Decision letter: Neuroendocrinology of the lung revealed by single-cell RNA sequencing

Article Figures and data Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Pulmonary neuroendocrine cells (PNECs) are sensory epithelial cells that transmit airway status to the brain via sensory neurons and locally via calcitonin gene-related peptide (CGRP) and γ- aminobutyric acid (GABA). Several other neuropeptides and neurotransmitters have been detected in various species, but the number, targets, functions, and conservation of PNEC signals are largely unknown. We used scRNAseq to profile hundreds of the rare mouse and human PNECs. This revealed over 40 PNEC neuropeptide and peptide hormone genes, most cells expressing unique combinations of 5–18 genes. Peptides are packaged in separate vesicles, their release presumably regulated by the distinct, multimodal combinations of sensors we show are expressed by each PNEC. Expression of the peptide receptors predicts an array of local cell targets, and we show the new PNEC signal angiotensin directly activates one subtype of innervating sensory neuron. Many signals lack lung targets so may have endocrine activity like those of PNEC-derived carcinoid tumors. PNECs are an extraordinarily rich and diverse signaling hub rivaling the enteroendocrine system. Editor's evaluation The authors present a comprehensive profile of signals and sensors expressed in mouse and human PNECs by single-cell RNA sequencing. Analyses revealed a myriad combination of neuropeptide, neurotransmitter, receptor and channel genes in PNECs. The authors also surveyed cognate receptors expressed in epithelial cells, endothelial cells, stromal cells, immune cells, and pulmonary sensory neurons, identifying potential local targets for the PNECs signals. They showed one new signal, angiotensin II, directly activates a subpopulation of innervating pulmonary sensory neurons that project to the brain. The scRNA-seq profile of a lung carcinoid tumor suggests that selected PNECs are susceptible to carcinoid transformation and together, these data indicate that PNECs serve as sentinels to perceive multiple airway stimuli and express a variety of signals that either act locally, through pulmonary sensory neurons, or potentially through circulation to regulate homeostasis. https://doi.org/10.7554/eLife.78216.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Pulmonary neuroendocrine cells (PNECs) are neuroepithelial cells scattered throughout the epithelium lining the airways, many as solitary cells and others in clusters of ~30 cells (called neuroepithelial bodies or NEBs) at airway branchpoints (Scheuermann, 1997). PNECs are thought to monitor airway oxygen and respiratory status, and rapidly signal this internal sensory information via afferent sensory neurons to the brain to regulate breathing (Garg et al., 2019; Nonomura et al., 2017). Such signaling may be mediated by the classical neurotransmitter serotonin (Lauweryns et al., 1973), but PNECs also produce γ-aminobutyric acid (GABA) (Schnorbusch et al., 2013) and several neuropeptides including calcitonin gene-related peptide (CGRP), which can signal locally to lung goblet (via GABA, γ-aminobutyric acid) and immune cells (via CGRP, calcitonin gene-related peptide) and may contribute to asthma (Barrios et al., 2019; Sui et al., 2018). PNECs also function as reserve stem cells that repair the surrounding epithelium after injury (Ouadah et al., 2019; Song et al., 2012; Stevens et al., 1997), and they can be transformed by loss of tumor suppressor genes into the deadliest human lung cancer, small cell lung cancer (SCLC) (Ouadah et al., 2019; Park et al., 2011; Song et al., 2012; Sutherland et al., 2011), and likely other neuroendocrine tumors such as lung carcinoids accompanied by systemic symptoms from signals secreted by the tumor (Davila et al., 1993). Despite the physiological and clinical significance of PNECs, molecular interrogation and understanding of PNEC function and diversity (Mou et al., 2021) has lagged because the cells are so rare, comprising just ~0.01% of human lung cells (Boers et al., 1996; Travaglini et al., 2020). Here we describe the isolation, expression profiling by single-cell RNA sequencing (scRNA-seq), and analysis of hundreds of mouse and human PNECs. The results reveal an extraordinary molecular diversity of these cells, including the expressed sensors along with dozens of neuropeptides and peptide hormones whose predicted targets indicate they transmit airway sensory information throughout the lung, to the brain, and potentially to the rest of the body. Results Enrichment and single-cell RNA sequencing of mouse pulmonary neuroendocrine cells Because PNECs are among the rarest of lung cell types, they are not found, or only poorly represented, in lung scRNA-seq studies (Han et al., 2018; Tabula Muris Consortium et al., 2018; Travaglini et al., 2020, https://www.lungmap.net/). We therefore genetically labeled PNECs using a Cre recombinase-dependent fluorescent reporter (ZsGreen) by tamoxifen induction of Ascl1CreER/+; Rosa26LSL-ZsGreen/+ mice, and then depleted other, abundant lung cell populations and enriched for the labeled PNECs prior to scRNA-seq (Figure 1A, Figure 1—figure supplement 1A and B). Poly-adenylated mRNA from each sorted ZsGreen + or control cell was reverse transcribed and PCR-amplified using Smart-seq2 protocol (Picelli et al., 2014), and the obtained cDNA was used to generate libraries sequenced to a depth of 105–106 reads/cell and quantified to determine expression levels of each gene in each cell. Cells with similar expression profiles were computationally clustered (Butler et al., 2018), and the cell type identity of each cluster was assigned based on expression of canonical lung cell type markers including Ascl1, Calca, and Chga for PNECs (Figure 1—figure supplement 1C). After filtering low quality cells and cell doublets, we obtained high quality transcriptomes (2025±450 genes (mean ± SD) detected per cell) of 176 PNECs. We also obtained 358 other pulmonary cells (Figure 1—figure supplement 1D) including Ascl1 lineage-labeled glial cells. Figure 1 with 1 supplement see all Download asset Open asset Single cell RNA sequencing of mouse PNECs reveals expression of dozens of peptidergic genes. (A) Strategy for labeling, enrichment, and scRNA-seq of exceedingly rare PNECs. Timeline (top) of tamoxifen (Tam) injections (gray arrowheads) of Ascl1CreER/+;Rosa26LSL-ZsGreen/+ mice beginning at embryonic day (E) 13 - E14 to permanently induce ZsGreen expression in pulmonary neuroendocrine cells (PNECs or NE cells). Lungs were dissected at indicated ages (black arrowheads, postnatal day (PN) 21, PN90, and PN120) and mechanically and enzymatically dissociated (dissoc.) into single cells. Endothelial (CD31+) and immune (CD45+) cells were depleted by magnetic-cell sorting (MACS) then PNECs enriched by fluorescence-activated cell sorting (FACS) EpCAM+/Zsgreen+ double-positive cells. Sorted cells were analyzed by plate-based scRNA-seq using SmartSeq2 protocol. (B) Most sensitive and specific PNEC markers identified by scRNA-seq, ranked by ratio of the natural logs of the average expression (ln (counts per million, CPM + 1)) of the marker in PNECs (NE cells) vs. non-PNEC (non-NE) lung epithelial cells in mouse lung cell atlas (Travaglini et al., 2020). *, previously known PNEC marker. (C) Violin plots showing expression of five new markers (Resp18, Pcsk1, Scg5, Chgb, Sez6l2) and three previously known markers (*; Calca, Syp, Chga) across 40 cell types from mouse lung cell atlas (Travaglini et al., 2020). From left to right (x-axis): (1) neuroendocrine (NE, PNEC), (2) club, (3) multiciliated, (4) basal, (5) goblet, (6) alveolar type 1, (7) alveolar type 2, (8) glial, (9) smooth muscle, (10) myofibroblast, (11) adventitial fibroblast, (12) alveolar fibroblast, (13) pericyte, (14) mesothelial, (15) chondrocyte, (16) artery, (17) vein, (18) capillary aerocyte, (19) capillary-general, (20) lymphatic, (21) B cells, (22) Zbtb32+ Bcells, (23) plasma, (24) CD8+ T, (25) CD4+ T, (26) regulatory T, (27) Ly6g5bt + T, (28) natural killer, (29) Alox5+ lymphocyte, (30) neutrophil, (31) basophil, (32) alveolar macrophage, (33) interstitial macrophage, (34) plasmacytoid dendritic, (35) myeloid dendritic type 1, (36) myeloid dendritic type 2, (37) Ccr7+ dendritic, (38) classic monocyte, (39) nonclassical monocyte, (40) intermediate monocyte. (D) Close-up of neuroepithelial