Circadian variation in pulmonary inflammatory responses is independent of rhythmic glucocorticoid signaling in airway epithelial cells

Abstract

The circadian clock is a critical regulator of immune function. We recently highlighted a role for the circadian clock in a mouse model of pulmonary inflammation. The epithelial clock protein Bmal1 was required to regulate neutrophil recruitment in response to inflammatory challenge. Bmal1 regulated glucocorticoid receptor (GR) recruitment to the neutrophil chemokine, CXC chemokine ligand 5 (CXCL5), providing a candidate mechanism. We now show that clock control of pulmonary neutrophilia persists without rhythmic glucocorticoid availability. Epithelial GR-null mice had elevated expression of proinflammatory chemokines in the lung under homeostatic conditions. However, deletion of GR in the bronchial epithelium blocked rhythmic CXCL5 production, identifying GR as required to confer circadian control to CXCL5. Surprisingly, rhythmic pulmonary neutrophilia persisted, despite nonrhythmic CXCL5 responses, indicating additional circadian control mechanisms. Deletion of GR in myeloid cells alone did not prevent circadian variation in pulmonary neutrophilia and showed reduced neutrophilic inflammation in response to dexamethasone treatment. These new data show GR is required to confer circadian control to some inflammatory chemokines, but that this alone is insufficient to prevent circadian control of neutrophilic inflammation in response to inhaled LPS, with additional control mechanisms arising in the myeloid cell lineage.-Ince, L. M., Zhang, Z., Beesley, S., Vonslow, R. M., Saer, B. R., Matthews, L. C., Begley, N., Gibbs, J. E., Ray, D. W., Loudon, A. S. I. Circadian variation in pulmonary inflammatory responses is independent of rhythmic glucocorticoid signaling in airway epithelial cells.

Keywords: circadian rhythms; endocrinology; inflammation.

Figure 1

Time-of-day variation in pulmonary LPS response is retained despite corticosterone clamp. A) Dose–responses used to establish optimal corticosterone (CORT) clamp concentration. Corticosterone concentration in serum samples from tail blood taken at the indicated time points, n = 5–7. Analysis was performed by 2-way ANOVA with Sidak’s multiple comparisons test between time points, and a significant main effect of treatment was observed. Individual dose plots are shown in Supplemental Fig. 2; a dose of 2.5 mg was used in subsequent experiments. B, C) Inflammatory measurements 5 h after nebulized LPS exposure at the indicated time points. Neutrophil counts in BAL fluid and BAL CXCL5 concentration in control animals (B) and CORT-clamped animals (C), median age 12 wk, n = 4–6/group. Analysis using Grubbs’ test revealed 3 outliers in the CXCL5 results in C (1 each from CT0, CT6, and CT18), which were removed before further analysis. Analysis was performed by 1-way ANOVA and post hoc tests with Tukey’s correction. Asterisk denotes significance from CT0 time point, except where indicated. D) mRNA expression of a panel of clock genes in whole lung of CORT-clamped and control animals after LPS challenge. Analysis was performed via 2-way ANOVA; all genes showed a significant effect of time but no differences between treatments. Data represent means ± se. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 2

Endogenous Gcs act through GR to modulate lung rhythmicity. A) Lung slices were taken from PER2::LUC mice and bioluminescence was recorded before and after treatment with Cort (100 nM; arrow denotes treatment time). DMSO was used as a control in all experiments; data are representative of 3–7 experiments. B, C) Sections were treated with GSK67a (10 nM) and RU486 (1 µM; B) or GSK69 (10 nM) and RU486 (C); data are representative of 3–7 experiments. D) Sections treated with Cort (100 nM), the MR-specific antagonist, spironolactone (Spir; 1 µM), or both Cort and Spir. Phase change was calculated relative to an untreated peak (positive being a delay, and negative being an advance). Analysis was performed with 1-way ANOVA and post hoc tests with Tukey’s correction, n = 3–7. Data represent means ± se. ****P < 0.0001 compared with DMSO control.

Figure 3

Deletion of GR in airway epithelial cells of Ccsp-GR^−/− mice is associated with reduced responsiveness to Dex-induced circadian synchronization. A) Immunofluorescence images of lung sections from GR^WT (top) and Ccsp-GR^−/− mice (bottom) stained for GR (red) and cell nuclei (DAPI). Green indicates background/auto fluorescence. B) Images of lung sections from GR^WT (top) and Ccsp-GR^−/− mice (bottom) probed for GR mRNA via in situ hybridization with relative quantification of GR (Nr3c1) mRNA abundance from multiple lung sections (n = 4/genotype, analyzed with 2-tailed Student’s t test). C) PER2-LUC bioluminescence (photon counts/min) over multiple days from GR^WT (left) and Ccsp-GR^−/− (right) lung sections. Dex was administered at the indicated time (arrow). Plot is representative of 3 independent replicates. D) Relative quantification of Bmal1, Per2, Nr1d1, and Nr1d2 gene expression in whole lung (n = 4–7, 2-way ANOVA with Sidak’s multiple comparisons test within genotypes). For Nr1d1, 1 outlier was removed from the GR^WT group at CT12 after using the Grubbs’ test to detect outliers. Data represent means ± se. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4

