Circadian Control of Pulmonary Endothelial Signaling Occurs via the NADPH Oxidase 2-NLRP3 Pathway

Abstract

Circadian rhythms are endogenous oscillations that occur with a 24-h periodicity and support organismal homeostasis. While the role of the circadian clock in systemic vasculature is well known, its role in pulmonary vasculature, specifically in the pulmonary endothelium, has remained unexplored. We hypothesized that the circadian clock directly regulates pulmonary endothelium to control lung inflammation. Using pulmonary artery segments and endothelial cells isolated from lungs of mPer2luciferase transgenic mice, we monitored circadian rhythms and observed that lipopolysaccharide (LPS) treatment disrupted rhythmicity. This disruption was mediated by reactive oxygen species (ROS) generated via NADPH oxidase 2 (NOX2). Remarkably, the pharmacologic inhibition of NOX2 before LPS exposure restored circadian rhythmicity in the pulmonary endothelium. In wild-type (WT) mice, LPS activated a NOX2-NLRP3 signaling axis that drove inflammation as evidenced by increased polymorphonuclear neutrophil (PMN) accumulation and intercellular adhesion molecule-1 (ICAM-1) expression on the pulmonary endothelium. In contrast, disruption of the clock using two different clock mutants (Bmal1^–/– and Cry1/2^–/–) resulted in a sustained baseline elevation of PMN and ICAM-1, which changed minimally with LPS. This effect was attributed to aberrant activation of the NLRP3 inflammasome at baseline in the clock mutants, as supported by lung transcriptomic data and reversal of the phenotype with an NLRP3 inhibitor. Importantly, these findings also reveal an intriguing bidirectional relationship: while the circadian clock modulates inflammatory responses, inflammatory stimuli in turn alter circadian rhythmicity via the NOX2 pathway. Together, our results identify a novel mechanism by which circadian control of pulmonary endothelial inflammation may be leveraged to mitigate the consequences of clock disruption in lung disease.

Graphical Abstract.

In the presence of the clock, naive endothelium is not inflamed. Inflammatory stimulus (LPS) causes endothelial ROS production, endothelial inflammation and an eventual resolution. In the absence of the clock, the naive endothelium is inflamed. An additional inflammatory stimulus has no effect on inflammation and its eventual resolution.

Circadian rhythms refer to endogenous oscillations that occur with a 24-h periodicity and are ubiquitous in mammalian cells (Feng and Lazar, 2012). The circadian clock influences almost all aspects of inflammation and immune response. The lungs form the first line of defense against airborne pathogens (Issah et al., 2021). The vascular endothelium plays a central role in inflammatory signaling, serving as a key integrative interface for blood, nutrient, and inflammatory signal transport, as well as immune cell recruitment (Chatterjee, 2018; Chatterjee et al., 2021b). Studies on the systemic vascular endothelium have established that endothelial cell signaling is subject to circadian regulation (Rudic et al., 2005; Shang et al., 2016; Facer-Childs et al., 2019). However, it remains unclear whether and how the circadian clock regulates lung endothelial function during inflammation. The pulmonary vascular endothelium is a key gatekeeper of inflammation in various lung diseases. Thus, uncovering the role of the circadian clock in pulmonary endothelial function may reveal novel insights into the inflammatory mechanisms underlying lung pathology. In this context, understanding the interplay between circadian rhythms and endothelial signaling is essential.

We have previously demonstrated that the circadian clock regulates lung inflammation during influenza infection (Sengupta et al., 2019). In other work, we have also reported that the lung endothelium responds to endotoxin exposure by assembling the endothelial NADPH oxidase 2 (NOX2) leading to the production of reactive oxygen species (ROS) (Lee et al., 2014; Orndorff et al., 2014). We showed that NOX2-driven ROS activates a signaling cascade that promotes pulmonary inflammation (Orndorff et al., 2014; Tao et al., 2016) and this process is amplified by NLRP3 inflammasome (Paul et al., 2024). However, the role of the circadian clock in regulating these inflammatory pathways remains unknown.

We hypothesized that pulmonary endothelial signaling is under circadian control via a NOX2 signaling pathway. To test this, we examined (a) the impact of inflammation on circadian rhythms in the lung endothelium and the effect of NOX2 blockade and (b) whether the heightened vascular inflammation observed in clock-disrupted mice is driven by NOX2 through the NLRP3 inflammasome pathway. Our findings reveal that inflammatory signaling disrupts circadian rhythms in the pulmonary endothelium. Furthermore, we find that an intact circadian clock mitigates inflammatory lung damage through endothelial signaling.

Materials And Methods

Materials

LPS (O111:B4), DPI, and apocynin were from Sigma-Aldrich (St Louis, MO, USA). MCC950 was from InvivoGen, San Diego, CA, USA). Antibodies against ICAM-1 and NLRP3 inflammasome were purchased from BD Biosciences (Franklin Lakes, NJ, USA). Anti-Ly6G (for PMN) was from ThermoFisher Scientific (Waltham, MA, USA). Anti-PECAM-FITC and DiIAcLDL were from BioLegend (San Diego, CA, USA) and Biomedical Technologies (Stoughton, MA, USA), respectively. Dynabeads pre-labeled with sheep anti-rat IgG were from Dynal (Oslo, Norway). Collagenase was from ThermoFisher Scientific (Waltham, MA, USA). Pierce BCA protein assay kit was from Thermo Fisher Scientific (Rockford, IL, USA).

ELISA kits were from the following: NLRP3: [LS BioSciences Inc. (Seattle, WA, USA)], Collagen: [Antibodies-Online (Aachen, Germany)], Myeloperoxidase (MPO): [ThermoFisher Scientific (Waltham, MA, USA)].

