Circadian regulation of lung repair and regeneration

Abstract

Optimal lung repair and regeneration are essential for recovery from viral infections, including influenza A virus (IAV). We have previously demonstrated that acute inflammation and mortality induced by IAV is under circadian control. However, it is not known whether the influence of the circadian clock persists beyond the acute outcomes. Here, we utilize the UK Biobank to demonstrate an association between poor circadian rhythms and morbidity from lower respiratory tract infections, including the need for hospitalization and mortality after discharge; this persists even after adjusting for common confounding factors. Furthermore, we use a combination of lung organoid assays, single-cell RNA sequencing, and IAV infection in different models of clock disruption to investigate the role of the circadian clock in lung repair and regeneration. We show that lung organoids have a functional circadian clock and the disruption of this clock impairs regenerative capacity. Finally, we find that the circadian clock acts through distinct pathways in mediating lung regeneration - in tracheal cells via the Wnt/β-catenin pathway and through IL-1β in alveolar epithelial cells. We speculate that adding a circadian dimension to the critical process of lung repair and regeneration will lead to novel therapies and improve outcomes.

Keywords: Influenza; Mouse models; Pulmonology; Respiration; Virology.

Figure 1. UK Biobank analyses.

(A) Description of study cohort from UK Biobank. (B) Hazard ratio for the risk of hospitalization for respiratory infections across the quintiles of relative amplitude (RA) of actigraphy as a measure of circadian rhythms. The lowest quintile represents individuals with the poorest circadian rhythms and serves as the comparison group. (C) Length of hospital stay defined across the quintiles of RA scores (median, IQR). (D) Hazard ratio of the risk of death within 30 days after discharge from above-defined hospitalization. All hazard ratios were adjusted for cofounding factors that included, age, sex, HbA1c, cancer diagnosis, FEV1, BMI, smoking status, history of previous respiratory illness requiring hospitalization, overall physical activity level, use of bronchodilators and inhaled corticosteroids, and self-rated overall health.

Figure 2. Infection time of day affects long-term lung repair and regeneration.

(A) Schematic of experimental design: Mice were infected with influenza A virus (IAV) at either dawn (ZT23/rest phase) or dusk (ZT11/active phase). (B) Left: Representative micrographs of H&E-stained lung sections 30 days post infection (p.i.). The red box denotes areas with maximal alveolar destruction — zones 3 and 4 in our scoring system. Scale bars: 5–6 mm. Right: Quantification of the severity of lung injury. **P = 0.0029 for time of infection by 2-way ANOVA. *P = 0.016 for zone 3 and **P = 0.04 for zone 4 on multiple comparison with Bonferroni’s correction. (C) Pulse oximetry on day 35 p.i. ****P < 0.0001 by unpaired 2-tailed t test. (D) Left: Representative micrographs of H&E-stained lung sections from Bmal1^creERT2/+ and cre^neg littermates 30 days p.i. Scale bars: 5–6 mm. Right: Quantification. *P = 0.0028 for time of infection, 2-way ANOVA. *P = 0.0171 for zone 3 and **P = 0.0056 for zone 4 on multiple comparison with Bonferroni’s correction. Data were pooled from 2–4 independent experiments and are represented as mean ± SEM. (E) Pulse oximetry on day 35 p.i. **P = 0.0061 by Mann-Whitney test. (F) Experimental design for organotypic assays. (G) Representative tracing of relative bioluminescence of Per2 expression in organoids grown from different levels of respiratory tree from mPer2^Luc mice. (H) Relative bioluminescence recording of tracheal organoids 4 days after seeding. Organoid data are summarized from 3–5 independent experiments with at least 3 technical replicates/experiment. Two-way ANOVA with Tukey’s multiple-comparisons test. **P = 0.009, 0.004, 0.005; hours of bioluminescence **P = 0.0019.

Figure 3. Deletion of Bmal1 reduces regenerative capacity in lung organoids.

Lungs from embryonic Bmal1^–/– and their Bmal1^+/+ littermates were used for organotypic assays. For postnatal deletion of Bmal1, Bmal1^creERT2/+ and their cre^neg littermates were treated with tamoxifen at 8 weeks of age. Representative images of tracheal organoids from (A) Bmal1^+/+ and Bmal1^–/–. (B) Regenerative capacity was quantified as colony-forming efficiency (CFE). (C) Bmal1^creERT2neg and Bmal1^creERT2/+ mice. (D) Quantification (CFE). (E) Representative images of CD104⁺ distal lung cell organoids grown from Bmal1^+/+ (WT) and Bmal1^–/– mice. (F) Quantification as CFE. (G) Bmal1^creERT2neg and Bmal1^creERT2/+. (H) Quantification as CFE. (I) Representative micrographs of H&E-stained lung sections from Scgb1a1^Cre/+ Bmal1^fl/fl mice and Scgb1a1^Creneg Bmal1^fl/fl littermates, infected as in Figure 2A and recovered until day 30. Scale bars: 5 mm. (J) Quantification. Each data point represents an individual animal, and data were pooled from 2 independent experiments (n = 7–8 mice per circadian time point) 30 days p.i. ****P < 0.0001 for genotype by 2-way ANOVA; ***P = 0.0002 for zone 3; *P = 0.015 for zone 4 on multiple testing. (K) Representative images of tracheal organoids from Cry1^–/– Cry2^–/– (Cry1,2–DKO) mice. Scale bar: 2000 μm. (L) AT2 organoids cocultured with Cry1,2–DKO fibroblasts. (M) Embryonic and (N) postnatal Bmal1-knockout mice. Organoid experiments: Data pooled from 3–5 independent experiments with at least 3 technical replicates/experiment and expressed as mean ± SEM. B: ***P = 0.005; D and F: *P = 0.01; H: *P = 0.04; J and L: *P = 0.01, ****P = 0.0001. Ordinary 1-way ANOVA. M: ^#P = 0.09; N: *P = 0.01. Unpaired 2-tailed t test with Welch’s correction.

