Bile acids inhibit cholinergic constriction in proximal and peripheral airways from humans and rodents

Abstract

Duodenogastroesophageal reflux (DGER) is associated with chronic lung disease. Bile acids (BAs) are established markers of DGER aspiration and are important risk factors for reduced post-transplant lung allograft survival by disrupting the organ-specific innate immunity, facilitating airway infection and allograft failure. However, it is unknown whether BAs also affect airway reactivity. We investigated the acute effects of 13 BAs detected in post-lung-transplant surveillance bronchial washings (BW) on airway contraction. We exposed precision-cut slices from human and mouse lungs to BAs and monitored dynamic changes in the cross-sectional luminal area of peripheral airways using video phase-contrast microscopy. We also used guinea pig tracheal rings in organ baths to study BA effects in proximal airway contraction induced by electrical field stimulation. We found that most secondary BAs at low micromolar concentrations strongly and reversibly relaxed smooth muscle and inhibited peripheral airway constriction induced by acetylcholine but not by noncholinergic bronchoconstrictors. Similarly, secondary BAs strongly inhibited cholinergic constrictions in tracheal rings. In contrast, TC-G 1005, a specific agonist of the BA receptor Takeda G protein-coupled receptor 5 (TGR5), did not cause airway relaxation, and Tgr5 deletion in knockout mice did not affect BA-induced relaxation, suggesting that this receptor is not involved. BAs inhibited acetylcholine-induced inositol phosphate synthesis in human airway smooth muscle cells overexpressing the muscarinic M3 receptor. Our results demonstrate that select BAs found in BW of patients with lung transplantation can affect airway reactivity by inhibiting the cholinergic contractile responses of the proximal and peripheral airways, possibly by acting as antagonists of M3 muscarinic receptors.

INTRODUCTION

Duodenogastresophageal reflux (DGER) is associated with reduced allograft survival after lung transplant (17) and with obstructive lung diseases (21). Microaspiration due to DGER exposes the airways to pathological levels of gastroduodenal contents, including acid, pepsin and other proteases, and bile acids (BAs). In lung transplant recipients, the presence of BAs in bronchoalveolar lavages (BAL) and bronchial washings (BW) has been associated with airway immunological derangement (3, 10). In fact, recipients presenting with BA aspiration experience airflow limitation and highly detectable infections, which are associated with impaired airway clearance and early lung allograft dysfunction, limiting patient survival (9, 11, 30, 46). Hence, the influence of DGER microaspiration in the airways is of significant interest in the realm of pulmonary disease and lung transplant. Although the presence of BAs in the airways has been shown to predict clinical outcomes, their effects on airway contractile responses have never been investigated (11).

Bile acids (BAs) are a chemically diverse group of amphipathic steroid molecules derived from cholesterol catabolism. They include conjugated and unconjugated primary BAs formed in the liver, which can be converted to secondary BAs by the gut microbiome (19). Despite having important roles in nutrient absorption and facilitation of colonic transit, they are also present in the systemic circulation and have roles as signaling molecules in several organs and systems (44). Thus, BAs regulate lipid and glucose metabolism and energy homeostasis (32), while playing immunological roles modulating cytokine release (14). However, whether BAs alter signaling functions in the airways, especially when they are exposed to pathological levels of BAs during DGER aspiration, has not been investigated.

The signaling effects of BAs are mediated by two major receptors: the nuclear farnesoid X receptor (FXR) and the Takeda G protein coupled receptor 5 (TGR5). However, there is evidence that BAs also modulate the activity of other well-known receptors that mediate important physiological functions including the cholinergic muscarinic receptors. Muscarinic receptors are major regulators of airway function, mediating the parasympathetic control of bronchoconstriction, mucus secretion, ciliary beat frequency, neurotransmission, and others. Consequently, we investigated for the first time the effects of 13 BAs detected in BALs and BWs from patients who had received lung transplants on proximal and peripheral airway contractility and their possible target receptors. Using different experimental approaches, including organ bath and precision-cut lung slices (PCLS), we discovered that most secondary BAs strongly and reversibly inhibit the cholinergic constriction of airways in both humans and rodents, possibly by interacting with muscarinic M3 receptors in airway smooth muscle cells. These results suggest that BAs are potential modulators of the cholinergic responses of the airways during lung disease.

MATERIALS AND METHODS

Chemicals.

A pH-buffered Hank’s Balanced Salt Solution (sHBSS) was prepared by diluting a 10X HBSS stock (Thermo Fisher Scientific, Waltham, MA) and supplementing with HEPES (20 mM) and NaOH to obtain pH 7.40 in the final solution. Bile acids (BAs) were purchased from Millipore-Sigma (St. Louis, MO). The following BAs were used and their abbreviation and chemical purity is indicated in parenthesis: primary BAs included cholic acid (CA, ≥98%), glycocholic acid (GCA, ≥97%), taurocholic acid (TCA, ≥95%), chenodeoxycholic acid (CDCA, ≥97%), glycochenodeoxycholic acid (GCDCA, ≥97%), taurochenodeoxycholic acid (TCDCA, ≥95); and secondary BAs included lithocholic acid (LCA, ≥95%), glycolithocholic acid (GLCA, ≥95%), taurolithocholic acid (TLCA, ≥97), deoxycholic acid (DCA, ≥98%), glycodeoxycholic acid (GDCA, ≥97%), taurodeoxycholic acid (TDCA, ≥95%) and ursodeoxycholic acid (UDCA, ≥99%). The specific TGR5 agonist (TC-G 1005) was from Tocris (Minneapolis, MN). BAs and TC-G 1005 were dissolved in dimethylsulfoxide (DMSO, MilliporeSigma) to prepare stock solutions 1,000-fold their final concentration and stored at −20°C. These stocks were diluted in sHBSS the same day of the experiment, and the final concentration of BAs did not exceed 0.1 mM. Acetylcholine (ACh), 5-hydroxitriptamine (5-HT), atropine, and tetrodotoxin were obtained from MilliporeSigma and dissolved in deionized water to prepare stock solutions at least 1,000-fold their final concentration. These stock solutions were stored at −20°C and diluted in sHBSS on the same day of the experiment. Caffeine (MilliporeSigma) was dissolved directly in sHBSS at a final concentration of 20 mM on the day of the experiment. The pH of the sHBSS containing diluted BAs and/or drugs at their final concentrations was 7.40.

Ethical approval.

