Report

A Critical Role for LTA4H in Limiting Chronic Pulmonary Neutrophilic Inflammation

See allHide authors and affiliations

Science  01 Oct 2010:
Vol. 330, Issue 6000, pp. 90-94
DOI: 10.1126/science.1190594

Abstract

Leukotriene A4 hydrolase (LTA4H) is a proinflammatory enzyme that generates the inflammatory mediator leukotriene B4 (LTB4). LTA4H also possesses aminopeptidase activity with unknown substrate and physiological importance; we identified the neutrophil chemoattractant proline-glycine-proline (PGP) as this physiological substrate. PGP is a biomarker for chronic obstructive pulmonary disease (COPD) and is implicated in neutrophil persistence in the lung. In acute neutrophil-driven inflammation, PGP was degraded by LTA4H, which facilitated the resolution of inflammation. In contrast, cigarette smoke, a major risk factor for the development of COPD, selectively inhibited LTA4H aminopeptidase activity, which led to the accumulation of PGP and neutrophils. These studies imply that therapeutic strategies inhibiting LTA4H to prevent LTB4 generation may not reduce neutrophil recruitment because of elevated levels of PGP.

Neutrophils are a critical component of the host’s defense against microorganisms (1, 2); however, cessation of neutrophil recruitment and clearance by apoptosis is mandatory to restore homeostasis and limit host tissue damage (3, 4). Chronic neutrophilic inflammation is observed in lung diseases such as cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD), mediating extensive tissue damage and contributing to organ dysfunction (5, 6). A classical chemoattractant for neutrophils is the glutamic acid–leucine–arginine (ELR+) motif containing CXC chemokines (7), such as interleukin (IL)–8 (CXCL8) in humans and keratinocyte-derived chemokine (KC) (CXCL1) and macrophage inflammatory protein (MIP)–2 (CXCL2) in mice. Proline-glycine-proline (PGP) is a tripeptide generated from the breakdown of extracellular matrix collagen and is specifically chemotactic for neutrophils in vitro and in vivo (8, 9). N-terminal acetylation of PGP (AcPGP), which occurs through an unknown mechanism, can enhance its chemotactic potential (10). PGP and AcPGP share sequence homology to key motifs found in the majority of ELR+ CXC chemokines and bind to their receptors, CXCR1 and CXCR2 (8, 9). PGP is generated from native collagen by the action of matrix metalloproteinase–8 (MMP-8) and/or MMP-9, followed by a secondary cleavage by prolyl endopeptidase (PE) (9). We have subsequently identified substantial quantities of PGP or AcPGP in clinical samples from patients with chronic lung diseases such as CF, COPD, and bronchiolitis obliterans syndrome (8, 9, 11, 12), where it functions to promote the maintenance of neutrophilic inflammation at a time of declining classical chemokine levels. Here, we investigate the role of PGP in acute pulmonary neutrophilic inflammation.

Influenza infection of mice elicits an acute neutrophilic inflammation peaking at 24 hours post-infection (fig. S1, A and B) (13), coinciding with peak KC and MIP-2 levels (fig. S1, C and D). MMP-9 protein expression and activity were elevated at 24 hours post-infection and persisted for several days before subsiding to baseline levels (fig. S1, E to G). PE activity (fig. S1H) and protein amounts (fig. S1I) were also rapidly elevated in bronchoalveolar lavage fluid (BALF) after influenza infection, peaking at 24 hours after infection. Thus, the entire enzyme repertoire required for PGP generation was present. Neutrophils were demonstrated to be a prominent source of both MMP-9 and PE (fig. S1, J and K) (13). Despite the presence of MMP-9 and PE, however, no PGP was detectable at any time point that we analyzed (fig. S2A) with the use of a highly sensitive mass spectrometry technique (8).

The absence of PGP during influenza infection led us to question whether an enzyme was present that degraded PGP. BALF from 24 hours post–influenza infection (maximal MMP-9/PE) exhibited a potent capacity to degrade exogenous PGP, as determined by mass spectrometry (Fig. 1A). Generation of free proline (from PGP degradation) by influenza BALF correlated with the loss of PGP by mass spectrometry (Fig. 1B), with one proline being cleaved per PGP molecule. N-terminal acetylation of PGP protected it from degradation (fig. S2B). BALF from different time points post–influenza infection was subsequently incubated with PGP, and degradation was assessed by mass spectrometry and free-proline generation (Fig. 1C). Naïve BALF had the capacity to degrade PGP, but activity was markedly increased after influenza infection (peaking at 24 to 48 hours) before subsiding to basal activity. Lung epithelial cells and neutrophils were shown to be prominent sources of this PGP-degrading activity (fig. S2, C to E) (13). Thus, both PGP-generating and -degrading enzymes are similarly up-regulated during influenza infection coinciding with neutrophilic inflammation.

