Attenuation of pulmonary ACE2 activity impairs inactivation of des-Arg⁹ bradykinin/BKB1R axis and facilitates LPS-induced neutrophil infiltration

Abstract

Angiotensin-converting enzyme 2 (ACE2) is a terminal carboxypeptidase with important functions in the renin-angiotensin system and plays a critical role in inflammatory lung diseases. ACE2 cleaves single-terminal residues from several bioactive peptides such as angiotensin II. However, few of its substrates in the respiratory tract have been identified, and the mechanism underlying the role of ACE2 in inflammatory lung disease has not been fully characterized. In an effort to identify biological targets of ACE2 in the lung, we tested its effects on des-Arg⁹ bradykinin (DABK) in airway epithelial cells on the basis of the hypothesis that DABK is a biological substrate of ACE2 in the lung and ACE2 plays an important role in the pathogenesis of acute lung inflammation partly through modulating DABK/bradykinin receptor B1 (BKB1R) axis signaling. We found that loss of ACE2 function in mouse lung in the setting of endotoxin inhalation led to activation of the DABK/BKB1R axis, release of proinflammatory chemokines such as C-X-C motif chemokine 5 (CXCL5), macrophage inflammatory protein-2 (MIP2), C-X-C motif chemokine 1 (KC), and TNF-α from airway epithelia, increased neutrophil infiltration, and exaggerated lung inflammation and injury. These results indicate that a reduction in pulmonary ACE2 activity contributes to the pathogenesis of lung inflammation, in part because of an impaired ability to inhibit DABK/BKB1R axis-mediated signaling, resulting in more prompt onset of neutrophil infiltration and more severe inflammation in the lung. Our study identifies a biological substrate of ACE2 within the airways, as well as a potential new therapeutic target for inflammatory diseases.

INTRODUCTION

Lung inflammation is an important pulmonary response to a variety of stimuli, both microbial and noninfectious (38). Although pneumonias, sepsis, aspiration, and trauma may all cause lung inflammation, the common instigating events remain unclear (42, 54). Lung inflammation is a common consequence of exposure to pathogen-associated molecular patterns such as endotoxin lipopolysaccharide (LPS). Indeed, LPS binding to myeloid differentiation factor 2 (MD-2) and activation of Toll-like receptor 4 (TLR4) play an important early role in the development and progression of inflammation in a variety of tissues and cells. In the lung, this is accompanied by neutrophil recruitment and injury of the airway and alveolar epithelium, leading to protein transudation (20, 39). Acute LPS exposure is sufficient to induce respiratory symptoms (e.g., airflow obstruction and increased airway hyperreactivity), inflammation (e.g., expression and secretion of proinflammatory cytokines), and neutrophil infiltration (27, 68).

Angiotensin I-converting enzyme 2 (ACE2) is a metallomonocarboxypeptidase that is generally considered a component of the renin-angiotensin system (RAS; 14, 23, 24). The peptidase cleaves a single-terminal residue from several bioactive peptides including angiotensin II (ANG II), neurotensin, dynorphin A (1-13), apelin-13, and des-Arg⁹ bradykinin (DABK) (15, 70). The functions of ACE2 in the lung have received attention in recent years because of the discovery that this ectoenzyme is also the receptor for the severe acute respiratory syndrome (SARS) and NL63 human coronaviruses (CoVs; 43). In addition, loss of pulmonary ACE2 function is thought to play an important role in the respiratory disease associated with acid aspiration, sepsis, and severe infections (including SARS-CoV and influenza viruses; 29, 40, 76, 80). Penninger and colleagues (29, 40) reported that loss of ACE2 catalytic function in association with infection or injury perturbs the pulmonary RAS (11), increasing inflammation and vascular permeability because of activation of the ANG II/angiotensin II type 1 receptor (AT1) axis. However, other biological substrates of pulmonary ACE2 and their contributions to inflammation are largely unknown and need to be addressed.

The kinin-kallikrein system (KKS) constitutes the precursor kininogen, the proteolytic kallikrein enzymes, and effector peptides bradykinin (BK-1-9 or BK) and its active metabolite [des-Arg⁹]-BK (BK-1-8 or DABK; 9, 50). These peptides recognize two pharmacologically distinct G protein-coupled receptors: the B1 receptor (BKB1R), whose main agonist is DABK; and the B2 receptor (BKB2R), whose ligand is BK (48, 49). BKB1R is a heptahelical protein and is distinct from its BKB2R counterpart in that its expression is extremely sensitive to inflammatory mediators such as LPS and interleukins (2, 49, 51, 61). In BKB1R-deficient mice, LPS-induced hypotension is blunted, and the number of polymorphonuclear leukocytes that accumulate in inflamed tissues is abnormally low (8, 11). Indeed, the BKB1R may be an important pharmacological target for the treatment of inflammatory disorders affecting the lungs.

Here, we provide the first evidence that DABK is a substrate of pulmonary ACE2 in vivo and that attenuation of ACE2 activity leads to impaired DABK inactivation and thus to enhanced BKB1R signaling. We also show that these effects coincide with release of the proinflammatory chemokine C-X-C motif chemokine 5 (CXCL5) and with neutrophil recruitment to the lung.

MATERIALS AND METHODS

Reagents

Unless indicated otherwise, reagent-grade chemicals were obtained from Sigma Chemical (St. Louis, MO).

Cell Culture

Well-differentiated airway epithelial cell cultures.

Primary airway epithelial cells were isolated from trachea or bronchi from human donors and cultured at the air-liquid interface on collagen-coated polycarbonate filters as described previously (36). All preparations used were well differentiated (>4 wk old; resistance >1,000 Ω·cm²). This study was approved by the Institutional Review Board at the University of Iowa. To stimulate cells in vitro with endotoxin, lipooligosaccharide complexed with MD-2 protein (LOS:MD-2, 100 ng/ml) was applied to the apical surface of airway epithelia as reported previously (34).

Cell lines.

Human embryonic kidney 293 (HEK293) cells were obtained from American Type Culture Collection. All cells were maintained at 37°C in 5% CO₂. Cells were maintained in the following culture media: 10% FBS and 1% penicillin-streptomycin in DMEM.

