An endothelial TLR4-VEGFR2 pathway mediates lung protection against oxidant-induced injury

Abstract

TLR4 deficiency causes hypersusceptibility to oxidant-induced injury. We investigated the role of TLR4 in lung protection, using used bone marrow chimeras; cell-specific transgenic modeling; and lentiviral delivery in vivo to knock down or express TLR4 in various lung compartments; and lung-specific VEGF transgenic mice to investigate the effect of TLR4 on VEGF-mediated protection. C57/BL6 mice were exposed to 100% oxygen in an enclosed chamber and assessed for survival and lung injury. Primary endothelial cells were stimulated with recombinant VEGF and exposed to hyperoxia or hydrogen peroxide. Endothelium-specific expression of human TLR4 (as opposed to its expression in epithelium or immune cells) increased the survival of TLR4-deficent mice in hyperoxia by 24 h and decreased LDH release and lung cell apoptosis after 72 h of exposure by 30%. TLR4 expression was necessary and sufficient for the protective effect of VEGF in the lungs and in primary endothelial cells in culture. TLR4 knockdown inhibited VEGF signaling through VEGF receptor 2 (VEGFR2), Akt, and ERK pathways in lungs and primary endothelial cells and decreased the availability of VEGFR2 at the cell surface. These findings demonstrate a novel mechanism through which TLR4, an innate pattern receptor, interacts with an endothelial survival pathway.

Keywords: HALI; TLR4; VEGF; lung injury.

Figure 1.

The role of compartment-specific TLR4 expression in protection against hyperoxic injury. Mice were exposed to continuous hyperoxia, and their survival proportions were compared. A) Bone marrow hematopoietic progenitors were transferred from WT or TLR4ko mice to their respective recipients. (WT→WT and TLR4→WT curves represent 10 and 26 mice, respectively. WT→TLR4ko and TLR4ko→TLR4ko curves represent 6 mice each.) *P = 0.0142; **P = 0.009. B) Progeny of lung epithelium-specific human TLR4 transgenic (CC10TLR4tg) mice crossed with TLR4ko mice were exposed to continuous hyperoxia. (Each curve represents 10 mice in 2 experiments.) C) TLR4ko mice were bred to Tie-2-hTLR4 transgenic mice, and their progeny were exposed to continuous hyperoxia (Each graph shows 10–12 mice from 2 experiments.) P = 0.0026. D) TLR4ko mice received vascular-specific hTLR^int lentivirus (lentivTLR4) or control (lenti-vCTRL) and were exposed to hyperoxia. (Each curve corresponds to 10 mice from 2 experiments.) *P = 0.0298. E) WT mice were infected with vascular-specific TLR4 shRNA (lenti-vTLR4sh), ubiquitin-TLR4shRNA (lenti-TLR4sh), or control lentivirus (lenti-vCTRL). (Each curve represents 8 or more mice from 2 separate experiments.) Data are expressed as means ± sem. *P = 0.0054; **P = 0.0017.

Figure 2.

Endothelial TLR4 protects lung against oxidant-induced injury. A–D) WT, Tie2-TLR4 transgenic (TLRtg), TLR4ko, and TLR4 reconstituted mice (TLR4ko+TLR4tg) were exposed to hyperoxia for 72 h. Protein extravasation (A) and LDH (B) were measured in their BAL fluids (n = 6 in each group). A) *P = 0.033, **P = 0.0014; B) *P = 0.0055, **P = 0.00766. C) Representative images of lung sections of mice stained by TUNEL for apoptotic cells. D) The number of TUNEL⁺ cells is expressed as the percentage of the total number of lung cells counted (n = 6 in each group). *P = 2.4 × 10⁻⁸; **P = 1.4 × 10⁻⁷. E–H) WT and TLR4ko mice received lenti-vTLR4 or lenti-vCTRL and were exposed to hyperoxia for 72 h. The level of LDH in their BAL fluids (E) and cellular apoptosis in their lungs (F) were measured as described in (B) and (D) (n = 6 or more from 2 experiments). E) *P = 0.012, **P = 0.0086; F) *P < 0.000001, **P = 0.00015. G) Representative images of H&E-stained lung sections at ×400 magnification show increased neutrophils, areas of atelectasis, and increased neutrophils in the interstitial space (arrowheads), alveolar disruption with occasional filling of the alveolar space (red circle) and alveolar septal thickening (arrows). H) Lung injury scores for the lungs described in (G) (n = 6 mice per group; 20 microscopic fields assessed per mouse). All values are expressed as means ± sem. *P = 0.015; **P = 0.03.