Mice: mPer2-luciferase on C57Bl6/J background was a kind gift of J Takahashi, UTSW, TX, USA. Bmal1^–/– mice were on a C57Bl6 background (Zhang et al., 2021b) CRY 1/2 double knockout mice (Cry1/2^–/–) were a gift from Katja Lamia, originally from Aziz Sancar (Thresher et al., 1998). Age and background-matched wild-type (WT) littermates were used as control. All animal studies were approved by the University of Pennsylvania Institutional Animal Care and Use Committee (IACUC) and met the stipulations of the Guide for the Care and Use of Laboratory animals.

Methods

Lung explants and isolation of pulmonary artery from mPer2-luciferase mice: Mice were anesthetized, and a tracheostomy was performed. Lungs were ventilated and perfused in Krebs-Ringer buffer to remove blood. Perfusion-cleared lungs were excised from the chest cavity to obtain lung explants. In separate experiments, a section of the pulmonary artery was removed from the excised lungs. For this, the lungs were dissected with micro-scissors (to remove superficial tissue) using a stereomicroscope (Leica Microsystems, Buffalo Grove, IL, USA), and the main pulmonary artery was removed. The adventitia was separated from the isolated arteries, and extreme care was taken to avoid any damage or stretching.

Isolation of pulmonary microvascular endothelial cells (PMVEC): Endothelial cells were isolated from the lungs of these mice as described by us previously (Chatterjee et al., 2006; Browning et al., 2014). Briefly, lungs from mPer2-Luc mice were removed and the tissue trimmed at the periphery. The trimmings were treated with collagenase (3 mg/ml) for digestion and then incubated with monoclonal antibody against platelet endothelial cell adhesion molecule (PECAM-1). This was followed by the addition of magnetic beads (Dynabeads, Dynal, Oslo, Norway) pre-labeled with sheep anti-rat IgG. The beads were separated on a magnetic rack, washed and plated on culture plates. Isolated endothelial cell islands were obtained in 1 to 2 weeks. A second round of immunoselection was carried out by flow sorting cells positive for the endothelial marker, platelet endothelial cell adhesion molecule (PECAM). PECAM⁺ cells were separated by sorting the isolated cells for PECAM-1 by anti-PECAM-FITC. The sorted cells were grown and the endothelial phenotype of the preparation was confirmed by evaluating cellular uptake of the endothelial-specific marker such as DiI-acetylated low-density lipoprotein (DiIAcLDL) and immunostaining for PECAM-1.

Monitoring of Circadian Oscillations: Pulmonary microvascular endothelial cells (PMVEC) of mPer2Luc were treated with 50 µg LPS and seeded in 35 mm dishes (at 80% confluence). Lung explants and pulmonary arteries were excised from perfused lungs for each mouse. Arteries were carefully cleaned from surrounding connective tissue. 4 to 5 mm pieces of explant or segments of 5 mm cut from each artery, were placed in tissue culture dishes and treated with LPS. Real-time oscillations of the circadian gene Per2 promoter activity were assessed by recording the bioluminescence (Lumicycle32, ActiMetrics) as described previously (Ramanathan et al., 2012). Data were analyzed using LumiCycle software (ActiMetrics).

Endotoxin Instillation: Mice (WT, Bmal1^–/– or Cry1/2^–/–) were anesthetized by intraperitoneal (IP) injection of ketamine (100 µg/g body wt), xylazine (4 µg/g) and acepromazine (1 µg/g). LPS (5 mg/kg per mouse; serotype O111:B4) dissolved in 50 μl sterile 0.9% NaCl was instilled intratracheally (IT) followed by 0.15 ml of air. LPS was instilled at ZT7. Immediately after instillation, the mice were kept upright for 10 min to allow the fluid to spread throughout the lungs. Mice were euthanized at various time points (6-72 h) after LPS instillation. Mice that received PBS were designated as control or naïve.

NOX2 and NLRP3 inflammasome inhibitors: For in vitro studies, PMVEC were pretreated with NOX2 inhibitor, diphenylene iodonium (DPI) (10 µM) for 1 h (Chatterjee et al., 2012). However, DPI may affect organs in vivo; thus, for in vivo studies another NOX2 inhibitor, apocynin (APO, 30 mg/kg weight) was administered IP (Kim et al., 2012). To block NLRP3 signaling, MCC950, a specific small molecule inhibitor of NLRP3 inflammasome, was administered (IP) at 40 mg/kg (Zhang et al., 2021a). Mice were pretreated for APO or MCC950 for 3 h followed either by sacrifice (naive +APO/MCC950) or by LPS instillation.

Measurement of ROS: ROS was monitored in LPS (50 µg/ml) and vehicle treated PMVEC by labeling cells with a ROS sensitive dye CellROX as described earlier (Chatterjee et al., 2019, 2021a). Briefly, PMVEC were incubated for 30 min with 5 μM Cell ROX and imaged by epifluorescence microscopy using a Nikon TMD epifluorescence microscope equipped with a Hamamatsu ORCA-100 digital camera and Metamorph imaging software (Universal Imaging, West Chester, PA, USA). Images were acquired at λ_ex = 488 nm. All images were acquired with the same exposure and acquisition settings as reported previously (Noel et al., 2013; Chatterjee et al., 2021a; Paul et al., 2024).

Immunofluorescence staining (ICAM-1, NLRP3, and PMN): Mice were sacrificed at various time points post LPS administration, lungs perfused, excised, and fixed in 4% paraformaldehyde. This was followed by dehydration by sequential sucrose treatment (10%, 20%, and 30% sucrose). The tissue was embedded in OCT blocks, and longitudinal sections were immunostained for ICAM-1 (1:250), NLRP3 (1:500), and lymphocyte antigen 6 complexes (Ly6G) (1:75) using anti-ICAM, anti-NLRP3, and NIMP antibodies, respectively. Secondary antibodies used were coupled to Alexa 488.

