated from control- and nicotine-treated iPNECs. (H) Immunofluorescence staining showing uptake of control- and nicotine-treated iPNEC-derived exosomes by neuronal cells. (I and J) Proteomic analysis of control- and nicotine-treated iPNECs: (I) representative plot showing the differentially expressed constituents in control- versus nicotine-treated iPNECs. The yellow-marked dot indicates significantly higher expression of serotransferrin in nicotine-treated iPNECs compared with the vehicle-treated control group. (J) Bar graph showing the enhanced expression of serotransferrin in nicotine-treated iPNECs compared with controls, as detected by proteomic analysis. (K) Representative bar graph showing levels of serotransferrin in iPNEC-derived exosomes compared with conditioned medium, as detected by ELISA. Data in bar graphs [(B), (J), and (K)] are presented as means ± SEM; n = 3 independent biological replicates. Technical replicates were averaged within each experiment. Statistics: two-tailed unpaired Student’s t test. Significance levels: *P < 0.05, **P < 0.01; control versus nicotine groups.

To identify candidate factors potentially contributing to the altered neuronal phenotypes, we performed proteomics analysis of control- and nicotine-treated iPNECs as an initial screen (Fig. 3I and fig. S8, A and B). The whole-cell proteomics served to nominate serotransferrin as a candidate, and its presence in exosomes was validated independently by immunogold labeling and enzyme-linked immunosorbent assay (ELISA). Serotransferrin, an iron-binding transport protein capable of facilitating transport of iron, was significantly up-regulated in nicotine-treated iPNECs (Fig. 3J), which was also increased in exosomes as compared with control iPNECs (fig. S9, A to C). Further, the ELISA-based analysis of conditioned medium as well as exosomes from iPNEC demonstrated that the secretion of serotransferrin was mainly through the release of exosomes (Fig. 3K). Thus, the augmented level of serotransferrin in exosomes from nicotine-treated iPNECs could be associated with phenotypes of neuronal cells. Elevated intracellular calcium has been associated with neuronal stress and vulnerability and in some contexts with neurodegeneration-related and cell death–associated pathways (47). The behavior of the nicotine-treated mice was found to be affected as evidenced by a decrease in the recognition index in novel object recognition (NOR) test (fig. S10, A and B), increase in latency time and number of nose poke errors by the mice in Barnes maze and damsel-in-distress test (fig. S10, C to E), a decrease in time required for the latency to fall by mice in mesh test (fig. S10, F and G, and movie S1), and increased time spent by mice in the center of an open arena as detected by open field test (fig. S10, H and I). These results indicate impaired recognition memory, and that locomotor activity and strength were also affected by the nicotine administration. There was also an increased level of α-synuclein (SNCA, a protein implicated in neurodegenerative disorders) in neuronal cells in the lung tissue of smokers and the nicotine-treated mice (fig. S11, A to D), and SNCA has been reported to be involved in dementia and neurodegeneration (48). In addition, mice subjected to nicotine (Nic)–iPNEC-exosome (EXO) treatment showed an up-regulation of Snca, glial fibrillary acidic protein, and Cd11b in brain tissue, reflecting neuroinflammatory responses and activation of glial cell populations, such as astrocytes and microglia (fig. S12, A to C), which may contribute to early vulnerability-associated neurobiological changes that could detrimentally affect brain health. Further, the recognition index was found to be reduced in mice injected with Nic-iPNEC-EXO compared with the control group (fig. S12D). In addition, the immunofluorescence staining of Snca in axonal projections of lung-innervating neurons showed enhanced accumulation of Snca in the terminal bronchioles of P301S tau transgenic mice, as compared with that in WT mice lungs (fig. S12E). These neurobiological alterations may reflect early changes relevant to cognitive functions, particularly recognition memory–related processes. Consistent with this, Nic-iPNEC-EXO–injected mice showed a reduced recognition index in the NOR test.

PNEC exosomes perturb neuronal iron handling and oxidative-stress responses

Accumulating evidence suggests that smoking cigarettes causes elevation of iron level in the lung, which has also been correlated with iron homeostasis imbalance. This iron dyshomeostasis further leads to oxidative stress and subsequent tissue injury (49). To examine how nicotine-treated iPNEC-derived exosomes affect the neuronal cells in their vicinity, we analyzed the expression of major iron transporters (TFR1, DMT1, and DCYTB) in neurons. Notably, the protein levels of TFR1, DMT1, and DCYTB were elevated in neurons upon treatment with exosomes from iPNECs exposed to nicotine, as detected by immunofluorescence imaging (Fig. 4, A to C) and Western blotting (Fig. 4D and fig. S13A). This implies that the enhanced level of iron transporters on the neuron may have a functional role. Intriguingly, serotransferrin has been reported to bind to TFR1 to facilitate the transport of iron across the cells (50), which led us to hypothesize that the augmented serotransferrin enriched in exosomes from nicotine-treated iPNECs could play a role in perturbed iron homeostasis in neurons. Therefore, we further explored the level of ferritin and oxidative stress in neurons. Ferritin levels were found to be significantly enhanced in neuronal cells treated with exosomes from iPNECs exposed to nicotine (Fig. 4E). In addition, the level of ATP was reduced (Fig. 4F), and oxidative stress was increased as evidenced by the CellROX reagent labeling (Fig. 4G and fig. S13B) and the decreased glutathione (reduced form)/oxidized glutathione (GSH/GSSG) ratio in neuronal cells exposed to exosomes from nicotine-treated iPNECs compared with iPNEC-EXO or vehicle control (Fig. 4H). Further, the neuronal cells showed a trend of up-regulated levels of TFR1 and oxidative stress upon treatment with nicotine exposed non-PNEC iLung cells–derived exosomes; however the increase is not significant (fig. S14, A and B). In addition, the SNCA was found to be elevated in neuronal cells exposed to exosomes from nicotine-treated iPNECs compared with the control (Fig. 4, I and J, and fig. S13C), suggesting a potential neuropathological phenotype trend, warranting future detailed investigation. The oxygen consumption rate (OCR) as analyzed by Seahorse showed a reduced rate of mitochondrial respiration, reflecting a decline in health and function of mitochondria (Fig. 4K) that was accompanied by lower maximal and basal respiration in neuronal cells treated with Nic-iPNEC-EXO versus control (Fig. 4L). Together, these findings support a model in which nicotine exposure increases the release of serotransferrin-enriched iPNEC exosomes, which in turn may contribute to iron dyshomeostasis–associated oxidative stress in neuronal cells.

Fig. 4. Exosomes from iPNEC perturb iron homeostasis and oxidative stress in neuronal cells.

(A to C) Immunofluorescence (IF) images and the quantitative violin plots, showing the expression and distribution of (A) TFR1, (B) DMT1, and (C) DCYTB in neuronal cells treated with exosomes from control- and nicotine-treated iPNECs. Scale bar, [(A) to (C)] 20 μm. (D) Immunoblots, showing the protein level of TFR1, DMT1, and DCYTB in neuronal cells treated with exosomes from control- and nicotine-treated iPNECs. (E and F) Bar graphs showing (E) elevation of ferritin level and (F) decrease of ATP level in neuronal cells exposed to exosomes from nicotine-treated iPNECs compared with iPNEC-EXO or vehicle-treated control. (G) Detection of oxidative stress in neuronal cells treated with exosomes from control- and nicotine-treated iPNECs; measured with CellROX reagent. (Scale bar, 50 μm). (H) The level of GSH/GSSG ratio in neuronal cells exposed to exosomes from nicotine-treated iPNECs compared with iPNEC-EXO or vehicle-treated control. (I and J) IF staining and the quantitative bar graph showing the expression of SNCA in neuronal cells treated with exosomes from control- and nicotine-treated iPNECs (Scale bar, 20 μm) and the corresponding bar graph. (K) Seahorse Mito stress test–based graph showing the time-dependent change in the OCR for neuronal cells treated with control- and nicotine-treated iPNECs-derived exosomes. (L) Bar graph showing the decrease in maximal and basal respiration in neuronal cells treated with nicotine-treated iPNECs derived exosomes compared with the control (N = 3 biological replicates). In (A) to (C), (E), (F), (H), (J), and (L), data are presented as means ± SEM; n = 3 independent biological replicates. Technical replicates were averaged within each experiment. Statistics: one-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparisons test. Significance levels: *P < 0.05, **P < 0.01; vehicle control versus iPNEC-EXO, iPNEC-EXO versus Nic-iPNEC-EXO.

