Published online before print January 22, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Immune Disease Institute and
Department of Pediatrics, Harvard Medical School, Boston, Massachusetts, USA;
Departments of Pathology and Immunology, Washington University School of Medicine, Saint Louis, Missouri, USA; and
Department of Pediatrics, University of Colorado School of Medicine and The Children’s Hospital, Denver, Colorado, USA
1 Correspondence: Immune Disease Institute, 800 Huntington Avenue, Boston, MA 02115, USA. E-mail: remold{at}cbr.med.harvard.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: cystic fibrosis • calcium ion • elastase • proteinase-3
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Increasing evidence implicates the hydrophilic surfactant proteins A and D (SP-A, SP-D) as vital "collaborators" with macrophages and neutrophils in pulmonary antimicrobial and anti-inflammatory defense (reviewed in refs. [2 3 ]). SP-A and SP-D are collagenous lectins (collectins) consisting of a short N-terminal domain, collagen domain, hydrophobic neck, and Ca2+-dependent (C-type) lectin domain. The basic subunit is a trimer consisting of an extended N-terminal "tail" that includes a triple helical collagen domain, coiled-coil neck region, and three C-terminal lectin "heads." The trimeric structure is stabilized by interchain disulfide bonds within the short N-terminal peptide. SP-D trimers can form dodecamers or higher-order structures that are stabilized further by intersubunit disulfide bonds involving the N-terminal cysteines. SP-D and SP-A are secreted by alveolar type II cells and nonciliated bronchiolar epithelial cells and are well-positioned to participate in pulmonary host defense. The lectin domains bind in a Ca2+-dependent manner to pathogen-associated molecular patterns on microbes, and clearance of the microbes by macrophages and neutrophils is thereby enhanced. SP-A and SP-D also enhance uptake of apoptotic neutrophils by alveolar macrophages [4 5 ], directly inhibit bacterial proliferation [6 ], and regulate pro- and anti-inflammatory cytokine production by macrophages [7 8 ].
Decreased levels of SP-A and SP-D have been found in inflammatory pulmonary disease, including bacterial pneumonia and acute respiratory distress syndrome [9 10 ]. The deficit is particularly marked in CF patients, especially for SP-D [11 12 ]. In a CF study that examined multiple disease parameters, SP-D levels were particularly low in patients with infection [13 ]. Several lines of evidence indicate that proteolytic degradation underlies, or contributes to, depletion. The pathogen is a possible source suggested by the finding of a metalloprotease and a trypsin-like protease from P. aeruginosa that cleaves SP-D in vitro [14 15 ]. Another source of SP-D-degrading activity is the neutrophil serine proteases (NSPs): elastase, cathepsin G, and proteinase-3. These are released from recruited neutrophils on degranulation or necrotic death. Indeed, human and rat SP-D were cleaved in vitro by low levels of the NSPs and generated N-terminal fragments lacking bacterial-aggregating activity [16 17 ]. Our preliminary studies of lavage from 22 patients showed considerable variation in the extent and pattern of cleavage. This prompted further investigation of the mechanisms of SP-D degradation.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
For the study of SP-D, aliquots of 22 CF BALF, 12 elastase-positive and 10 elastase-negative, were each randomly drawn from the larger panel of specimens. Eight control (non-CF) BALF were also randomly drawn; these samples consisted of aliquots of BAL performed for clinical indications in patients with chronic cough (three), asthma (two), immunodeficiency (one), chronic wheezing (one), and neuromuscular disease (one). The median age for the CF patients was 7.9 years (range, 1.2–24.0 years) and for the control patients, 10.3 years (range, 1.4–14.6 years). For the CF patients, older individuals were disproportionately represented in the elastase-positive subgroup (Supplementary Table S1). Urea levels were similar in both groups; the mean was 0.42 mg/dl and range, 0.1–1.1 for the CF patients; and 0.40 mg/dl and 0.1–1.2 for the control patients, indicating similar dilution factors in the two groups. Six of the CF BALF specimens and four of the control BALF were further studied for cleavage of SP-D in vitro, fractionation on two-dimensional (2D) gel electrophoresis, and bacterial aggregation. The major criterion for sample selection was availability of sufficient volume of the specimen for study; other criteria (i.e., elastase content, culture status) are named in the accompanying text and figure legends.