Detection of mRNA Expression by Quantitative Reverse Transcriptase-PCR

Total RNA was isolated from cultured primary human airway epithelia or mouse lung tissue and reverse transcribed using QuantiTect Reverse Transcription Kit (Qiagen). The expression levels of human BKB1R or mouse ACE2 were determined relative to the housekeeping gene RPLO by quantitative real-time PCR using the Bio-Rad CFX96 real-time system as described previously (18). The following primers were used to detect respective mRNA levels: human BKB1R forward primer AGGCCAATTTGTTCATCAGC and reverse primer AGGCCAGGATGTGGTAGTTG. Primers for mouse ACE2 are as follows: forward, TCTGCCACCCCACAGCTT; reverse, GGCTGTCAAGAAGTTGTCCATTG. Human and mouse RPLO forward primer is as follows: forward, GGCGACCTGGAAGTCCAACT; reverse, CCATCAGCACCACAGCCTTC.

Lung Wet Weight-to-Dry Weight Ratio Measurement

Mouse lungs from indicated experimental groups were collected, and a portion of the lung was subjected to wet weight-to-dry weight ratio measurement. Briefly, mouse lungs were removed, and wet weights were measured, followed by placing the lung tissues in an oven at 65°C for 72 h. The dry weights of lung tissues were measured. The wet weight-to-dry weight ratio (W/D) was calculated to evaluate lung edema: W/D = wet weight (mg)/dry weight (mg).

Immunohistochemistry

Cells were washed with PBS and fixed in 4% paraformaldehyde for 5 min at room temperature. The cells were then again washed with PBS, after which 5% BSA in PBS was used to block nonspecific binding. An anti-BKB1R rabbit antibody (1:100 dilution, kind gift from Dr. Rejean Couture, University of Montreal) was applied to the cells (both sides in the case of air-liquid interface cultures), and after incubation at 37°C for 1 h the cells were washed with PBS. A secondary rabbit anti-mouse antibody (conjugated with FITC) was incubated overnight at 4°C. On day 2, cells were washed with PBS three times, and a cyanin 3 (Cy3)-labeled mouse anti-β-tubulin IV monoclonal antibody was applied for 1 h at 37°C. Finally, cells were washed with PBS three times and mounted with 4′,6-diamidino-2-phenylindole (DAPI; VectaShield). In a subset of samples, nuclei were stained with To-pro-3 (Molecular Probes). Immunofluorescent staining of lung tissues was performed on 4% paraformaldehyde-fixed, 5-μm-thick paraffin sections. The sections were first warmed to 56°C in a vacuum incubator (Isotemp Vacuum Oven, Fisher Scientific) and then washed immediately twice in xylene, gradually redehydrated in ethanol (100, 95, and 70% and water), and then processed for antigen retrieval in citrate buffer (10 mM, pH 6.0) and microwave (1,000 W, 6 min). Samples were then washed with PBS, blocked with 1% BSA-5% donkey serum (1 h, room temperature), then incubated overnight at 4°C with primary antibodies (1:200 dilutions in 0.5% BSA), washed three times with PBS, and incubated with appropriate fluorescently labeled secondary antibodies (1:1,000 dilution in 0.5% BSA; Life Technologies) as well as the nuclear marker DAPI (BioLegend), and slides were then mounted using Gelvatol (Sigma-Aldrich) solution before imaging using a Zeiss LSM 710 confocal microscope (Carl Zeiss, Jena, Germany) under appropriate filter sets. Antibodies used in mouse lung immunostainings are cleaved caspase-3 (rabbit anti-rat/mouse; Biocare Medical), DAPI (Life Sciences), and inducible nitric oxide synthase (iNOS, mouse anti-rat/mouse; BD Biosciences).

Measurement of Activator Protein-1 Activation

A commercially available pAP1-TA-Luc plasmid (Clontech, Palo Alto, CA) was used for transfection and as a template for generating recombinant adenovirus vector (Adv-AP1-Luc; 52). The fragment containing the firefly luciferase (9) gene, the minimal TATA-box promoter from the herpes simplex virus thymidine kinase gene (TA) upstream of it, and six tandem copies of the activator protein-1 (AP-1) enhancer (drive expression in response to treatment with phorbol ester) was released by XhoI and NotI double digestion. It was inserted into a promoterless adenoviral shuttle plasmid (pAd5mcspA), and Adv-AP1 virus was generated by homologous recombination as previously described (52) and stored in 10 mM Tris with 20% glycerol at −80°C. The particle titer of the adenovirus stock was determined by spectrophotometry (absorbance 260 nm); the functional titer was determined by plaque titration in HEK293 cells and the analysis of expression of the encoded protein. AP-1 luciferase activity was measured using a luciferase assay kit according to the manufacturer’s instructions (Promega). The readings, in relative light units (RLU), were normalized to the protein concentration in each tested sample.

Measurement of Catalytic Activity of ACE2

ACE2 activity in cell washes and bronchoalveolar lavage (11) fluid was determined by measuring the fluorescence intensity of ACE2 substrate Mca-Y-V-A-D-A-P-K(Dnp)-OH (cat. no. ES007; R&D Systems). In brief, cell surface washes and bronchoalveolar lavage fluid (BALF) were diluted in assay buffer [50 mM 2-(N-morpholino)ethanesulfonic acid (MES), 300 mM NaCl, 10 nM ZnCl₂, and 0.01% Brij 35, pH 6.5] and incubated with ACE2 substrate, with or without the ACE2 inhibitor DX600 (1 nM, cat. no. 62337; AnaSpec), at 37°C for 45 min. The fluorescence intensity was determined using a multiplate fluorescence reader at excitation wavelength of 330 nm and emission wavelength of 400 nm.

ELISA for Mouse ACE2 Protein Quantification

Sandwich ELISA analysis was performed according to manufacturer’s instructions (cat. no. IRKTAH5291; Innovative Research). Briefly, capture antibody was incubated on 96-well flat-bottomed plates overnight. Plates were washed and blocked with 5% BSA (1 h, room temperature), and samples were added to the plate, incubated overnight (4°C), washed extensively, and then incubated with biotinylated detection antibody (2 h, room temperature). Following washes, streptavidin-alkaline phosphatase was added to the wells. The enzymatic reaction was stopped after 30 min by the addition of an equal volume of 0.2 N sulfuric acid, and the color change was read on a spectrophotometer (450 nm; Molecular Dynamics). Data were normalized to total protein concentration and quantified using GraphPad Prism.