To assess the in vitro observations in mouse and human lung samples, hematoxylin and eosin (H&E) stain, and immunofluorescence staining were performed along with scRNA-seq data analysis. The scRNA-seq data (GEO accession: GSE122960) (51)–based Uniform Manifold Approximation and Projection (UMAP) analysis revealed an increased number of PNECs in the lungs of smokers compared with nonsmokers (Fig. 5A). Violin plots demonstrated elevated expression of TSG101 and FLOT1 (exosome biogenesis genes), as well as CP and HFE (iron homeostasis–related genes involved in transferrin/iron handling), in PNECs from smoking individuals relative to nonsmokers (Fig. 5, B and C). H&E staining showed thicker alveolar wall in smoking or nicotine-exposed lungs, likely due to inflammation and fibrosis. Moreover, the increased number of cells indicates the increased cellularity owing to smoking or nicotine exposure (Fig. 5, D and E). Immunofluorescence staining in human lungs (smoking group) and mouse lung (nicotine-treated group) showed an enhanced level of major iron transporters including TFR1, DMT1, and DCYTB (Fig. 5, F to H, and fig. S15, A to C). This was also accompanied by the increase in the level of calcium-sensing receptor, consistent with engagement of calcium-signaling components (fig. S16, A to D). Further, UMAP and violin plot analyses showed expression of TFRC, SLC11A2, and CYBRD1 in lung neuronal cells from both nonsmoking and smoking individuals, with higher levels observed in the smoking group (Fig. 5, I to N). These scRNA-seq data are consistent with an association between smoking/nicotine exposure and transcriptional changes related to iron handling and oxidative stress in lung neuronal populations in humans and mice. Consistently, markers associated with cognition, oxidative stress, and iron homeostasis were up-regulated in neuronal cells from smoking human lungs, as further supported by analyses within the publicly available scRNA-seq dataset (GSE122960) (fig. S17A to D) (51). Together, these findings support a model in which PNEC-derived exosomal signaling may modulate iron homeostasis–linked metabolic stress in lung neurons.

Fig. 5. Effect of smoke or nicotine on the iron homeostasis in neurons innervating PNEC in human and mouse lungs.

(A) UMAP shows the enhanced number of PNECs in the lungs of smoking individuals as compared with that in nonsmoking individuals. (B and C) Violin plots show the enhanced level of (B) TSG101 and FLOT1 (genes associated with exosome biogenesis), and (C) CP and HFE (genes associated with iron regulation), in the PNECs of smoking individuals as compared with that in nonsmoking individuals. (D and E) H&E staining of lung tissue from (D) nonsmoking versus smoking humans, and (E) WT control versus nicotine-injected mice. (Scale bar, 100 μm). (F to H) Immunofluorescence images and the quantitative bar graphs showing the costaining of proteins (F) DAPI, CGRP, TRPV1, and TFR1; (G) DAPI, CGRP, TRPV1, and DMT1; and (H) DAPI, CGRP, TRPV1, and DCYTB in the lung tissue of nonsmoking versus smoking humans (Scale bar, 50 μm). (I to N) UMAPs and violin plots show the enhanced expression level of [(I) and (L)] TFRC, [(J) and (M)] SLC11A2, [(K) and (N)] CYBRD1 in BDNF⁺ neuronal cells in the lung of nonsmoking and smoking subjects. In bar graphs in (D) to (H), data are presented as means ± SEM; n = 4 human donors and n = 6 mice. Multiple fields per donor or mouse were quantified as technical replicates and averaged to obtain one value per biological sample. Statistics: two-tailed unpaired Student’s t test. Significance levels: *P < 0.05, **P < 0.01; smoking versus nonsmoking.

TFR1 inhibition attenuates PNEC-EXO–induced oxidative stress and partially normalizes iron homeostasis

As noted above, serotransferrin binds to TFR1 and mediates the transport of iron. We examined the effect of blocking TFR1 on neuronal cells by using ferristatin-II, an iron uptake inhibitor, which interferes with TFR1-mediated iron delivery by inducing internalized degradation (52, 53). When neuronal cells were treated with exosomes from iPNECs exposed to nicotine along with ferristatin-II (Fig. 6A), the level of ferritin was significantly reduced (Fig. 6B), the level of ATP was significantly increased (Fig. 6C), and the ratio of GSH/GSSG was significantly increased (Fig. 6D), compared with the condition without ferristatin-II treatment. This change indicates that iron dyshomeostasis and oxidative stress induced in neuronal cells by exosomes from nicotine-exposed iPNECs can be alleviated by blocking the iron transporter TFR1, thereby restoring redox homeostasis.

Fig. 6. Blocking iron transport alleviates ferritin accumulation and oxidative stress in neuronal cells exposed to exosomes derived from iPNECs or SYN⁺ immunocaptured lung exosomes from mice.

(A) Experimental schematic showing neuronal treatment with nicotine-exposed iPNEC-derived exosomes, with or without ferristatin-II, an iron uptake inhibitor. (B to D) Bar graphs depict ferritin levels, ATP content, and GSH/GSSG ratios, respectively. (E and F) Outline showing immunocapture of SYN⁺ exosomes from ex vivo nicotine-treated mouse lung cultures, followed by neuronal exposure with or without ferristatin-II. (G to I) Data indicate ferristatin-II treatment decreases ferritin accumulation, increases ATP production, and raises GSH/GSSG ratios, reflecting reduced oxidative stress. (J) Neuronal cells, including those with TFR1 KD were treated with immunocaptured PNEC-derived exosomes. (K) Immunoblot verifies efficient TFR1 KD. (L to N) Bar graphs show reduced ferritin, elevated ATP, and increased GSH/GSSG ratios in TFR1-deficient neurons, confirming oxidative stress reduction. (O) Schematic illustrating neuronal exposure to exosomes from control iPNECs, GW4869-treated iPNECs (exosome biogenesis inhibitor), or iPNECs with TF KD. (P and Q) Immunoblots confirm GW4869 and TF KD effects on exosome protein content. (R to U) Quantification reveals reduced exosome release and mitigated ferritin, ATP, and oxidative stress markers in neurons treated with modified iPNEC exosomes. (V) Expression of Snca and NeuN in lungs of control and P301S tau transgenic mice, with (W) quantified Snca expression. (X) Spatial transcriptomics [STARmap PLUS (69)] indicate enhanced neuronal vulnerability and neurodegenerative gene signatures in P301S tau transgenic mouse brains. In (B) to (D), (G) to (I), (L) to (N), (R) to (U), and (W), data are presented as means ± SEM; n = 3 independent biological replicates. Technical replicates were averaged within each experiment. Statistics: two-tailed unpaired Student’s t test for two-group comparisons, and one-way ANOVA followed by Tukey’s multiple-comparisons test for comparisons involving more than two groups. Significance levels: *P < 0.05, **P < 0.01.