body (NEB) in PN155 wild type (C57BL/6NJ) mouse lung probed by multiplex single molecule RNA fluorescence in situ hybridization (smFISH) to detect expression of indicated PNEC markers, with DAPI nuclear counterstain. Dashed circles, individual PNECs (numbered); dashed line (basement membrane). Scale bar, 10 μm. Quantification (right) of clustered PNECs that express indicated markers (n=76 cells scored in left lobe and right lower lobe). Note classic marker Calca (CGRP) was not detected in 6% of Resp18+Scg5+double-positive PNECs. (E) Quantification of peptidergic gene expression in PNECs by scRNA-seq. Bars show percent of profiled PNECs (NE cells, n=176) with detected expression of the 43 peptidergic genes indicated; values above bars are log-transformed mean gene expression (ln (CPM + 1)) among expressing cells. Black dots, expression values from a second PNEC dataset (filled circles, n=92 PNECs) in which PNECs were genetically labeled using CgrpCreER;Rosa26LSL-ZsGreen mice, sorted, and isolated on a microfluidic platform (Ouadah et al., 2019). (Comparison by Fisher’s exact test (two-tailed) of the proportions of PNECs detected in the two scRNA-seq datasets is provided in Supplementary file 4, with caveat that the comparison is of results from different techniques on different samples.) *, Previously known mouse PNEC peptidergic genes; blue highlight, classic hormone genes. (F) Micrograph of NEB from PN90 Ascl1CreER/+;Rosa26LSL-ZsGreen/+ mouse lung immunostained for CGRP and newly identified PNEC neuropeptide CARTPT. White arrowheads, CGRP+ PNECs; red arrowheads, CARTPT+ PNECs; *, CGRP- CARTPT- double-negative PNEC. Right panel, quantification of CGRP and CARTPT staining in PNECs defined by Ascl1-CreER-lineage label (n=237 PNECs scored in three PN60 Ascl1CreER/+;Rosa26LSL-ZsGreen/+ mice). (G) Quantification of PNEC (NE cell) expression of the indicated peptidergic genes by scRNA-seq (black bars, n=176 cells) and multiplex smFISH (grey bars and dashed extensions, n=100cells scored in NEBs from 2 mice, see Figure 2). Grey bars, cells with high expression (>5 puncta/cell); dashed extensions, cells with low expression (1–4 puncta/cell). Fisher’s exact test (two-tailed) gave p=1 (not significant) for all comparisons of proportions of expressing PNECs for each gene as detected by smFISH (>5 puncta/cell) vs. scRNA-seq (black bars); when the comparisons included cells with 1–4 puncta/cell by smFISH, differences were significant (p<0.05) for Chga, Cartpt, Agt, Pomc, Nmb, and Adcyap1 but not for Scg5 and Calca (p=0.9 for both). Figure 1—source data 1 Corresponds to Figure 1G. https://cdn.elifesciences.org/articles/78216/elife-78216-fig1-data1-v1.xlsx Download elife-78216-fig1-data1-v1.xlsx We identified dozens of mouse PNEC markers (Figure 1B and C, Supplementary file 1) beyond the canonical four (Calca (encoding calcitonin gene-related peptide CGRP), Ascl1, Syp (synaptophysin), and Chga (chromogranin A)) (Supplementary file 2) using Wilcoxon rank sum (Seurat v2.3.4) to find differentially expressed genes selectively expressed by PNECs, then prioritizing those with high sensitivity (expressed in >85% of PNECs), specificity (<5% of the other lung cell types), and fold-difference in mean expression over other lung cell types. Many of the identified markers were expressed at higher levels (Resp18, Pcsk1, Scg5) and were more sensitive (Resp18, Pcsk1) and specific (Chgb, Sez6l2) than the canonical PNEC markers (Figure 1C). Among the top reported tracheal neuroendocrine cell (TNEC) markers, just over half (54% (19/35) Montoro et al., 2018; 63% (31/49) Plasschaert et al., 2018 were also top PNEC markers, defining a set of shared airway neuroendocrine markers (Supplementary file 1). But there were also notable differences between TNECs and PNECs (e.g.) TNEC-selective Cib3, Snca, Cxcl13, Mthfd2) and PNEC-selective (Ptprn, Piezo2, Igf2, Meg3) (Figure 1—figure supplement 1F, Supplementary file 1), consistent with their different development, distribution, innervation, and functions (Kuo and Krasnow, 2015; Montoro et al., 2018; Mou et al., 2021; Ouadah et al., 2019; Song et al., 2012). Selective expression of the most robust PNEC markers was validated by single molecule in situ hybridization (smFISH) (Resp18, Scg5, Figure 1D) and immunohistochemistry (Pcsk1, see below). PNECs express dozens of neuropeptide and peptide hormone genes PNECs were discovered by their hallmark secretory vesicles (Scheuermann, 1997) and later shown to secrete serotonin (Lauweryns et al., 1972), GABA (Barrios et al., 2017), and several neuropeptides including CGRP (Johnson et al., 1988; Scheuermann et al., 1987) and chromogranin A (Lauweryns et al., 1987; Supplementary file 2). To determine the full set of PNEC neurotransmitters and neuropeptides, we searched our scRNA-seq dataset for neurotransmitter biogenesis genes (Supplementary file 3) and neuropeptide and peptide hormone genes ("peptidergic" genes, Supplementary file 4) expressed in PNECs. This confirmed that PNECs are GABAergic because many (28%) expressed biosynthetic gene Gad1. Although only rare (1%) PNECs express the serotonin biosynthetic gene Tph1 and none expressed Tph2, many (24%) expressed the reuptake transporter Slc6a4, suggesting they can modulate serotonin signaling through serotonin imported from other sources such as circulating serotonin (Herr et al., 2017). Some PNECs may also use other neurotransmitters because rare cells expressed key dopaminergic (Th, Ddc), glutamatergic (Slc17a6), cholinergic (Slc18a3/VAChT), or histaminergic (Hdc) genes (Supplementary file 3). PNECs expressed dozens of genes that encode neuropeptides and/or peptide hormones. Thirty-one peptidergic genes were expressed in our dataset, and 30 of those plus an additional 12 were found in our second, independent dataset (see below), totaling 43 (47%) of the 91 peptidergic genes in mice (Figure 1E, Supplementary file 4). The full set includes previously known Calca (encodes CGRP) and Chga but also some of the most biologically and medically significant neuropeptide and peptide hormone genes (e.g. Pomc (pro-opiomelanocortin), Hcrt (hypocretin/orexin), Agt (angiotensinogen), Gnrh1 (gonadotropin releasing hormone), Oxt (oxytocin), Lhb (luteinizing hormone), Crh (corticotropin-releasing hormone), Adm (adrenomedullin), Ghrl (ghrelin), Igf2 (insulin-like growth factor), Inha, Inhba and Inhbb (inhibins), Edn1 and Edn3 (endothelins), Nppa (atrial natriuretic peptide)). The seven granin genes (Chga, Chgb, Scg2, Scg3, Scg5, Pcsk1n, Vgf) and 16 others (Calca, Calcb, Cartpt, Agt, Iapp, Adcyap1, Igf1, Igf2, Lhb, Inha, Inhba, Inhbb, Adm, Oxt, Pomc, Crh) have also been detected in neuroendocrine (NE) cells outside the lung, but this is the first description of NE cell expression for 17 PNEC peptidergic genes (Dbi, Nppa, Nppb, Nppc, Nmb, Npb, Npff, Npw, Pnoc, Prok2, Gal, Gnrh1, Hcrt, Ucn2, Agrp, Thpo, Uts2). Some of the peptidergic genes including Calca, Igf2, and granins (Chga, Chgb, Scg2, Scg3, Scg5, Pcsk1n) were expressed in most PNECs analyzed, but others were detected in minor subpopulations (Figure 1E, Supplementary file 4). We analyzed peptidergic gene expression in our second scRNA-seq dataset of 92 PNECs that used a different strategy for PNEC labeling and capture (Ouadah et al., 2019) and obtained a similar distribution of peptidergic gene expression, with the exception of Pcsk1n (expressed in 14% of PNECs vs. 62% in original dataset) and Ucn2 (34% vs 0%) (Figure 1E, Supplementary file 4). We also validated expression of eight peptidergic genes and determined the distribution and abundance of the expressing cells in vivo by immunostaining and multiplex smFISH. The canonical mouse PNEC marker and neuropeptide CGRP (Calca) was detected in 95% of PNECs by scRNA-seq, 95% of PNECs by immunostaining (n=237 scored cells in 3 mice) (Figure 1F), and 94% by smFISH (n=100 scored cells in 2 mice) (Figure 1D). Cartpt was detected in 34% of PNECs by scRNA-seq and twice that by immunostaining (68%) (Figure 1F) and smFISH (67%) (Figure 1G). smFISH also confirmed PNEC expression of the six other peptidergic genes examined (Scg5, Chga, Agt, Pomc, Nmb, Adcyap1) but generally gave higher percentages of expressed PNECs especially for the less abundantly expressed genes, implying smFISH is more sensitive than scRNA-seq in detecting gene expression (Figures 1G and 2B). Indeed, when PNECs with very low expression (1–4 mRNA puncta detected per cell) were not included