Endotoxin-Induced Lung Inflammation Model

C57BL/6 mice of 6–12 wk of age were purchased from the National Cancer Institute (Bethesda, MD). ACE2^−/− mice are a kind gift from Dr. Josef Penninger of the Institute of Molecular Biotechnology, Vienna, Austria. BKB1R^−/− mice were generously shared by Dr. Michael Bader of Max Delbrück Center for Molecular Medicine, Berlin, Germany. Both ACE2^−/− and BKB1R^−/− mouse lines are on C57BL/6 background. This study was approved by the institutional animal care and use committees at Johns Hopkins University and at the University of Iowa. Mice were anesthetized by intraperitoneal injection of ketamine-xylazine (2.5 mg/kg) and were then administered 50 µg of LPS (Escherichia coli 055: B5; Sigma) in 50 µl of normal saline by nasal instillation. DX600 (1 mg/kg) and [Leu⁸]-DABK (15 μg/mouse) were also administered by nasal instillation.

Bronchoalveolar Lavage

Mice were euthanized, the chest was opened by midline incision, and the lungs were lavaged in situ via polyethylene (PE)-90 tubing inserted into the exposed trachea. Lavage was performed with 0.5-ml volumes of sterile saline (total lavage volume, 4 ml/mouse). The recovered volumes of each BALF sample were recorded to normalize protein concentration in BALF. The lavage fluid was centrifuged for 10 min at 200 g, and the cell pellet was resuspended in 1 ml Hanks’ buffered salt solution. The cells were then counted and spun onto a glass slide. Diff-Quik stain was applied to the cells on the slides using a standard technique, and the percentage of neutrophils and macrophages was determined as described previously (53). Protein contents in BALF were determined by bicinchoninic acid assay (no. 23225; Pierce). Murine CXCL5, C-X-C motif chemokine 1 (KC), macrophage inflammatory protein-2 (MIP2), and TNF-α abundances were assessed by ELISA (R&D Systems) following the manufacturer’s instructions.

Statistical Analysis

Where indicated, data were analyzed for statistical significance by two-tailed Student’s t-test or analysis of variance (ANOVA) using GraphPad Prism software. Statistical significance was determined as having a P value of <0.05, and data are represented as means ± SE as indicated. All experiments were repeated at least twice, with at least four pups per group for assessment.

RESULTS

des-Arg⁹ Bradykinin Is an ACE2 Substrate in Airway Epithelia

DABK was shown to be a substrate of ACE2 in a biochemical assay; however, no evidence has confirmed that it is a biological substrate of ACE2 in the lung. To test this hypothesis, we assessed the effects of DABK on the activity of human BKB1R, using a luciferase assay measuring AP-1 activity as a surrogate. AP-1 activation is a well-established response to BKB1R signaling (4, 58, 59). Initially, we double-transfected HEK293 cells with vectors expressing AP-1 luciferase and human BKB1R or triple-transfected cells with AP-1 luciferase, human BKB1R, and human ACE2 expression vectors. We next stimulated cells with increasing concentrations of DABK and assessed luciferase activity. As shown in Fig. 1A, cotransfection of HEK293 cells with BKB1R and ACE2 significantly blunted the AP-1 luciferase activity induced by DABK, consistent with ACE2 reducing DABK activity. The observation was further supported by a second experiment in which DABK was applied with apical washes recovered from air-liquid interface cultures of human airway epithelial cells; we previously reported that such airway surface liquid (ASL) contains catalytically active soluble ACE2 (34). As shown in Fig. 1B, ASL inhibited DABK-induced AP-1 luciferase activation, and this effect was likely specific to ACE2 because it was ameliorated when the ACE2 inhibitor DX600 was added with DABK and ASL.

Fig. 1.

des-Arg⁹ bradykinin (DABK) is a substrate of angiotensin I-converting enzyme 2 (ACE2) in human airway epithelia. Luciferase assays were used to assess bradykinin receptor B1 (BKB1R) activity under several conditions. A: BKB1R activity in HEK293 cells in the presence of DABK (across a range of doses) with or without ACE2 (n = 5). HEK293 cells were transfected with plasmids expressing activator protein-1 (AP-1) luciferase and human BKB1R (B1), with or without a human ACE2 (hACE2) construct. RLU, relative light units. B: BKB1R activity under the same transfection conditions as in A. Additional interventions included stimulation with DABK alone, DABK application in the presence of airway surface liquid (ASL) collected from air-liquid interface cultures of human airway epithelial cells (contains ACE2), or application of DABK, ASL, and ACE2 inhibitor DX600 (n = 6). C: BKB1R activity in air-liquid interface (ALI) cultures of human airway epithelia. The cultures were transduced with an adenovirus expressing the AP-1 luciferase reporter at multiplicity of infection = 1. DABK was applied apically, either alone or in the presence of ACE2-neutralizing antibody (α-ACE2) or isotope goat IgG (n = 6). D: BKB1R activity in cultured airway epithelia as described in C [DABK (ALI)], but ACE2 abundance was reduced by covering air-liquid interface cultures with media for 7 days (DABK subm; n = 5). All comparisons are undertaken between control (CTL) and treated groups, unless otherwise indicated in the figure. *P < 0.05, **P < 0.01.

We next asked whether DABK could be cleaved by ACE2 in primary epithelial cells from the human airways. In this case, air-liquid interface cultures of human airway epithelia that had been transduced with an adenovirus expressing the AP-1 luciferase reporter (multiplicity of infection = 1) were treated apically with DABK (30 nM). As shown in Fig. 1C, the apically applied DABK caused little change in AP-1 luciferase activity. To explore the possibility that this failure of apically applied DABK to induce AP-1 luciferase activity was due to inactivation by ACE2, we applied an ACE2-neutralizing antibody to the apical side of the cells for 1 h before DABK stimulation. As shown in Fig. 1C, the addition of DABK in this context induced a significant increase in AP-1 luciferase activity; however, treatment of the cells with the antibody alone was ineffective, consistent with the notion that ACE2 derived from airway epithelia inactivates DABK. As a second approach to reduce the available ACE2, we covered the apical surface of the epithelia with culture medium for 7 days, an intervention known to cause the loss of ACE2 expression (33). We then applied DABK to the cells apically. As shown in Fig. 1D, this intervention significantly enhanced DABK-induced AP-1 luciferase activity. Collectively, these results from both transfected HEK293 cells and primary cultures of human airway epithelia suggest that DABK is a biological substrate of ACE2 in the airways.