Loss of GR in the bronchiolar epithelium does not affect circadian gating of neutrophilia after LPS challenge. GR^WT and Ccsp-GR^−/− mice (median age, 10 wk) exposed to nebulized LPS at CT0 or CT12, with samples collected 5 h later. A) Quantification of total cells (left) and neutrophils (right) in BAL fluid. B) BAL CXCL5 (left) and total protein (right) concentration. C, D) Expression of clock genes (C) and inflammatory genes (D) in lung homogenate after challenge. For all panels, n = 8–10; data were analyzed with 2-way ANOVA, followed by Sidak’s multiple comparisons test for effects within genotype. Data represent means ± se.*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 5

Ccsp-GR^−/− mice retain sensitivity to anti-inflammatory effects of Dex during LPS challenge. GR^WT and Ccsp-GR^−/− mice (median age, 12 wk) were exposed to nebulized LPS at CT0 and culled 5 h later. The intraperitoneal injection of either Dex (1 mg/kg) or saline (vehicle) took place 1 h before LPS exposure. A, B) Quantification (A) of total cells (left) and neutrophils (right) in BAL fluid. Expression (B) of Cxcl5, Cxcl1, and Il6 in lung homogenate after challenge. n = 6–10, and data were analyzed with 2-way ANOVA, followed by Sidak’s multiple comparisons test for effects within genotype. C) Hierarchical clustering of proinflammatory cytokines/chemokines from Nanostring analysis (see Supplemental Fig. 6 for full heat map of Dex-responsive genes, and Supplemental Table 1 for list of all genes analyzed). Data represent means ± se. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6

Airway epithelium-specific GR loss affects gene expression during homeostasis. RNA-sequencing from laser-microdissected distal bronchiolar epithelial cells at ZT14. A) Volcano plot of gene expression in bronchiolar epithelial cells from Ccsp-GR^−/− mice relative to GR^WT littermate controls. B) Assessment of rhythmic expression of Cxcl5 and Cxcl15 in whole lung harvested at the indicated time points (n = 3–6, 2-way ANOVA with Sidak’s multiple comparisons test within genotype). Data represent means ± se. **P < 0.01.

Figure 7

LysM-GR^−/− mice retain time of day variation in response to nebulized LPS and Dex sensitivity. GR^WT and LysM-GR^−/− mice (median age, 12 wk) were exposed to nebulized LPS at the indicated time points and culled 5 h later. A) Quantification of total cells (left) and neutrophils (right) in BAL fluid. B) Concentrations of CXCL5 (left) and total protein (right) in BAL. C) Concentrations of IL-6 (left) and TNF-α (right) in BAL. D) The intraperitoneal injection of Dex (1 mg/kg) or saline (vehicle) took place 1 h before LPS exposure at CT0. BAL was collected 5 h after LPS challenge, and total cells (left) and neutrophils (right) were quantified. For all panels, n = 7–10; data were analyzed with 2-way ANOVA and Sidak’s multiple comparisons test within genotypes. Data represent means ± se. **P < 0.01, ***P < 0.001.

Figure 8

The gated LPS response. WT: Upon stimulation with LPS, inflammatory signaling pathways in the lung are engaged, resulting in production of proinflammatory chemokines and neutrophil influx through the pulmonary circulation (42, 43). In WT animals, the clock regulates access to modulatory regions of chromatin, including an enhancer region in the Cxcl5 sequence to which GR has been shown to bind with inhibitory effects upon transcription. At CT12, GR binding is permitted, and cell influx is suppressed relative to the challenge at CT0. A strong signal (CXCL5) combined with more neutrophils in the circulation during the day (34) leads to an oscillation in infiltrating cell count (3). Cxcl5^−/−: CXCL5 predominates in driving neutrophil influx to the alveolar space in the nebulized LPS model (30). Mice can no longer produce Cxcl5 upon LPS stimulation, resulting in low numbers of invading neutrophils at both time points. Some cells are still found in lavage fluid, indicating successful migration without a CXCL5 signal (3). Ccsp-Bmal1^−/−: Without Bmal1 in the bronchial epithelial cells, the inhibitory binding of GR no longer occurs at CT12 and gating of CXCL5 production is lost. Production of CXCL5 is enhanced and increased neutrophilia occurs at both time points without a day/night variation in infiltrating cell number (3). Ccsp-GR^−/−: Without GR in the bronchial epithelial cells, the inhibitory binding of GR no longer occurs at CT12, and gating of CXCL5 production is lost despite the presence of an intact circadian clock. However, an intact oscillation in cell influx remains. A nonrhythmic signal (CXCL5), combined with oscillations in neutrophil count in the bloodstream (34), still results in an oscillation in infiltrating cell count.