Excess neuropeptides in lung signal through endothelial cells to impair gas exchange

Abstract

Although increased neuropeptides are often detected in lungs that exhibit respiratory distress, whether they contribute to the condition is unknown. Here, we show in a mouse model of neuroendocrine cell hyperplasia of infancy, a pediatric disease with increased pulmonary neuroendocrine cells (PNECs), excess PNEC-derived neuropeptides are responsible for pulmonary manifestations including hypoxemia. In mouse postnatal lung, prolonged signaling from elevated neuropeptides such as calcitonin gene-related peptide (CGRP) activate receptors enriched on endothelial cells, leading to reduced cellular junction gene expression, increased endothelium permeability, excess lung fluid, and hypoxemia. Excess fluid and hypoxemia were effectively attenuated by either prevention of PNEC formation, inactivation of CGRP gene, endothelium-specific inactivation of CGRP receptor gene, or treatment with CGRP receptor antagonist. Neuropeptides were increased in human lung diseases with excess fluid such as acute respiratory distress syndrome. Our findings suggest that restricting neuropeptide function may limit fluid and improve gas exchange in these conditions.

Keywords: lung function; neuropeptides; pediatric disease; pulmonary neuroendocrine cells.

Figure 1.. Establishing the Nkx2–1 Mutant Model of NEHI using CRISPR/Cas9 Genome Editing.

(A) Schematic of CRISPR/Cas9 strategy to generate Nkx2–1^R161L (Nkx2–1^L) allele. Aside from engineered changes, a silent C→G change also arose in the CRISPR/Cas9 repair process, and is noted in green. (B) RFLP-based genotyping. Introduction of BspEI site in the mutant allele led to digestion of 503bp PCR product into 326bp and 177bp fragments. (C) PNEC quantification of whole left lobe at E18.5 (n=3 for each group), statistical analysis was performed on total PNEC number (airway and alveoli combined). (D) qPCR of PNEC markers Ascl1 and Calca at P22 (n=3 for each group). (E-G) Representative longitudinal section of the main airway in P22 lung for each group with anti-CGRP staining labeling PNECs (arrows). Scale bar: 100μm. (E’-G’) Magnified view of boxed areas in E-G, respectively. Scale bar: 50μm. (H-J) CGRP+ cells (arrowheads) were ectopically detected in the alveolar region of Nkx2–1^L/L mutants. Scale bar: 50μm. (K) NEHI patient biopsy with anti-GRP/Bombesin labeling of PNECs showing large NEB at bronchiole (dashed box) and ectopic solitary PNECs in alveolar region (arrowheads). (L) Representative body size at P22. (M) Growth curve from P5 to P22 of 12 pups nursed by the same outbred foster mother (n=4 for each group). Statistical analysis was carried out separately at each time point. (N)Oxygen saturation at P22 (n≥6 for each group). One-way ANOVA Tukey’s multiple comparisons test was used for C, D, M and N. N.S. not significant, p≥0.05. * for p<0.05, ** for p<0.01, *** for p<0.001. Error bars represent means ± SD. See also Figure S1.

Figure 2.. PNEC Hyperplasia in Nkx2–1^L/L Mutants is Associated with Increased Sox2, Decreased Notch Signaling and Increased Specification rather than Proliferation.

(A) qPCR showing increased PNEC marker Ascl1 at E13.5 when PNECs are first specified (n=3 for each group). (B) Rate of epithelial cell proliferation quantified by EdU staining at E13.5 (n=3 for each group). (C-F) Representative longitudinal section of the main airway of E18.5 lungs with anti-CGRP antibody staining labeling PNECs. Scale bar: 100μm. (C’-F’) Magnified view of boxed areas in C-F, respectively. Scale bar: 50μm. (G) qPCR of Sox2 and Notch pathway genes Notch1, Notch2 and Hes1 at E13.5 (n=3 for each group). (H-J) Representative anti-SOX2 antibody staining at P22. Scale bar: 100μm. (K) qPCR of Sox2 at P22 (n=3 for each group). (L) Schematics of a molecular mechanism linking Nkx2–1^L mutation to PNEC hyperplasia, as supported by data here and previous publications (Domyan et al., 2011; Ito et al., 2000; Li et al., 2013; Que et al., 2007; Xu et al., 2014). Student’s t test was used for A and K. One-way ANOVA Tukey’s multiple comparisons test was used for B and K. N.S. not significant, p≥0.05. * for p<0.05, ** for p<0.01, **** for p<0.0001. Error bars represent means ± SD. See also Figure S2.