Fluorescence Imaging: For ROS, images were acquired on a Nikon EpiFluorescence (Nikon Diaphot TMD, Melville, NY, USA) microscope. For ICAM-1, NLRP3 and Ly6G, lung sections were imaged on a Leica TCS SP8 Confocal equipped with LASX software for image acquisition. Fluorescence controls of non-immune IgG and no-antibody-treated samples were used to obtain acquisition settings. Images were acquired at excitation laser line of λ_ex 488 nm. All images were acquired at the same settings and the fluorescence quantified over 3 to 4 fields using either Image J (National Institutes of Health) or Metamorph Software (Molecular Devices, Downington, PA, USA). For NLRP3 and ICAM-1, the field was scanned for pulmonary vessels and the fluorescent signal along the vessel perimeter (vascular wall) integrated using outlining tools of the Software as reported earlier (Noel et al., 2013; Paul et al., 2024). Colocalization of ICAM-1 and NLRP3 with the endothelial layer was carried out by co-staining lung sections for endothelial marker PECAM-1 (using anti-PECAM antibody) and NLRP3 or ICAM-1 ( Supplemental Figures 1, 2 ). However, co-stained sections were not employed to quantify fluorescence intensity to avoid the inclusion of bleed-through fluorescent signals (Waters, 2009; J and Waters, 2019). Fluorescence (ROS, NLRP3, ICAM-1) was quantitated from single color fluorescence images as arbitrary fluorescence units and normalized to area of the vessel perimeter/cells. This allowed for normalization of fluorescent units to the area where vessels or cells are present.

PMN accumulation. This was quantified in two ways: (a) Number of fluorescently labeled (PMN) cells/field obtained from automated counting of large dots in the field and (b) Activated PMN obtained from total fluorescence (emanating from the PMN) of the entire field. Image acquisition and analysis were carried out blindly.

ELISA of lung lysate: NLRP3, MPO and Collagen were quantified by ELISA. Lungs perfused and cleared of blood were minced in protein extraction reagent. The tissue was homogenized (on ice) and centrifuged at 5000 × g for 5 to 10 min. The supernatant was assayed for both the moiety of interest and protein content (Pierce BCA protein assay kit standardized to bovine serum albumin (BSA) according to the manufacturer’s protocol). All samples were run in triplicate. Values were normalized to protein content.

Lung Histology: Following euthanasia, lungs were removed, fixed in 10% formalin and stained with hematoxylin and eosin (H&E) in accordance with the protocol previously described (Pourfathi et al., 2018; Fisher et al., 2021). Slides were assessed for lung injury as described by us earlier (Pourfathi et al., 2018; Siddiqui et al., 2019; Pourfathi et al., 2020; Fisher et al., 2021). Briefly, for each mouse lung, 2 to 3 axial planes were assessed. Digital images were acquired with an Aperio Image Scope (Leica Biosystems, Buffalo Grove, IL) and were visualized at magnifications ranging from 1× to 60×. Injury scores were given based on image assessment (at 60×) by computationally deriving 10 fields from each axial plane. The parameters for injury assessment were: a. cellular infiltration on a scale of 0 to 2 with 0 = no infiltrate, 1 = infiltrate in the perivascular compartment, and 2 = infiltrate in perivascular and alveolar compartments); b. proteinaceous alveolar exudate, using a scale of 0 to 5 (as a percentage of the area involved; 0 = no exudate; 5 = exudate in 80% of alveolar area).

RNA Sequencing: Reads were aligned using STAR (Dobin et al., 2013) v2.7.6a to the GRCm38 genome with Ensembl v102 annotations. Reads were quantified and normalized using PORT (https://github.com/itmat/Normalization) v0.8.5e-beta. Differential expression between genotypes was tested using limma-voom (Law et al., 2014) v3.44.3. Pathway enrichment was assessed using Ingenuity Pathway Analysis (QIAGEN Inc., https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis).

Statistics: For all in vivo and in vitro experiments at least 3 independent biological experiments were performed. For each assay, multiple replicates of the same sample were assessed. Data were expressed as mean ± SD and one-tailed paired t-tests were used to determine statistical significance.

Results

Nox2 derived ROS regulates circadian rhythmicity

Inflammatory Stimulus (LPS) Alters the Circadian Rhythm in Lung Tissues, Pulmonary Arteries, and Pulmonary Endothelial Cells

We determined the effect of sterile pulmonary inflammation on circadian rhythms by exposing explants from freshly harvested lungs (of mPer2-luciferase mice) to LPS. Prior to this, circadian rhythms in the explants were synchronized with dexamethasone (Dex). Addition of LPS to lung explants altered the circadian rhythm as determined by the bioluminescence reading over several days (Figure 1a). Analysis of the rhythm showed that LPS significantly increased the circadian amplitude although the period was not affected (Figure 1b and 1c). Next, to determine if inflammation affected the circadian rhythms in pulmonary vessels, pulmonary arteries were dissected out of these mice and then exposed to LPS. Following LPS treatment, the circadian rhythms of the pulmonary arteries (Figure 1d) showed a significant increase in amplitude and decrease in period (Figure 1e and 1f).

Figure 1.