Reagents
Recombinant human [20
] and rat [20
] SP-D dodecamers were expressed in Chinese hamster ovary cells and purified by maltosyl-agarose affinity chromatography and gel filtration. Trimeric human neck plus C-type carbohydrate recognition (lectin) domain (NCRD) of SP-D was expressed in bacteria, isolated from inclusion bodies, and purified by sequential nickel affinity and gel filtration chromatography [21
]. Human neutrophil elastase from purulent sputum (Elastin Products, Owensville, MO, USA) and human neutrophil cathepsin G and proteinase-3 from blood leukocytes (Athens Research and Technology, Athens, GA, USA) were stored as aliquoted stocks (1 mg/ml) at –20°C. Recombinant human monocyte/neutrophil elastase inhibitor (MNEI; SerpinB1) was expressed in insect cells and purified [22
]. Recombinant secretory leukocyte protease inhibitor (SLPI) was obtained from R&D Systems, Inc. (Minneapolis, MN, USA), reconstituted with PBS, and stored as 1 mg/ml aliquoted stocks at –20°C. Rabbit polyclonal P13 antibody prepared using the collagenase-resistant C-terminal region of human proteinosis SP-D [23
], goat C18 antibody to a C-terminal peptide of human SP-D (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and rabbit AB3434 antibody generated to mouse SP-D and cross-reactive with rat SP-D (Chemicon, Temecula, CA, USA) were used in Western blots. Mouse mAb to human SP-D (The Antibody Shop, Copenhagen, Denmark) was used for immune precipitation. Diisopropyl fluorophosphate (DFP), PMSF, phosphoramindon, and E64 were from Sigma Chemical Co. (St. Louis, MO, USA). Z-Phe-Ala-diazomethylketone (Z-Phe-Ala-CHN2), a cysteinyl protease inhibitor, was from Bachem (King of Prussia, PA, USA). Leupeptin was from Roche (Indianapolis, IN, USA).
Quantitation of active elastase
Elastase activity was quantified in BALF specimens by measuring hydrolysis of MeO-suc-Ala-Ala-Pro-Val-p-nitroanilide (Sigma Chemical Co.) against human neutrophil elastase standard as change in absorbance at 410 nm (A410 nm) over time [18
], with or without the inhibitor MeO-Suc-Ala-Ala-Pro-Val-CH2Cl. The cutoff, i.e., the level at which the coefficient of variation of repeat assays is less than 15%, is 0.5 µg/ml. Although the two elastinolytic NSPs, proteinase-3 and elastase, share reactivity with some substrates, the specificity of this assay for neutrophil elastase was demonstrated in preliminary experiments in which purifed proteinase-3 produced no cleavage under conditions of the assay (Supplementary Fig. S1).
Evaluation of SP-D in BALF by Western blot
For quantitation of SP-D, aliquots of BALF were combined with 5x SDS solution (10% SDS, 25% glycerol, 0.1 mg/ml bromphenol blue, 5% ME), heated at 100°C for 2 min, electrophoresed on 4–20% polyacrylamide gels (Invitrogen, Carlsbad, CA, USA), and transferred onto 0.2 µm nitrocellulose. The membranes were blocked with 20% milk solids in PBS-Tween (0.05%) and stained for 1 h with 1:1000 dilution of P13 rabbit anti-human SP-D antibody in PBS-Tween-0.1% milk solids. The blots were washed with PBS-Tween and incubated with 125I-labeled goat anti-rabbit IgG. Full-length bands (43 kDa) were detected by phosphorimaging with the Storm 860 scanner (Molecular Dynamics, Sunnyvale, CA, USA) and quantified relative to recombinant human SP-D standards of known concentration electrophoresed in parallel. Results are the mean of triplicate assays. For the detection of SP-D fragments, BALF aliquots were solubilized with and without 5% ME.
SP-D proteolysis reactions
CF BALF samples (15 µl) were incubated in the presence or absence of protease inhibitors DFP, MNEI, EDTA, phosphoramindon, PMSF, leupeptin, E64, Z-Phe-Ala-CHN2, or SLPI for 5 min at 37°C, and recombinant rat SP-D (6 ng) was added to a final vol of 20 µl and incubated at 37°C for 0–120 min. The concentration of Ca2+ in lavage specimens was estimated at 200 µM; this level was not further supplemented in proteolysis reactions. Recombinant human SP-D (250 ng) was incubated in TBS (20 mM Tris-HCl, pH 7.4, 150 mM NaCl) with varying protease concentrations in a final vol of 20 µl at 37°C for 10 min (200 µM CaCl2) or 120 min (2 mM CaCl2). For analysis of fragments, recombinant human SP-D (1 µg) or the NCRD fusion protein (1 µg) was incubated in TBS with varying protease concentrations as described above in the presence or absence of protease inhibitors DFP, MNEI, or PMSF. The reactions were stopped by adding DFP (2 mM), incubating at room temperature for 2 min, and solubilizing with 5x SDS solution (BALF reactions) or 2x SDS solution with or without ME.