Animal studies were approved by the Columbia University Animal Care and Use Committee (AC-AAAT6472). Human lungs were provided by an Organ Procurement Organization agreement with the Lung Transplant Program at New York-Presbyterian/Columbia University Irving Medical Center, and these studies were approved by the Institutional Review Board (IRB-AAAR2681).

Preparation of mouse and human precision-cut lung slices.

Mouse lung slices were prepared as described previously (33). Briefly, male C57BL/6 mice (Charles River, MA) were maintained in our animal facility with free access to food and water under a 12-h light cycle. Approximately 12-wk-old mice of both sexes were euthanized with sodium pentobarbital (120 mg/kg ip), the chest cavity was opened, and the lungs were inflated with 1.3 mL of 2% agarose (low-melting point agarose, Thermo Fisher Scientific) in sHBSS (37°C) and ~0.2 mL of air. The agarose in the lungs was gelled by placing the mouse at 4°C for 20 min. Lungs and heart were removed from the animal, and the lung lobes were cut in 130-μm-thick slices using a tissue slicer (Compresstome VF-300, Precisionary Instruments, Greenville, NC) in a laminar flow cabinet under sterile conditions.

Human lungs from 18–60-yr-old donors of both sexes with no documented lung disease were collected and brought to the laboratory within 4–12 h of harvest. The lungs were warmed by submersion in sHBSS at 37°C, and two lobes were cannulated through the main bronchi. The lung lobes were inflated by instilling ~1 liter of 2% agarose in sHBSS at 37°C using constant positive pressure, and gentle manipulation of the tissue was applied to facilitate homogeneous distribution of the agarose. Agarose infusion was terminated when the lung periphery reached a firm consistency upon gentle pressure. The inflated lungs were immediately transferred to ice-cold sHBSS and kept at 4°C for 1 h to allow agarose solidification. Small cubical pieces (~1.5 mL) of agarose-inflated lungs were cut from peripheral regions and mounted in the tissue holder of the compresstome as per manufacturer’s instructions. The lung tissue was sectioned in 150-μm-thick slices and transferred to sHBSS.

Precision-cut lung slices (PCLS) from both human and mice, containing small terminal airways, were transferred to 60-mm Petri dishes containing 10 mL of low-glucose Dulbecco's Modified Eagle Medium (DMEM, Thermo Fisher Scientific) supplemented with antibiotics and were incubated for 10– 16 h at 37°C and 10% CO₂ in a cell culture incubator. PCLS selected for the experiments contained airways with a lumen diameter of 100–250 μm, completely lined by active ciliated epithelial cells and fully attached to the surrounding lung parenchyma.

Measurement of the contractile response of airways.

The contractile response of airways in PCLS was measured using phase-contrast video microscopy as previously described (33). Briefly, PCLS were mounted on a cover glass in a custom-made perfusion chamber and held in place with a small sheet of nylon mesh with a narrow opening to allow for airway imaging. Perfusion of the PCLS was performed by dripping solution at one end of the chamber and removing it by suction at the opposite end by means of a custom-made, gravity-fed, computer-controlled perfusion system consisting of eight syringe tubes connected to individual electronic solenoid valves (The Lee Company, Westbrook, CT) and to an 8-way manifold. The PCLS were continuously superfused with one of the solutions at ~800 μL/min, and solution changes were made by switching between solutions at preprogrammed times. Exposure of PCLS to two or more drugs/chemicals (e.g., ACh + TLCA) was made by superfusing a single solution containing all drugs/chemicals at the final concentration. The chamber was placed on the stage of an inverted phase-contrast microscope and lung slices were imaged with a ×10 objective. Digital images were recorded to a hard drive in time lapse (0.5 Hz), using a CCD camera, frame grabber, and image acquisition software (Video Savant, IO Industries, Ontario, Canada). The airway cross-sectional luminal area (lumen area) was calculated from each image using a custom-written script in Video Savant that distinguishes the airway lumen from the surrounding tissue. The lumen area was normalized to the initial area (before stimulation) and the changes in lumen area were plotted versus time using graphics software. To study the effects of BAs on airways preconstricted with cholinergic stimuli, we choose ACh because it is the natural agonist for muscarinic receptors in airway smooth muscle cells. In contrast to other cholinergic agonists such as methacholine, ACh is sensitive to inactivation by acetylcholine esterase that is active in the PCLS. However, under our experimental conditions of continuous superfusion of PCLS, the airways are exposed to fresh ACh and able to maintain a sustained constriction until ACh washout.

Electrical field stimulation-induced contractions of guinea pig trachea.

The effect of selected BAs on proximal airway contractions elicited by electrical field stimulation (EFS) was evaluated in guinea pig tracheal rings as previously described (23). Briefly, Hartley male guinea pigs were euthanized with 100 mg/kg ip sodium pentobarbital, and the tracheas were removed and then dissected into closed rings consisting of two cartilaginous rings from which mucosa, connective tissue, and epithelium were removed. The tracheal rings were suspended in wire myograph organ baths containing physiological salt solution (composition: 118 mM NaCL, 5.6 mM KCl, 0.5 mM CaCl₂, 0.2 mM MgSO₄, 25 mM NaHCO₃, 1.3 mM NaH₂PO₄, and 5.6 mM d-glucose) with 10 μM indomethacin (dimethyl sulfoxide vehicle, final concentration of 0.1%). After equilibration, the guinea pig tracheal rings were precontracted twice with a range of ACh concentrations (0.1 μM to 1 mM) to obtain a series of contractile responses from no constriction to maximal constriction. After extensive ACh washout with buffer exchanges, the resting tension was reset at 1 g. EFS was delivered via two platinum electrodes situated on opposite sides of the preparation and separated ~0.8 cm. The electrical signal was generated using a Grass RPS 107 stimulator (Grass-Telefactor) and consisted of trains of pulses, each with a duration of 0.5 ms and frequency of 32 Hz. Each train duration was 5 s and repeated every 80 s. After waiting a minimum period of 60 min, and when the EFS-induced contraction force had become constant, cumulative doses of BAs were added to the baths to determine the 50% inhibitory concentration (IC₅₀) of BA at inhibiting EFS-induced contractions. The effect of BAs on the EFS-induced contraction was expressed as a percentage of the EFS-induced force before BAs were added.

Inositol phosphate assays.