LPS alters the circadian rhythm in synchronized lung explants and pulmonary microvascular endothelial cells (PMVEC). (a) Representative graphs of bioluminescence data from PER2::LUC tissue explants monitored in culture for 12 days. (b) Circadian Period with and without LPS treatment. (c) LPS increases the amplitude of rhythms. Data are mean ± SD of N = 3 independent experiments. For each experiment, N = 3 technical replicates were performed. *p < 0.05 as compared to control. (d) LPS alters the circadian rhythm in dissected pulmonary artery that is synchronized with Dex. Representative graphs of bioluminescence data from PER2::LUC pulmonary artery with LPS and without LPS (Control or CTL). (e, f) LPS alters the circadian period and amplitude of rhythms. Data are mean ± SD of N = 4 independent experiments. For each experiment, N = 3 technical replicates were performed. *p < 0.05 as compared to control. (g) LPS-induced circadian rhythm in PMVEC is NADPH oxidase 2(NOX2)-dependent. PMVEC were isolated from the lungs of Per2Luc mice. Cells were expanded and endothelial lineage characterized to obtain primary cultures. Naïve cells and LPS-treated cells that are synchronized show circadian oscillation as monitored from the luminescence of the reporter. LPS induces alteration in circadian rhythmicity of synchronized PMVEC. A phase shift and an increased amplitude are induced with LPS treatment of PMVEC. Treatment with NADPH oxidase 2 inhibitor DPI resulted in restoration of the amplitude. (h) LPS decreases the period and (i) LPS treatment increases the amplitude of rhythms which are restored with DPI treatment. Data are mean ± SD of N = 3 independent experiments. For each experiment, N = 3 technical replicates were performed. *p < 0.05 as compared to control. (j) LPS treatment causes ROS production in Per2Luciferase PMVEC. Cells were labeled with ROS sensitive dye Cell ROX (5 µM) for 20 min and imaged at λ_ex 488 nm by epifluorescence microscopy. Scale bar = 10 µm. (k) The fluorescence intensity of the images was integrated and normalized to field area. After a gradual increase until 36 h, ROS levels decrease to baseline values. ROS production post-LPS is significantly high as compared to untreated controls and DPI pretreated cells. *p < 0.05 as compared to LPS + DPI and DPI alone values. Data are Mean ± SD of N = 3-4 independent experiments. For each experiment, 4 to 5 fields were assessed. *p < 0.05 as compared to LPS + DPI and DPI.

Blocking NOX2 Activation Restores Circadian Rhythms in Pulmonary Endothelial Cells

Given the critical role of the endothelia in pulmonary vascular function, we were particularly interested in studying the circadian rhythms of the endothelial cells. Thus, we isolated endothelial cells from mPer2Luc mice and expanded them into large primary cultures (henceforth referred to as pulmonary microvascular endothelial cells or PMVECs). We observed that (following synchronization) PMVEC showed robust circadian rhythms at baseline, which were altered in the presence of LPS (Figure 1g). We reported earlier that LPS exposure led to the activation of NADPH oxidase 2 (NOX2), which, in turn, causes ROS production in PMVEC (Lee et al., 2013; Lee et al., 2014; Orndorff et al., 2014) and that blocking NOX2 assembly reduced ROS (Paul et al., 2024). To investigate if NOX2-ROS alters the circadian rhythms of the PMVEC, we blocked NOX2 using the NOX2 assembly inhibitor diphenylene iodonium (DPI) (Figures 1g-i). Pre-treating PMVEC with DPI 1 h prior to LPS treatment restored the amplitude of the circadian cycle to control (naïve) values. However, DPI treatment further decreased the period (Figure 1h). Interestingly, increased ROS production and its blockade paralleled the altered amplitude of the circadian rhythm and its restoration to baseline values, respectively (Figure 1j and 1k). Overall, these results support the role of NOX2 in affecting circadian rhythms of the pulmonary endothelium following inflammation.

The clock regulates NOX2-dependent signaling

Circadian Regulation of ICAM-1 and PMN

Given the effect of NOX2 on circadian rhythmicity, we next investigated how the clock affects NOX2 signaling using two models of clock disruption. These involved the positive and negative limbs of the core clock, that is, Bmal1^–/– and Cry1/2^–/–, respectively. We investigated how disruption of the circadian clock affects NOX2 signaling, specifically intercellular adhesion molecule-1 (ICAM-1) expression and polymorphonuclear leukocyte (PMN) accumulation at the pulmonary endothelium in vivo. ICAM-1 expression was quantified along the vessel wall (Orndorff et al., 2014).

In WT mice, LPS instillation led to an initial increase in ICAM-1 (0-24 h) and a decrease thereafter (24-48 h) (Figure 2a). This pattern was absent in Bmal1^–/– and Cry1/2^–/– mice, where ICAM-1 increased with LPS exposure but did not decline after 48 h. Interestingly, even at baseline, both clock mutants showed significant ICAM-1 expression in blood vessels (Figure 2b). Next, we checked the effect of NOX2 blockade (using APO) on LPS-treated wild-type and naïve clock mutants and observed abrogation of ICAM-1 [Figures 2b (rightmost column) and 2c]. The ICAM-1 expression was predominantly along the vessel wall as noted by the colocalization of ICAM-1 in PECAM-1 positive layer and along the nucleated cells of the lumen ( Supplemental Figure 1A-D ).

Figure 2.