SP-D products were detected by Western blots on 4–20% acrylamide gels (unless otherwise noted), which were stained with rabbit P13 (1:1,000), goat N14 (1:500), or goat C18 (1:250) anti-human SP-D antibodies or with rabbit AB3434 anti-mouse SP-D antibody (1:1000). The membranes were probed with 125I-labeled goat anti-rabbit antibody as described above or with HRP-conjugated goat anti-rabbit (Upstate, Charlottesville, VA, USA) or donkey anti-goat antibodies (Santa Cruz Biotechnology) and developed using WestPico reagents (Pierce Biotechnology, Rockford, IL, USA), according to the manufacturer’s instructions.
To obtain N-terminal sequence of fragments, SP-D products generated by incubation with 0.25 µg/ml elastase in 200 µM CaCl2 for 10 min were separated on 18% polyacrylamide-reducing gels and transferred to polyvinylidene difluoride (PVDF). The 11-kDa and 8-kDa fragments were visualized by staining with Coomassie blue and separately excised for automated Edman degradation.
2D gels
Samples were prepared for isoelectrophoresis and run according to the Zoom IPGRunner system (Invitrogen). Briefly, BALF sample (15 µl) was added to rehydration buffer containing 8 M urea and DTT and incubated with a pH 3–10 strip overnight and electrofocused according to system protocol. The strips were equilibrated with SDS sample buffer, separated on 4–20% Tris-glycine Zoom gels, transferred onto 0.2 µ PVDF (Millipore Corp., Bedford, MA), and stained with P13 anti-SP-D antibody.
Aggregation assays
Escherichia coli strain BL21, which is derived from K-12 and is of the rough type [24
], was grown in Luria broth to stationary phase. Bacteria were pelleted and resuspended in PBS containing 0.2 mM CaCl2. Aliquots of patient BALF and/or rat SP-D were diluted to 200 µl with PBS and preincubated for 2 h at 37°C with inhibitors and additives as indicated in figure legends. The preincubated BALF were combined with the bacterial suspension in a total vol of 1.0 ml with final density adjusted to A700 nm of 0.8 and were incubated at room temperature. Aggregation was followed by the decrease of A700 nm measured at 20 min intervals over 3 h.
Immunoprecipitation of SP-D fragments
Protein G-Sepharose (GammaBind Plus Sepharose, Amersham Biosciences, Piscataway, NJ, USA; 15 µl pelleted resin) was incubated with 5 µg mAb at room temperature for 1 h and was washed by centrifugation in TBS. Elastase-treated SP-D fragments were diluted to 250 µl with TBS and incubated for 2 h with the resin-bound mAb. The mixture was centrifuged and the supernatant removed and combined with 2x SDS solution. The pelleted resin was washed with TBS, and bound material was eluted with 1x SDS at 100°C for 5 min. Pellet and supernatant were analyzed by Western blot.
Statistics
Statistical testing of SP-D levels between groups was done using Student’s t-test (nonpaired). A P value less than 0.05 was considered significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (19K): [in a new window] |
Figure 1. Quantitation of SP-D in BALF of CF patients. SP-D was quantified by Western blot (reducing conditions) and densitometry of the 43-kDa monomer band (inset). Each symbol represents the mean of triplicate assays for an individual patient. Data for control patients (left), culture-negative CF patients (middle), and culture-positive (>300 organisms/ml) CF patients (right). *, The single culture-positive control patient (left group). •, Elastase-positive (NE+; >0.5 µg/ml) CF specimens;  , elastase-negative (NE–) CF and control specimens.
|
 View larger version (88K): [in a new window] |
Figure 2. SP-D fragments in BALF of CF patients. (A) Schematic of SP-D (355 amino acids). Shown are the N-terminal region with interchain disulfide bonds (vertical bars) that stabilize dodecamers and higher-order species, the collagen domain, hydrophobic neck region, and lectin domain with intrachain disulfide bonds. The bracket indicates the collagenase-resistant NCRD region used as immunogen for P13 antibody. (B) Western blots with P13 antibody. Representative BAL samples were electrophoresed on 4–20% acrylamide under reducing conditions (upper) or nonreducing conditions (lower). The upper one-third of the reduced blot was blank and is not shown. In lavage from control patients, P13 detected a major reduced species of 43 kDa, consistent with monomeric SP-D and under nonreducing conditions, detected species consistent with oligomers that migrate more slowly than the 120-kDa standard, dimers ( 80 kDa), and monomers. Arrows on the right of the reduced CF specimens indicate fragments of 35 kDa, 27 kDa, and 8 kDa in Lanes 5, 6, 8–10, and 12–14. In nonreduced CF specimens, varying amounts and sizes were found of higher molecular weight-cross-linked species, the 35-kDa fragment was decreased in relative amount or absent, and a 27-kDa fragment was noted as well as the smallest fragment (arrow), which migrated at 11 kDa.
