25-Hydroxycholesterol exacerbates vascular leak during acute lung injury

Abstract

Cholesterol-25-hydroxylase (CH25H), the biosynthetic enzyme for 25-hydroxycholesterol (25HC), is most highly expressed in the lung, but its role in lung biology is poorly defined. Recently, we reported that Ch25h is induced in monocyte-derived macrophages recruited to the airspace during resolution of lung inflammation and that 25HC promotes liver X receptor-dependent (LXR-dependent) clearance of apoptotic neutrophils by these cells. Ch25h and 25HC are, however, also robustly induced by lung-resident cells during the early hours of lung inflammation, suggesting additional cellular sources and targets. Here, using Ch25h-/- mice and exogenous 25HC in lung injury models, we provide evidence that 25HC sustains proinflammatory cytokines in the airspace and augments lung injury, at least in part, by inducing LXR-independent endoplasmic reticulum stress and endothelial leak. Suggesting an autocrine effect in endothelium, inhaled LPS upregulates pulmonary endothelial Ch25h, and non-hematopoietic Ch25h deletion is sufficient to confer lung protection. In patients with acute respiratory distress syndrome, airspace 25HC and alveolar macrophage CH25H were associated with markers of microvascular leak, endothelial activation, endoplasmic reticulum stress, inflammation, and clinical severity. Taken together, our findings suggest that 25HC deriving from and acting on different cell types in the lung communicates distinct, temporal LXR-independent and -dependent signals to regulate inflammatory homeostasis.

Keywords: Cell stress; Inflammation; Innate immunity; Mouse models; Pulmonology.

Figure 1. Ch25h deletion reduces microvascular leak and cytokines in an acute lung injury model.

(A) BAL total leukocytes (WBCs) were quantified at various time points following a high-dose (3 mg/mL) LPS aerosol exposure (n = 4–9 per genotype per time point). (B–D) BALF protein (B), albumin (C), and IgM (D) were quantified at the indicated times following high-dose LPS aerosol (n = 4–6 per genotype per time point). (E) BALF cytokines were quantified by multiplex assay 24 hours after high-dose LPS aerosol (n = 5 per genotype). (F and G) Mice (n = 4–8 per genotype) underwent lung infection with K. pneumoniae by oropharyngeal aspiration. BALF protein (F) and cytokines (G) were quantified 24 hours after infection. (H) BALF protein was quantified in WT and LXR-null mice (n = 5 per genotype per time point, repeated twice) 48 hours after high-dose LPS inhalation. Data are mean ± SEM and are representative of 2–3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 by unpaired 2-tailed t test.

Figure 2. 25HC induces endothelial activation and injury.

(A) Serum angiopoietin-2 was measured in mice at the indicated time points relative to high-dose LPS aerosol inhalation (n = 13–14 per genotype). (B) Lung tissue from Ch25h^+/+ and Ch25h^–/– mice was analyzed by quantitative PCR (qPCR) for the targets shown following LPS inhalation (n = 4–5 per genotype). (C) Lung homogenates were evaluated by immunoblot for VE-cadherin and β-actin (loading control) in Ch25h^+/+ and Ch25h^–/– mice at baseline and 48 hours or 72 hours after high-dose LPS aerosol inhalation. Immunoblot results from independent mice are shown in the left panel, and densitometry for actin-normalized VE-cadherin signal in the right panel (n = 3–8 per condition). The dashed line indicates juxtaposition of nonadjacent portions of the original gel. (D) Lung tissue from WT mice was analyzed by qPCR for Vcam1 following i.p. treatment with 25HC or vehicle (n = 10 per treatment). (E) Mouse pulmonary microvascular endothelial cells were cultured for 24 hours in 50% medium/50% BALF (0 or 48 hours after inhaled LPS) collected from mice of the indicated genotype and then analyzed by qPCR for Vcam1. *P < 0.05, **P < 0.01 by unpaired 2-tailed t test. FC, fold change.

Figure 3. Ch25h in non-hematopoietic cells regulates lung injury.

(A–C) Mice chimeric for Ch25h expression in hematopoietic and/or non-hematopoietic cells were produced by bone marrow transfer from Ch25h^+/+ (WT) or Ch25h^–/– (KO) mice into WT or KO recipients following sublethal irradiation (graphs show donor > recipient). Chimeras were then exposed to inhaled LPS. At 72 hours after exposure, protein (A) and albumin (B) were measured in BALF, and 25HC was measured in BALF, lung, or serum (C) (n = 4–10 per condition, representative of 2 independent experiments). (D) Endothelial cells (CD45^–EpCAM^–CD31⁺) were FACS-sorted from mouse lung 72 hours after saline or LPS inhalation. RNA was prepared and qPCR performed for Ch25h (n = 2 saline, n = 6 LPS). In A–C, all chimeras with statistically significant differences in 25HC compared with WT>WT are indicated with asterisks. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 compared with WT>WT by 1-way ANOVA with Dunnett’s post hoc test. FC, fold change.