in the smFISH quantification, there was excellent agreement between smFISH and scRNA-seq values (Figure 1G). The broadly expressed peptidergic genes Calca, Scg5, Chga, and Cartpt were detected in both solitary and clustered PNECs, and the four genes detected in smaller subpopulations (Agt, Pomc, Nmb, Adcyap1) were most commonly detected in NEBs at bronchial branchpoints, with local clustering observed within a NEB of the Pomc-expressing cells (Figure 2—figure supplement 1A). Figure 2 with 1 supplement see all Download asset Open asset PNECs express myriad combinations of peptidergic genes. (A) Peptidergic genes expressed in individual PNECs (n=176) from scRNA-seq. Histogram (top) shows number of PNECs expressing each of the 154 observed combinations of expressed peptidergic (NP) genes (dots, bottom). Values at right are average expression level of peptidergic genes among expressing cells; values at bottom are number of expressed peptidergic genes for each combination (NPs/cell). Each PNEC expressed 7.2±1.9 (mean ± SD) peptidergic genes (range 2–13, median 7, mode 6). (B) Micrograph of NEB from adult PN155 wild type mouse lung probed by multiplex smFISH (RNAscope) for PNEC marker Resp18 and the indicated peptidergic genes including ones detected by scRNA-seq in most (Scg5, Calca), some (Cartpt, Chga), or rare (Agt, Pomc, Nmb, Adcyap1) PNECs. Epithelial cells are outlined (white dots), basement membrane is indicated by dashed line, and individual PNECs numbered (n=11) with close-up of the indicated PNEC shown at right and PNEC 1 at bottom. Scale bar, 2μm. Filled colored circles above cells, PNECs with high expression of gene (≥5 puncta; filled circles); open circles, PNECs with low expression of gene (1–4 puncta, open circles). Schematic (bottom) shows summary of expression of the eight peptidergic genes (using colored circles as above) in the 11 PNECS in this NEB optical plane. Grey cells, other (non-PNEC) epithelial cells in field of view. Bar, 10μm. (C) Percent of PNECs expressing each of the 30 observed combinations of the eight peptidergic genes probed by smFISH (n=100 PNECs scored in NEBs from two wild-type mice). Filled circles, expressed peptidergic gene. (D) Percent of PNECs expressing each of the 24 observed combinations of the same eight peptidergic genes in scRNA-seq dataset (n=176 PNECs). We conclude that PNECs express dozens of peptidergic genes, nearly half (47%) of all annotated peptidergic genes and over an order of magnitude more than previously known, with most expressed in rare PNEC subpopulations. Statistical modeling of our PNEC sampling in scRNA-seq indicated we likely achieved saturation of PNEC peptidergic genes (Figure 2—figure supplement 1B), however the value of 43 expressed genes is a lower limit because even more sensitive methods of detecting gene expression in individual cells such as smFISH could identify additional PNEC peptidergic genes. PNECs express myriad combinations of peptidergic genes The scRNA-seq analysis showed that individual mouse PNECs expressed 7.2 ± 1.9 (mean ± S.D.) peptidergic genes, with some cells expressing up to 13 (range 2–13). Remarkably, almost every PNEC expressed a different combination of peptidergic genes: 154 peptidergic patterns were identified among the 176 PNECs analyzed (Figure 2A). Because the number of combinations detected by scRNA-seq could be artificially inflated by ‘technical dropout’ (failure to detect an expressed gene), we also scored peptidergic combinations by the more sensitive technique of multiplex smFISH (Figure 2B). For the eight peptidergic genes probed in 100 PNECs, we identified 30 different cellular patterns of expression, similar to the 24 patterns detected by scRNA-seq for the same eight genes (Figure 2C and D). Thus, each PNEC expresses multiple peptidergic genes and in an extraordinary number of combinations. The number and diversity of neuropeptides and peptide hormones expressed by PNECs is further expanded by post-transcriptional processing. Some of the expressed genes are alternatively spliced to produce transcripts encoding different neuropeptides or hormones with distinct expression patterns and physiological functions. For example, Calca, which encodes the classical PNEC