ACE2 Activity Is Reduced by Endotoxin, Coincident with Neutrophil Infiltration

We selected LPS inhalation to further investigate ACE2 biology and its possible role in mediating inflammation. Mice received 50 µg LPS at time 0, and the catalytic activity of ACE2 was measured at the times indicated in the presence and absence of DX600 (Fig. 2A). Bronchoalveolar lavage (BAL) cell counts (Fig. 2B) and percentage of neutrophil and macrophage in BALF (Fig. 2C) were also determined. As shown in Fig. 2, B and C, the abundance of neutrophils in the BAL increased significantly within 6 h of LPS exposure. This coincided with a decline in ACE2 catalytic activity, reaching a nadir at the 6-h time point (Fig. 2A). Thus LPS-induced neutrophil infiltration is temporally associated with a reduction in the catalytic activity of ACE2.

Fig. 2.

Endotoxin inhalation reduces pulmonary angiotensin I-converting enzyme 2 (ACE2) activity and induces neutrophil infiltration into the lung. Mouse lung bronchoalveolar lavage fluid (BALF) measured for ACE2 enzymatic activity, plus or minus ACE2 inhibitor DX600 (A), bronchoalveolar lavage total neutrophil [polymorphonuclear neutrophils (PMN)] and macrophage (MAC) counts (B), and percentage of PMN and macrophages in BALF (C). RLU, relative light units. For each time point and group, n ≥ 5. *P < 0.05, **P < 0.01.

Endotoxin Reduces ACE2 Protein Abundance in the Lung

We next tried to determine the mechanism that underlies endotoxin exposure-associated ACE2 activity impairment. To test the possibility that LPS per se could interfere directly with ACE2 activity, we measured catalytic activity of 10 ng recombinant human ACE2 with or without coincubating with 10 μg LPS at 37°C for 60 min. As shown in Fig. 3A, endotoxin does not attenuate ACE2 activity per se, indicating an endotoxin-mediated regulation of ACE2 activity in the lung. To explore whether endotoxin could affect ACE2 abundance and thus impact ACE2 activity at the transcriptional level, we isolated RNA from mouse lung exposed to endotoxin for varying lengths of time and conducted quantitative reverse transcriptase-PCR (qRT-PCR) to measure ACE2 mRNA abundance. Surprisingly, as demonstrated in Fig. 3B, endotoxin does not reduce ACE2 mRNA abundance; instead, it induces ACE2 gene expression in the lung promptly. Next, we quantified ACE2 protein level in mouse lung with or without endotoxin inhalation and found that endotoxin reduces ACE2 protein abundance in a time-dependent manner within our experimental time frame (Fig. 3C), suggesting that endotoxin impairs ACE2 activity in the lung by attenuating ACE2 abundance translationally or posttranslationally. To explore the signaling pathway that mediates LPS-induced ACE2 alteration in the lung, we pretreated mice with or without a NF-κB antagonist, Bay11-7082 (50 μM, 50 μl), by inhalation, 30 min before LPS instillation through nose. As shown in Fig. 3, D and E, inhibiting NF-κB signaling by Bay11-7082 restored ACE2 activity (Fig. 3D) and protein abundance (Fig. 3E) in the lung, indicating that the endotoxin-induced ACE2 attenuation in the lung is, in part, NF-κB signaling dependent.

Fig. 3.

Endotoxin induces reduction of angiotensin I-converting enzyme 2 (ACE2) protein abundance in the lung. A: enzymatic activity of recombinant human ACE2 (rhACE2) protein alone (10 ng) or coincubated with 10 μg endotoxin at 37°C for 60 min was measured by fluorometric-based substrate assay (n = 5). N.S, not significant. B: ACE2 mRNA (mACE2) expression in mouse lung was detected by quantitative reverse transcriptase-PCR (qRT-PCR). Control (CTL) mouse lung or mouse lung exposed to LPS for various times as indicated was collected, and total RNA was extracted for qRT-PCR (n ≥ 5). C: ACE2 protein abundance in control mouse lungs or mouse lungs that received LPS through inhalation was measured by ELISA (n ≥ 5). D: ACE2 activity in mouse lungs repressed by LPS inhalation for 6 h was restored by Bay11-7082 (Bay11), an antagonist for NF-κB signaling (n ≥ 5). RLU, relative light units. E: representative micrograph of Western blot showing that Bay11-7082 partially blocked the reduction in ACE2 protein abundance in mouse lungs that inhaled LPS for 6 h. p-IKKα/β, phospho-IKKα/β. *P < 0.05, **P < 0.01.

A Reduction in Pulmonary ACE2 Activity Facilitates Neutrophil Infiltration

To further investigate the link between reduced ACE2 activity and neutrophil infiltration post-LPS exposure, we treated mice with ACE2 inhibitor DX600 intranasally before instilling LPS. As presented in Fig. 4B, DX600 pretreatment resulted in markedly increased neutrophil infiltration 3 h after LPS exposure, and the rapid neutrophil infiltration is in parallel with reduced ACE2 abundance in the lung (Fig. 4C), suggesting that functional ACE2 may modulate inflammatory cell influx. A similar result was also observed in ACE2 knockout mice, which showed a significant increase in neutrophil presented in BALF compared with wild-type counterparts when the mice were given LPS for 3 h (Fig. 4D). Of note, the influxes of neutrophil in ACE2^−/− and in wild-type pretreated with DX600 mouse lungs did not differ from those in wild-type mice from 6 to 72 h post-LPS challenge. Taken together, the results imply that lacking active ACE2 in the lung facilitates neutrophil infiltration induced by LPS exposure more promptly rather than more intensely. However, the more rapid neutrophil influx due to lack of active ACE2 at an early stage of endotoxin exposure has its consequences. As shown in Fig. 4, E–J, mouse lungs treated with ACE2 inhibitor before endotoxin inhalation exhibited more severe pathology as manifested by hematoxylin and eosin staining (Fig. 4E), enhanced lung permeability, and edema as evidenced by elevated protein content in BALF (Fig. 4F) and wet weight-to-dry weight ratio of the lungs (Fig. 4G). Moreover, proinflammatory cytokine iNOS and apoptosis marker cleaved caspase-3 immunohistochemistry (Fig. 4H) and cytokine levels in BALF (Fig. 4, I and J) are all further proofs of exaggerated lung inflammation in mouse lungs in this setting. In seeking to understand the underlying mechanisms, we found that pretreatment with ACE2 inhibitor before LPS exposure further boosted endotoxin-induced IKKα/β phosphorylation at 3 h post-LPS inhalation and the strengthened IKKα/β phosphorylation was faded at 6 h post-LPS inhalation (Fig. 4K), suggesting that inhibition of ACE2 facilitates endotoxin-induced NF-κB signaling transiently and exaggerates subsequent lung inflammation and injury.