Fig. 1

Influenza infection induces the release of PGP-degrading enzymes. (A and B) Undiluted, diluted by 1/10, or diluted by 1/100 BALF from 24 hours after influenza infection was incubated with PGP, and degradation was assessed by mass spectrometry (A), expressed as percentage degradation relative to peptide alone, or (B) generation of free proline. (C) BALF diluted by 1/10 from indicated times post–influenza infection was incubated with PGP, and degradation was assessed after 2 hours by mass spectrometry or free-proline generation. Data [error bars indicate mean ± SD (A and B) or mean ± SEM (C)] are representative of three experiments with triplicates (A and B) or four mice per group (C). Abs, absorbance.

The enzyme responsible for degrading PGP was isolated from influenza BALF and demonstrated to be leukotriene A4 hydrolase (LTA4H) (fig. S3 and table S1) (13). LTA4H acts as an epoxide hydrolase to catalyze the conversion of LTA4 to the proinflammatory mediator leukotriene B4 (LTB4) (14, 15). It has also been suggested to have aminopeptidase capacity, but no physiological substrate has been identified (16, 17). We incubated recombinant human LTA4H with PGP with and without bovine serum albumin (BSA)—because its aminopeptidase activity is reportedly enhanced by albumin (18)—and assessed PGP degradation (Fig. 2, A and B). LTA4H had a potent capacity to degrade PGP, which was augmented at lower enzyme concentrations by BSA. Recombinant LTA4H did not degrade AcPGP (fig. S4A). LTA4H displayed a significantly greater catalytic constant (kcat) for a PGP substrate than for LTA4, but it also possessed a greater Michaelis constant (KM). Consequently, the catalytic efficiency (kcat/KM) of LTA4H for PGP was ~1 × 105 s−1 M−1, which compares favorably with reported specific activities for LTA4H with LTA4 (fig. S4, B to D, and table S2) (13, 17). Murine recombinant LTA4H was expressed and purified (fig. S5, A to C) and demonstrated to possess a comparable capacity to degrade PGP (fig. S5, D to G).

Fig. 2

Leukotriene A4 hydrolase degrades PGP. (A and B) Degradation of PGP by 1, 0.1, and 0.01 μg/ml recombinant human LTA4H with or without BSA, as determined by (A) mass spectrometry or (B) generation of free proline. (C) Generation of PGP from type II collagen by BALF from naïve and influenza-infected (24 hours) mice, with or without bestatin. (D) PGP in BALF and (E) neutrophil numbers in lung tissue 24 hours after influenza infection in mice that received intratracheal phosphate-buffered saline (untreated), bestatin (Best.), or amastatin (Amast.). *P < 0.05; **P < 0.01. Data [error bars indicate mean ± SD (A) to (C) or mean ± SEM (D) and (E)] are representative of three experiments with triplicates (A) to (C) or four mice per group (D) and (E).

LTA4H selectively cleaved the N-terminal amino acid from tripeptides and was inhibited by the tripeptide aminopeptidase inhibitor bestatin, but not other general aminopeptidase inhibitors such as amastatin (fig. S6, A and B) (13). Biological samples containing MMP-9/PE can generate PGP from collagen ex vivo (9). Although BALF incubated with collagen failed to generate PGP, co-incubation with bestatin yielded small quantities of PGP with naïve BALF and substantial quantities with BALF from 24 hours post–influenza infection (Fig. 2C and fig. S2C). Bestatin or amastatin was subsequently administered intratracheally to influenza-infected mice. Although we did not detect PGP in the BALF from untreated mice at 24 hours post-infection, those administered bestatin generated substantial concentrations of PGP in their BALF (Fig. 2D) and possessed significantly greater numbers of neutrophils in their lung tissue (Fig. 2E). It would, therefore, appear that LTA4H degrades collagen-derived PGP, limiting neutrophilic inflammation.