[³H]inositol phosphate synthesis was measured using the method of Wedegaertner et al. (47) with some modifications, as we previously described (20). Briefly, we used immortalized human airway smooth muscle cells stably transfected to express the human M3 muscarinic receptor, as we described earlier (45). These cells were grown in 24-well plates in media M199 with 10% FBS, 1 ng/mL FGF, 0.25 ng/mL EGF, 0.17 μM insulin, 6.9 nM transferrin, 3.9 nM selenium, and antibiotics (100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B) until they reached confluence. Subsequently, the medium was replaced with DMEM containing 10 μCi/mL myo-[³H]inositol (specific activity 20 Ci/mmol; Perkin Elmer, Waltham, MA) on the day before the assay. The next day, the loading media was aspirated and the wells were washed twice with 500 μL warm (37°C) assay buffer (sHBSS supplemented with 10 mM LiCl). Assay buffer (300 μL) was added to each well, and cells were pretreated for 15 min at 37°C with increasing concentrations of GCDCA, TLCA, or vehicle (0.1% DMSO). Cells were then left untreated (basal inositol phosphate synthesis) or were stimulated with 0.15 μM ACh for 30 min at 37°C. The reaction was terminated by the addition of 330 μL of cold methanol. Then, 660 μL of chloroform was added and the samples were transferred to an Eppendorf tube. The phases of the samples were separated by centrifugation at 820 g for 10 min at 4°C. Four hundred fifty microliters of the upper aqueous phase were transferred to a new glass tube. Cold 50 mM formic acid (300 μL) and 100 μL of 3% ammonium hydroxide were added, and total [³H]inositol phosphates were separated from free myo-[³H]inositol by chromatography and quantified by liquid scintillation.

RT-PCR of TGR5 expression in lung slices.

PCLS from TGR5 knockout (KO) and wild-type (WT) mice were prepared as described above and stored at −80°C. Total RNA was obtained from these PCLS using Trizol Reagent, and cDNA was synthesized using SuperScript VILO reagents (Thermo Fisher Scientific). Two micrograms of RNA were used for each 20 μL RT-PCR reaction. PCR was then performed (40 cycles) using 1 μL of cDNA product. All primer sets were designed to flank at a large intron to avoid confounding amplification of genomic DNA. Two-step PCR was used with a denaturing temperature of 94°C for 10 s and an annealing/extension temperature of 68°C for 1 min. Mouse whole brain served as a positive control, and PCR reaction mixtures devoid of cDNA served as RT-PCR negative controls. All reagents were from Life Technologies (Carlsbad, CA).

Quantitative RT-PCR of muscarinic receptors expression in peripheral airways.

PCLS of 350 μm in thickness were prepared from C57BL/6 mice and incubated overnight in a cell culture incubator as described above. Peripheral airways in cross-section were microdissected from PCLS under a dissecting microscope (Nikon, Japan), using a pair of 20G needles to detach them from the surrounding lung parenchyma. Special care was taken to assure that the peripheral airways were also separated from the adjacent pulmonary arteries. A total of ~15 microdissected airways from each mouse were collected in a 0.5 mL conical tube containing 100 μL of lysis buffer. In addition, a sample of mouse brain cortex was prepared in parallel. The latter served as a calibration reference for the expression of muscarinic receptors. Total RNA was isolated using the Arcturus PicoPure RNA Isolation Kit (ThermoFisher) according to the manufacturer's recommendations. RNA purity and quantity were measured using the NanoDrop One (ThermoFisher). Complementary DNA synthesis was performed with SuperScript VILO cDNA Synthesis Kit (ThermoFisher) following the manufacturer's recommendations and using 200 ng total RNA in a final volume of 20 μL. Quantitative RT-PCR was performed with the mouse primers shown in Table 1, and amplification was carried out using the PowerSYBR Green PCR Master Mix (ThermoFisher) and the 7500 Real-Time PCR System (ThermoFisher) following manufacturer recommendations. The relative expression of the muscarinic receptor subtypes was calculated using the double-delta Ct method (26) with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as the reference gene and mouse brain as the calibrator sample.

Table 1.

Sequence-specific primers for mouse muscarinic receptor subtypes M1, M2, and M3 and GAPDH

Gene	Accession ID	Forward and Reverse Sequence of the Primer (5′ to 3′)	cDNA Product Size, bp	gDNA Product Size, bp
Chrm1	NM_007698	F: AGGCCCCCGGAGAAGCACTGAA	104	13,508
		R: AGCCCCTTCCTCCAGTCACAAGATT
Chrm2	NM_203491	F: CGCTCGCTCCCAAACCGGTCCAA	136	134,554
		R: GTGTTCAGTAGTCAAGTGGCCAAAGAAACAT
Chrm3	NM_033269	F: GCTCAGTGGACTGTGGATTGAGTGAACCATA	114	102,469
		R: GAATGTCACGTGCTTGGTCACTTGGTCAGAA
Gapdh	NM_008084	F: CCGTAGACAAAATGGTGAAGGTCGGTGTGAA	120	1,954
		R: CAATGAAGGGGTCGTTGATGGCAACAAT

Statistics.

Statistical values are expressed as mean ± SE. Student’s t test or one-way analysis of variance (ANOVA) followed by Dunnett's Test comparisons were used to evaluate the significance between means from two or more groups, respectively.

RESULTS

Select BAs relax human and mouse peripheral airways precontracted with ACh.