To further examine in vivo condition, we isolated exosomes from ex vivo–cultured lungs of C57BL/6 mice that had received nicotine intraperitoneally; these exosomes originate from a heterogeneous population of different lung cell types. Therefore, to obtain highly enriched exosomes from PNECs, SYN⁺ exosomes (SYN⁺ EXO) were immunocaptured and pooled from ex vivo lung cultures of mice, using anti-SYN antibody (Ab)–coated magnetic beads (Fig. 6E), and then eluted and used to treat neuronal cells without and with ferristatin-II (Fig. 6F). The immunocapture and enrichment of SYN⁺ exosomes was validated by AFM force-displacement analysis between exosomes immobilized on the AFM disc and the anti-SYN Ab–functionalized AFM tip, which demonstrated significantly increased amount of SYN⁺ exosomes after immunocapture (fig. S18, A to D). The level of ferritin was significantly reduced (Fig. 6G), the level of ATP was significantly increased (Fig. 6H), and the ratio of GSH/GSSG was significantly increased (Fig. 6I) in neuronal cells cotreated with SYN⁺ exosomes and ferristatin-II, compared with the condition without ferristatin-II treatment. A similar trend was obtained when the genetic ablation of TFR1 was performed (Fig. 6, J to N). To assess nonspecific effects unrelated to exosome exposure, we examined the ferritin accumulation in neuronal cells treated with ferristatin-II or shRNA alone, which did not show significant changes (fig. S18, E and F). Overall, iron dyshomeostasis and oxidative stress in neuronal cells by the immunocaptured SYN⁺ exosomes from mice, injected with nicotine, can be mitigated by blocking of iron transporter, TFR1. The inhibition of exosome biogenesis via GW4869, and the knockdown (KD) of TF were carried out in PNEC, followed by the evaluation of their effect on the release of EVs, amount of ferritin accumulation, and the production of ATP and oxidative stress (Fig. 6, O to U), which demonstrated the role of PNEC-derived exosomes carrying serotransferrin on the regulation of iron homeostasis in neurons.

To explore the in vivo relevance of the iron dyshomeostasis–associated oxidative-stress/vulnerability signature in lung neurons, we examined SNCA together with NeuN in lungs from WT and P301S mice. SNCA immunoreactivity was increased in lung neurons of P301S mice compared with WT controls (Fig. 6, V and W). In addition, spatial transcriptomic analyses of Alzheimer’s disease (AD) brain datasets revealed enrichment of iron-handling and oxidative stress–related gene signatures (Fig. 6X and fig. S19, A to C), along with spatial patterns of expression consistent with these programs (fig. S20, A to E). These observations align with the stress-associated neuronal signatures identified in our lung analyses.

DISCUSSION

PNECs are rare airway sensory epithelial cells that are innervated by neurons in the lung. They play a vital role in monitoring the airway environment and regulating airway neuron conduction. PNECs are thought to be involved in a variety of lung diseases, including asthma, chronic obstructive pulmonary disease (COPD), and bronchopulmonary dysplasia. A growing body of research is investigating the role of PNECs in lung health and disease. One study found that PNECs are more numerous in the lungs of patients with COPD than in healthy individuals (54). Given that smoking is a primary cause of COPD, this suggests a potential link between nicotine exposure and PNEC proliferation or activation. While much research has been done on PNECs, there is still much that we do not know about these cells. For example, it is not fully understood how PNECs sense changes in the airway environment and how they communicate with neuronal cells (24, 25, 55). Accordingly, it is pertinent to examine the distribution and pattern of proximity between PNEC and neuronal cells. Our investigation is consistent with previous findings, indicating that PNECs are mostly situated on the epithelial layer, either alone or in a group of several PNECs, which are also referred to as NEBs, in human and mouse lungs (25). PNECs lie in close proximity to and are innervated by neuronal cells (Fig. 1, A to F, and fig. S1, A and B), which was also found to be consistent with previous reports (25). Overall, PNECs are epithelial cells and innervated by neurons in the lung.

Previous reports as well as our observations suggest that PNECs are a rare cell type in the lung (Fig. 1, A to F) (24, 25, 37, 55). The limited availability and rare nature of PNECs has posed a challenge to research on these cells. In addition, modeling PNECs of human origin is daunting as PNECs are rare airway epithelial cells, and only a few methods have been demonstrated to grow and propagate the primary cells. Therefore, we generated PNECs via directed differentiation of hPSCs, namely iPNECs, using our previously reported technique (37). Briefly, hPSCs were efficiently differentiated into PNECs through a stepwise-directed differentiation process using differentiation mixtures I to VI. This process involves differentiating hPSCs into DE, AFE, and LP cells, followed by further differentiating the LP cells into PNECs using mixture VI and N-[(3,5-Difluorophenyl)acetyl]-L-alanyl-2-phenyl]glycine-1,1-dimethylethyl ester (DAPT) (Fig. 2, A to C, and fig. S2, A and C). Further, our scRNA-seq analysis showed its similarity to human native PNECs in terms of its functional biomarkers (Fig. 2, D to I, and fig. S3, A and B) and their applicability for use as a research tool (fig. S4, A to D). This technique offers an ample source of PNECs, which could be used for delineating the role of exosome-based communication between PNEC and neuron under nicotine exposure.

As the PNECs were found to lie in the proximity of neuronal cells, it is intriguing how they might interact with each other and what could be the potential reason behind the neuroinnervation of PNECs. Previous reports have found that PNECs interact with other lung structural cells, including lung-innervating neurons. Apparently, the definitive test of the role of neural innervation on PNEC specification and differentiation requires new genetic tools to comprehensively study the communication between PNECs and innervating neurons in both healthy and diseased states (24). Other groups have demonstrated that innervation of PNECs is required for the increased production of GABA, suggesting that stimulating efferent neurons that innervate PNECs may be a mechanism that can activate PNECs (56, 57). Overall, the communication between the PNECs and neurons in lung may be important. Recently, exosome-based intercellular communication has gained massive attention (32, 35); however it has not been investigated in the context of PNECs so far. In this study, we investigated the effect of nicotine, a major element of tobacco smoke, on the iPNECs, and observed that exosomes are released in higher amounts from PNECs treated with nicotine as compared with the control PNECs. Also, the exosomes were found to be of heterogeneous sizes, as analyzed using various techniques (Fig. 3, A to G), and exosomes were taken up by the neuronal cells (Fig. 3H). Furthermore, the PNEC-derived exosomes were found to be enriched with serotransferrin (Fig. 3, I to K), which plays crucial role in iron metabolism (58, 59). The expression of major iron transporters including TFR1, DMT1, and DCYTB was found to be enhanced in neuronal cells treated with exosomes from nicotine-treated iPNECs (Fig. 4, A to D) and were associated with increased ferritin levels and oxidative stress in neuronal cells. Subsequently, the levels of ATP and OCR were found to be reduced in neuronal cells treated with exosomes from nicotine-treated iPNECs (Fig. 4, E to L), suggesting the reprogramming of iron homeostasis and oxidative stress. Evidently, exosomes from nicotine-treated iPNECs are enriched predominantly with serotransferrin, which could trigger the ferritin accumulation and enhance oxidative stress in neuronal cells. This was further supported by human and mouse lung tissue samples (Fig. 5, A to N, and fig. S15, A to C). The pharmacological inhibition of TFR1 in neuronal cells decreases the PNEC-EXO–dependent oxidative stress, which also led to the restoration of iron homeostasis in neuronal cells (Fig. 6, A to D). Further, we intended to examine using mouse PNEC-EXO; however, it is challenging to isolate PNEC-specific exosomes from mouse. Here, we also isolated PNEC-enriched (SYN⁺) exosomes from mouse via ex vivo culture of lung, followed by their treatment to the neuronal cells without or with TFR1 inhibitor, or lenti-TFR1-shRNA (Fig. 6, E to N).