Circadian regulation of Intercellular Adhesion Molecule (ICAM-1) with LPS. (a) LPS (administered i.t.) causes an increased expression of ICAM-1 along the vessel wall. Mice were treated with LPS and sacrificed at 6 to 72 h later. Lung sections were immunostained with anti-ICAM antibody. Representative images of cross sections of vessels and magnified inset of vessel wall are shown. Scale bar = 50 µm. Right: Quantitation of ICAM-1 expression from fluorescence intensity normalized to vessel wall area. (b) Circadian regulation of ICAM-1 is NOX2-dependent. In WT lungs, LPS led to an initial increase with endothelial ICAM-1 (24 h) followed by a decrease to baseline after 48 h. Bmal1^–/– and Cry1/2^–/– showed high ICAM-1 at baseline that did not alter with LPS. APO (rightmost column) and shows abrogation of ICAM-1 by APO (inhibitor of NOX2 activation). WT were administered APO 3 h prior to LPS administration. APO treatment of naïve Bmal1^–/– and Cry1/2^–/– (which show high baseline ICAM-1) led to reduction of ICAM-1. Cross-sections of vessels (as outlined by dotted lines) were visualized and imaged. Representative fluorescent images are shown. Magnified inset shows endothelial cells (arrow) along the vessel wall. Scale bar = 50 µm. (c) Quantitation of ICAM-1 expression from fluorescence intensity. The intensity of green signal along the vessel wall was quantitated over several sections and normalized to vessel area using Image J software. Data are Mean ± SD for n = 4 WT mice, n = 4 Bmal1^–/– and n = 3 Cry1/2^–/– mice. For each mouse, at least 4 to 5 sections were immunostained and imaged. Within each section 2 to 3 fields were imaged to obtain an average for each section. *p < 0.05 between groups.

PMN (and other immune cells such as monocytes) that adhered/extravasated into lung tissue were imaged (Figure 3a) and assessed by either (a) counting the number of positive signals using Ly6G stained lung sections (Figure 3b) or (b) total fluorescent signaling emanating from the sections to represent activated PMN (Figure 3c). Total PMNs in WT lungs were low in naïve lungs but increased significantly following LPS instillation. In contrast, naïve Bmal1^–/– and Cry1/2^–/– exhibited significantly higher baseline PMN levels compared to WT, but showed no further increase in response to LPS.

Figure 3.

Circadian regulation of Polymorphonuclear Neutrophils (PMN) accumulation in murine lungs is NOX2-dependent. (a) Representative images of PMN in lung sections. Lungs of mice treated with LPS were sectioned and immunostained for PMN (Ly6G) by anti-Ly6G6C antibody (also called NIMP). Secondary antibody was tagged to Alexa 488 (green). Rightmost column: In separate experiments, APO (inhibitor of NOX2 activation) was administered 3 h prior to LPS in WT mice; in Bmal1^–/– and Cry1/2^–/– APO was administered to naïve mice. Scale bar = 40 µm. (b) PMN accumulation or number of PMN (from the fluorescent green dots that represent PMN) per field was quantitated over several sections using Image J Software. (c) Total PMN activation captured via integrated fluorescence Intensity of Ly6G+. *p < 0.05 between groups. For B and C. Data are Mean ± SD for n = 4 WT mice, n = 4 Bmal1^–/– and n = 3 Cry1/2^–/–. For each mouse, at least 4 to 5 sections were immunostained and imaged. 3 fields were imaged per section to obtain an average for each section. Images were quantified by semi-automated image analysis that counts particles. *p < 0.05 between groups.

These findings suggest that circadian clock disruption elevates baseline inflammation while impairing the host’s capacity to respond to additional inflammatory stimuli. NOX2 blockade with apocynin (APO) reduced PMN accumulation following LPS and also suppressed the elevated PMNs in naïve clock mutant lungs (Figure 3a, rightmost column), indicating that the heightened baseline inflammation in clock-deficient lungs is mediated via the NOX2 pathway.

Circadian Regulation of Lung Inflammatory Injury and Remodeling

We assessed myeloperoxidase (MPO) that is released by PMN. MPO generates free radicals that directly damage the lungs (Lee et al., 2014; Fisher et al., 2021). As shown in Figure 4a, MPO increased following LPS instillation in WT lungs. In contrast, both Bmal1^–/– and CRY1/2^–/– lungs had higher baseline MPO (Cry1/2^–/– was significantly higher, p < 0.01) compared to WT with no significant increase after LPS exposure. Next, histological analyses of lung tissue 24 h post-LPS revealed inflammatory injury characterized by immune cell infiltration alveolar damage, and thickened septa (Figure 4b). Semi-quantitative scoring of cellular infiltration and alveolar damage (Table 1) showed that WT mice developed acute inflammatory injury after LPS, which was partially resolved by 48 h. In comparison, lungs of Bmal1^–/– and Cry1/2^–/– mice displayed inflammatory infiltrates at baseline, with minimal additional response to LPS.

Figure 4.

Circadian regulation of pulmonary injury and remodeling. (a) Myeloperoxidase (MPO) in WT, Bmal1^–/– and Cry1/2^–/– lungs. WT, Bmal1^–/– and Cry1/2^–/– mice treated with LPS i.t. 48 h post LPS, mice were sacrificed, lungs were cleared of blood and lung tissue were homogenized and homogenate centrifuged; the supernatant thus obtained was measured for MPO by ELISA according to manufacturer’s instructions. #p < 0.01 as compared to their respective control. (b) H&E staining of lung sections from WT, Bmal1^–/– and Cry1/2^–/– mice 0,24 and 48 h post LPS instillation. Magnified inset is shown in the next column. The arrowhead indicates alveolar septal thickening. The arrow indicates neutrophils in the interstitial space and the waved arrow indicates neutrophils in the alveolar space. Scale bar = 50 µm. (c) Collagen as measured by ELISA in homogenate from lungs of WT, Bmal1^–/– and Cry1/2^–/– mice up to 48 h post LPS instillation. Control represents mice instilled with PBS and sacrificed 24 h later. #p < 0.01 as compared to control.

Table 1.

Average lung injury scores for the different groups. Results are the mean ± SD for n = 3 each of WT, Bmal1^–/– and Cry1/2^–/–.

