To the Editor:

The enzyme leukotriene A4 hydrolase (LTA4H) is classically known for its epoxide hydrolase activity that converts leukotriene A4 (LTA4) to the neutrophil chemoattractant LTB4 [1]. In 2010, our group published a study in Science that demonstrated that during an influenza model of acute airway inflammation, LTA4H was released from cells to degrade proline-glycine-proline (PGP), a non-canonical CXCR1 and -2 agonist of polymorphonuclear neutrophil (PMN) recruitment and activation [2], thereby attenuating PMN inflammation [3].

Neutrophil elastase (NE) is a well-described serine protease that has been shown to have multiple biological activities [4] and is a biomarker of pulmonary inflammation in both α1-antitrypsin (α1-AT) deficiency and cystic fibrosis (CF) [5, 6]. Its dysregulation propagates inflammation and is believed to disrupt the pulmonary architecture through damage of structural proteins in both diseases [5, 7]. A previous manuscript demonstrated that NE is capable of degrading LTA4H in vitro [8], but the impact of this cleavage on enzymatic activity and impact of these proteolytic degradation products in patients with chronic lung disease has not been shown. Therefore, we examined if NE-specific activity would alter both structure and function of LTA4H in people with CF.

Sputum samples from CF outpatients and normal control individuals were collected after approval by the University of Alabama at Birmingham institutional review board (IRB#140414004). Human LTA4H (#10007817, Cayman), CF sputum, or NE (#324681, Sigma) were co-incubated at 37°C in the presence of NE inhibitor α1-AT (A2003-1, rPeptide) and evaluated on protein blot. Ala-pNA was used as a substrate for the aminopeptidase activity of LTA4H. ELISA assays for NE levels in CF sputum were used. Tandem mass spectrometry analyses for determining N-terminal sequencing of LTA4H fragments were performed with a q-TOF-2 mass spectrometer (Micromass, Manchester, UK) using electrospray ionisation. The predicted model of NE cleavage sites on the structure of LTA4H (PDB ID: 4MS6) and images were generated by PyMol (Schrödinger, Inc., New York, NY, USA). The statistical analyses are detailed in the figure caption.

The sputum from CF patients during periods of stability and exacerbation showed multiple LTA4H fragments by immunoblot (69, 50, 40 and 28 kDa) (figure 1a). The aminopeptidase activity was increased in the patients with exacerbation (figure 1b), and this finding was verified with direct PGP degradation (figure 1c). Both effects correspond with the emergence of the 28 kDa band on immunoblot (figure 1a). Sputum NE level was elevated in stable and further increased in exacerbation samples from people with CF (figure 1d). When purified human NE was incubated with recombinant LTA4H, the fragmentation profile of LTA4H (figure 1e) was comparable to that obtained with CF sputum with the presence of the most characteristic bands at 69, 50, 40 and 28 kDa (figure 1a). Compared to the LTA4H without NE co-incubation, NE-cleaved LTA4H showed increased aminopeptidase activity and PGP degradation (figure 1f and g, respectively). LTA4H fragmentation underwent dose-dependent cleavage by NE (figure 1h) and demonstrated increased LTA4H aminopeptidase activity, corresponding with LTA4H cleavage (figure 1i). These findings reflect a selective enhancement of aminopeptidase activity in extracellular LTA4H caused by NE cleavage. Further, the addition of the endogenous NE inhibitor α1-AT [9] showed that α1-AT was able to counteract LTA4H fragmentation and attenuated the increase in aminopeptidase activity (figure 1j and k, respectively).