Figure 3.. PNEC Hyperplasia is Responsible for Lung Fluid Increase and Gas Exchange Deficiency.

(A, B) Lung W/D ratio at P22 (n≥8 for each group) and at P39 (n≥5 for each group). (C) Representative body size at P22. (D) Growth curve from P5 to P22 of 12 pups nursed by the same outbred foster mother (n=3 for each group). Statistical analysis was carried out separately at each time point. (E) Lung W/D ratio at P22 (n≥8 for each group). (F) Oxygen saturation as measured by SpO₂ at P22 (n≥7 for each group). One-way ANOVA Tukey’s multiple comparisons test was used for A, B and two-way ANOVA Tukey’s multiple comparisons test was used for D-F. N.S. not significant, * for p<0.05, ** for p<0.01, *** for p<0.001. Error bars represent means ± SD. See also Figure S3.

Figure 4.. PNEC Product CGRP Contributed to Lung Fluid Increase and Gas Exchange Deficiency.

(A) CGRP concentration in BAL as measured by ELISA for each genotype (n=8 for first three groups, n=4 for Nkx2–1^L/L; Ascl1^cKO mutants). (B) Wild-type mice were intranasally administrated with 10μl 1pg/μl CGRP or saline at P10, P12, P14 and P16. This concentration was selected based on physiological range measured by BAL ELISA. (C) Lung W/D ratio measured at P18 (n=4 for saline-treated group, n=3 for CGRP-treated group). (D) CGRP concentration in BAL as measured by ELISA for each genotype (n=4 for each group). (E) Representative body size at P22. (F) Growth curve from P5 to P22 of 12 pups nursed by the same outbred foster mother (n=3 for each group). Statistical analysis was carried out separately at each time point. (G) Lung W/D ratio at P22 (n≥7 for each group). (H) Oxygen saturation at P22 (n≥10 for each group). One-way ANOVA Tukey’s multiple comparisons test was used for A. Student’s t test was used for C, and two-way ANOVA Tukey’s multiple comparisons test was used for F-H. N.S. not significant, p≥0.05. * for p<0.05, ** for p<0.01, *** for p<0.001, **** for p<0.0001. Error bars represent means ± SD. See also Figure S4.

Figure 5.. Excess CGRP Produced by PNECs Signals to Pulmonary Endothelial Cells and in turn Leads to Excess Fluid and Reduced Oxygenation in the NEHI Mouse Model.

(A) qPCR of CGRP downstream cAMP signaling genes Creb1, Atf2 (also termed Creb2) and Creb3 in P22 whole lungs (n=3 for each group). (B) Integrated uniform manifold approximation and projection (UMAP) plot of scRNA-seq data of CD31+ endothelial cells isolated from P22 Nkx2–1^+/+ and Nkx2–1^L/L lungs. (C) Dot plot showing marker genes for each population. (D) Dot plot of Rras expression in individual endothelial population showing a decrease in vein, lymphatics and both capillary populations in the Mut (Nkx2–1^L/L) compared to Ctrl (Nkx2–1^+/+) lungs. Numbers on the left denote expression levels while numbers on the right denote percentages of cells within each equivalent cell types expressing Cldn5. (E) CDH5 protein level as assayed by western blot of P22 lungs. Dashed line denotes a cut from the original blot. (F) Quantification of CDH5 protein level after normalization to internal control beta-actin level (n=3 for each group). (G) qRT-PCR confirming increased Calca and showing decreased Cldn5 in P22 mutant versus control lungs. (H) CLDN5 protein level assayed by western blot of P22 lungs. Dashed line denotes a cut from the original blot. (I) Quantification of CLDN5 protein level in whole lungs after normalization to internal control beta-actin level (n=3 for each group). (J-M) Anti-CLDN5 antibody staining (magenta) and anti-EMCN antibody staining for vein cells (green) showed reduced CLDN5 expression in Nkx2–1^L/L mutants compared to Nkx2–1^+/+ controls (J, K). Nkx2–1^L/L; Ascl1^cKO and Nkx2–1^L/L; Calca^−/− compound mutants showed similar CLDN5 pattern as Nkx2–1^+/+ controls (L, M). Scale bar: 50μm. (N) Percentage of the vessel circumference outlined by EMCN staining that was also CLDN5+ (n=3 for each group). (O) Lung Evans Blue level as a ratio of dye retained in lung versus level in serum (n≥6 for each group). (P, Q) Representative whole mount angiogram images of right middle lobes of Nkx2–1^+/+ (P) and Nkx2–1^L/L (Q) mice at P22 where leakage was observed in the mutant lung. Magnification: 15X. Leaking regions were marked by yellow arrowheads in Q. (R) Lung W/D ratio at P22 (n≥6 for each group). (S) Oxygen saturation at P22 (n≥5 for each group). (T, U) Representative anti-CD31 antibody (green) and anti-CLDN5 antibody double staining (red) showed a decrease of CLDN5 in human NEHI patient biopsies (U) compared to normal lung biopsies (T). Scale bar: 100μm. (V) CLDN5/CD31 intensity ratio showing a trending decrease of ratio in NEHI biopsies compared to control (n=4 for each group). (R) Schematics of a molecular mechanism linking increased CGRP and disrupted endothelial integrity in Nkx2–1^L/L mutants, as supported by data here and previous publications (Aiyar et al., 1996; Jang et al., 2011; Perrot et al., 2018; Taddei et al., 2008; Travaglini et al., 2020; Wang et al., 2020). Student’s t test was used for A, F, G, I and V. One-way ANOVA Tukey’s multiple comparisons test was used for N and O. Two-way ANOVA Tukey’s multiple comparisons test was used for R and S. N.S. not significant, p≥0.05. ** for p<0.05, ** for p<0.01, *** for p<0.001, **** for p<0.0001. Error bars represent means ± SD. See also Figure S5 and Table S1.