Samples	Cellular Infiltration	Alveolar Exudate
Wild type (Naïve)	1.2 ± 0.6	1.8 ± 0.6
Wild Type + LPS (24 h)	2.05 ± 0.4*	3.6 ± 1.0*
Wild Type + LPS (48 h)	1.6 ± 0.3	2.8 ± 0.8
Bmal1–/– (Naïve)	1.9 ± 0.3	2.9 ± 1.0
Bmal1–/– + LPS (24 h)	2.1 ± 0.5	2.2 ± 1.7
Bmal1–/– + LPS (48 h)	2.4 ± 0.8	2.6 ± 1.0
Cry1/2 –/– (Naïve)	3.0 ± 1.1	4.2 ± 0.8
Cry1/2–/– + LPS (24 h)	2.5 ± 1.0	3.4 ± 1.2
Cry1/2–/– + LPS (48 h)	2.1 ± 0.9	3.4 ± 1.6

^*p < 0.05 as compared to naïve of the same genotype.

Deposition of Collagen I, a marker of lung remodeling, was quantified to further characterize tissue injury in response to LPS. As shown in Figure 4c, collagen I in the lungs of WT peaked within 24 h following lung injury and resolved over 48 h, indicating a well-regulated resolution phase in tissue response. In contrast, naïve Bmal1^–/– mice had significantly more collagen I in their lungs than WT with no significant change after LPS treatment.

Overall, these findings confirm that the disruption of the circadian clock results in heightened lung injury and remodeling at baseline. In WT mice, the clock enables a dynamic, time-sensitive response to inflammatory injury—characterized by an initial increase in inflammatory and remodeling signals followed by timely resolution. This adaptive response is absent in clock-deficient mice. Thus, the circadian clock plays a critical role in regulating NOX2-induced pulmonary inflammation, injury, and remodeling.

Transcriptomic profiling of lungs from clock-disrupted mice also supports the heightened state of inflammation

Next, to test the underlying mechanisms for this increased inflammation in the clock mutant mice, even at baseline, we performed transcriptomic profiling of lungs of Bmal1^–/– to that of its WT littermates (harvested between ZT 1 and 2). We found 1181 differentially expressed genes (DE) between the two—of which, 627 genes were upregulated in the Bmal1^–/– lungs and 554 genes were upregulated in WT mice ( Supplemental Table 1 ). As expected, several circadian genes were differentially expressed in these two groups (Nr1d1, Nr1d2, Ror1, Rorc, Cry1, Cry2, Per2, Per3). We also noted that chemokines, Cxcl3 and Cxcl5, were significantly upregulated in the Bmal1^–/– lungs relative to the WT lungs. This could explain the preponderance of PMNs in naïve Bmal1^–/– lungs versus WT lung. Furthermore, several genes associated with NLRP3 synthesis and regulation were differentially regulated between the WT and Bmal1^–/– (Table 2). Specifically, a circadian gene, Nr1d1 [that has been reported to control NLRP3 expression directly (Pourcet et al., 2018)] and other NLRP3 regulating genes viz. Nek7, (He et al., 2016) Rras2 (Wu et al., 2020), CTSB (Bai et al., 2018) and Adamts8 (Lu et al., 2021) were differentially expressed in the Bmal1^–/– versus WT. Furthermore, genes that are key regulators of endothelial inflammation such as Ndrg1 (N-myc downstream-regulated gene 1) are significantly downregulated in the Bmal1^–/– compared to WT (Zhang et al., 2023). Interestingly, pathway analyses based on these DE genes included both those that suggested both immune activation (“Neutrophil degranulation,” “Integrin-cell surface integration” and ”acute phase response”) as well as lung remodeling (“Pulmonary Fibrosis Idiopathic Signaling Pathway”) (Figure 5). Overall, our studies show that both the inflammatory and lung remodeling phenotypes seen in the clock mutant mice are strongly correlated with underlying changes in their transcriptome.

Table 2.

Differentially expressed genes between lung tissue of Bmal1^–/– and WT littermates that regulate the NLRP3 inflammasome.

Gene	Gene Symbol	Function
nuclear receptor subfamily 1, group D, member 1	Nr1d1	Governs circadian rhythmicity of the NLRP3 inflammasome pathway
Never in Mitosis A-related kinase 7	Nek7	Mediator of NLRP3 inflammasome activation
RAS Related 2	Rras2	Induces NLRP3 subunit expression
The cathepsin B	CTSB	Regulates NLRP3 expression
ADAM metallopeptidase with thrombospondin type 1 motif 8	Adamts8	Activates NLRP3
N-myc downstream-regulated gene 1	Ndrg1	Regulates endothelial inflammation and vascular remodeling and engages in crosstalk with the NLRP3 inflammasome
C-X-C motif chemokine ligand 5	Cxcl5	Promotes a pro-inflammatory environment that is also associated with NLRP3 activation
C-C motif chemokine ligand 21	Ccl21a	Expression correlated with NLRP3 activation

Figure 5.

Differentiated regulatory pathways of Bmal1^–/– versus WT lung tissue based on transcriptomic profiling of lung tissue from the naïve Bmal1^–/– and WT littermates.

The clock regulates NOX2-induced pulmonary endothelial inflammation via the NLRP3 inflammasome

Based on our transcriptomic data, we hypothesized that NOX2-derived ROS modulates circadian rhythms in pulmonary endothelium by regulating NLRP3 inflammasome. We assessed NLRP3 inflammasome activation by monitoring the expression of the NLRP3 subunit along the pulmonary vessel wall, as we have previously demonstrated its role in driving pulmonary endothelial inflammation l (Paul et al., 2022, 2024).

In naïve WT lungs, the expression of NLRP3 along the vessel wall was minimal but rose sharply following LPS treatment (Figure 6a), peaking at 36 h and returning to baseline at 48 h (Figure 6b). Similarly, NLRP3 levels in lung homogenates increased significantly in WT mice at 36 h post-but remained unchanged in the Bmal1^–/– and Cry1/2^–/– lungs (Figure 6c). Comparable responses were observed in vessel wall staining as NLRP3 expression increased post-LPS in WT. However, this was not observed in clock-deficient mice (Figure 6d-e).