To verify a role for LTA4H in PGP degradation in vivo, we next assessed the amount of PGP in the BALF of wild-type (WT) (fig. S7, A to D) (13) and Lta4h–/– mice. Naïve WT mice possessed no PGP, whereas Lta4h–/– mice exhibited low but reproducible concentrations of PGP (Fig. 3A). This phenotype was more pronounced at day 3 post–influenza infection (peak neutrophil infiltrate) (Fig. 3B). Lta4h–/– mice failed to produce LTB4 but also possessed lower amounts of KC and MIP-2 in BALF relative to wild types (Fig. 3, A and B). There was a small yet significant increase in neutrophil numbers in the lungs of Lta4h–/– mice at day 3 of influenza infection (Fig. 3C); however, the contribution of PGP may be partially masked because of the reduced amounts of other chemoattractants. To address this, we measured neutrophil chemotaxis ex vivo to WT and Lta4h–/– BALF (day 3 post–influenza infection) in the presence of a PGP-neutralizing antibody. BALF from both groups of mice was chemotactic for neutrophils, and chemotaxis was suppressed by neutralization of KC/MIP-2 (fig. S8A). Neutralization of PGP, however, caused a marked reduction in the chemotactic potential of Lta4h–/– BALF only (Fig. 3D). The BALF of WT mice could degrade exogenous PGP, which was markedly increased with influenza infection; however, degradation was absent with BALF from Lta4h–/– mice (Fig. 3, E and F). Furthermore, whereas BALF from influenza-infected WT mice generated negligible levels of PGP from collagen ex vivo (fig. S8, B and C), BALF from Lta4h–/– mice generated significant quantities. PGP generation was significantly reduced by co-incubation with inhibitors to MMP-9 and PE, verifying the importance of these enzymes in the generation of the tripeptide (fig. S8, B and C).

Fig. 3

Lta4h–/– mice lose capacity to degrade PGP. The concentrations of KC, MIP-2, LTB4, and PGP in BALF of (A) naïve or (B) influenza-infected (day 3) WT and Lta4h–/– mice are shown. (C) Neutrophil numbers in the lungs of naïve or influenza-infected (day 3) WT or Lta4h–/– mice. (D) Inhibition of WT and Lta4h–/– BALF induced neutrophil chemotaxis ex vivo by PGP neutralization. (E and F) BALF, diluted by 1/10, from naïve or influenza-infected (day 3) WT and Lta4h–/– mice incubated with PGP and degradation assessed after 2 hours by (E) mass spectrometry or (F) free-proline generation. *P < 0.05; **P < 0.01. Data [error bars indicate mean ± SEM (A to C) or mean ± SD (D to F)] are from four mice per group (A) to (C) or triplicates (D) to (F).

Failure to detect PGP owing to degradation of the tripeptide was not unique to influenza infection, because we obtained comparable results in response to Streptococcus pneumoniae (fig. S9) (13). We next questioned why PGP was present in chronic neutrophilic conditions such as COPD (12), because LTA4H so potently degrades it. Chronic instillation of AcPGP into the lungs of mice results in neutrophil recruitment and an emphysematous lung phenotype (8, 19). We demonstrated that cigarette smoke condensate (CSC) could N-terminally acetylate PGP (Fig. 4A), not only protecting PGP from degradation by LTA4H but also enhancing its chemotactic capacity (10). Exposure to CSC also dose-dependently inhibited the capacity of LTA4H to degrade PGP (Fig. 4, B and C). CSC appeared to exhibit selective inhibition of the aminopeptidase activity of LTA4H, with minimal inhibition of the hydrolase activity observed (Fig. 4D). Furthermore, no difference was observed in LTB4 generation in response to the chemoattractant formyl-methionine-leucine-phenylalanine when neutrophils were in the presence of CSC (fig. S10). Selective inhibition of aminopeptidase activity is feasible because residues have been delineated that are critical to one activity but have no role in the other (20, 21), and selective modulation has previously been observed in response to biological and chemical mediators (18, 22, 23).

Fig. 4

Cigarette smoke inhibits PGP degradation by LTA4H. (A) Acetylation of PGP by 0.5, 1, 2, and 4% CSC or vehicle control (2% dimethyl sulfoxide). (B and C) Degradation of PGP by LTA4H alone or in the presence of 0.5, 1, or 2% CSC or vehicle control, as determined by (B) mass spectrometry or (C) generation of free proline. (D) Inhibition of hydrolase and aminopeptidase activities by 2% CSC. (E to G) Human lung epithelial cells cultured with 2% CSC or vehicle control for 24 hours. (E and F) Degradation of PGP after 2 hours by apical supernatant (diluted 1/10), as assessed by (E) mass spectrometry or (F) free-proline generation. (G) Western blot for LTA4H in apical supernatant from 2% CSC or vehicle control–treated cells. Each lane represents apical supernatant from a different experiment. (H and I) BALF, diluted 1/10, from vehicle control– or CSE-treated mice (24 hours post-treatment) incubated with PGP and degradation assessed after 2 hours by (H) mass spectrometry or (I) free-proline generation. AcPGP (J) and neutrophil numbers (K) in BALF of mice 24 hours post-administration of vehicle control or CSE. *P < 0.05; **P < 0.01. Data [error bars indicate mean ± SD (A) to (F) or mean ± SEM (H) to (K)] are representative of at least triplicates (A) to (F) or three experiments with four mice per group (H) to (K).