To study the effects of BAs on peripheral airways, we exposed PCLS to 30 μM TLCA for 8 min and found no changes in airway lumen area of either human or mouse PCLS (Fig. 1, A and D). A concentration of 30 μM was selected to test the efficacy of the BAs on the contractile responses of the airways based on the affinity of BA receptors with EC₅₀ in the range of 0.1 μM to 20 μM and previous reports investigating the signaling functions of BAs in other organs and systems (8) as well as our concentration-response curves presented later in this work. Then we tested whether TLCA had any effects on airways precontracted with 0.3 μM ACh, which induces submaximal airway constriction (34a). ACh reduced the luminal area of human and murine airways by 48.4 ± 8.2% and 59.3 ± 4.5%, respectively [Fig. 1, B and E, Supplemental Video S1 (https://doi.org/10.7916/d8-4wpn-wr76) and Supplemental Video S2 (https://doi.org/10.7916/d8-0pha-mw03)]. The subsequent addition of 30 μM TLCA, in the continuous presence of 0.3 μM ACh, caused airway relaxation in both human and mouse airways, as evidenced by the significant increase in airway lumen area. However, this TLCA-induced airway relaxation occurred at slower rates in human airways than in mouse airways (maximal relaxation rates were 2.6 ± 0.4%/min and 15.4 ± 1.8%/min, respectively; P < 0.001; unpaired t test), and reached 36.9 ± 3.7% and 51.0 ± 4.1% after 10 min of TLCA exposure, respectively. TLCA-induced relaxation was reversible upon TLCA washout, as the airways recontracted by superfusing the solution containing ACh alone, although mouse airways recontracted faster than human airways (Fig. 1, B and E). The subsequent washout of ACh increased the airway lumen area nearly to the resting condition in both human and mouse airways, indicating that the overall integrity and functional state of the airways was maintained following the exposure of the PCLS to ACh and TLCA. To investigate the action of other BAs on airway reactivity, we performed similar experiments with select BAs on human airways and with the complete panel of 13 BAs on mouse airways. We found that, like TLCA, the other BAs tested caused no significant changes in the basal lumen area of unstimulated airways (Table 2). However, in airways precontracted with 0.3 μM ACh, some BAs (all BAs tested at 30 μM) caused relaxation, whereas others had no significant effects (Fig. 1, C and F). All of the secondary conjugated BAs (TLCA, GLCA, TDCA, and GDCA), the primary unconjugated BA CDCA, and the secondary unconjugated BAs LCA and UDCA induced significant relaxation in mouse airways precontracted with ACh (Fig. 1F). In contrast, the primary BAs CA, GCA, TCA, GCDCA, and TCDCA and the secondary BA DCA did not have significant effects. Because of limited availability of human lungs for these experiments, we only tested TLCA, GDCA, and CA in human airways. Similar to mouse airways, TLCA and GDCA significantly relaxed human airways precontracted with ACh, whereas CA had no significant effects (Fig. 1C). Altogether, these results suggest that BAs have no significant effects on unstimulated airways in PCLS, yet several BAs induce strong relaxation on airways contracted by ACh. Furthermore, the similar effects of BAs in human and mouse airways indicate that the mouse is a suitable model to study the role and mechanisms of BAs in the peripheral airways. Accordingly, some of our subsequent studies were performed on mouse models because of the ease of tissue preparation and availability.

Fig. 1.

Bile acids (Bas) induce relaxation of human and mouse peripheral airways precontracted with ACh. A and D: representative traces showing exposure of human and mouse precision-cut lung slices (PCLS) to 30 μM taurolithocholic acid (TLCA) in HBSS. TLCA induced neither constriction nor relaxation of unstimulated (resting) airways. B and E: images and traces showing the changes in cross-sectional lumen area of peripheral airways in human and mouse PCLS constricted with 0.3 μM ACh and relaxed with 30 μM TLCA in the continuous presence of ACh. Drugs were continuously superfused at the times indicated by the lines on top of the traces; HBSS alone was superfused otherwise. The representative images were obtained at the times indicated by numbers below the traces. C and F: summary showing the relaxation induced by the indicated BAs (tested at 30 μM) or by 0.1% DMSO of airways preconstricted with ACh. Data were obtained from experiments similar to those shown in B and E: relaxation was calculated by dividing the lumen area value 10 min after BA addition by the lumen area value after ACh but before BA addition. Data are means ± SE of 3–5 PCLS from 3 human lungs or 6–8 PCLS from 3 mice; *P < 0.05 vs. DMSO, ANOVA with post hoc Dunnett’s test. Visualization of ACh-induced airway constriction and TLCA-induced relaxation in human and mouse PCLS can be watched in Supplemental Video S1 (https://doi.org/10.7916/d8-4wpn-wr76) and S2 (https://doi.org/10.7916/d8-0pha-mw03), respectively.

Table 2.

BAs induce neither contraction nor relaxation in unstimulated (resting) airways

BA	Human	Mouse
DMSO	−2.6 ± 2.7	−1.9 ± 0.2
CA	−5.3 ± 1.8	−0.5 ± 0.3
GCA	ND	−1.3 ± 0.7
TCA	ND	0.4 ± 0.4
CDCA	ND	−1.7 ± 1.0
GCDCA	ND	−0.4 ± 0.4
TCDCA	ND	0.1 ± 0.5
LCA	ND	−0.2 ± 3.2
GLCA	ND	6.4 ± 1.2
TLCA	−2.7 ± 3.7	1.0 ± 0.5
DCA	ND	−1.1 ± 0.5
GDCA	−5.2 ± 1.4	−0.6 ± 0.4
TDCA	ND	−0.3 ± 0.4
UDCA	ND	−0.2 ± 0.4

Values are means ± SE of 3–5 precision-cut lung slices (PCLS) from 3 human lungs or 6–8 PCLS from 3 mice. Summary of decrease (+) or increase (−) in airway lumen area (%) induced by exposure of the PCLS to bile acids (BA; tested at 30 μM) or 0.1% DMSO (vehicle) in human and mouse peripheral airways at rest. The data were obtained from experiments similar to those shown in Fig. 1, A and D. There were no significant differences between groups; α = 0.05, one-way ANOVA. CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; GCA, glycocholic acid; GCDCA, glycochenodeoxycholic acid; GDCA, glycodeoxycholic acid; LCA, lithocholic acid; ND, not determined; TCA, taurocholic acid; TCDCA, taurochenodeoxycholic acid; TDCA, taurodeoxycholic acid; TLCA, taurolithocholic acid; UDCA, ursodeoxycholic acid.

Concentration-dependence of TLCA- and CDCA-induced relaxation in ACh-precontracted airways.

To better understand the effects of BAs on airway relaxation, we precontracted mouse airways with ACh and subsequently exposed them to repetitive challenges with increasing concentrations of TLCA or CDCA. As shown in Fig. 2, both TLCA and CDCA induced a concentration-dependent airway relaxation that was reversible upon BA washout. TLCA-induced airway relaxation was 21.0 ± 7.4% at 1 μM, increased gradually up to 80.2 ± 2.3% at 30 μM, and had an inhibitory concentration fifty (IC₅₀) of 3.2 ± 0.6 μM (Fig. 2C). CDCA was less potent than TLCA to relax ACh-precontracted airways and did not reach saturation at the maximal concentration tested (100 μM). Consequently, we did not calculate the maximal relaxation and IC₅₀ values for CDCA. Since the effects of TLCA were readily reversible and maximal at 30 μM, we choose this BA at this concentration for our subsequent studies to investigate the action mechanisms of BA in the peripheral airways. Similarly, a concentration of 30 μM was used to investigate the effects of the different BAs tested in the studies presented earlier in Fig. 1.

Fig. 2.