Iron dyshomeostasis can lead to oxidative stress and neurotoxicity, ultimately contributing to neurodegenerative diseases such as AD (60). In the context of COPD, scRNA-seq data have been analyzed to identify the cell types and mechanisms associated with the development of the disease, highlighting immunological dysregulations and oxidative stress as major factors (61, 62). Oxidative stress is a known driving mechanism of COPD, leading to chronic inflammation and cellular damage. The accumulation of iron and the generation of reactive oxygen species (ROS) due to abnormal iron transporter mechanisms have been linked to COPD (63). The link between iron dyshomeostasis, oxidative stress, and neurodegenerative diseases such as AD is further supported by the association of ferroptosis, an iron-dependent form of cell death, with AD and Parkinson’s disease (64). The role of iron dyshomeostasis–dependent oxidative stress in the pathology of these neurodegenerative diseases is also supported by our results of scRNA-seq and spatial transcriptomic analysis (Fig. 6, O to X, and figs. S19, A to C, and S20, A to E).

The present study puts forward a previously unknown way of studying exosome-based intercellular communication using hPSC-derived PNECs as well as integrating with magnetic immunocapture of cell type–specific exosomes. Overall, the nicotine-exposed PNECs release a significantly higher number of exosomes, enriched with serotransferrin. These PNEC-derived exosomes are internalized by the lung neuronal cells, which can induce iron dyshomeostasis. The levels of TFR1, DMT1, and DCYTB were elevated in neuronal cells exposed to PNEC-derived exosomes, consistent with increased iron uptake, ferritin accumulation, ATP depletion, and oxidative-stress responses. Figure S21 provides a conceptual working model that integrates these observations with prior literature; in this model, cell-stress and cell-death pathways (including apoptotic signaling) are depicted as potential downstream consequences (fig. S21, A to D). In addition, the altered iron balance—reflected by ferritin accumulation and transporter changes—may involve noncanonical extracellular vesicle (EV)–mediated modulation of iron regulatory protein and iron responsive element interactions, highlighting the need for future assessment of labile-iron pools and TF-loading states.

Although our in vitro inhibition and KD experiments support a causal role for the exosomal transferrin-TFR1 axis, the in vivo human and mouse transcriptomic and staining results remain associative; future in vivo EV-blocking and/or TF-depletion rescue experiments will be required to test causality directly. In the future, iPNECs could also be used to develop personalized treatments for patients with lung disease. By differentiating a patient’s own hPSCs into iPNECs, scientists could create an experimental model for the patient’s lung disease. This model could then be used to test different treatments and identify the one that is most likely to be effective.

MATERIALS AND METHODS

Ethics statement

All procedures were in accordance with the animal protocol (ACUP ID: 72747) approved by the Institutional Animal Care and Use Committee at the University of Chicago, Illinois, USA. Human lung tissue samples were obtained from BiocoreUSA Corp (IORG# IORG0012023) according to their approved Institutional Review Board (IRB) registration number IRB00014226, and OMB no. 0990-0279. The P301S tau transgenic mice–related experiments were conducted at the Hong Kong Baptist University following the approved animal protocol.

Generation of PNECs from hESC culture and differentiation

Culture and differentiation of human embryonic stem cells (hESCs) to produce PNECs were conducted as per our previously published protocol (23, 37). All embryonic stem cell studies were approved by the IRB at the University of Chicago. hESC line, RUES2 (National Institutes of Health approval no. NIHhESC-09-0013, registration no. 0013; passage 7–10) was cultured on irradiated mouse embryonic fibroblasts (Global Stem; catalog no. GSC-6001G) at a density of 20,000 to 25,000 cells per cm² in a medium of Dulbecco’s modified Eagle’s medium with nutrient mixture F-12 (DMEM/F12), 20% knockout serum replacement (Life Technologies), 0.1 mM β-mercaptoethanol (Sigma-Aldrich), and fibroblast growth factor–2 (FGF-2, 20 ng/ml; R&D Systems), and medium was changed daily. hESC cultures were maintained in an undifferentiated state at 37°C in a 5% CO₂/air environment until stem cells reached ∼90% confluence.

hESC differentiation into endoderm was performed in serum-free differentiation (SFD) media of DMEM/F12 (3:1; Life Technologies) supplemented with N2 (Life Technologies), B27 (Life Technologies), ascorbic acid (50 μg/ml; Sigma-Aldrich), glutamax (2 mM; Life Technologies), monothioglycerol (0.4 μM; Sigma-Aldrich), and 0.05% bovine serum albumin (BSA, Life Technologies) at 37°C in a 5% CO₂, 5% O₂, and 90% N₂ environment. hESCs were treated with Accutase (Stemcell Technologies) and plated onto low-attachment six-well plates (Corning Inc.), resuspended in endoderm induction media containing Y-27632 (10 μM; R&D Systems), human BMP4 (0.5 ng/ml; R&D Systems), human basic FGF (β-FGF, 2.5 ng/ml; R&D Systems), and human activin A (100 ng/ml; R&D Systems) for 72 to 84 hours dependent on the formation rates of endoderm cells. On day 3 or 3.5, the embryoid bodies were dissociated into single cells using 0.05% trypsin/0.02% EDTA and plated onto fibronectin (Sigma-Aldrich)–coated, 24-well tissue culture plates (∼100,000 to 150,000 cells per well). For induction of AFE, the endoderm cells were cultured in SFD medium supplemented with 1.5 μM dorsomorphin dihydrochloride (R&D Systems) and 10 μM SB431542 (R&D Systems) for 36 to 48 hours and then switched to 36 to 48 hours of 10 μM SB431542 and 1 μM IWP2 (R&D Systems) treatment.

For induction of early-stage LP cells (days 6 to 15), the resulting AFE was treated with CHIR99021 (3 μM; R&D Systems), human FGF10 (10 ng/ml; R&D Systems), human FGF7 (10 ng/ml; R&D Systems), human BMP4 (10 ng/ml; R&D Systems), and all-trans retinoic acid (ATRA; 50 to 60 nM; Sigma-Aldrich) in SFD medium for 8 to 10 days. The day-10 to -15 cultures were maintained in a 5% CO₂/air environment. On days 15 and 16, the LP cells were replated after brief 1-min trypsinization onto fibronectin-coated plates in the presence of SFD containing either a combination of five factors [CHIR99021, 3 μM; human FGF10 (10 ng/ml); human FGF7 (10 ng/ml); human BMP4 (10 ng/ml); and ATRA, 50 nM], or three factors [CHIR99021, 3 μM; human FGF10 (10 ng/ml); and human FGF7 (10 ng/ml)] for days 14 to 16. Day-16 to -25 cultures of late-stage LP cells were maintained in SFD media containing CHIR99021 (3 μM), human FGF10 (10 ng/ml), and human FGF7 (10 ng/ml) in a 5% CO₂/air environment.