Figure 6.

Circadian regulation of NLRP3. (a) Representative images of expression of the NLRP3 subunit along lung vessel wall. Mice treated with LPS (intratracheal instillation) were sacrificed at 6 to 72 h and lungs cleared of blood, fixed and sectioned. Sections were immunostained with anti-NLRP3 antibody. Secondary antibody was tagged to Alexa 488. Inset is magnified to show the endothelial layer (white arrow). Scale bar = 50 µm. (b) Quantitation of NLRP3 along the vessel wall. Data are Mean ± SD of three independent (N = 3-4 mice) experiments. For each mouse, 5 sections were immunostained and imaged. For each section, 4 to 5 fields were assessed. Arbitrary fluorescence units were normalized to the vessel area. Scale bar = 50 µm. (c) NLRP3 expression in whole lung homogenate as measured by ELISA. The concentration is normalized to lung protein content. (d) Circadian regulation of the NLRP3 inflammasome is NOX2-dependent. NLRP3 expression in WT, Bmal1^–/– and Cry1/2^–/– mouse lungs. Representative fields of fluorescence (NLRP3) as imaged by confocal microscopy. White arrow shows the endothelial layer. Rightmost column shows the effect of APO treatment on expression of NLRP3. Scale bar = 50 µm. (e) Quantitation of the NLRP3 subunit of the inflammasome. Data are Mean ± SD for n = 4 WT mice, n = 4 Bmal1^–/– and n = 3 Cry1/2^–/– mice. For each mouse, 4 sections were immunostained; 3 fields were imaged for each section. *p < 0.05 between groups as indicated.

Notably, NOX2 blockade with apocynin (APO) significantly suppressed LPS-induced NLRP3 expression in WT lungs and reduced elevated baseline NLRP3 levels in clock mutants (Figure 6d, rightmost column) indicating that NOX2 activity drives NLRP3 expression and that this regulation is circadian clock–dependent. We also noted that NLRP3 subunit expression was largely along the endothelial layer ( Supplemental Figure 2A-D ) of the lumen.

To further examine this pathway, we inhibited NLRP3 using MCC950 (Coll et al., 2015) and evaluated its effect on ICAM-1 expression and PMN accumulation. NLRP3 blockade markedly reduced both ICAM-1 (Fig. 7a,b) and PMN infiltration (Fig. 7c,d) in WT lungs. In clock-deficient mice, MCC950 also reduced elevated baseline ICAM-1 and PMN levels, confirming that the heightened inflammation in the absence of the clock is NLRP3-dependent.

Figure 7.

The circadian regulation of ICAM-1 expression and PMN accumulation is NLRP3-dependent.

Naïve and LPS-treated WT, and naïve Bmal1^–/– and Cry1/2^–/– mice were administered with MCC950 (40 mg/kg) and evaluated for ICAM-1 and PMN. Lung sections were immunostained for A. ICAM-1 and C. PMN. Scale bar = 50 µm. B. Total ICAM-1 and D. PMN activation captured via integrated fluorescence intensity of ICAM-1 and Ly6G+, respectively. Data are expressed as Mean values ± SD of n = 4 WT, n = 4 Bmal1^–/– and n = 3 Cry1/2^–/– mice. For each mouse, 4 to 5 sections were immunostained and average fluorescence intensity normalized to field area imaged. *p < 0.01, #p < 0.05 as compared to MCC950.

Together, these findings demonstrate that the circadian clock regulates NOX2-driven pulmonary inflammation through the NLRP3 inflammasome.

Discussion

Inflammation itself can directly affect circadian rhythmicity; conversely, the circadian clock is known to regulate inflammation (Putker and O’Neill, 2016). The importance of the clock in inflammation was evident more than half a century ago when studies showed that the mortality rate of mice exposed to the bacterial endotoxin (lipopolysaccharide, LPS) depended on the time of exposure (Halberg et al., 1960; Feigin et al., 1969, 1972; Shackelford and Feigin, 1973), that is, LPS challenge given at the end of the rest time resulted in a mortality rate of 80%, while that given in the middle of the active time resulted in a mortality rate of only 20% (Halberg et al., 1960). However, this relationship is opposite for the lung—wherein infection at the beginning of the active period (dusk) resulted in 3-fold higher mortality than infection at the beginning of the rest period or dawn (Sengupta et al., 2019; Issah et al., 2021).

This indicates that circadian regulation of lung inflammation is distinct from systemic inflammation. However, a comprehensive understanding of circadian regulation in pulmonary inflammation requires examining its various cellular components. To date, most research has focused on immune, epithelial cells (Wang et al., 2016; Issah et al., 2021; Oyama et al., 2022) as well as alveolar cells. These investigations have revealed that the clock in epithelial and alveolar cells control diurnal variation of pulmonary inflammatory responses (Oyama et al., 2022). However, the pulmonary endothelium—despite its central role in inflammatory signaling—has received limited attention.

The enzyme NADPH oxidase (NOX2) is essential for production of ROS which is crucial entity in initiating inflammation signaling specifically in pulmonary endothelial cells. While link between the NOX2 and the clock has been reported in microglial cells (that participate in the inflammatory responses) (Muthukumarasamy et al., 2023), the complex relationship between a functional or disrupted clock, NOX2 and inflammation is not clear. Our study provides novel insights into the mechanisms by which the clock regulates pulmonary endothelial inflammation, highlighting a previously unexplored mechanism involving the NOX2-NLRP3 signaling axis.