|
Although information is not available for proteases produced by A. xylosoxidans and other CF pathogens, the P. aerugininosa proteases that have been studied in vitro generated N-terminal SP-D fragments indistinguishable based on size from NSP-generated fragments [14 15 ]. We therefore used sensitivity to protease inhibitors as a means to characterize the endogenous SP-D-cleaving proteases. Cleavage of recombinant rat SP-D by protease in CF BALF was inhibited by the broad-spectrum serine protease inhibitors DFP and PMSF and by SerpinB1/MNEI, an inhibitor of the NSPs [22 ], and was not inhibited by the metalloprotease inhibitor EDTA or by phosphoramindon, an inhibitor of the metalloelastase secreted by P. aeruginosa [26 ] (Supplemental Fig. S2, left panels) or by inhibitors of cysteinyl proteases: leupeptin, E64, and Z-Phe-Ala-CHN2 (right panels). As anticipated, the inhibitors had no direct effect on SP-D (bottom panels). These findings indicate that NSPs in a subset of CF BALF cleave recombinant SP-D in vitro.
To determine which of the NSPs in CF BALF cleave recombinant SP-D, protease-positive BALF of two CF patients were evaluated for sensitivity to SLPI, which inhibits elastase and cathepsin G but not proteinase-3 [27 ]. We first verified that SerpinB1/MNEI completely inhibited rat SP-D cleavage by the two tested CF BALF, whether examined after short or long incubation (Supplementary Fig. S3). In contrast, SLPI partially inhibited SP-D cleavage when examined after short incubation (30 min for CF1 and 10 min for CF2), indicating that SP-D was cleaved in vitro by endogenous SLPI-sensitive protease (elastase and/or cathepsin G) and by endogenous SLPI-insensitive protease, i.e., proteinase-3. When the incubation time was extended, recombinant SP-D was completely insensitive to SLPI (Supplementary Fig. S3), indicating that on longer incubation, the amount of proteinase-3 in CF BALF was sufficient to cleave all added SP-D (Supplementary Fig. S3).
Cleavage of human SP-D by NSPs liberates disulfide-cross-linked C-terminal fragments
To characterize the novel, immunoreactive SP-D fragments in CF lavage, we used 2D isoelectrofocusing. SP-D of control BALF migrated at 43 kDa as a series of isoforms that focused in the neutral through basic range (Fig. 3A
, upper). In CF BAL samples positive for infection and elastase, the 43-kDa species were largely replaced by a 35-kDa, more-basic species (at pH 9) and an acidic species that focused at pH 4 (Fig. 3A
, lower), consistent with the derivation of these fragments from the N terminus and C terminus, respectively. To determine whether the same fragments, particularly the novel, small fragment, could be generated in vitro by NSPs, recombinant human SP-D was briefly incubated with these proteases. As all three of these enzymes can be present in CF airway fluids [28
29
30
], each was evaluated separately. Elastase treatment generated a 35-kDa fragment and an 11-kDa fragment (reducing conditions); the latter decreased in amount, as an 8-kDa fragment accumulated, suggesting that 11 kDa is further cleaved to 8 kDa (Fig. 3B
, upper). Under nonreducing conditions, there were large, disulfide-bonded fragments, but the only small fragment migrated at 11 kDa (not shown). The elastase-generated 11- and 8-kDa fragments electrofocused at pH 4 (Fig. 3B
, lower). Edman sequencing identified the N terminus of the 11-kDa fragment as AGFVKPFTE, indicating cleavage after Thr-247 within the lectin domain. Two sequences were found for the 8-kDa fragment, AKNEAA and FLSMTDSKTE, indicating cleavage after Val-285 and Ala-291. These findings indicate that elastase cleaves human SP-D in vitro after Thr-247 to generate a fragment of 11,582 Da and a predicted isoelectric point of 4.4, which is further cleaved after Val-285 or Ala-291 to generate subfragments that remain covalently attached via an intrachain disulfide bond (Fig. 3C)
. Extrapolation to the CF lavage specimens strongly suggests that the small C-terminal fragment that persists in BAL is the two-chain, disulfide-bonded, 11,582 Da species beginning with Ala-248 with internal cleavage after Val-285 (subfragments of 3868 and 7714 Da) or Ala-291 (4452 and 7130 Da).