Taurolithocholic acid (TLCA) and chenodeoxycholic acid (CDCA) induce concentration-dependent relaxation of ACh-precontracted airways with different potencies. A and B: representative traces of changes in airway lumen area of mouse precision-cut lung slices (PCLS) contracted with 0.3 μM ACh and exposed to gradual increases in concentration of TLCA or CDCA, as labeled above each trace. C: summary of airway relaxation as a function of TLCA or CDCA concentrations, obtained from experiments similar to those shown in A and B. Data are means ± SE of 5–7 PCLS from 3 mice. BA, bile acid.

TLCA inhibits EFS-induced, endogenous ACh-mediated constrictions in guinea pig trachea rings.

To corroborate that TLCA inhibits cholinergic responses of the airways, we investigated the effects of this BA on the muscle force generated by isolated guinea pig tracheal rings in response to electrical field stimulation (EFS). We first confirmed that EFS-induced airway constriction was mediated by endogenous ACh-release from parasympathetic nerve terminals using tetrodotoxin and atropine. These agents inhibit neuronal sodium channels preventing neural release of acetylcholine and postsynaptic muscarinic receptors in the airway smooth muscle, respectively. As shown in Fig. 3A, both tetrodotoxin and atropine rapidly and completely inhibited EFS-induced transient contractions, confirming the cholinergic nature of this response. Next, we found that TLCA dose-dependently attenuated the magnitude of the transient airway smooth muscle contractions induced by EFS. In contrast, the primary BA GCDCA had a nonstatistically significant effect at any concentration tested (Fig. 3, B and C). Thus, these findings in proximal airways are consistent with those obtained in PCLS and support the hypothesis that TLCA, but not primary BAs, inhibits cholinergic airway constriction.

Fig. 3.

Taurolithocholic acid (TLCA) inhibits muscle force induced by electrical field stimulation (EFS) and mediated by endogenous release of acetylcholine in guinea pig trachea rings. A and B: representative traces show the inhibitory effects of 1 μM tetrodotoxin and 1 µM atropine (A) as well as vehicle (0.1% DMSO) and increasing concentrations of TLCA or glycochenodeoxycholic acid (GCDCA) (B) on EFS-induced transient airway constrictions. C: summary data show the concentration-dependent inhibition of EFS-induced airway constriction caused by TLCA and GCDCA. Data are means ± SE of n = 6 from 3 animals; ***P < 0.001 with respect to DMSO (ANOVA with post hoc Dunnett’s test).

TLCA specifically inhibits airway constriction induced by ACh but not by other noncholinergic bronchoconstrictors.

Since TLCA and other BAs relaxed airways preconstricted with ACh, we tested whether TLCA would also prevent airway constriction induced by ACh and by other noncholinergic bronchoconstrictors. To this aim, we tested constriction before and after exposure to TLCA, DMSO (negative control), and atropine, a muscarinic receptor antagonist (positive control for ACh stimulation). For these experiments we chose the noncholinergic bronchoconstrictors 5-hydroxitriptamine (5-HT) and caffeine because they induce narrowing of mouse peripheral airways independently of muscarinic receptor activation, as previously demonstrated (34a). Furthermore, to test the effects of TLCA on 5-HT-induced airway constriction, we choose 0.3 μM as the testing 5-HT concentration because it induces a submaximal airway constriction. As shown in Fig. 4A, an initial control exposure to 0.3 μM ACh induced an airway contraction that was reversed by ACh washout. Subsequent exposure to 30 μM TLCA did not have an effect on airway luminal area alone but strongly inhibited the contraction induced by a second exposure to ACh. On average, the airway constriction induced by the second ACh stimulus was 50.4 ± 11.2% of that induced by the first stimulus in slices exposed to TLCA, whereas it was 107.3 ± 7.9% in slices exposed to DMSO and −3.1 ± 2.3% in slices exposed to atropine (Fig. 4B). Importantly, a similar constriction induced by the first and second ACh stimulation before and after DMSO exposure indicates that there is no desensitization of ACh receptors in our experimental conditions and that the reduced ACh-induced airway constriction after exposure to TLCA was not due to such mechanism. In contrast to the inhibitory effects of TLCA and atropine on ACh-induced airway constriction, the results in Fig. 4, C–F show that neither of these inhibitors prevented the constriction induced by 0.3 μM 5-HT or 20 mM caffeine. These results suggest that TLCA specifically inhibits airway constriction triggered by cholinergic stimulation and that this BA had lower efficacy and potency than atropine.

Fig. 4.

Pre-exposure of airways to taurolithocholic acid (TLCA) specifically inhibits cholinergic bronchoconstriction. A, C, and E: representative traces showing airway constriction induced by exposure of mouse precision-cut lung slices (PCLS) to 0.3 μM ACh (A), 0.3 μM 5-hydroxitriptamine (5-HT) (C), or 20 mM caffeine (Caff) (E) before and after exposure to 30 µM TLCA as labeled above traces. TLCA preexposure reduced the magnitude of airway constriction induced by ACh but not by 5-HT or caffeine. B, D, and F: summary data of the effects of 0.1% DMSO (control), 1 μM atropine (cholinergic antagonist), and 30 μM TLCA preexposure on airway contraction induced by ACh (B), 5-HT (D), or caffeine (F), obtained from similar experiments to those shown in the representative traces on the left. Data are means ± SE of 5–8 PCLS from 3 mice and represent the proportion of the airway contraction induced by second bronchoconstrictor exposure (postinhibitor addition) with respect to the first exposure (previous to the inhibitor). *P < 0.05 (paired t test) when testing differences between the first and second constrictions for each agent.

Primary BAs do not affect airway constriction induced by noncholinergic agonist 5-HT.

Since primary BAs are the most abundant in lung transplant patients with DGER aspiration (30, 46), we further investigated whether these BAs have any effects on airway contractility. We preexposed the peripheral airways in mouse PCLS to the noncholinergic contractile agonist 5-HT and subsequently added the primary BAs (or DMSO as a control). We used 5-HT at a concentration that produces a submaximal airway constriction as previously described (34a). As shown in Fig. 5, the addition of primary BAs had little effect on 5-HT-induced airway constriction, and these effects were not significantly different from those induced by the vehicle (DMSO). These results suggest that primary BAs have neither relaxing nor potentiating effects on 5-HT-induced airway contraction.

Fig. 5.