Figure 3.

The effect of TLR4 knockdown on VEGF-mediated lung protection. WT and VEGF transgenic (VEGF TG) mice received lentiviral siRNA against TLR4 (TLR4si) or control siRNA (Ctrlsi) and were exposed to hyperoxia (HO) or room air (RA) for 72 h. A, B) Protein levels (A) and LDH (B) were measured in the BAL fluids (n = 8 mice from 2 experiments). A) *P < 0.04, **P < 0.02; (B) *P = 0.03, **P = 0.005. C) Representative images of TUNEL staining on the lungs of mice exposed to hyperoxia. D) Apoptotic cells were quantified as described in Fig. 2D. Data are expressed as means ± sem (n = 6 mice from two experiments). *P < 2 × 10^–6; **P < 9 × 10^–7).

Figure 4.

The role of endothelial TLR4 in regulating VEGF-mediated protection in the lung. VEGF transgenic (VEGF TG), TLR4ko (TLR4^−/−) and VEGF transgenic/TLR4ko mice (VEGFTG/TLR4ko) received lenti-vTLR4 or lenti-vCTRL vectors. VEGF expression was induced by adding doxycycline to the drinking water, and mice were exposed to 72 h of hyperoxia. A, B) Protein levels (A) and LDH (B) were measured in the BAL fluid. C) Apoptotic cells were quantified by TUNEL staining as described in Fig. 2D) (Each column represents 10 or more mice from 2 experiments.)

Figure 5.

TLR4 protects endothelial cells against oxidant-induced injury. A) MLECs from WT and TLR4ko mice were transduced with lenti-vTLR4 or lenti-vCTRL vectors, stimulated with VEGF(10 ng/ml), and exposed to hyperoxia. Apoptosis was determined by staining for annexin V and flow cytometry (n = 6 from two experiments). *P = 0.0017; **P = 0.042; ***P = 1.2 × 10–8; ****P = 0.00052. B) HUVECs were transfected with TLR4 siRNA or control siRNA, stimulated with recombinant human VEGF (10 ng/ml), and exposed to H₂O₂. Apoptosis was determined as in (A) (n = 6 in two experiments). Data are expressed as means ± sem. *P = 3.7 × 10⁻⁵; **P = 0.00057; ***P = 0.001627; ****P = 2.3 × 10⁻⁵.

Figure 6.

The role of TLR4 effect in VEGF signaling in endothelial cells. A) MLECs were transfected with TLR4 siRNA or control (Ctrl) siRNA, starved for 24 h, and stimulated with recombinant human VEGF (10 ng/ml). Cell lysates were harvested at the indicated time points, and activated and total protein fraactions were visualized by Western blotting with the corresponding antibodies. β-Actin antibody was used as the protein loading control. B) Total and activated VEGFR2 levels in the lungs of VEGF transgenic animals were visualized by Western blot as described in (A). C) HUVECs were treated as in (A) and harvested after 10 min to assess the phosphorylation of VEGFR2 Tyr¹¹⁷⁵.

Figure 7.

The effect of TLR4 on VEGFR2 trafficking. A) Endothelial (CD31⁺, CD45⁻) VEGFR2 levels in the lungs of WT, VEGFtg, TLR4ko, and VEGFtg/TLR4ko animals were measured by flow cytometry (n = 6 in each group). *P = 0.00354; **P = 0.000479. B) HUVECs were treated as in Fig. 6A, stained with anti-VEGFR2 antibody, and examined by flow cytometry. The graph shows the mean fluorescence intensity (MFI) (n = 6 from 2 experiments). Data are means ± sem.*P = 1.423 × 10⁻⁶; **P = 2.697 × 10⁻¹⁰; ***P = 1.925 × 10⁻⁷.