 View larger version (58K): [in a new window] |
Figure 3. Characterization of endogenous and in vitro-generated SP-D fragments. (A) 2D gels of non-CF control (upper) and CF (lower) BAL specimens fractionated by isoelectrofocusing and SDS electrophoresis (reducing conditions) and stained with P13 antibody. The dotted oval outlines the 8-kDa fragment, which was found only in CF samples and focused at pH 4. The 2D gels are representative of two control BALF and three CF BALF positive for elastase and infection. (B) Recombinant human (rh)SP-D treated with the indicated concentration of elastase (NE) for 10 min in TBS with 200 µM CaCl2 and examined by SDS-PAGE (upper); the 2D gel (lower) shows human SP-D after treatment with 0.25 µg/ml elastase. Both gels were stained for protein with Aurodye (gold stain). (C) Schematic of the C-terminal lectin domain showing the disulfide bonds and sites of cleavage by elastase. (D) Immune precipitation of SP-D fragments. Recombinant human SP-D treated with 0.1 µg/ml elastase was immune-precipitated with Hyb246-08 mAb to human SP-D or irrelevant isotype control (M14) mAb. The precipitated proteins (P) and supernatants (S) were examined by Western blot (reducing conditions) stained with P13. NS, An unidentified nonspecific component introduced during immune precipitation.
|
Pattern of SP-D cleavage by each of the NSPs depends on calcium ion concentration
Treatment of recombinant human SP-D with protease-3 and cathepsin G also generated fragments of 35, 11, and 8 kDa (reduced gels; Fig. 4A
, middle and bottom). To quantify the relative capacity of the three NSPs for cleaving SP-D, the concentration of protease for 50% SPD cleavage was determined by plotting the amount of remaining SP-D (determined by densitometry of duplicate blots) versus protease concentration. The capacity to cleave SP-D was greatest for (50% cleavage by less than 0.04 µg/ml elastase) followed by proteinase-3 (50% cleavage by 0.08 µg/ml). Cathepsin G had the weakest SP-D cleaving activity (0.37 µg/ml required for 50% cleavage).
 View larger version (64K): [in a new window] |
Figure 4. Effect of protease concentration and calcium ion concentration on SP-D cleavage by each of the NSPs. Recombinant human SP-D (250 ng in 20 µl) was treated with the indicated concentrations of elastase (top), proteinase-3 (middle), and cathepsin G (bottom) for (A) 10 min in the presence of 200 µM CaCl2 or (B) 2 h in the presence of 2 mM CaCl2. The blots (reducing conditions) were stained with P13 antibody.
|
Generation of the 27-, 21-, and 14-kDa fragments by elastase and proteinase-3 in 2 mM CaCl2 was inhibited by serine protease inhibitors DFP (Fig. 5A , left), PMSF, and SerpinB1/MNEI (not shown). The three fragments migrated more slowly under nonreducing conditions (data not shown), consistent with the presence of the C-terminal intrachain disulfide bonds. Of note, neither the 35-kDa nor the 11-kDa fragments were detected, indicating that calcium levels qualitatively alter the cleavage pattern. Blots with antibody to a C-terminal 18-mer peptide (C18 antibody) detected all three fragments (Fig. 5A , right), indicating that each contains C-terminal epitopes and thus, that they are alternative or sequential products. On staining of select CF BALF specimens, C18 antibody detected the 27-kDa fragment in patient lavage (Fig. 5B , right), indicating the presence of the C terminus and suggesting that the 27-kDa CF species and the in vitro-generated, 27-kDa fragment are identical. In a preliminary experiment, the two distinct cleavage patterns were largely reproduced when aliquots of SP-D-depleted CF BALF were used as the source of protease. On incubation for 2 h in 200 µM CaCl2, recombinant human SP-D was degraded extensively by CF BALF, and only the 35-kDa species was detected. Incubation in 2 mM CaCl2 caused less-extensive SP-D degradation, and fragments of 27 and 14 kDa but not 21 kDa were observed (data not shown). Molecular size suggests that the elastase-generated 27-kDa and 21-kDa fragments result from cleavage in the collagen domain and 14 kDa from cleavage in the neck region. We were unable to isolate adequate amounts of the fragments for sequencing. However, we compared the SP-D products with elastase-generated products of a trimeric fusion protein of the NCRD of human SP-D [21 ]. Elastase cleaved the 26-kDa fusion protein, releasing the NCRD (18 kDa). Subsequent cleavages within the NCRD produced a 14-kDa fragment that comigrated with the 14-kDa fragment from SP-D under reducing (Fig. 5C) and nonreducing (not shown) conditions, strongly indicating that these are identical. Edman sequencing of the 14-kDa fragment of the NCRD revealed ELFP, indicating cleavage after Val-231 in the neck region and a predicted mass of 13,358 Da.