For differentiation of mature lung cells (LCs) (days 25 to 55), cultures were replated after brief trypsinization onto 3.3% Matrigel-coated 24-well plates in SFD media containing maturation components, including 3 μM CHIR99021, human FGF10 (10 ng/ml), human FGF7 (10 ng/ml), and DCI {50 nM dexamethasone, 0.1 mM 8-bromo-cAMP (Sigma-Aldrich), and 0.1 mM IBMX [3,7-dihydro-1-methyl-3-(2-methylpropyl)-1H-purine-2,6-dione; (Sigma-Aldrich)]}. DAPT or dibenzazepine (5 to 10 μM; Sigma-Aldrich) was added to the maturation media for induction of PNECs.

Development of YFP reporter cells under CGRP promoter

The human CALCA promoter (HPRM32750) and EYFP were inserted into the pLenti-BSD vector, and the recombinant vector was validated by sequencing. The plasmid DNA was purified using the ZymoPURE II Plasmid Midiprep Kit and expanded into Stbl3 chemically competent Escherichia coli cells. This process yielded high-quality plasmid DNA suitable for lentiviral vector construction and subsequent viral production.

Lentiviral constructs were packaged in human embryonic kidney (HEK) 293T cells for efficient transduction of hESC. Briefly, HEK293T cells and hESCs were prepared in parallel. Lentiviral plasmids, along with appropriate packaging and envelope plasmids, were cotransfected into HEK293T cells using a lipid-based transfection reagent according to the manufacturer’s instructions. Culture medium was refreshed 24 hours posttransfection. Viral supernatants containing lentiviral particles were harvested at 48 and 72 hours posttransfection, clarified by centrifugation, filtered through a 0.45-μm filter to remove cellular debris, and either immediately used for hESC transduction or aliquoted and stored at −80°C.

For hESC transduction, lentivirus was diluted 1:1 with mTeSR1 medium and applied to hESC cultures, followed by incubation at 37°C for 24 hours. Subsequently, the virus-containing medium was carefully removed and discarded safely in 10% bleach, and fresh mTeSR1 medium was added. Cells were allowed to recover and integrate the viral constructs for 3 to 4 days before initiating positive antibiotic selection. Positive selection was performed for 5 to 7 days, during which cell cultures were closely monitored to prevent overconfluency, ensuring regular medium replacement as needed to maintain cell pluripotency. Following successful antibiotic selection and cell recovery, hESCs were either banked for future use, directed for differentiation, or subjected to further analysis.

FACS of YFP reporter cells under CGRP promoter from mixed lung culture

Day 55 to 65 are the best time for conducting experiments with hPSC-derived iPNECs. The medium supplemented with growth factors was changed every alternate day until the day of the experiment. The plate containing the cells was taken out of the 37°C incubator, and the medium was collected from each well into a 15-ml tube. The remaining medium was aspirated from the wells, and trypsin was added to just cover the bottom. The plate was incubated at 37°C for 8 min. The collected medium was added back to each well. A 1-ml pipette tip was used to scrape the cells, and the medium was pipetted up and down 15 to 20 times in each well to create a single-cell suspension. The cells were collected in the 15- or 50-ml tube and centrifuged at 400g for 3 min. The supernatant was carefully aspirated, and the cells were washed once with 1× phosphate-buffered saline (PBS) wash buffer. Cells were dissociated and filtered through a 35-μm strainer. YFP⁺ cells were sorted directly by native YFP fluorescence on a BD FACSAria with appropriate FSC/SSC singlet gating and YFP channel compensation at the flow cytometry core facility of the University of Chicago. The fluorescence image of YFP⁺ cells was acquired using Pseudo-green color for better visibility.

Nicotine treatment in iPNECs

iPNECs were treated with nicotine by supplementing the medium at the concentration of 1 μg/ml based on a previous study (65).

Proteomic analysis

Proteins from control- and nicotine-treated iPNECs samples were purified by acetone/trichloroacetic acid precipitation method. Then, the proteins were reduced, alkylated, and digested with trypsin according to an optimized protocol. Digested peptides were desalted on C18 columns and then subjected to mass spectrometry (MS) analysis. We searched data against a human database. Proteomic data were reported using visualization software called Scaffold Viewer, which is a scaled-down version of the complete Scaffold software. Proteomics analysis of iPNECs and nicotine-treated iPNECs samples was performed by the Northwestern University Proteomics Core Facility, generously supported by NCI CCSG P30 CA060553 awarded to the Robert H. Lurie Comprehensive Cancer Center, instrumentation award (S10OD025194) from NIH Office of Director, and the National Resource for Translational and Developmental Proteomics supported by P41 GM108569.

Metabolomic analysis

For metabolomic analysis, the iPNEC cells were washed with cold 0.9% NaCl and pelleted, and the supernatant was carefully aspirated. Cold 80% acetonitrile was then added at a defined cell density ratio (1000 to 5000 cells/μl) with a minimum volume of 40 μl, adjusting for any residual fluid from earlier steps. Cells were lysed by vortexing, and lysates were incubated at −20°C overnight to precipitate proteins. Samples were then centrifuged at 20,000g for 30 min at 4°C. The supernatant (containing metabolites) was transferred to liquid chromatography–MS (LC-MS) vials or clean tubes. All samples were stored at −80°C until LC-MS analysis, which was performed at the Metabolomics Core, Robert H. Lurie Comprehensive Cancer Center at Northwestern University.

Isolation of exosomes

Exosomes were isolated from control- and nicotine-treated iPNECs using a total exosome isolation reagent. Briefly, extracellular medium from cultured iPNECs was first centrifuged at 2000g for 30 min to remove cells and debris. The resultant supernatant was mixed with a half-volume of a total exosome isolation reagent and incubated at 4°C overnight, followed by centrifugation at 10,000g for 1 hour at 4°C. The supernatant was aspirated and discarded, and the exosome pellets were resuspended in 1 ml of cold PBS and repelleted under the same conditions. This PBS wash was repeated three times (≥3× pellet volumes per wash) to minimize potential carryover of residual nicotine or soluble factors. Vehicle/buffer controls that underwent the same precipitation and wash procedures were included in neuronal-cell assays. The final exosome pellet was resuspended in 1× PBS for downstream processing. Exosome inputs were normalized on the basis of protein content quantified using a bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, catalog no. 23225).

Analysis of exosomes by NTA

The number and size distribution of exosomes were characterized by nanoparticle tracking analysis (NTA) via a Malvern NanoSight NS300 instrument (66). Briefly, a monochromatic laser beam at 405 nm was applied to 200 μl of exosome solutions loaded into the sample chamber. Three video captures for exosome movements within a 30-s duration were recorded and further analyzed by NTA software (version 2.2, NanoSight) through the optimization for the identification and tracking of exosomes on a frame-by-frame basis. The number of exosomes released from iPNECs with various conditions was calculated by NTA analysis.

Analysis of exosomes by TEM and immunogold-EM

The size and morphology of exosomes were detected by TEM analysis (67). Briefly, exosomes were fixed with 2.5% paraformaldehyde (PFA) for 30 min, washed twice with PBS, dissolved in PBS/0.5% BSA, deposited onto formvar carbon–coated 200 mesh EM grids (Electron Microscopy Sciences, catalog no. 215-412-8400), and exposed for 10 min in a dry environment. Then, exosomes on the grids were washed five times (3 min each) with PBS/0.5% BSA. Afterward, fixed exosomes on the grid were washed twice with PBS/50 mM glycine (3 min each) and lastly with PBS/0.5% BSA (10 min), stained with 2% uranyl acetate for 5 min, and then viewed by an electron microscope (FEI/Philips Tecnai 12 BioTWIN at University of Chicago EM core facility).