In this study, we demonstrated the presence of a functional circadian clock in the pulmonary endothelium and evaluated its role in modulating pulmonary inflammation. We used two distinct genetic models of clock disruption—Bmal1^–/– and Cry1/2^–/– to investigate the link between circadian disruption and pulmonary endothelial inflammation signaling.

Our data on the in vitro PMVEC model demonstrated that inflammatory stimulation with endotoxin alters circadian rhythms. Inhibition of NOX2-derived ROS using DPI, restored circadian amplitude but shortened the period—likely due to the circadian clock’s sensitivity to ion flux alterations induced by DPI. NOX2-derived ROS modulates the amplitude of circadian rhythms in pulmonary endothelium.

Other studies have also shown that cellular clocks are susceptible to ROS (Wang et al., 2012; Tamaru et al., 2013), though the specific source of ROS was not identified. Given NOX2’s influence on circadian rhythmicity, we next investigated the reverse: whether the circadian clock regulates downstream NOX2 signaling.

Understanding this bidirectional relationship between NOX2-ROS signaling and the circadian clock has significant implications. It reveals a potential feedback loop in which inflammation disrupts circadian rhythms and vice versa, compromising the temporal regulation of inflammatory responses. In the context of lung disease, where tightly regulated inflammation and its resolution are essential to prevent excessive injury, this crosstalk provides a compelling target for therapeutic intervention. By restoring circadian alignment or targeting ROS signaling pathways in a time-sensitive manner, we may improve outcomes in inflammatory lung disorders and other diseases characterized by circadian disruption and oxidative stress. Here, we hypothesized that the circadian clock modulates NOX2-driven inflammation. To test this, we evaluated several indices of inflammation including ICAM-1 along vessel wall, PMN accumulation, and markers of lung injury/remodeling. Notably, we use ICAM-1 expression along the endothelium, since this is a more specific marker of inflammation than total ICAM-1 content, as it reflects active translocation of the molecule to the vascular surface in response to inflammatory stimuli (Nourshargh and Alon, 2014; Orndorff et al., 2014). Our findings show that NOX2 regulation of ICAM-1 is lost when core circadian genes are deleted, consistent with previous reports of circadian control of ICAM-1 in systemic inflammation (Gao et al., 2014; He et al., 2018).

Similarly, NOX2-dependent PMN accumulation was also lost with clock deletion which elevated ICAM-1 and PMNs at baseline and did not increase further with LPS. Interestingly, NOX2 blockade with APO normalized this heightened baseline inflammation, suggesting that NOX2 signaling is normally restrained by rhythmically expressed clock proteins. Without the oscillations in clock proteins, NOX2 activity remains constitutively upregulated, thereby leading to persistent inflammation, both at baseline and in response to stimuli.

Beyond acute inflammation, we found that clock genes are essential for regulating tissue damage and remodeling. In WT mice, LPS-induced inflammatory injury and collagen 1 deposition peaked and resolved in a time-dependent manner-reflecting a normal tissue repair dynamic. However, this pattern was absent in Bmal1^–/– and Cry1,2^–/– mice, where high levels of MPO and Collagen I persisted, underscoring a failure of resolution. These findings suggest that the circadian clock not only modulates the onset of inflammation but also orchestrates its resolution and tissue repair. We have previously shown the role of the lung epithelial clock in tissue repair in a flu model (Naik et al., 2023). However, the current study establishes another regulatory node in the pulmonary vascular endothelium.

Our results extend previous work linking NOX2 activity to NLRP3 inflammasome activation injury (Leszczynska et al., 2022; Paul et al., 2022, 2024) by demonstrating that this axis is under circadian control in the pulmonary endothelium. While the role of BMAL1 and CLOCK in regulating NLRP3 in immune cells has been reported in a few studies ((Pourcet et al., 2018; Wang et al., 2018; O’Siorain et al., 2024)), its role in the endothelial cells was previously unexplored. Our transcriptomic and histological analyses revealed that NLRP3 expression in WT lung vessels followed a dynamic, time-sensitive pattern—minimal at baseline, peaking at 36 h post-LPS and declining by 48 h. This pattern mirrored the inflammatory and remodeling response. In contrast, Bmal1^–/– and Cry1/2^–/– mice displayed high baseline NLRP3 expression that failed to resolve, indicating a breakdown of temporal control. Furthermore, the sequential reduction of inflammation following NLRP3 blockade further supports its pivotal role downstream of NOX2 in driving endothelial inflammation.

Interestingly, while inflammation altered circadian rhythms in lung explants and endothelial cells, rhythmicity was not abolished—consistent with previous findings in endotoxemia models (Haspel et al., 2014). This suggests that inflammatory signals modulate but do not eliminate the molecular clock. The observed increase in rhythm amplitude post-LPS may represent an adaptive mechanism facilitating resolution. In contrast, in circadian-deficient mice, elevated inflammation and injury persisted throughout the experimental window, indicating that clock integrity is essential for recovery.

Collectively, our findings reveal a critical role for the circadian clock in modulating pulmonary endothelial inflammation through the NOX2-NLRP3 axis. Clock genes appear to sequester pro-inflammatory signaling, enabling a balanced immune response and timely resolution. In their absence, unchecked NOX2-NLRP3 activation leads to sustained inflammation and impaired repair. These insights expand our understanding of circadian biology in lung disease and suggest that targeting circadian-regulated inflammatory pathways may offer novel therapeutic avenues for inflammatory lung disorders (Schema, Figure 8).

Figure 8.

Schema showing the molecular cross talk between circadian clock and NOX2-NLRP3 axis in pulmonary endothelial inflammation. Increased NOX2-NLRP3 initiates pulmonary inflammation but the circadian clock sequesters this axis leading to resolution of inflammation (as monitored by ICAM-1 and PMN). In the absence of the circadian clock, NOX2-NLRP3 remains high and thus inflammation cannot be controlled.