 View larger version (44K): [in a new window] |
Figure 5. SP-D fragments generated by elastase in the presence of 2 mM CaCl2. (A) Recombinant human SP-D was incubated for 2 h, with or without 20 µg/ml elastase (NE) and DFP (2 mM), and the products were examined by Western blot (reducing conditions). (A, left) Staining with P13 antibody; (right) staining with C18 antibody to a peptide at the extreme C terminus. (B) Western blot (reducing conditions) of BALF specimens from a non-CF control individual (N) and two CF patients (CF1, CF2) stained with P13 antibody (B, left) and C18 (right). Note that C18 detects the 27-kDa CF fragment and as anticipated, also detects the 8-kDa but not the 35-kDa fragment. (C) Comparison of the 14-kDa fragment of SP-D with fragments of NCRD. SP-D and a NCRD fusion protein were treated with 0 or 7 µg/ml NE for 2 h and examined by P13 Western blot. Note that the 14-kDa fragment of SP-D comigrated with the 14-kDa-sequenced fragment from the NCRD, corresponding to cleavage after Val-231.
|
100 ng SP-D; Fig. 6A
). Aggregation induced by control BALF was abrogated by the Ca2+ chelator EDTA (2 mM final) or maltose (20 mM), which binds to the lectin domain, consistent with SP-D-dependent aggregation (Fig. 6B)
. Preincubation with elastase or proteinase-3 also abrogated bacteria-aggregating activity (Fig. 6B)
, consistent with the preliminary finding that each of the NSPs cleaves SP-D of control BALF (not shown). The control BALF shown is representative of two specimens (578 and 630 ng/ml SP-D) with similar results.
 View larger version (17K): [in a new window] |
Figure 6. Aggregation of E. coli by control (non-CF) BALF but not by CF BALF. BALF or rat SP-D (100 ng) or PBS was preincubated with or without additives at 37°C for 2 h, combined with log-phase E. coli in PBS with 200 µM CaCl2. Aggregation was measured over 3 h at room temperature as the decrease of A700 nm. (A) Control (non-CF) BALF causes bacterial aggregation. (B) Aggregation induced by control BALF was inhibited by EDTA (2 mM), maltose (malt; 20 mM), and by pretreatment with elastase (NE) or proteinase-3 (PR-3; 4 µg/ml during preincubation). (C) CF BALF, positive for elastase, fail to induce aggregation and abrogate aggregation induced by rat SP-D. Shown are CF BALF (200 µl 1:40 dilution) or rat SP-D (100 ng) preincubated individually or together or with DFP (2 mM) or MNEI/SerpinB1 (12.5 µg/ml). Undiluted CF BALF also lacked bacteria-aggregating activity (data not shown). All experiments were performed twice; average results are shown.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
As shown for the first time here, the presence of the NSP elastase as active enzyme is the best correlate for the absence or near-absence of SP-D in the lavage of CF patients. We do not expect these findings to be specific for CF. As previous studies have shown that elastase-positivity correlates with infection [18
], it is likely that infection status, which was previously correlated with decreased SP-D [12
13
16
], is a surrogate for neutrophil-dominated inflammation with the accumulation of active elastase. Importantly, the amount of elastase used for our in vitro studies (0.1–20 µg/ml) is less than that reported for CF patients. Whereas no free elastase is found in airway epithelial lining fluid of healthy subjects, mean levels of 8.2 µM (200 µg/ml) were found for CF patients with moderate lung function impairment [32
] and 2.3 µM (60 µg/ml) for patients with mild impairment [33
].
The three NSPs are colocalized in neutrophils and released in response to degranulating stimuli, and all three are found in lung fluids of CF patients [28 29 30 ]. Although only elastase was quantified in the present study, cathepsin G and proteinase-3 have been quantified for BALF specimens from the same parent panel. The content of active cathepsin G was 3.4 ± 1.7 µg/ml (mean±SEM) for 11 CF samples with 9.1 ± 2.0 µg/ml active elastase [19 ], and active proteinase-3 content was 0.42 ± 0.14 µg/ml for 17 CF samples with 8.7 ± 2.3 µg/ml active elastase (J. Cooley, F. J. Accurso, E. Remold-O’Donnell, manuscript in preparation). These values are for elastase-positive BALF. With minor exceptions, BALF that lacked active elastase lacked cathepsin G and proteinase-3. The combined data indicate that active forms of these proteases, when present in CF, are in the approximate ratio 10:4:0.5 for elastase:cathepsin G:proteinase-3. Thus, proteinase-3 and cathepsin G may contribute to SP-D loss, but potency and concentration in BALF indicate that elastase, the conventionally monitored protease, is the pre-eminent agent of SP-D depletion.