Primary bile acids (BAs) do not affect 5-hydroxitriptamine (5-HT)-induced peripheral airway contraction. A and B: representative traces show the contractile responses of peripheral airways to 5-HT and subsequent addition of 0.1% DMSO (control) (A) or 30 μM chenodeoxycholic acid (CDCA) (B). C: summary of airway relaxation induced by the primary BAs and DMSO (control). Data are means ± SE of 5–10 precision-cut lung slices from 3 mice; there were no significant differences between DMSO and the primary BAs, α = 0.05 (one-way ANOVA). CA, cholic acid; GCA, glycocholic acid; GCDCA, glycochenodeoxycholic acid; TCA, taurocholic acid; TCDCA, taurochenodeoxycholic acid.

TGR5 does not mediate TLCA-induced relaxation of ACh-precontracted airways.

TGR5 is the major BA-specific G protein coupled receptor mediating the acute signaling effects of BAs in several organs and systems (5, 14, 44). To test whether TGR5 mediates the effects of TLCA on the airways, we first studied the effects of TC-G 1005, a potent activator of TGR5, on airways precontracted with ACh. Since TC-G 1005 has an EC₅₀ of 6.2 nM for the activation of mouse TGR5 (13), we used 30 μM TC-G 1005 in our assays to assure that we were using a concentration that was high enough to stimulate TGR5 in our PCLS preparation. We found that 30 μM TC-G 1005 did not cause significant relaxation, whereas TLCA strongly and significantly relaxed ACh-precontracted airways in experiments performed in parallel (Fig. 6A and B). These results suggest that TGR5 does not play a role in smooth muscle relaxation in peripheral airways. Second, we studied the effects of TLCA on ACh-precontracted airways in PCLS prepared from homozygote Tgr5⁻/⁻ knockout (Tgr5-KO) mice (kindly donated by Dr. Higuchi at Columbia University Medical Center) and from wild-type control mice. Using PCR, we confirmed that TGR5 receptors were absent in the lung slices from TGR5-KO mice but present in those from WT mice (n = 6 slices from 3 mice). However, we found that TLCA caused a similar airway relaxation in both TGR5-KO and WT mice (Fig. 6, C and D) suggesting that the TLCA-induced airway relaxation was not mediated by TGR5.

Fig. 6.

Takeda G protein-coupled receptor 5 (TGR5) does not mediate taurolithocholic acid (TLCA)-induced relaxation of ACh-contracted airways. A: representative trace of changes in airway lumen area in mouse precision-cut lung slices (PCLS) exposed to 30 μM TC-G 1005 in the continuous presence of 0.3 μM ACh. B: summary of airway relaxation induced by 30 μM TC-G 1005 and 30 μM TLCA tested in experiments similar to those shown in A. Data show mean ± SE from 6-8 PCLS from 3 mice. ***P < 0.001, paired t test, ACh alone vs. ACh + treatments. C: representative traces showing airway constriction induced by 0.3 μM ACh and relaxation induced by 30 μM TLCA in PCLS obtained from both TGR5-knockout (KO) and wild-type (WT) mice. D: percentage of TLCA-induced relaxation of ACh-contracted airways in both wild-type and TGR5-KO mice. Data show mean ± SE of 6 PCLS from 3 mice in each group; nonsignificant (NS), P > 0.05, unpaired t test.

TLCA inhibits ACh-induced inositol phosphate synthesis in human airway smooth muscle cells overexpressing M3 receptors.

Altogether, our results in ex vivo tissue preparations strongly suggest that the inhibitory effects of TLCA on airway contraction are specific for cholinergic agonists and are not mediated by TGR5 and support the alternative hypothesis that TLCA may specifically inhibit muscarinic receptor activation to cause airway relaxation. To evaluate this hypothesis, we used human airway smooth muscle cells stably overexpressing M3 muscarinic receptors to study the effects of TLCA on ACh-induced inositol phosphate synthesis. We found that TLCA but not GCDCA concentration-dependently inhibited the ACh-induced inositol phosphate synthesis (Fig. 7). These results support the hypothesis that TLCA may interact with muscarinic M3 receptors in airway smooth muscle to cause inhibition of inositol phosphate synthesis and airway relaxation. The contrasting outcomes between TLCA and GCDCA also indicate that the effects of BAs on inositol phosphate synthesis are consistent with their effects on airway relaxation in PCLS.

Fig. 7.

Taurolithocholic acid (TLCA) inhibits ACh-induced inositol phosphate synthesis in cultured human airway smooth muscle cells overexpressing muscarinic M3 receptors. Inositol phosphate synthesis was measured in response to 30 min of stimulation with 0.15 μM ACh in [³H]-myo-inositol-loaded cells and preincubated with either TLCA, glycochenodeoxycholic acid (GCDCA; 0.1 μM to 100 μM), or vehicle (0.1% DMSO) for 15 min before the addition of ACh. Data are means ± SE of n = 6. **P < 0.01 with respect to DMSO (ANOVA with post hoc Dunnett’s test). BA, bile acid.

Expression of muscarinic receptor subtypes M2 and M3 in mouse peripheral airways.

To confirm the expression of muscarinic receptors M2 and M3 in the peripheral airways, we microdissected airways from mouse PCLS and performed quantitative RT-PCR using sequence-specific primers (Table 1). We found that the expression of M2 and M3 receptors in the peripheral airways was ~0.018-fold that in brain cortex (Table 3). In contrast, M1 receptor expression in peripheral airways was not accurately detectable.

Table 3.

Relative expression of muscarinic receptor subtypes (M1 to M3) in mouse peripheral airways

	M1	M2	M3
Brain	20.19 (1.0)	22.94 (1.0)	22.27 (1.0)
Airways 1	34.61 (0.000)	29.00 (0.029)	30.18 (0.008)
Airways 2	Undetectable	29.93 (0.018)	30.52 (0.007)
Airways 3	34.36 (0.000)	30.38 (0.015)	28.78 (0.028)
Airways 4	34.68 (0.000)	29.85 (0.019)	28.80 (0.025)
Airways 5	34.93 (0.000)	29.83 (0.016)	28.66 (0.022)
Average (1–5)	(0.000)	(0.019)	(0.018)
SE	(0.000)	(0.003)	(0.004)

Quantitative RT-PCR data shown are C_T values followed by the fold difference (in parentheses) in gene expression (2^−ΔΔC_T) obtained with the ΔΔC_T method using GAPDH as the reference gene and brain tissue as the calibrator sample. Each sample contained ~15 peripheral airways microdissected from several precision-cut lung slices (PCLS) from a different mouse (n = 5 mice) as described in materials and methods.