Fig. 4.

Lack of pulmonary angiotensin I-converting enzyme 2 (ACE2) activity promotes neutrophil infiltration of the lung and exacerbates LPS-induced lung inflammation and injury. A: schematic depiction of treatment time lines. B: wild-type mice received the ACE2 inhibitor DX600 (1 mg/kg inhalation) both before and after LPS treatment via nasal instillation. Neutrophil infiltration of the lung at indicated times post-LPS inhalation was measured by polymorphonuclear neutrophil (PMN) counts in bronchoalveolar lavage fluid (BALF). CTL, control. C: ACE2 contents in mouse lung from the same experimental groups as in B are measured by ELISA. D: PMN numbers in BALF of wild-type (WT) and ACE2^−/− mice at an array of LPS exposure times were counted (n ≥ 5). E–G: inhibition of ACE2 activity exacerbates LPS-induced lung injury as manifested by representative images of hematoxylin and eosin staining (E), lung permeability (F), and edema (G). H–J: lack of active ACE2 enhances lung inflammation response to LPS stimulation as demonstrated by representative immunohistochemical images of inducible nitric oxide synthase (iNOS, cyan; 43) and cleaved caspase-3 (CC3, red) in the lung (H) and proinflammatory cytokine levels [C-X-C motif chemokine 1 (KC), I; and TNF-α, J] in BALF. K: impact of ACE2 inhibition on LPS-induced NF-κB activation illustrated by Western blot. W/D ratio, wet weight-to-dry weight ratio; DAPI, 4′,6-diamidino-2-phenylindole; p-IKKα/β, phospho-IKKα/β. In all experimental groups, n ≥ 5. Comparison is conducted as indicated in B–D and between the LPS-alone group and the LPS-plus-ACE2 inhibitor DX600 group (F–I). *P < 0.05, **P < 0.01, ***P < 0.001. Scale bar = 50 μm.

BKB1R Activity Contributes to the Neutrophil Recruitment Triggered by Endotoxin Exposure

The results in Fig. 1 indicated that DABK is a substrate of pulmonary ACE2, and those in Fig. 4 suggested that a reduction in ACE2 function contributes to LPS-induced neutrophil infiltration of the lung. We next tested the hypothesis that the attenuation of ACE2 activity facilitates neutrophil infiltration in part by reduced inactivation of DABK, and thus enhanced signaling by the BKB1R pathway. We first assessed BKB1R expression in air-liquid interface cultures of human airway epithelia. As demonstrated in Fig. 5, A–C, under resting conditions, BKB1R expression is low, at both the mRNA and protein levels. Following LPS (LOS:MD-2) stimulation, the levels of both BKB1R mRNA and protein were significantly increased (Fig. 5, A–C).

Fig. 5.

Endotoxin-induced des-Arg⁹ bradykinin (DABK)/bradykinin receptor B1 (BKB1R) activation contributes to pulmonary neutrophil infiltration. A–C: expression of BKB1R in air-liquid interface primary airway epithelial cell cultures, controls (CTL) or treated with lipooligosaccharide-myeloid differentiation factor 2 (LOS:MD-2). A: human BKB1R (hBKB1R) mRNA expression as assessed by quantitative reverse transcriptase-PCR (qRT-PCR), in the absence (CTL) or presence of LOS:MD-2 stimulation (n = 6). B: BKB1R protein expression as assessed by immunoblot. Cells from three different donors were used per treatment group (CTL or LOS:MD-2 treatments). C: expression of BKB1R in well-differentiated airway epithelia minus or plus LOS:MD-2 treatment. Blue, To-pro-3 (nuclei); red, β-tubulin; green, BKB1R. D–F: neutrophil infiltration of the lung with or without BKB1R antagonist treatment. D: bronchoalveolar lavage (BAL) neutrophil abundance in LPS-treated mice in the presence of BKB1R antagonist [Leu⁸]-DABK. BKB1R inhibition reduced neutrophil infiltration 6 h post-LPS instillation. E: BAL polymorphonuclear neutrophil counts in wild-type (WT) and BKB1R^−/− mouse lungs exposed to LPS for 6 h. N.S, not significant. F: neutrophil abundance in mouse lungs treated with BKB1R antagonist and/or angiotensin I-converting enzyme 2 (ACE2) inhibitor. Treatment with BKB1R antagonist reduces accelerated neutrophil infiltration 3 h post-LPS challenge in mice pretreated with ACE2 inhibitor DX600. For each experimental group, n = 6–9. **P < 0.01, ***P < 0.001. Scale bar = 50 μm.

To extend these in vitro results to an in vivo model of inflammation, we administered the BKB1R antagonist [Leu⁸]-DABK to mice 15 min before and 3 h after LPS stimulation and assessed the effects on BAL neutrophil abundance at 6 h following LPS exposure. As demonstrated in Fig. 5D, antagonism of BKB1R reduced LPS-induced neutrophilic infiltration, suggesting that the DABK/BKB1R axis contributes to LPS-induced neutrophil influx. The observation was further confirmed in BKB1R-null mice, in which LPS induces little or no neutrophil infiltration in the lung 6 h post-LPS inhalation (Fig. 5E). To further explore the possibility that attenuated ACE2 function leads to neutrophil infiltration via activation of the DABK/BKB1R axis, we pretreated wild-type mice with the BKB1R antagonist [Leu⁸]-DABK, followed by treatment with the ACE2 inhibitor DX600, and then instilled LPS. BAL neutrophil abundance 3 h post-LPS exposure was significantly lower in the group pretreated with the BKB1R antagonist (Fig. 5F). This finding suggests that reduced pulmonary ACE2 activity contributes to the neutrophil infiltration in response to LPS, at least in part because of a failure to counteract DABK/BKB1R signaling.