a) 24 μL of sputum from stable, exacerbation cystic fibrosis (CF) subject and leukotriene A4 hydrolase (LTA4H) control (LA) were run on SDS-PAGE and probed for LTA4H via polyclonal antibody (Cat.# 160250). Black arrow corresponds to intact LTA4H (69 kDa), orange band corresponds to ∼50 kDa, green band corresponds to ∼40 kDa and blue band corresponds to ∼28 kDa. b) The aminopeptidase activity of sputum collected from patients with CF (Sta: stable, n=24; Exa: exacerbation, n=29) or healthy control (Hea, n=8) was measured by using Ala-PNA substrate. *: p=0.0001 compared to stable patients; #: p=0.0003 compared to healthy control by one-way ANOVA with Tukey's multiple comparison post-test. c) Sputum was collected from patients with CF (Sta: stable, n=24; Exa: exacerbation, n=29) or healthy control (Hea, n=5), co-incubated with the proline-glycine-proline (PGP) peptide for 24 h and PGP degradation was assessed by liquid chromatography with tandem mass spectrometry (LC-MS/MS). The percentage of PGP peptide degraded was determined relative to control samples of 1.0 mg·mL−1 PGP alone. *: p<0.0001 compared to healthy control by one-way ANOVA with Tukey's multiple comparison post-test. d) CF sputum neutrophil elastase (NE) levels were determined by using commercial ELISA kit (R&D Systems). *: p<0.001 compared to stable patients by Mann–Whitney test. e) For LTA4H cleavage by exogenous NE, human recombinant LTA4H in different doses was co-incubated with 12 μg·mL−1 recombinant NE for 30 min at 37°C. The reactions were resolved by SDS-PAGE under reducing conditions and visualised by immunoblotting. Black arrow corresponds to intact LTA4H (69 kDa), orange band corresponds to ∼50 kDa, green band corresponds to ∼40 kDa and blue band corresponds to ∼28 kDa. f) The aminopeptidase activity was measured by using Ala-PNA substrate. g) Recombinant LTA4H (LA, 50 μg·mL−1) was co-incubated with recombinant NE (12 μg·mL−1) for 30 min at 37°C, and the capacity of LTA4H to degrade PGP was subsequently determined by incubation with the PGP peptide for 1 h; PGP degradation was assessed by LC-MS/MS. The percentage of degraded PGP peptide was determined relative to control samples of 1.0 mg·mL−1 PGP alone. *: p<0.0001 compared to healthy control by one-way ANOVA with Tukey's multiple comparison post-test. n=4–5 per group. h) LTA4H (30 μg·mL−1) was incubated in the presence of varying concentrations of NE. The reactions were resolved by SDS-PAGE under reducing conditions and visualised by immunoblotting. i) The corresponding aminopeptidase activity was measured by using Ala-PNA substrate. j) Degradation of LTA4H by exogenous NE is blocked by NE inhibitor. LTA4H (30 μg·mL−1) was incubated with NE (12 μg·mL−1) in the presence of varying concentrations of α1-antitrypsin (α1-AT) (H: 125.28 μM; L: 31.32 μM). The reactions were resolved by SDS-PAGE under reducing conditions and visualised by immunoblotting. k) The corresponding aminopeptidase activity was measured by using Ala-PNA substrate. l) Degradation of LTA4H by CF sputum is blocked by α1-AT. LTA4H (30 μg·mL−1) was incubated with CF sputum (PS) in the presence of varying concentrations of α1-AT (H: 125.28 μM; M: 62.64 μM; L: 31.32 μM). The reactions were resolved by SDS-PAGE under reducing conditions and visualised by immunoblotting. m) Corresponding aminopeptidase activity was measured by using Ala-PNA substrate. The findings shown in figure 1a and e–m are illustrative of at least three independent experiments. n, o) To identify LTA4H fragments that result that from NE cleavage, bands of NE (24 μg·mL−1) cleaved LTA4H (1280 μg·mL−1) were excised from Coomassie blue stained SDS-PAGE gel and underwent trypsin digestion for N-terminal sequencing analysis via mass spectrometry to determine NE-specific cleavage site of LTA4H fragment. To map the NE cleavage site on the structure of LTA4H, the coordinates of human LTA4H (PDB ID: 4MS6) were acquired from the Protein Data Bank (https://www.rcsb.org). n) Surface model of intact LTA4H with the amino acid residues N-terminal of the 367/368 cleavage site shaded in cyan (41 kDa band), while residues C-terminal are yellow (28 kDa band) (top); and cleaved N-terminal domains with the active site region shaded in purple and bound PGP analogue in stick model (bottom). o) Active site of cleaved LTA4H modeled with the PGP peptide. In this panel, N-terminal residues are shown in cyan shades with PGP in orange carbon base. Zinc, water and chlorine are shown as purple, blue and greyed yellow spheres, respectively. Potential hydrogen bonds are shown as dashed lines. Residues within 5 Å of PGP are shown as sticks. Active site residues are shown with carbon in darker blue shade. A water molecule that sits as a potential proton donor is denoted with a red asterisk.

As CF sputum has increased NE expression (figure 1d), we next determined if CF sputum could generate fragmented LTA4H ex vivo. Using CF sputum, we identified that when incubated with recombinant whole LTA4H, the LTA4H fragments were generated in a similar pattern observed in CF sputum. However, in the presence of α1-AT, there was little degradation of LTA4H and reduced aminopeptidase activity (figure 1l and m, respectively), indicating that the cleavage activity of CF sputum was due to NE.