Although the principal fragments in the lavage specimens of CF pediatric patients, 35 kDa, 11 kDa, and 27 kDa, could all be replicated in vitro by treating recombinant human dodecamers with NSPs, they were products of distinct cleavage pathways that are dependent on calcium concentration. The rapid SP-D cleavage pathway that generates the 35-kDa N-terminal fragments [16 17 ] and the 11-kDa C-terminal fragment identified here is operative at low calcium, which is represented in this study by 200 µM CaCl2, the estimated concentration in CF BALF specimens. Immune precipitation, which allowed analysis of proteolytic products under nondenaturing conditions, suggests that the 11-kDa C-terminal fragment is released by cleavage to survive as an independent species in CF lavage fluid.
Using 2 mM CaCl2, which is presumed to approximate the levels in normal epithelial lining fluid, cleavage of SP-D by NSPs was substantially retarded, confirming an important role for calcium in protecting SP-D from NSPs. Although the inverse relationship of calcium ion concentration to the rate of SP-D cleavage was previously reported [16 17 25 ], its impact on the pattern of cleavage has not been described. At 2 mM calcium, elastase and proteinase-3 cleaved within the collagen domain to generate sequential or alternative fragments of apparent 27 kDa and 21 kDa and cleaved near the C-terminal end of the neck to generate a 14-kDa fragment corresponding to the C-terminal carbohydrate recognition domain. Identical mobility on SDS electrophoresis and shared reactivity with a C-terminal antibody suggest that the 27-kDa fragment in CF BALF and the 27-kDa fragment generated in vitro under physiological calcium conditions are identical. A minor site identified for SP-D cleavage by neutrophil elastase at Pro-111 [16 ] would generate a predicted 24,494-Da fragment, consistent with the properties of the 27-kDa fragment identified in this study. Although some lavage samples showed minor species near the expected position of the 21-kDa fragment, no 14 kDa fragments were observed. Whether these fragments are generated in larger amounts in vivo and fail to survive in CF lung fluids requires further study. We also cannot exclude the possibility that components of lavage, such as surfactant lipids, further alter the precise pattern of cleavage.
We showed previously that varying Ca2 + from 0 to 2 mM had no effect on the activities of any of the NSPs, indicating that the altered pattern of SP-D cleavage by NSPs is a result of calcium effects on SP-D [17 ]. A coordinated calcium ion is integral to the lectin domain of most C-type lectins [34 ], and two or three additional calcium ion-binding sites have been identified for SP-D [35 36 ]. Moreover, calcium-dependent alterations of the circular dichroism spectrum of SP-D have been described [37 ]. Our findings suggest that variation of calcium alters the structure of the lectin region and that structural alterations account for the calcium ion dependence of SP-D susceptibility to protease. The findings suggest that SP-D is exposed to active elastase in a low-calcium environment in vivo. Decreased blood concentrations of ionized calcium are common among intensive care patients and sepsis syndrome, and [38 39 ] SP-D levels are at least transiently decreased in these settings [40 ]. Although monovalent ion concentrations in epithelial lining fluid have been studied intensively in CF [41 ], it is not known whether perturbations in calcium occur in lungs of CF patients.
The enhanced susceptibility of SP-D to cleavage by NSPs under conditions of decreased calcium ion concentration suggests that SP-D will be susceptible to ex vivo degradation in lavage specimens containing active NSP. Lavage is usually performed in the absence of added calcium. In the present study, the lavage specimens were immediately placed on ice, aliquoted, stored at –80°C, and maintained in the frozen state until assayed, thus minimizing the potential for post-lavage cleavage.