We next tested effects of blocking BKB1R on LPS-induced inflammation, measuring BAL neutrophil counts and body weight at various times post-LPS challenge. The three treatment regimens used are illustrated in Fig. 6A. Regardless of whether the BKB1R antagonist was applied pre- or post-LPS challenge, it significantly reduced weight loss (Fig. 6B) and BAL neutrophil counts (Fig. 6C). The protective effect of BKB1R blockade is associated with less reduced ACE2 abundance (Fig. 6D) and diminished IKKα/β phosphorylation at 24 h post-endotoxin exposure (Fig. 6E) in the lung, indicating a possibility that the inhibition of DABK/BKB1R signaling alleviates LPS-induced repression of ACE2 by blocking the NF-κB signaling pathway. The beneficial effects of blocking the DABK/BKB1R pathway in LPS-induced lung inflammation and injury are further evidenced by lessened lung pathology (Fig. 6F), alleviated lung permeability (Fig. 6G), reduced wet weight-to-dry weight ratio (Fig. 6H), extenuated apoptosis (Fig. 6I), and mitigated inflammatory response (Fig. 6, J–L).

Fig. 6.

Effects of bradykinin receptor B1 (BKB1R) antagonist treatment on acute lung injury induced by LPS exposure. A: time lines outline the experimental protocol. B: inhibiting BKB1R improved outcomes of LPS-mediated reduction of body weight (n ≥ 5). C: inhibiting BKB1R alleviated LPS-induced neutrophil infiltration in mouse lung (n ≥ 5). D: blocking BKB1R partially restored LPS-mediated reduction of angiotensin I-converting enzyme 2 (ACE2) protein content in mouse lung. E: blockade of BKB1R reduces LPS-induced NF-κB activation as assessed by Western blot. As indicated in separated lanes, lane 1, control (CTL) mouse lung; lane 2, mouse lung exposed to LPS for 24 h; lane 3, mouse lung that inhaled BKB1R inhibitor for 24 h; and lane 4, mouse lung treated with BKB1R inhibitor before LPS inhalation. The panels at top show phospho-IKKα/β (p-IKKα/β) protein expression; the panels at bottom display housekeeping protein β-actin expression. F–H: inhibition of BKB1R mitigated LPS-induced lung injury as manifested by representative images of hematoxylin and eosin staining (F), reduced lung permeability (G), and lung edema (H). I–K: lack of BKB1R activation alleviated lung inflammation response to LPS inhalation as demonstrated by representative immunohistochemical images of inducible nitric oxide synthase (iNOS; I) and cleaved caspase-3 (CC3, red; J) in the lung and proinflammatory cytokine levels [C-X-C motif chemokine 1 (KC), K; and TNF-α, L] in bronchoalveolar lavage fluid (BALF). W/D ratio, wet weight-to-dry weight ratio; DAPI, 4′,6-diamidino-2-phenylindole. In all experimental groups, n ≥ 5. Comparison analysis is conducted among LPS vs. trial A (*), LPS vs. trial B (+), and LPS vs. trial C (#). *+# P < 0.05, ** ++ P < 0.01, *** +++ ### P < 0.001. Scale bar = 50 μm.

Endotoxin Enhances, and BKB1R Blockade Attenuates, the Induction of CXCL5 Expression

In an effort to identify the mechanism that underlies activation of the DABK/BKB1R axis and neutrophil infiltration in response to LPS stimulation, we measured several neutrophil-recruiting chemokine/cytokine contents in BALF (Fig. 7A) and then focused on the role of CXCL5 in LPS-induced neutrophil recruitment in the lung. CXCL5 is an airway epithelium-derived chemokine involved in neutrophil recruitment following LPS exposure (10, 31). As shown in Fig. 7A, CXCL5 abundance in mouse BALF increased in a time-dependent manner after LPS stimulation and reached its highest levels 6 h post-LPS challenge. This correlates with the time when pulmonary ACE2 activity is at its lowest (Fig. 2A), suggesting a link between reduced ACE2 activity and CXCL5 induction by LPS. As previous reports indicated that activation of the DABK/BKB1R axis leads to enhanced production of CXCL5 (16, 17, 75), we next pretreated mice with the BKB1R antagonist [Leu⁸]-DABK and challenged them with LPS. As demonstrated in Fig. 7B, BKB1R antagonism was associated with reduced CXCL5 production in BALF after LPS treatment. As shown in Fig. 7C, we immunoneutralized CXCL5 levels with an antibody, and this decreased neutrophil infiltration in response to LPS.

Fig. 7.

Endotoxin enhances airway epithelial C-X-C motif chemokine 5 (CXCL5) expression, and bradykinin receptor B1 (BKB1R) blockade attenuates the induction. A: LPS-induced neutrophil-recruiting chemokine production in bronchoalveolar lavage fluid (BALF), as assessed by ELISA. MIP2, macrophage inflammatory protein-2; KC, C-X-C motif chemokine 1. B: LPS-induced CXCL5 production in BALF is inhibited by BKB1R antagonist, 6 h post-LPS challenge. CTL, control. C: effects of CXCL5 depletion with a neutralizing antibody (α-CXCL5) on LPS-induced neutrophil infiltration 6 h post-LPS challenge. PMN, polymorphonuclear neutrophils, MAC, macrophages, N.S, not significant. D: inhibition of ACE2 activity in airway epithelia enhances CXCL5 production in response to des-Arg⁹ bradykinin (DABK) treatment. DABK was applied to air-liquid interface cultured human airway epithelial cells apically for 6 h to activate BKB1R. Airway surface liquid ACE2 activity was inhibited by DX600. In all experimental groups, n > 6. *P < 0.05, **P < 0.01, ***P < 0.001.

To better understand the sequence of events that underlie LPS-induced neutrophil infiltration and the role of ACE2, we again used air-liquid interface cultures, adding DABK and DX600, individually and in combination, to the apical side of airway epithelia and measuring the CXCL5 concentration in the apical washes. When ACE2 activity was inhibited, DABK-induced CXCL5 production on the apical surface was significantly enhanced (Fig. 7D). Together with the results shown in Fig. 4, this suggests that LPS exposure reduces pulmonary ACE2 activity and induces DABK/BKB1R activation and that active DABK enhances CXCL5 production in airway epithelial cells, thereby promoting neutrophil recruitment.