Our data indicate that NE in the CF lung digests LTA4H into different fragments with the unique observation that this digestion can lead to increased aminopeptidase activity with emergence of a consistent 28 kDa band. In order to better understand the mechanism by which NE cleavage enhances LTA4H aminopeptidase activity, we utilised mass spectrometry to sequence the LTA4H fragments and a structural model with LTA4H in association with the PGP analogue (PDB ID: 4MS6 [10]) to visualise the structure of LTA4H fragments (figure 1n and o). LTA4H is comprised of three domains (N-terminal amino acids (1–207), catalytic (208–450) and C-terminal (461–610)) that associate to form a large cleft to accommodate enzymatic substrates [11]. The NE cleavage site on the LTA4H fragment exists between residues 367/368. The resultant N- and C-terminal fragments of LTA4H (41 and 28 kDa, respectively) are shown in figure 1n (cyan and yellow). Removal of the C-terminal fragment of LTA4H uncovers 3751 Å2 of surface area on the N-terminal segment of the LTA4H (22.2% of the total surface area of the N-terminal domain), in turn revealing the substrate binding pocket (figure 1o), where the active aminopeptidase site is left intact. Thus, the site of LTA4H aminopeptidase activity would appear to have greater substrate accessibility for LTA4H aminopeptidase-based cleavage of PGP and shows intact binding in the aminopeptidase site after cleavage. Therefore, although the enhanced aminopeptidase activity resides in the N-terminal portion of the molecule, the consistent presence of the 28 kDa C-terminal portion serves as a unique biomarker for NE-related cleavage.

To assess if PGP would interact with the active site cleft in the aminopeptidase region, we carried out a molecular dynamics simulation of PGP in the binding pocket of the cleaved LTA4H (3–367). The coordinates of LTA4H were manually truncated, and the PGP analogue was removed from the structure. PGP peptide was taken from the PDB and docked into the aminopeptidase active site. This structure was solvated and energy minimised with YASARA [12]. The structure of the truncated LTA4H–PGP complex is shown in figure 1o. Many of the interactions in the full-length structure are maintained while new interactions are formed to enhance stability of this interaction. One key difference between the intact and cleaved structures is the loss of an interaction between Tyr-383 and a main chain carbonyl at the PGP cleavage site. Tyr-383 has been described as an essential proton donor in the peptidase reaction mechanism [13]. In the modelled structure of LTA4H (3–367), a water molecule is bound at the location previously occupied by the hydroxyl from Tyr-383 [10]. Given the apparent increased activity of the cleaved LTA4H, the water has the potential to serve as a proton donor for the cleaved enzyme.

Our results demonstrate that LTA4H, a critical enzyme in pulmonary lung inflammation, is endogenously cleaved in CF. Further, this cleavage leads to the generation of a biological fragment with enhanced aminopeptidase activity. This observation goes against conventional wisdom that cleaved enzymes lead to reduced activity and highlights the importance of understanding an enzyme's structure as a feature of its function. Although this cleavage enhances activity, it is important to note that this PGP degradation process is running in parallel with another very robust proteolytic pathway that enhances PGP generation [14] and we hypothesise that this very active PGP-generating pathway allows for the presence of PGP peptides to be observed in diseases such as CF.

The intact LTA4H is comprised of three domains that fold together to form in part the aminopeptidase binding cleft and the hydrolase active site, and represents a unique enzyme in that it harbours both an aminopeptidase active site and epoxy hydrolase active site adjacent to each other [11]. NE cleaves LTA4H within the middle catalytic domain, splitting the enzyme into two pieces. The aminopeptidase active site remains intact and exists on the N-terminal cleavage product. Assuming the C-terminal portion of LTA4H partially or completely dissociates from the rest of the protein, as much as 3751 Å2 of buried surface area on the N-terminal portion of the protein (PGP binding site) would be exposed, providing greater accessibility to the aminopeptidase active site, an interaction previously shown by Stsiapanava et al. [10].

The unique dual-enzymatic activity of LTA4H makes it an attractive therapeutic target for small molecule targeting, but important consideration should be paid to the state of the enzyme in vivo. It is possible that the enhancement of LTA4H aminopeptidase activity by cleavage may serve as an endogenous process for attenuating neutrophilic inflammation in chronic lung disease and, as such, caution should be considered in completely neutralising NE as a disease-related therapy.