For the immunogold labeling with Ab, fixed exosomes on the grid were incubated with 5% BSA for 30 min at room temperature, washed five times with PBS/0.5% BSA (3 min), transferred to a drop of the anti-CD63 or anti-Tf Ab (details in table S2) in PBS/0.5% BSA, and incubated for 2 hours at room temperature. Afterward, exosomes in the grids were washed five times with PBS/0.5% BSA (3 min), incubated with goat anti-rabbit IgG H&L Gold (10 nm) preadsorbed (Abcam, catalog no. ab270555) in PBS/0.5% BSA for 1 hour at room temperature, and then washed five times (3 min) with PBS/0.5% BSA. Last, exosomes on the grids were stained with UranyLess contrast EM solution and then viewed under an electron microscope.

Exosome uptake assay

Neuronal cells (approximately 30,000 cells per well) were cultured on the chamber slide (Lab-Tek, Thermo Fisher Scientific, USA) for 24 hours with normal growth medium at 37°C in a 5% CO₂ incubator. On the next day, the cells were washed twice with PBS and replenished again with normal growth medium supplemented with exosomes from control or nicotine-treated iPNECs (250 μg), with exosome preparations quantified by NTA to ensure comparable particle input across conditions. Exosomes were labeled with Exo-Green Exosome Protein Fluorescent Label (System Biosciences, catalog no. EXOG200A-1) (100 μl), washed to remove unbound dye, and further maintained for 24 hours. This exosome amount supported efficient neuronal uptake under our culture conditions and did not alter baseline neuronal morphology or cell density. Afterward, the neuronal cells were washed three times with PBS and fixed with 4% PFA on ice for 30 min. Fixed neuronal cells on the slide were stained with anti–β-III-tubulin Ab at room temperature for 1 hour to detect neuronal cytoskeleton, followed by washing with PBS. Then, stained neuronal cells with iPNEC-derived exosomes on the slide were mounted with Vectashield medium containing 4′,6-diamidino-2-phenylindole (DAPI) and observed under a laser scanning confocal microscope.

Neuronal cell culture and differentiation

HD10.6, a human DRG neuronal cell line, was used as an in vitro model of neuronal cells. The cells were cultured and differentiated as described previously (68). Undifferentiated HD10.6 cells were cultured on fibronectin-coated flasks in DMEM-F12 supplemented with growth factors and antibiotics [1× GlutaMAX, 1× B-27 supplement, prostaglandin E1 (10 ng/ml), β-FGF (0.5 ng/ml), and G418 solution (50 μg/ml)]. To induce differentiation, the cells were seeded onto plates coated with 2% Matrigel and poly-d-lysine (10 μg/ml). The medium was then changed to NeuralQ Basal Medium supplemented with 1× GlutaMAX; 1× B-27 Plus neuronal supplement; nerve growth factor (50 ng/ml); ciliary neurotrophic factor, glial cell line–derived neurotrophic factor, and NT-3 (25 ng/ml each); tetracycline (1 μg/ml), and 25 μM forskolin. The medium was changed every 3 to 4 days. The cells were maintained at 37°C in an incubator with 5% CO₂/air environment.

Genetic KD of TFR1 in neuronal cells

The lentiviral vector designated as pLKO.1-shTFRC (Addgene, catalog no. 169882), which encodes for the shRNA-targeting TFR1, was used for KD of TFR1 in neuronal cells. This specific construct was then meticulously cotransfected into HEK-293 T cells, in conjunction with lentiviral Helper plasmids, namely psPAX2 and pMD2G. Following the cotransfection process, lentivirus particles were generated within the HEK-293 T cells. Subsequently, these lentivirus particles were subjected to concentration using polyethylene glycol 6000. The resulting concentrated lentiviral preparation was then introduced to neuronal cells and incubated for a duration of 24 hours. Following this incubation period, the treated neuronal cells were cultured in a medium supplemented with puromycin (at a concentration of 5 μg/ml) for an additional 48 hours, ultimately yielding a population of stable, puromycin-resistant cells. The KD of TFR1 in neuronal cells was validated via Western blot. The control neuronal cells (with empty lenti-shRNA) and neuronal cells with TFR1 KD (lenti-TFR1-shRNA), were further treated with immunocaptured SYN⁺ exosomes for further study.

Single-cell dataset and analysis

scRNA-seq data (accession numbers SRR24392606–SRR24392629) were processed using the Seurat (v4.3.0) package in R (v4.2.2). After merging samples into a unified dataset, low-quality cells were removed on the basis of mitochondrial gene content >10% and hemoglobin gene content >3%. The data were normalized, and highly variable genes were identified. Batch correction and sample integration were performed using the Harmony (v0.1.1) algorithm, followed by dimensionality reduction with principal components analysis and t-SNE. Clustering was conducted using a graph-based approach to identify distinct cell populations. Cell type annotation was performed using canonical marker–based assignments and manual curation, complemented by reference-based label transfer using SingleR (v1.12.0) with a published human lung single-cell atlas as the reference. Clusters annotated as NE cells were further filtered to extract high-confidence PNECs, defined as cells with detectable expression of both CALCA and ASCL1. Gene expression patterns were visualized using t-SNE feature plots generated by ggplot2 (v3.4.2) and patchwork (v1.1.2), and the proportions of marker-positive cells (e.g., TUBB3, CHGA, and TRPV1) were quantified. Additional packages used included dplyr (v1.1.0) and Matrix (v1.5._4.1).

Spatial transcriptomic data analysis

The spatial transcriptomics of control and AD mouse brain were analyzed via STARmap PLUS sequencing data (Single Cell Portal of Broad Institute, Study# SCP1375) and Zenodo (DOI: 10.5281/zenodo.7332091) (69). The spatial transcriptomics of the brain of control and patient with AD were analyzed via STOmics of the brain tissue of control (dataset ID: STDS0000242) (70) and patient with AD (dataset ID: STDS0000188) (71).

Animal

All experiments were done with the approval of the University of Chicago’s Institutional Animal Care and Use Committee. Procedures also followed the guidelines for the care and use of animals provided by the National Institutes of Health Office of Laboratory Animal Welfare. All efforts were made to minimize animal distress and suffering during experimental procedures. Adult male (6 to 8 weeks old) C57BL/6 mice (the Jackson Laboratories or bred in-house) were group housed in a colony room (at 22°C) on a standard 12-hour light-dark cycle. Upon arrival, the mice were undisturbed for at least 72 hours to allow acclimatization to the colony room. Water and food were available ad libitum, and cages were changed twice per week. Experiments were conducted during the animals’ light period. P301S tau transgenic mice, purchased from the Jackson Laboratory, were handled at the Hong Kong Baptist University as per protocols approved by the Committee on the Use of Human and Animal Subjects in Teaching and Research.

Nicotine treatment in mice

Adult male C57BL/6 mice (6 to 8 weeks old) were randomly assigned to each group (n = 6 per group). Mice in the experimental group received intraperitoneal nicotine (0.1 mg/kg body weight) every alternate day for 2 weeks, while control mice received vehicle injections. Behavioral testing and outcome assessments were performed by investigators blinded to treatment allocation.