DISCUSSION

This is the first study to demonstrate that DABK is a biological substrate of ACE2 in airway epithelia and that lack of functional ACE2 results in increased neutrophil recruitment to the lung in the setting of LPS challenge. It also reveals that the mobilization of neutrophils is due, in part, to DABK and BKB1R-mediated induction of chemokines such as CXCL5 by the lung epithelium.

Although many ACE2 functions are attributed to its regulation of ANG II and its metabolite ANG (1-7), other ACE2 substrates may contribute to physiologic responses in a tissue-specific manner. In this study, we sought to determine whether other biological substrates of ACE2 are present and physiologically relevant in airway epithelia, under resting conditions and/or during inflammatory processes, as well as how ACE2 attenuation mediates these effects. We chose to investigate DABK because it is a well-known pulmonary inflammatory mediator and known ACE2 substrate (19, 21, 71, 72). In particular, there is strong evidence that BKB1R activation is key to the initiation and progression of a variety of inflammatory conditions, including those involving the lung. Our results indicate that DABK is a biological substrate of ACE2 in airway epithelia and that reducing ACE2 function in epithelia leads to impaired DABK inactivation, leaving this protein free to bind its receptor and initiate the activation of a subsequent signaling cascade.

Previous studies of ACE2-deficient mice showed that loss of ACE2 function in the lung leads to worsening of acid aspiration or LPS-induced inflammation (29, 40). However, little is known regarding how inflammatory stimuli affect ACE2 function in the lung. Following LPS delivery to the mouse lung, we found that ACE2 enzymatic activity was reduced over time.

Intestinally, in an effort to understand the mechanism underlying endotoxin-induced attenuation of ACE2 enzymatic activity in the lung, we found in the present study that LPS does not interfere with ACE2 activity per se; rather, it regulates ACE2 expression at transcriptional and posttranscriptional levels, in part through the NF-κB signaling pathway. LPS reduces ACE2 protein abundance in the lung in a time-dependent manner, which is in parallel with reduced ACE2 activity, an observation that could partially explain why ACE2 activity is dampened by LPS. However, it induces ACE2 mRNA expression rapidly, indicating a possible compensational effect in response to reduced ACE2 protein abundance and enzymatic activity by LPS exposure. The segregation of mRNA and protein levels in the lung under LPS insult underscores the importance of investigating posttranscriptional regulation of ACE2. Reports accumulated in recent years have confirmed that an array of micro-RNAs can regulate ACE2 expression posttranscriptionally (7, 28, 41) in animal models of endotoxin-induced tissue injury including the lung. Our results could further empower the hypothesis that micro-RNAs are involved in the pathogenesis of acute lung injury by targeting the ACE2 gene. Takahashi et al. reported an inhibitory autoantibody against ACE2 in patients with systemic sclerosis (66). So far, no other reports have identified the autoantibodies. Even though we do not know whether endotoxin could induce this kind of antibodies in our mouse model, the possibility warrants further investigation. Moreover, reports indicate that ACE2 undergoes shedding and internalization posttranslationally (30, 34) in settings of LPS challenge and other stimulation. Endocytosis of ACE2 by LPS reduces cell surface protein and thus impairs its activity. Although still controversial, shedding of ACE2 might enhance circulating activity but repress its function at the local level. Substantial evidence indicates that LPS induces TNF-α-converting enzyme (TACE) expression and activity (3, 46), and TACE is a protease responsible for ACE2 shedding. So, it is possible that the reduced ACE2 activity in our LPS inhalation model is due, in part, to enhanced shedding/internalization. Furthermore, neutrophil elastase and other metalloproteases might also affect ACE2 activity. However, in our present LPS inhalation setting, ACE2 activity started to decline even before neutrophil readily infiltrated, suggesting that neutrophil elastase might play a minor role in reducing ACE2 activity at an early stage in this setting.

Notably, the decrease in ACE2 activity coincided with neutrophil influx into the lung, suggesting a causal role. To further investigate this cause-effect question, we treated mice with an ACE2 inhibitor before LPS challenge. Our finding that reducing ACE2 activity in the lung significantly accelerated LPS-induced neutrophil influx suggests that loss of ACE2 activity contributes to neutrophil infiltration. The more severe and prolonged lung inflammation and injury in mice preconditioned due to a lack of ACE2 further underscore a protective role of active ACE2 in this disease setting. Although it is not clear why more prompt neutrophil infiltration upon endotoxin exposure is induced by lack of active ACE2 has such a profound impact on lung inflammation and injury. An exacerbated and persistent proinflammatory cytokine release might be attributed to it. In addition, several lines of study suggest a theory of two phases of neutrophil recruitment that proposes an early phase of recruitment of mature neutrophils from peripheral reserves and a second phase of recruitment of neutrophils from bone marrow, in which the neutrophils are less mature and exhibit increased oxidant production and decreased chemotaxis and diapedesis to proinflammatory cytokines (60). It is possible that the lack of active ACE2 might trigger the second phase of recruitment prematurely. Interestingly, we found that inhibiting ACE2 activity before endotoxin exposure facilitates LPS-induced IKKα/β phosphorylation, although transiently, which provides further evidence and a possible mechanism by which ACE2 negatively regulates the inflammation process so as to alleviate the detrimental effect of the host response to insults.

Kinins are important regulators of cardiovascular homeostasis, inflammation, and nociception (64, 67). In BKB1R-deficient mice, LPS-induced hypotension is blunted, and there is a reduced accumulation of polymorphonuclear leukocytes such as neutrophils in inflamed tissues (5, 17, 55, 56). We showed that BKB1R is expressed in airway epithelia, although the expression level is low under native conditions. LPS exposure increased BKB1R mRNA and protein expression. Interestingly, once induced, this receptor was localized predominantly on the apical surface of airway epithelial cells, the same cellular distribution as ACE2 (32, 33). This would create a microenvironment for ACE2 to influence the activity of the DABK/BKB1R axis.