DISCUSSION

This study investigates, for the first time, the effects of the 13 BAs detected in BAL and BW from post-lung transplant patients with DGER aspiration on proximal and peripheral airway contractility in both humans and rodents. We found that BAs had no acute effects on contraction of unstimulated (resting) airways but strongly, dose-dependently, and reversibly relaxed peripheral airways precontracted with ACh in both humans and mice. The secondary BA litocholic acid and its glycine and taurine conjugates (i.e., LCA, GLCA, and TLCA) as well as the taurine conjugate of deoxycholic acid (TDCA) had the strongest effects on airway relaxation, whereas only CDCA among the primary BAs had small but significant effects at the concentration tested (30 μM). The inhibitory effects of BAs were specific for ACh-induced airway constriction, but they did not inhibit constriction induced by noncholinergic bronchoconstrictors. BAs also inhibited airway constriction mediated by endogenous ACh-release in proximal airways of guinea pigs stimulated by EFS, and in this preparation, secondary conjugated BAs (i.e., TLCA) also had stronger effects. Accordingly, we propose that BAs have anticholinergic effects in both peripheral and proximal airways. Furthermore, we found that TC-G 1005, a highly specific agonist for the BA receptor TGR5, failed to relax ACh-precontracted airways. Additionally, the relaxing effect of TLCA was not affected by the genetic deletion of TGR5 in mice. These results suggest that TGR5 is not the receptor mediating the acute effects of BAs in airway smooth muscle. Finally, we showed that TLCA but not the primary BA GCDCA inhibited ACh-induced inositol phosphate synthesis in human airway smooth muscle cells that overexpress muscarinic M3 receptors. Altogether, our results suggest that select BAs have the potential to inhibit the cholinergic contractile responses of the proximal and peripheral airways, possibly by acting as M3 receptor antagonists.

We found that neither primary nor secondary BAs had any acute effects on contraction of unstimulated (resting) airways in peripheral lung slices from human and mice (see Fig. 1). Furthermore, the BAs did not potentiate airway constriction in lung slices exposed to low concentrations of the cholinergic agonist ACh, but instead most secondary BAs caused relaxation. Similarly, pre-exposure of lung slices to secondary BA TLCA reduced the airway constriction stimulated by ACh and had no effects on the contractile response of the airways to the noncholinergic agonist 5-HT (see Fig. 4). Furthermore, primary BAs did not potentiate constriction or cause relaxation of airways precontracted with 5-HT (see Fig. 5). These results suggest that BAs do not cause acute constriction of unstimulated airways or of those exposed to contractile agonists. These results were surprising, as lung transplant patients with chronic BA aspiration from DGER are at increased risk for increased airflow limitation and decreased forced expiratory volume in 1 s (FEV₁) (30, 31), and suggest that this association cannot be explained by an acute direct effect on airway smooth muscle.

We found that secondary BAs inhibited cholinergic airway constriction with higher potency than primary BAs. TLCA at 1 μM and 30 μM relaxed airways by ~20% and ~80% respectively, whereas most primary BAs caused no significant airway relaxation at 30 μM in PCLS (see Figs. 1 and 2). Furthermore, the primary BAs caused neither relaxation nor further contraction of peripheral airways precontracted with the noncholinergic agonist 5-HT (see Fig. 5). These results were interesting, as the most abundant BAs are of the primary category. Yet prandial states and microbial conversion impose a high level of complexity in the measurement of individual BA levels at a given time. In fact, following a meal, total BAs could reach the low millimolar range (0.2–1.0 mM) in the intestine. Ranges as wide as 30–700 μM DCA or 1–450 μM LCA have been measured, implying large variability among individuals (16). Upon aspiration, BAs are diluted in variable volumes of saliva and airway mucus and likely have a differential local distribution along the airway tree. Therefore, no correlation can be noted between BA concentrations in the gastrointestinal system, serum, and in the lung (11). To note is that total, primary, and secondary BAs breakdowns have been identified in BAL and BW of patients with lung transplants and terminal disease, with primary BAs being the most abundant (9, 11, 30, 46). Assayed as a marker of DGER aspiration in post-lung transplant surveillance BAL, BAs have a median of 0.3 μM (25th-75th percentile range 0–16 μM) in healthy recipients and of 1.6 μM (25th–75th percentile range 0–32 μM) in those with failed allograft (46). These values originate from BAL samples that present with a low saline-to-lung-volume ratio and an undefined dilutive magnitude that do not allow for a precise calculation of BA concentration in the airways. Since the concentrations of individual BAs in the liquid phase of the airways are presently unknown, we first performed concentration-response curves for TLCA and CDCA, which showed consistent and substantial contractile responses between 1 and 30 μM. Hence, our choice to use 30 μM to test airway reactivity was determined by our experimental observations, as BAs concentrations in the airways upon DGER microaspirations cannot be accurately estimated. Although a direct comparison of the BA (e.g., TLCA) concentrations that produced contractile effects in the peripheral airways with their concentration in the patient airways cannot be established, we think our findings are important when considering the possible mechanisms of the association between the presence of BAs in patients with DGER microaspirations and the decline in airway function.

TGR5 has been identified as the primary cell membrane receptor that mediates the acute signaling effects of BAs in a variety of cells in several organ systems (5, 18, 25, 29, 44). Upon BAs binding, TGR5 initiates a signaling cascade through G_s by activating adenylyl cyclase and causing elevation of intracellular cAMP. In airway smooth muscle, agonists that cause an increase in intracellular cAMP (e.g., β₂-adrenoceptor agonists) also induce airway relaxation (15). Accordingly, we first hypothesized that cAMP-coupled TGR5 receptors could mediate the BA-induced airway relaxation. In fact, TGR5 receptors have been found expressed in gastric smooth muscle, and activation with a TGR5-selective ligand, oleanolic acid, was found to elevate intracellular cAMP and induce relaxation (38). However, we found that this hypothesis was not supported in airway smooth muscle by several of our experimental results. First, the BA-induced inhibition of airway constriction was observed only when the airways were stimulated with ACh but not with other noncholinergic bronchoconstrictors (i.e., 5-HT and caffeine). In contrast, β₂-agonists are known to similarly inhibit airway constriction stimulated by either cholinergic (e.g., ACh, methacholine) or noncholinergic bronchoconstrictors, including 5-HT, endothelin-1 (ET-1), histamine, and cysteinyl leukotrienes (2, 22, 34a–36). Yet, receptors for all the latter bronchoconstrictors are all G_q-coupled, suggesting that β₂-agonists act downstream of receptor activation (4, 22, 34a, 36, 48). In sharp contrast, because BAs only inhibited airways stimulated by ACh and not by 5-HT or caffeine, we believe that they may be blocking muscarinic receptor activation rather than activating TGR5. Second, TC-G 1005 not causing relaxation suggests that TGR5 receptors may not be functionally expressed in the smooth muscle cells of the peripheral airways. Third, genetic deletion of TGR5 receptors in the TGR5-KO mice did not affect the TLCA-induced airway relaxation. Consequently, our results do not support a role of TGR5 receptors in the airway relaxation induced by BAs.