Other proteases present in CF airways are also potential contributors to proteolysis. For example, a Pseudomonas-derived metalloelastase and a trypsin-like protease were shown to cleave SP-D in vitro [14 15 ]. However, for the current study, a major role for these proteases seems unlikely. BALF specimens from Pseudomonas-negative patients that contained active neutrophil elastase were depleted of endogenous SP-D and were active in cleaving recombinant rat SP-D in vitro. The effects of inhibitors on recombinant SP-D cleavage by CF lavage also rule against this possibility. First, abrogation of cleavage by DFP strongly implicates serine proteases. Second, cleavage was not inhibited by EDTA, an inhibitor of metalloproteases, or phosphoramindon, a specific inhibitor of P. aeruginosa metalloelastase [26 ]. A role for the P. aeruginosa trypsin-like serine protease is unlikely, as SerpinB1/MNEI, which abrogated SP-D cleaving activity, is an efficient inhibitor of elastase-like and chymotrypsin-like serine proteases and is inactive against trypsin-like proteases [22 ]. Other host proteases that can be present include metalloproteases and cysteinyl proteases [42 43 ]. In this respect, cleavage was also not inhibited by leupeptin, E64, and Z-Phe-Ala-CHN2, ruling against contributions of cysteinyl proteases to SP-D degradation by CF BALF.
On the other hand, host-derived cysteinyl- and/or metalloproteases, present in inflammatory lung fluids, may play an indirect role in SP-D depletion by degrading endogenous serine protease inhibitors, and fostering the accumulation of active NSP, this effect would not be detected ex vivo. Similarly, virulence factors secreted by P. aeruginosa and A. xylosoxidans could increase the supply of host NSPs within diseased airways, e.g., via increased neutrophil chemotaxis, activation, or necrosis. Such an effect was recently demonstrated in a mouse model of increased neutrophil NSP activity generated by deletion of the inhibitor SerpinB1 (murine MNEI) [44 ]. On lung infection with P. aeruginosa, SerpinB1-deficient mice had increased levels of inflammatory cytokines and NSPs and increased neutrophil necrosis compared with infected wild-type mice; proteolytic depletion of SP-D was detected only in the SerpinB1-deficient mice.
The two proteolytic pathways identified here should generate SP-D species that are defective in a variety of functions, including microbial aggregation, uptake of microbes, and clearance of apoptotic neutrophils. Although it is true that SP-D molecules are degraded in the absence and presence of calcium, different spatial and temporal patterns of cleavage have profoundly different implications for function of the cleaved protein. Cleavage within the lectin domain (low-calcium pathway) generates N-terminal species that remain disulfide-bonded, likely as dodecamers or higher-order species. As these lack a functioning lectin domain, they will not bind pathogen but may still have the capacity to activate proinflammatory pathways in macrophages [8 ]. At physiological calcium, the lectin domain remains quite resistant to NSPs, and the major sites of cleavage are within more N-terminal domains, predicted to generate incompletely oligomerized molecules with active lectin domains. Previous studies also suggest that the state of oligomerization can alter cellular responses to SP-D, including the response to bacterial products such as LPS [8 ]. Notably, CF BALF, unlike control BALF, failed to induce Ca2+ -dependent aggregation of bacteria, suggesting that the NSPs abrogate at least one important, protective function of SP-D. Also, purified rat SP-D was cleaved by CF BALF and its bacterial-aggregating activity destroyed. Both effects were abrogated by inhibitors of NSPs, strongly implicating NSPs as effectors of SP-D depletion and destruction of bacterial-aggregating activity in CF patients. These findings provide a link between the presence of elastase and other NSPs, decrease of intact SP-D, presence of N- and C-terminal fragments of SP-D, and loss of protective function represented by bacterial-aggregating activity.
The new findings strengthen the evidence implicating host neutrophil protease in the in vivo depletion of SP-D in inflammatory lung disease. The data indicate that at least some of the resulting SP-D fragments are long-lived in CF lung fluid, where they may adversely modify the responses of host defense cells. As depletion of SP-D by NSPs can contribute to the loss of innate antimicrobial defense capacity in the setting of inflammatory lung disease, these findings indicate that treatment with NSP inhibitors may be beneficial to improve host defense through preservation of SP-D function.
Received October 10, 2007; revised December 11, 2007; accepted December 20, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
or calreticulin/CD91, lung collectins act as dual function surveillance molecules to suppress or enhance inflammation Cell 115,13-23[CrossRef][Medline] 1-antitrypsin treatment for cystic fibrosis Lancet 337,392-394[CrossRef][Medline]-helical coiled-coil and the three lectin domains of human lung surfactant protein D Structure 7,255-264[Medline]This article has been cited by other articles:
 |
|
 |
|
  T. HARA, F. OGAWA, K. YANABA, Y. IWATA, E. MUROI, K. KOMURA, M. TAKENAKA, K. SHIMIZU, M. HASEGAWA, M. FUJIMOTO, et al. Elevated Serum Concentrations of Polymorphonuclear Neutrophilic Leukocyte Elastase in Systemic Sclerosis: Association with Pulmonary Fibrosis J Rheumatol, January 1, 2009; 36(1): 99 - 105. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