Of note, DABK is not the only substrate of ACE2 in the lung. ANG II plays a critical role in LPS-induced lung inflammation and injury, which has been extensively studied (13, 44, 45, 73, 74, 79). Interestingly, KKS and RAS have long been thought to participate in several biological and pathological processes and interact at various levels (63). This concept has been validated not only by their functional interaction but also by the identification of a structural network (65). Therefore the impact of the lack of ACE2 in the lung that we demonstrated in the present study is quite possibly a joint effort of KKS and RAS, even though we focused on KKS in this paper.

A key finding of the present study is that inhibition of BKB1R activity reduced pulmonary neutrophil infiltration following LPS exposure and mitigated the neutrophil influx in response to LPS, which was exacerbated by the lack of ACE2 enzymatic activity. These results support the idea that reduced pulmonary ACE2 activity may allow more activation of the DABK/BKB1R axis. In support of this notion, our results further demonstrate that the mice in which the DABK/BKB1R axis was blocked were better protected, as manifested by reduced neutrophil influx and less weight loss, in addition to alleviated lung pathology, cytokine production, and apoptosis of the lung cells. These results shed light on the functions of pulmonary ACE2 and offer an additional mechanism by which supplementing recombinant ACE2 protein can reverse many of the phenotypes associated with experimental pulmonary inflammation (25, 26, 29, 40, 57, 62, 80). Furthermore, our finding that antagonizing BKB1R activation diminished LPS-induced NF-κB signaling 24 h post-endotoxin exposure might be a result of inhibition of NF-κB signaling by DABK/BKB1R, secondary to initial endotoxin stimulation.

Neutrophil recruitment contributes significantly to the pathology of pulmonary inflammatory disorders (6, 77). Excessive recruitment to the alveolar compartment in particular has been associated with progressive lung damage and increased mortality. CXCL1, CXCL2, and CXCL5 are major chemokines that direct neutrophil recruitment to the murine lung (22, 37, 47, 69, 78). In light of the report from Ahluwalia and Perretti (1), we tested the role of CXCL5 in LPS-induced neutrophil infiltration and observed that BKB1R-induced neutrophil recruitment is, at least partly, due to CXCL5 expression. In cultured human umbilical vein endothelial cells, stimulation of endothelial BKB1R induces CXCL5 expression, an effect that is blocked by antagonizing BKB1R (12, 17, 55). Our results revealed that LPS-induced CXCL5 production in mouse lung BALF is coincident with an LPS-induced reduction in ACE2 activity. Moreover, the increased CXCL5 production was reduced by applying a BKB1R antagonist, and this translated into less neutrophil infiltration of the lung following LPS exposure. These results confirm the finding of Ahluwalia et al. that the increased neutrophil influx arising from BKB1R activation was due, in part, to the induction of CXCL5. The contributions of other chemokines, such as CXCL1 and CXCL2, in reducing ACE2 activity and DABK/BKB1R-mediated neutrophil infiltration will require further investigation. The novel observation that attenuating ACE2 function in airway epithelial cells enhanced DABK-induced CXCL5 production is intriguing, as it suggests that respiratory epithelia contribute to the sequence of events promoting inflammation in response to LPS. Thus we propose that attenuation of ACE2 function in the lung may contribute to excessive neutrophil infiltration through activation of the DABK/BKB1R axis, as depicted in Fig. 8.

Fig. 8.

Schematic of how attenuated pulmonary angiotensin I-converting enzyme 2 (ACE2) activity may influence endotoxin-induced des-Arg⁹ bradykinin (DABK)/B1 receptor signaling and neutrophil infiltration. ACE2 is expressed on the apical surface of well-differentiated ciliated epithelia. There it inhibits DABK/bradykinin receptor B1 (BKB1R) activation by inactivating DABK, cleaving a single amino acid residue from its carboxyl terminus. Following exposure to infectious or inflammatory stimuli, ACE2 activity is impaired, leaving the DABK/BKB1R axis more active. This promotes the production and release of chemokines such as C-X-C motif chemokine 5 (CXCL5) from airway epithelial cells. By binding to receptors such as C-X-C motif chemokine receptor 2 (CXCR2) on neutrophils, these chemokines recruit neutrophils from the bone marrow or other peripheral reservoirs to the lung. Exacerbated neutrophil infiltration of the lung contributes to the pathogenesis of acute lung inflammation.

We acknowledge that our LPS inhalation model has its limitations. Clinically, it is not such a relevant model and does not represent effects of endotoxins from gram-positive bacteria. Our hypothesis needs to be tested in more clinically relevant models, such as bacterial pneumonia mouse models.

In summary, our findings demonstrate the following: 1) DABK is a biological substrate of ACE2 in airway epithelia; 2) an LPS-triggered reduction in ACE2 function leads to impaired inactivation of DABK; 3) the resulting activation of the BKB1R signaling cascade enhances the production of neutrophil-recruiting chemokines, such as CXCL5, in airway epithelial cells; and 4) CXCL5 produced in response to BKB1R signaling promotes neutrophil recruitment into the lung. These findings provide insights into the biology of pulmonary ACE2 and define a new role for ACE2 in the pathogenesis of acute lung inflammation through its interactions with DABK.

GRANTS

We acknowledge the support of National Institute of Allergy and Infectious Diseases Grant 60699 (to P. B. McCray, Jr.). We acknowledge the support of the Cell Culture and Gene Transfer Vector Cores, partially supported by the Center for Gene Therapy for Cystic Fibrosis (National Institute of Diabetes and Digestive and Kidney Diseases Grant 54759) and the Cystic Fibrosis Foundation, for preparing airway epithelial cultures and adenoviral vectors.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.P.S., C.W.-L., P.B.M.J., and H.J. conceived and designed research; C.P.S., C.W.-L., Y.Y., T.P., W.B.F., S.W., and H.J. performed experiments; C.P.S., C.W.-L., Y.Y., W.B.F., and H.J. analyzed data; C.P.S., C.W.-L., Y.Y., P.B.M.J., M.C., D.J.H., and H.J. interpreted results of experiments; C.P.S., C.W.-L., and H.J. prepared figures; C.P.S., P.B.M.J., M.C., D.J.H., and H.J. edited and revised manuscript; C.P.S., D.J.H., and H.J. approved final version of manuscript; C.W.-L., P.B.M.J., and H.J. drafted manuscript.