Immunolabeling, 3D iDISCO tissue clearing, and light-sheet microscopy

Lung lobes were immunolabeled and cleared using the 3D iDISCO protocol as described previously (72). Briefly, for immunolabeling, the lung samples were incubated in a solution of PBS/0.2% Triton X-100/20% dimethyl sulfoxide (DMSO)/0.3 M glycine at 37°C overnight, followed by blocking in PBS/0.2% Triton X-100/10% DMSO/6% donkey serum at the temperature of 37°C. Lung samples were washed in PBS/0.2% Tween-20 with heparin (10 μg/ml) (PTwH) for 1 hour, two times, and then incubated in primary Ab dilutions (ratio of 1: 50) in PTwH/5% DMSO/3% donkey serum at 37°C for 3 days. Samples were then washed in PTwH for 1 day and then incubated in secondary Ab dilutions (ratio of 1: 100) in PTwH/3% donkey serum at 37°C for indicated time. Samples were finally washed in PTwH for 2 days before clearing and imaging.

For tissue clearing, the lung samples were incubated overnight in 10 ml of 50% v/v tetrahydrofuran (THF)/H₂O (Sigma-Aldrich, catalog no. 186562). The samples were then incubated for 1 hour in 10 ml of 80% THF/H₂O and twice for 1 hour in 100% THF and then in dichloromethane (Sigma-Aldrich, catalog no. 270997) until they sank to the bottom of the vial. Last, the samples were incubated in 18 ml of dibenzyl ether (DBE) (Sigma-Aldrich, catalog no. 108014) until clear (∼2 hours) and then stored in DBE at room temperature. For imaging, lung samples were imaged on a LaVision light-sheet microscope (available at the Imaging Core Facility of Biological Science Division at the University of Chicago) equipped with an sCMOS camera.

Human sample

Lung tissue slides from four healthy controls (nonsmokers), four smoking patients, and four patients with AD were purchased from BioCoreUSA Corp, Philadelphia, PA. The patients’ treatment history and background of diseases are presented in table S1.

Immunofluorescence staining

Living cells in culture were directly fixed in 4% PFA for 25 min, followed by permeabilization in 1% Triton X-100 for 15 min, and immunostaining. Histology was performed on paraffin-embedded or frozen sections of mouse lung tissues and P301S mouse brain tissues using standard tissue-processing procedures as previously described (37, 72). Tissues were either fixed overnight in 10% buffered formalin and transferred to 70% ethanol, followed by paraffin embedding, or snap-frozen in optimal cutting temperature compound (Thermo Fisher Scientific) and fixed in 10% buffered formalin, followed by paraffin embedding. For immunofluorescence, cells or tissue sections were immunostained with antibodies and counterstained with DAPI. Adjacent sections stained with H&E were used for comparison. Noninnervated and innervated PNECs were counted from each immunofluorescence image based on the presence or absence of colocalization of CGRP with neuronal markers, respectively (two fields per sample). Two nonoverlapping fields were acquired per sample as technical replicates, and field-level measurements were averaged to obtain a single value per donor or mouse for statistical analysis. Innervation scoring was performed manually by two independent blinded raters. A PNEC was classified as “innervated” when the CGRP⁺ signal showed direct contact or clear colocalization with TUBB3⁺ or TRPV1⁺ neuronal fibers; otherwise, it was classified as “noninnervated.” Discrepancies were resolved by consensus. All fluorescence images were acquired using a Leica DMi8 Inverted Microscope at 40× magnification. Image processing and analysis were carried out using ImageJ or Adobe Photoshop CS6. Similarly, the lung tissue sections of P301S mice (73) were analyzed to determine the expression level of SNCA and NeuN. The details of antibodies used here are mentioned in table S2. The immunofluorescence images were quantified by ImageJ-based corrected total cell fluorescence analysis by subtracting the product of the area of the selected cell and the mean fluorescence of the background from the integrated density.

Western blotting

Neuronal cells (control and with different treatments) were harvested and subjected to lysis using radioimmunoprecipitation assay lysis buffer, supplemented with phenylmethylsulfonyl fluoride and a cocktail of protease and phosphatase inhibitors. Following lysis, the concentration within the cell lysates was quantified using the BCA protein assay kit. Subsequently, 20 μg of aliquots of the protein were loaded onto a 4 to 12% SDS–polyacrylamide gel electrophoresis precast gel, followed by a separation process conducted at 80 V over a span of 3 hours, using MOPS SDS Running buffer. Electrophoretic transfer of proteins onto polyvinylidene difluoride membranes ensued at 180 mA for a duration of 1.5 hours, post which the membranes were blocked using 4% BSA solution at ambient temperature for an hour. This was followed by overnight incubation with primary antibodies at 4°C. The details of specific antibodies used here are mentioned in the table S2. Following several washes with 1× TBST buffer, the membranes were exposed to horseradish peroxidase–conjugated goat anti-rabbit or anti-mouse antibodies (diluted 1:5000) for 2 hours at room temperature. Detection of signals was carried out with the SuperSignal West Femto Maximum Sensitivity Substrate using Azure Biosystems Gel Documentation system (C600). The bands of target proteins were quantified via ImageJ analysis based on the intensity, normalized with β-actin control.

Ex vivo culture of mouse lungs and exosome isolation

Mice were anesthetized and euthanized, followed by collection of the lungs, which were placed in DMEM/F-12 supplemented with penicillin (100 U/ml) and streptomycin (100 μg/ml). The ex vivo culture was maintained in the medium for 1 day, followed by replacing the normal medium with exosome-depleted medium, and maintained at 37° C with 5% CO₂, for further downstream analysis.

Enrichment of SYN⁺ exosomes via magnetic bead-based immunocapture

Magnetic beads were conjugated with anti-SYN Ab as described previously (74). After incubation of anti-SYN Ab–conjugated magnetic beads (100 μl, bead diameter 4.5 μm, 1 × 10⁷ beads/ml) in 500 μl of isolation buffer (0.1% BSA in PBS), conglomerates of magnetic beads were formed and concentrated using a magnetic rack for 1 min. Thereafter, a 100-μl ex vivo lung culture supernatant sample was added and mixed on a thermomixer at 4°C for 18 hours. Subsequently, the magnetic rack was again used to concentrate EVs captured on the magnetic beads, and the supernatant was removed without disturbing the beads. Two final washings were completed with isolation buffer before either lysis directly for molecular content analysis or elution for whole-particle assays. EVs bound to magnetic beads were eluted by incubation with 100 μl of PBS containing 1% formic acid for 10 to 30 min. The neutralization/buffer-exchange steps were used to eliminate residual acid before functional experiments. Further, the eluted EVs were prepared for TEM or diluted 100 times in PBS for NTA analysis.

Calcium assay

For calcium assay, HD10.6 neuronal cells were allowed to grow to 90% confluence in a 96-well plate. After washing them with Hanks’ balanced salt solution (HBSS) buffer, neuronal cells were loaded with the Ca²⁺ indicator, 50 μl of 4 μM Fluo-4 AM (Invitrogen, catalog no. F14201), in HBSS buffer. Next, the neuronal cells were incubated at room temperature for 1 hour and kept under dark conditions. After washing them with HBSS buffer, the continuous measurement of fluorescence kinetics was performed with or without exosome treatment (excitation: 485 nm, emission: 525 nm) in a microplate reader.

FluoVolt assay

Electrical activity was recorded using the FluoVolt voltage-sensitive dye kit. A single vial containing the dye was diluted in 100 μl of PowerLoad (a solubilizing agent provided in the FluoVolt kit) and then resuspended in 10 ml of experimental medium warmed at 37°C according to the manufacturer’s protocol. The background suppressor furnished in the kit was not added because our conditions did not require it. After removal of the culture medium and two washes with 100 μl of experimental medium, 50 μl of each dye was applied on the cells in distinct wells for 15 min at 37°C. After two washouts of the dyes with 100 μl of experimental medium supplemented with Ctrl-iPNEC-EXO or Nic-iPNEC-EXO, recordings were carried out with the microplate reader.