Figure 6.. Antagonizing CGRP Signaling Reversed Molecular Changes, Alleviated Excess Fluid and Reduced Oxygenation in the NEHI Mouse Model.

(A) Regime of intranasal BIBN-4096 administration. The dose was selected based on published literature (Aubdool et al., 2014). (B) Representative body size comparison at P22. All individuals were nursed by the same foster mother. (C) Lung W/D ratio at P22 (n≥3 for each group). (D) Oxygen saturation at P22 (n≥3 for each group). (E) qPCR of Cldn5 at P22, n=3 for each group. (F-H) CDH5 and CLDN5 protein levels in P22 lungs as assayed by western blot. Dashed line denotes a cut from the original blot. Quantifications were shown in G and H, respectively. n=3 for each group. (I, J) Representative anti-CLDN5 antibody staining (magenta) and anti-EMCN antibody staining for vein cells (green) showed restored CLDN5 level in BIBN-treated Nkx2–1^L/L mutants compared to DMSO-treated Nkx2–1^L/L mutants. Two-way ANOVA Tukey’s multiple comparisons test was used for C and D. One-way ANOVA Tukey’s multiple comparisons test was used for E. Student’s t test was used for G and H. N.S. not significant, p≥0.05. * for p<0.05, *** for p<0.001. Error bars represent means ± SD. See also Figure S6.

Figure 7.. Human ARDS Lungs Exhibit Increased Proportion of CGRP+ PNECs and Reduced Tight Junction Protein Expression.

(A-H) Examples of anti-GRP staining (green, pan-PNEC marker) and anti-CGRP staining (magenta) showed differential proportion of CGRP+ PNECs in different groups of donor lungs. Ctrl 1: controls from lobectomy tissues without notable lung pathology (A, B), Ctrl 2: controls from autopsy tissues without notable lung pathology (C, D), COVID-19 ARDS samples (E, F), and non-COVID-19 ARDS samples (G, H). Asterisks indicate CGRP-negative PNECs; arrows indicate CGRP-positive PNECs. (I) Percentages of CGRP+ PNECs in all GRP+ PNECs in noted sample groups. Each dot represents averaged data from one donor across multiple sections. (J-O) Representative images of anti-ERG staining (magenta, endothelial marker) and anti-CLDN5 staining (white) showed decrease of CLDN5 level in non-COVID-19 ARDS samples. Ctrl 1 (J, K), Ctrl 2 (L, M) and ARDS (N, O) groups are the same as described in A-H. (P) Percentages of CLDN5+ cells in ERG+ endothelial cells in noted sample groups. Each dot represents averaged data from one donor across multiple sections. (Q) A model for PNEC function in the control of lung fluid balance. In NEHI, PNEC hyperplasia led to increased neuropeptides such as CGRP, which signal to pulmonary endothelial cells to disrupt barrier integrity, leading to excess extravascular lung fluid as well as gas exchange deficiency. One-way ANOVA Tukey’s multiple comparisons test was used for P and I. N.S. not significant, p≥0.05. * for p<0.05, ** for p<0.01, *** for p<0.001, **** for p<0.0001. Error bars represent means ± SD. See also Table S2.