Figure 4. Single-cell transcriptome analysis of Bmal1^–/– lung cells reveals downregulation of Wnt-associated pathways.

(A) Schematic of experimental design for isolation of lung cells for single-cell sequencing from Bmal1^+/+, Bmal1^–/–, Bmal1^creERT2/+, and Bmal1^creERT2neg mice. DARQ7^–CD31^–CD45^– lung cells from uninfected mice underwent single-cell transcriptomic analysis. Data were integrated from Bmal1^+/+, Bmal1^–/–, Bmal1^creERT2neg, and Bmal1^creERT2/+ mice. (B) Uniform manifold approximation and projection (UMAP) visualization dimension reduction analysis of integrated scRNA-seq data generated using a Seurat pipeline. (C) Venn diagram depicting the number of differentially expressed genes. (D) Enriched ontology clusters of differentially expressed genes in both the knockout models.

Figure 5. Disruption of the circadian clock decreases proliferation after IAV infection.

(A) Sftpc^CreERt2/+ Bmal1^fl/fl mice, Cry1,2–DKO, and Scgb1a1^Cre/+ Bmal1^fl/fl and their cre^neg or WT littermates were moved to constant darkness (DD) 2 days prior to administering IAV at either CT23 or CT11. (B) Left: Representative Ki67-stained images from Scgb1a1^Creneg Bmal1^fl/fl mice and Scgb1a1^Cre/+ Bmal1^fl/fl littermates. Right: Quantification of Ki67⁺ cells/high power field (HPF). P = 0.0939 for genotype, P = 0.0091 for time of infection, and P = 0.0094 for interaction by 2-way ANOVA; **P = 0.0012 for cre^neg group and P > 0.999 for cre⁺ group with Bonferroni’s correction for multiple comparisons. (C) Left: Sftpc^CreERT2neg Bmal1^fl/fl mice versus Sftpc^CreERt2/+ Bmal1^fl/fl littermates. Right: Quantification of Ki67⁺ cells/HPF. P = 0.0314 for genotype, P = 0.0265 for time of infection, and P = 0.0035 for interaction by 2-way ANOVA; **P = 0.0030 for cre^neg group and P = 0.75 for cre⁺ group with Bonferroni’s correction for multiple comparisons. (D) Images and quantification Ki67⁺ AT2 cells in Sftpc^CreERT2neg Bmal1^fl/fl and Sftpc^CreERt2/+ Bmal1^fl/fl mice. (E) Cry1,2–DKO mice and quantification. *P = 0.0129 by unpaired 2-tailed t test. Each point represents an animal and data are expressed as mean ± SEM. Scale bars: 100 μm.

Figure 6. Activation of Wnt signaling in tracheal organoids rescues the regenerative defect in the absence of Bmal1.

(A) Gene expression of Wnt3a from Bmal1^creERT2neg lungs harvested at different time intervals determined by qPCR (n = 3–4 mice per time point, q = 0.014). (B) Representative immunoblot of β-catenin expression from whole-lung extracts from Bmal1^creERT2neg and Bmal1^creERT2/+ mice. (C) Quantification of β-catenin expression normalized to β-actin from 2 independent experiments (n = 4–5, *P < 0.01). (D) ChIP assay of BMAL1 occupancy on the Wnt3a promoter. mPer2 primers were used as positive controls for the analysis. Data are expressed as percentage of input level normalized to IgG control (n = 4, pooled from 2 independent experiments). *P = 0.04, unpaired 2-tailed t test with Welch’s correction. Tracheal organoids were supplemented with Wnt3a in DMSO. (E) Bmal1^+/+ WT littermates, Bmal1^–/– DMSO, Bmal1^–/– Wnt3a. (G) Bmal1^creERT2neg littermates DMSO, Bmal1^creERT2/+ DMSO, and Bmal1^creERT2/+ Wnt3a CFE for (F) Bmal1^–/– and (H) Bmal1^creERT2/+. *P = 0.01,**P = 0.001, ****P = 0.0001. (I) AT2 organoids were supplemented with IL-1β (10 ng/mL) and DMSO (0.05% final concentration). (J) CFE: *P = 0.04, **P = 0.008, ***P = 0.0004. Scale bars: 2000 μm. Data pooled from 3–5 independent experiments with at least 3 technical replicates/experiment expressed as mean ± SEM. Kruskal-Wallis test with Dunn’s multiple-comparison test.