neuropeptide CGRP, can be alternatively spliced to generate transcripts encoding the thyroid hormone calcitonin that regulates calcium homeostasis (Amara et al., 1980; Amara et al., 1982). Mapping of PNEC scRNA-seq reads at the Calca genomic locus revealed that Calca transcripts are alternatively spliced in PNECs (Figure 3A); most cells expressed both calcitonin and CGRP transcripts, although the ratio varied greatly across individual PNECs and some cells exclusively expressed CGRP (20% of cells) or calcitonin (10%) (Figure 3B, Figure 3—figure supplement 1A). Similar results obtained by co-staining for CGRP and calcitonin proteins (Figure 3C, Figure 3—figure supplement 1B). Figure 3 with 1 supplement see all Download asset Open asset Additional PNEC peptidergic diversity from post-transcriptional processing. (A) Sashimi plots (top) showing mapped scRNA-seq reads for Calca gene and deduced alternative splicing patterns for three representative PNECs (cells 2, 9 and 19 in B), and the resultant mRNAs structures (bottom) encoding either CGRP or calcitonin, with exons numbered (black fill, coding exons). Arrowhead, translation start site. Note cell 2 expresses only CGRP, cell 19 expresses only calcitonin mRNA, and cell 9 expresses both. (B) Quantification of alternative splicing of Calca mRNA as above for 20 randomly selected Calca-expressing PNECs (sashimi plots in Figure 3—figure supplement 1). (C) Fluorescence super-resolution (AiryScan SR) confocal micrograph (left) of neighboring PNECs in adult PN68 wild type mouse lung immunostained for CGRP (green), calcitonin (red), and E-cadherin (Ecad, white) to show epithelial cell boundaries, with DAPI nuclear counterstain (blue). Close-ups of boxed regions (right) show split and merged channels of apical (1) and basal (2) regions of a PNEC. Most PNECs express both of these peptides (see quantification at right), but note that the neuropeptide (CGRP) and hormone (calcitonin) localize to separate vesicles. Scale bar, 2μm (inset 2μm). Right, quantification of immunostaining (n=326 PNECs scored in 3mice). (D) Classical post-translational processing scheme for pro-opiomelanocortin (POMC) in anterior pituitary, showing cleavage sites (arrowheads) of endopeptidases proprotein convertase subtilisin/kexin type 1 (PCSK1, orange) and PCSK2 (blue), and carboxypeptidase E (CPE, purple), and modification sites (arrowheads) of peptidyl-glycine–amidating monooxygenase (PAM amidation site, red) and N-acetyltransferase (NAT acetylation site, green) (Harno et al., 2018). ACTH, adrenocorticotrophic hormone; β-LPH, lipotropin, MSH, melanocyte stimulating hormone, Endo, endorphin. Gray box, junctional peptide; red bar, antigen (residues 27–52) of POMC antibody in panel G. In pituitary, other PCSK2 cleavage events produce additional peptides (γ3-MSH, ACTH (1-17), γ-LPH, β-endorphin) although γ−MSH and β-MSH are likely not produced in mouse due to absence of those dibasic cleavage sites. (E) Heatmap of expression of POMC processing genes in individual PNECs (n=176) from scRNA-seq dataset. Note expression of Pcsk1, Cpe, and Pam in most (94%, 65%, 74%), Pcsk2 in some (16%), and Nat2 and Nat3 in rare (1%) PNECs; none expressed Nat1. This predicts production of all the major pituitary peptides in PNECs, with individual PNECs producing different sets of POMC peptides due to differential expression of the processing enzymes. (F) Micrograph of NEB in adult PN90 wild type (CD-1) mouse lung immunostained for PCSK1 (red), CGRP (green), and E-cadherin (Ecad, white). Right, quantification of immunostaining (n=123 PNECs scored in three mice). Most PNECs co-express PCSK1 and CGRP, although rare PNECs (~10%) express only PCSK1 (asterisks in micrograph). Scale bar, 10μm. (G) Super-resolution confocal micrograph of NEB in adult PN155 wild type (CD-1) mouse lung immunostained for CGRP (green), POMC (red), and E-cadherin (Ecad, white), with DAPI counterstain (blue). Bottom panel, close-up of boxed region showing separate POMC and CGRP puncta. Quantification (n=347 distinct puncta scored in 8 PNECs co-expressing CGRP and POMC) showed 157 CGRP+ puncta, 190 POMC+ puncta, and no puncta with co-localization of the two. Bar, 10μm (inset 2μm). Many neuropeptide and peptide hormone