Figure 4. 25HC decreases transendothelial resistance.

(A and B) Human pulmonary artery endothelial cells (HPAECs) were cultured in medium with 2% FBS supplemented with vehicle or the indicated concentrations of 25HC (A) or cholesterol (CH; a negative control) (B), and transendothelial electrical resistance (TER) measurements were performed over a 70-hour time period. (C) HPAECs were incubated with vehicle or heat-killed S. aureus (HKSA; 2 × 10⁸ particles/mL) for 30 minutes followed by treatment with vehicle or 25HC (20 μM) and TER monitored over 70 hours. (D) HPAECs were incubated with vehicle or LPS (100 ng/mL) for 30 minutes followed by treatment with vehicle or 25HC (20 μM) and TER monitored for 70 hours. Data are mean ± SEM and are representative of 2–3 independent experiments. The number of replicates per condition is given in the figure. Intercurve differences at 60 hours were analyzed by 1-way ANOVA with all pairwise post hoc comparisons and Tukey’s adjustment. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 5. CH25H/25HC induces ER stress in the injured lung.

(A) Lung tissue from Ch25h^+/+ and Ch25h^–/– mice was analyzed by qPCR for Chop 72 hours after LPS inhalation (n = 10 per genotype). (B) Lung tissue from WT mice was analyzed by qPCR for the indicated targets after i.p. treatment with 25HC or vehicle (n = 8–10 per condition). (C) Mouse pulmonary microvascular endothelial cells were cultured for 24 hours in 50% medium/50% BALF (0 or 48 hours after inhaled LPS) collected from mice of the indicated genotypes and then analyzed by qPCR for the indicated targets. (D) Mice of the indicated genotypes were treated with 100 mg/kg 4-phenylbutyric acid (PBA) or PBS vehicle (Veh.) i.p. at –1, +6, and +24 hours in relation to LPS inhalation and then had BALF IgM and albumin quantified at 72 hours after LPS (n = 15–18 per condition). (E) Mice were treated as in D and then had the indicated cytokines quantified in BALF (n = 15 per condition). Data are mean ± SEM and are representative of 2–3 independent experiments. ^ΨP = 0.06, *P < 0.05, **P < 0.01, ***P < 0.001 by unpaired 2-tailed t test. FC, fold change.

Figure 6. AM CH25H expression and 25HC BALF concentrations are associated with higher BALF total protein and von Willebrand factor levels.

AM CH25H mRNA (quantified by microarray) (n = 30) and BALF 25HC levels (quantified by mass spectrometry) (n = 81) were measured in patients within 48 hours of ARDS diagnosis from a therapeutic trial of omega-3 fatty acids. Depicted are the individual values and linear regression lines showing associations between AM CH25H expression and (A) total BALF protein and (B) von Willebrand factor levels, as well as between BALF 25HC levels and (C) total BALF protein and (D) von Willebrand factor levels. The estimates and P values are adjusted for age, sex, treatment group, and APACHE II severity of illness scores.

Figure 7. AM CH25H and airspace 25HC track with lung inflammation and injury in human ARDS.

AM CH25H mRNA (quantified by microarray) (n = 30) and BALF 25HC (quantified by mass spectrometry) (n = 81) were measured in patients within 48 hours of ARDS diagnosis from a therapeutic trial of omega-3 fatty acids. (A and C) Multiple linear regression, adjusted for age, sex, APACHE II score, and treatment group, was used to test the relationship between normalized log₂ CH25H probe intensity (A) or log₂ BALF 25HC concentration (C) and the Berlin definition of ARDS severity as assessed by P_aO₂/F_IO₂ (P/F) ratio (P/F = 200–300: mild; P/F = 100–200: moderate; P/F < 100: severe). (B and D) A t test was used to compare normalized log₂ AM CH25H probe intensity (B) or log₂ 25HC BALF concentration (D) in patients with mild-to-moderate versus severe ARDS as assessed by the Murray Lung Injury Score (LIS) definition (LIS = 0.1–2.5: mild-to-moderate; LIS > 2.5: severe). Depicted are the individual patient values, mean, and SD.