The nuclear BA receptor farnesoid X receptor alpha (FXR-α) is the mediator of the genomic actions of BAs in hepatocytes and many other cell types (18, 24, 44) and has been found expressed in the lung (51). FXR-α receptors mediate most of the endocrine effects of BAs on glucose and lipid metabolism, including the feedback regulation of BA synthesis (42). However, the acute airway relaxation induced by BAs occurred quickly, initiating within 1–3 min after exposure, suggesting that it could not be mediated by FXR-α-activated alterations of gene expression. In addition, CDCA and its glycine and taurine conjugates constitute the most potent FXR-α ligands with an EC₅₀ of ~5–10 μM (28, 34), yet they had little or no effect on airway constriction. Another alternative is that cytotoxicity or inflammatory cytokines are causing BA-induced airway relaxation. In fact, exposure of an immortalized bronchial epithelial cell line to BAs during 48 h caused cell death and IL-8, IL-6, and GM-CSF release (1). However, because of the brief exposure to BAs and reversibility of their effects in airways, we do not believe such alternatives are applicable in the results we presented. Nevertheless, prolonged exposure of airways to BAs may have additional effects on reactivity that could be mediated by activation of FXR-α receptors in airway smooth muscle or by cytokines, but these alternatives require further investigation. Altogether, our data suggest that neither TGR5 nor FXR-α receptors are involved in the acute effects of BAs on airway contraction.

Muscarinic M3 receptors have been suggested as mediators of the anticholinergic effects of select BAs in secretory and other nonmuscle cells. In dispersed chief cells from guinea pig stomach, TLCA, but not TCA or taurine, strongly inhibited pepsinogen secretion induced by the cholinergic agonist carbachol (40). Furthermore, in this preparation, TLCA alone inhibited the binding of the muscarinic receptor radioligand N-methyl-[³H]scopolamine to chief cells and induced a small but significant pepsinogen secretion that was inhibited by atropine. These pioneer findings lead to the suggestion that TLCA interact with muscarinic receptors (as partial agonists) in chief cells (40). Subsequent studies by the same group confirmed this hypothesis, showing that LCA, GLCA, TLCA, DCA, GDCA, and TDCA, but not the primary BAs, inhibited N-methyl-[³H]scopolamine binding in Chinese hamster ovary (CHO) and human colon cancer cell lines expressing recombinant M3 receptors (7, 39). Our results in proximal and peripheral airways along with human airway smooth muscle cells expressing M3 receptors appear to support the hypothesis that select BAs inhibit cholinergic contractile responses of the airways by interacting with M3 receptors in airway smooth muscle due to the similar selectivity and the concentration-dependence of their relaxation. In sharp contrast, our results do not support that BAs are partial agonists of M3 receptors in airway smooth muscle, as they did not induce significant constriction in peripheral airways. There are several possible reasons why BAs do not induce airway constriction but induce pepsinogen secretion. First, airway smooth muscle cells express both M2 and M3 receptors (27, 41), and we found these receptors expressed in the peripheral airways, whereas chief cells express M1 and M3 receptors. The evidence with transgenic mice suggests that both M1- and M3-receptor subtypes mediate pepsinogen secretion (49); however, it is unknown which mediate the stimulatory and inhibitory effects of BAs in chief cells. Binding assays in CHO and human colon cancer cells expressing recombinant M3 (but not M1) receptors (7, 39) support that at least the inhibitory effects of BA are mediated by M3 receptors. If the stimulatory effects of BA in chief cells are mediated by M1 receptors, then the absence of BA-induced airway constriction could be explained by the lack of M1-receptor expression in airway smooth muscle and our inability to detect significant expression of M1 receptors in the peripheral airways despite evident expression of M2 and M3 receptors (see Table 2). Secondly, BAs could be partial agonists for M3 receptors; however, the downstream cell-signaling mechanisms required for airway constriction may differ from those for pepsinogen secretion. In airway smooth muscle, contraction requires the activation of two independent signaling pathways initiated by phospholipase C beta (PLCβ) and RhoA/Rho Kinase (50), whereas in chief cells, pepsinogen secretion is associated only with PLCβ-initiated signaling (49). It is possible that BAs alone only partially activate PCLβ signaling to induce secretion, but not the RhoA/Rho Kinase signaling ultimately required for airway smooth muscle contraction. In addition, the lack of BA-induced relaxation in resting airways is consistent with the inability of several other bronchodilators, including β₂-agonists, bitter taste receptor agonists, nitric oxide, and hydrogen sulfide, to cause relaxation in unstimulated airways (2, 6, 35, 43).

In conclusion, our data indicate that most secondary BAs acutely inhibit the cholinergic constriction of the proximal and peripheral airways in vitro in humans and rodents and that these effects are likely mediated by inhibition of muscarinic M3 receptors in the airway smooth muscle. While these current results alone do not account for the detrimental effects of chronic microaspiration of BAs on transplanted lungs, we think that these novel acute effects of antagonism of muscarinic receptor signaling are an important component of unraveling the likely multifactorial effects of aspirated BAs on multiple cell types in the lung. For example, by inhibiting cholinergic bronchoconstriction, we speculate that aspirated BAs in patients with lung transplants may further affect the innate defense of the lung, avoiding clearance of bronchial secretions, in turn stimulating inflammatory cytokine release, deranging surfactant protein and lipids, and facilitating airway infections as previously shown (3, 11, 12, 46).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

A.U., F.D., D.X., C.W.E., N.W.B., and J.F.P.-Z. conceived and designed research; A.U., D.X., C.W.E., and J.F.P.-Z. performed experiments; A.U., D.X., C.W.E., and J.F.P.-Z. analyzed data; A.U., F.D., D.X., C.W.E., and J.F.P.-Z. interpreted results of experiments; A.U., C.W.E., and J.F.P.-Z. prepared figures; A.U., F.D., C.W.E., and J.F.P.-Z. drafted manuscript; A.U., F.D., D.X., C.W.E., N.W.B., and J.F.P.-Z. edited and revised manuscript; A.U., F.D., D.X., C.W.E., N.W.B., and J.F.P.-Z. approved final version of manuscript.