We have previously demonstrated that lung epithelial cells release LTA4H (fig. S2C). Exposure of primary human bronchial epithelial cells to a nonlethal dose of CSC significantly reduced the aminopeptidase activity of the released LTA4H (Fig. 4, E and F), while substantially increasing the detectable amount of total LTA4H protein released (Fig. 4G). It is important to note that, in this instance, although the apparent percentage inhibition of activity is less than 30%, the amount of LTA4H protein released by the CSC-stimulated epithelial cells is well in excess of that released by control-treated cells, so the actual inhibition is far greater. The increase in total LTA4H protein with CSC is in excess of 10-fold; thus, the actual relative inhibition is likely to be in excess of 95%. Furthermore, single exposure of mice to intranasal cigarette smoke extract (CSE) markedly reduced the capacity of BALF (24 hours later) to degrade PGP ex vivo (Fig. 4, H and I) and led to significant concentrations of AcPGP in BALF (Fig. 4J) and neutrophil recruitment (Fig. 4K) in vivo.

In this study, we propose disparate functions of LTA4H in generating one neutrophil chemoattractant (LTB4) while degrading another (PGP). LTA4H is generally assumed to be a cytosolic enzyme, but we report an extracellular role, corroborating detection of this enzyme previously in the plasma of various mammals and in human BALF (24, 25). The peptidase, but not the epoxide hydrolase, activity of LTA4H is augmented in the presence of albumin (18) and chloride ions (22). The disparate concentrations of Cl and albumin, intracellularly and extracellularly, support the notion that the epoxide hydrolase activity of LTA4H operates intracellularly, and the aminopeptidase capacity occurs extracellularly. Neutrophils are a source of enzymes that both generate and degrade PGP and are therefore capable of promoting and then restricting their own recruitment.

Our results also suggest that cigarette smoke shifts the emphasis of LTA4H, which has dual pro- and anti-inflammatory functions, toward a proinflammatory phenotype. Consequently, a combination of LTB4 and PGP, both implicated in the pathogenesis of COPD (8, 12, 19, 26), could drive the observed inflammation. Such selective modulation of the disparate activities of LTA4H may not be unique to COPD. Extracellular levels of chloride ions, which selectively activate the aminopeptidase activity, are diminished in CF patients owing to defective CF transmembrane conductance regulators (27) and could explain the PGP observed in these patients (9).

Finally, LTB4 is implicated in acute and chronic inflammatory diseases (2831). The development of drugs that seek to inhibit LTA4H activity and LTB4 generation to alleviate pathologies of inflammatory disease is under way. When developing these drugs, it will be critical to consider the potential repercussions of preventing PGP degradation.

Supporting Online Material

www.sciencemag.org/cgi/content/full/science.1190594/DC1

Materials and Methods

SOM Text

Figs. S1 to S10

Tables S1 and S2

References and Notes

  1. For supplementary figures, methods, and text, see the supporting material on Science Online.
  2. This project was supported by grants from the Wellcome Trust (082727/Z/07/Z to R.J.S.), the National Heart, Lung, and Blood Institute (HL07783, HL090999, and HL087824 to J.E.B. and HL102371-A1 to A.G.), the Cystic Fibrosis Foundation (GAGGAR07 to A.G.), the National Institute of Diabetes and Digestive and Kidney Diseases (1K23DK075788 and 1R03DK084110-01 to S.M.R.), the Flight Attendant Medical Research Institute (Young Clinical Scientist Award to Y.M.S.) and the National Heart, Lung, and Blood Institute (K08HL091127 to Y.M.S.), and the Medical Research Council (P171/03/C1/048 to T.H.). The Univ. of Alabama Birmingham (UAB) Lung Health Center Pulmonary Proteomics Laboratory is funded through the NIH (by grants RR19231, P30CA13148, P50 AT00477, U54CA100949, P30AR050948, and P30 DK079337). We thank R. Moore and L. Wilson of the UAB Targeted Metabolomics and Proteomics Laboratory for their technical assistance with mass spectrometry, D. Muccio of the UAB Chemistry Department for the use of his fast protein liquid chromatography system, G. Xia for technical assistance, and D. Saliba of the Kennedy Institute of Rheumatology, Imperial College London, for his technical assistance in the generation of murine LTA4H. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute or the NIH.
View Abstract

Navigate This Article