mRNAs are translated as larger pre-pro-peptides that are proteolytically processed and modified to generate up to eight different neuropeptides and/or peptide hormones, typically with different expression patterns and functions (Supplementary file 5), which would further increase the PNEC signaling repertoire. A classic example is POMC, cleaved by proprotein convertases PCSK1 and PCSK2 to generate ACTH (adrenocorticotrophic hormone), MSH (melanocyte stimulating hormone), β-endorphin as well as others by additional processing events (Figure 3D). We surveyed expression of the processing enzyme genes in our scRNA-seq dataset (Figure 3E) and found that nearly all PNECs expressed Pcsk1 (Figures 1C and 3F) and a subset of those expressed Pcsk2, predicting that POMC is processed in PNECs into multiple neuropeptides and peptide hormones including ACTH, α3-MSH, and β-endorphin. Confocal imaging of PNEC immunostains for POMC, CGRP, and calcitonin showed that each co-expressed neuropeptide or peptide hormone localized largely to its own secretory vesicles (Figure 3G), even ones expressed from the same gene (CGRP and calcitonin, Figure 3C), implying distinct vesicular targeting and packaging pathways. Thus, post-transcriptional and post-translational processing expands the number and diversity of peptidergic signals expressed by each PNEC, and their separate vesicular packaging raises the possibility of differentially regulated release and impact on targets. Diverse targets of PNEC signals To predict the direct targets of the PNEC signals, we searched our molecular cell atlas of the lung, comprising the full expression profiles of nearly all lung cell types (Travaglini et al., 2020) and pulmonary sensory neurons (Liu et al., 2021), for cells that express the cognate receptors (Supplementary file 5). The only previously defined targets of PNEC signals are a subpopulation of immune cells (IL5 lineage-positive) proposed to be attracted to PNECs by secreted CGRP in a mouse model of asthma (Sui et al., 2018), and goblet cells, which are increased in macaque and mouse models of inflammation through neurotransmitter GABA from PNECs (Barrios et al., 2019; Sui et al., 2018). Recently, CGRP from tracheal TNECs has been found to support tracheal epithelial cells following hypoxic injury (Shivaraju et al., 2021). Receptors for serotonin and GABA, the two major PNEC neurotransmitters, are expressed by the two pulmonary sensory neuron (PSN) subtypes that innervate NEBs (PSN4 (Olfr78+) and PSN7 (Calb1+); Figure 4A), identifying the first signals from PNECs to these afferent fibers that communicate pulmonary sensory information to the brain. GABA receptor genes were also expressed in goblet cells, supporting the conclusion from macaque and mouse models and identifying the specific GABA receptor subunit as GABRP, as well as in club cells, epithelial cells that neighbor PNECs (Figure 4—figure supplement 1). Receptors for the neurotransmitters glutamate, dopamine and histamine predicted by our transcriptomic data to be produced by rare PNECs (see above) were also expressed in innervating PSNs (Figure 4A) as well as in plasmacytoid dendritic cells (glutamate receptor Grm8) and basophils (histamine receptor Hrh4; Figure 4—figure supplement 1B). Figure 4 with 2 supplements see all Download asset Open asset Predicted lung cell targets of PNEC signals from expression of their receptors. (A) scRNA-seq dot plot of mean expression level (dot heatmap, ratio relative to maximum value for 10 pulmonary sensory neuron (PSN) subtypes, Liu et al., 2021) and percent of cells in population with detected expression (dot size) in the NEB-innervating PSN subtypes (PSN4, PSN7) of genes encoding receptors for each of the indicated PNEC neurotransmitters. (B) scRNA-seq dot plot of expression in PSN4 and PSN7 (as above) and mean level of expression (dot intensity) and percent of cells in population with detected expression (dot size) in cell types (arranged by tissue compartment) in mouse lung cell atlas Travaglini et al., 2020 of genes encoding receptors for indicated PNEC peptidergic ligands (shown above corresponding receptor genes). Ligands with expressed receptors are colored; receptors without detectable expression