ATP assay

Total ATP production was measured through an ATP assay kit (Abcam, catalog no. ab83355). Briefly, neuronal cells were harvested, washed with PBS, and then resuspended in 100 μl of ATP assay buffer. Cells were homogenized and then centrifuged (4°C at 13,000g) to remove any insoluble material. The supernatants were then collected and incubated with the ATP probe. Absorbance was detected at 580 nm using a microplate reader. The results are presented as the ratio between the test and control values.

Atomic force microscopy

Biosensing single molecular interaction between surface antigens of immobilized exosomes on the AFM disc and anti-SYN Ab functionalized in the sensing tip was conducted by using fluidic force microscopy. The spring constant of AFM silicon nitride (Si₃N₄) tip was calibrated for the detection of exosomal proteins. Exosomes were immobilized on the mica-based AFM disc as described previously (34, 74). Two hundred microliters of exosome solution in PBS (50 μg/ml) was first added to the sample discs, incubated for 10 min, and replaced with 1 ml of fresh PBS by mild decantation. Immobilized exosomes on the surface of the AFM disc were further confirmed and analyzed by AFM scanning.

To determine exosomal SYN levels by the measurement of intermolecular force between antigens and Abs, the Si₃N₄ AFM tip was functionalized with anti-SYN Ab. Briefly, primary anti-SYN Ab (dilution 1:50) was covalently attached to the Si₃N₄ tip of AFM via thiol ester linkage. The probe tip was washed with PBS, incubated in blocking solution (1% BSA-PBS) for 1 hour, followed by a series of washes with PBS.

All measurements of exosomal proteins with AFM were recorded in PBS. Separation forces between exosomal SYN and anti-SYN functionalized on the sensing tips were measured by AFM ramp mode. Exosomal SYN level was determined and analyzed by the maximum peak of the AFM force-distance curve. Each AFM force curve was obtained from at least three independent experiments.

Oxidative stress test

Intracellular oxidative stress was determined via ROS production measurement using a CellROX reagent (Invitrogen, catalog no. C10422) according to the manufacturer’s guidelines, followed by fluorescence microscopy.

Glutathione redox (GSH/GSSG) assay

GSH/GSSG ratio detection assay kit (Fluorometric-Green, Abcam, catalog no. ab138881) was used to determine the GSH/GSSG ratio in neuronal cells according to the manufacturer’s protocol. Briefly, the whole-cell lysates were diluted to 1:80 for GSH analysis, and a serial dilution of GSH and GSSG stock standards were prepared as standards. A one-step fluorometric reaction of samples with respective assay buffer and probes was incubated for 1 hour protected from light at room temperature. Then, fluorescence intensity was monitored at EX/EM of 490/520 nm. GSH was calculated from the standard curve, and GSSG = (total glutathione − GSH)/2.

OCR and Mito-stress test by Seahorse XFe analyzer

OCR was measured using the Seahorse XFe Analyzer (Agilent Technologies). Briefly, at 48 hours before the assay, neuronal cells (4 × 10⁴ cells per well) were seeded in three wells per group in a Seahorse plate and incubated at 37°C with 5% CO₂. Then, 24 hours before the assay, one group of neuronal cells were treated with an equal amount of iPNEC-EXO, and the other group was treated with Nic-iPNEC-EXO. On the day of assay, the cells were equilibrated with prewarmed Seahorse XF assay medium [DMEM with 10 mM glucose, 2 mM l-glutamine, and 1 mM sodium pyruvate without NaHCO₃ (pH 7.4)] at 37°C for 1 hour. The OCR of the cellular monolayer was measured before (basal level) and after the sequential injection of 1 μM oligomycin (blocks proton translocation through complex F0/F1, inhibiting ATP production, and reducing OCR), 1 μM carbonyl cyanide p-trifluoromethoxy phenylhydrazone (a potent uncoupler of mitochondrial oxidative phosphorylation), and 0.5 μM rotenone and 0.5 μM antimycin A (shut down the mitochondrial respiration inhibiting complex I and III, respectively, and reducing OCR to a minimal value). The cellular bioenergetic parameters determined were basal OCR, ATP-linked OCR, proton leak, maximal OCR, and spare capacity.

ELISA for ferritin

The level of ferritin was measured in neuronal cells with the human ferritin ELISA kit (Abcam, catalog no. ab108698)

Mouse behavior tests

NOR test

The C57BL/6 mice with or without nicotine treatment were evaluated for memory function and recognition memory using an NOR test. Before the test, the animals were given 1 day to familiarize themselves with the test environment by freely exploring the box. On day 1, the mice were trained on two known objects placed at an equal distance from the center. On day 2, one of the objects was replaced with a novel object of a different shape. The animal’s exploring behavior on the novel object was then recorded for 5 to 10 min using an online ANYMAZE real-time video tracking system (Stoelting). The recognition index percentage was calculated with ANYMAZE software (Stoelting Co., Wood Dale, IL, USA). Similarly, the NOR was conducted with mice injected with control-iPNEC-EXO or Nic-iPNEC-EXO.

Open field test

To assess locomotor and exploratory behavior in mice with or without nicotine treatment, an open field test was conducted. In brief, each mouse was placed in the center of an open field arena for an observation period of 5 min. Between trials, the testing area was thoroughly sanitized with ethanol to eliminate olfactory cues. ANYMAZE software from Stoelting was used to track and analyze various behavioral parameters. Locomotion was quantified by measuring the total distance traveled as well as the frequency of center and wall rears. To evaluate exploratory and anxiety-related behaviors, the time spent by the mice in the central area was measured.

Barnes maze test

The Barnes maze test was conducted by placing a circular board on the floor with 20 equally spaced holes in it, which was elevated off the ground. An escape box (goals) was placed beneath one random hole in the board, with the mice being placed in the center of the board. The room lights were turned off, and the mice were allowed to freely explore the board and look for the escape hole. A bright light was placed over the board to aid the mice in locating the correct hole, after which the latency (time taken) to escape and the number of errors the mice made in attempting to find the escape hole were recorded.

Mesh test

The mice with or without nicotine treatment were placed in the center of a mesh apparatus, and its exploratory behavior was observed. The number of plane and vertical head dips, the time spent in each one of the chambers of the apparatus and the latency to enter an open chamber were recorded. The distance traveled, the amount of time spent in the center of the mesh, the time spent traversing the walls of the mesh, the number of trips among chambers, and the number of attempts at escaping from the apparatus were measured. The results were analyzed; any changes such as latency to exit the apparatus, the distance traveled, and the number of attempts to overtake the mesh indicated anxiety-like behavior.

Statistical analysis

Quantitative data are presented as means ± SEM unless indicated otherwise. In all experiments, n denotes biological replicates (independent human donors, individual mice, or independent cell-culture experiments performed on different days). Technical replicates (e.g., multiple imaging fields/images, multiple cells per image, replicate wells, or NTA video captures) were averaged to generate one value per biological replicate before statistical testing. For two-group comparisons, two-tailed unpaired Student’s t test was used; for comparisons involving more than two groups, one-way analysis of variance (ANOVA) followed by an appropriate multiple-comparisons test was used. Exact statistical tests and multiple-comparison procedures are specified in the corresponding figure legends. Statistical significance was defined as P < 0.05.

Supplementary Materials

The PDF file includes:

Figs. S1 to S27

Table S2

Legend for table S1

Legend for movie S1

Other Supplementary Material for this manuscript includes the following:

Table S1

Movie S1