© 2003 American Thoracic Society DOI: 10.1165/rcmb.4591 Selective Inhibition of Large-to-Small Surfactant Aggregate Conversion by Serine Protease Inhibitors of the bis-Benzamidine TypeDepartment of Internal Medicine, Justus-Liebig-University Gießen, Gießen; and Center for Vascular Biology and Medicine, Friedrich-Schiller-University Jena, Erfurt, Germany Address correspondence to: A. Günther, M.D., Medizinische Klinik II, Klinikstrasse 36, D-35385 Gießen, Germany. E-mail: andreas.guenther{at}innere.med.uni-giessen.de
Conversion of the biophysically active large surfactant aggregate subtype (LA) of alveolar surfactant into the less surface active small surfactant aggregates (SA) occurs in vivo and is reproduced under conditions of cyclic surface area changes in vitro. A serine-active carboxyl esterase has been suggested as the responsible enzymatic activity, although the exact mechanisms underlying the conversion process are presently unclear. We investigated the influence of exogenous serine proteases and synthetic and natural serine protease inhibitors on the conversion kinetics of natural rabbit surfactant, obtained as bronchoalveolar lavage fluid (BALF). In vitro cycling of BALF was performed for various time periods in the absence or presence of increasing amounts of several serine proteases (trypsin, plasmin, thrombin, tryptase), and one natural (aprotinin) and 25 synthetic serine protease inhibitors (including regular benzamidines [group A], 3-amidinophenylalanine derivatives [group B], bis-benzamidines [group C], and analogs of naphthylsulfonyl-glycyl-4-amidinophenylalanine piperidide [group D]). LA were separated from SA by 48,000 x g centrifugation. Surface activity of the LA fraction was measured by means of the pulsating bubble surfactometer. None of the "classical" serine proteases forwarded any acceleration of the LA-to-SA conversion kinetics. Some of the serine protease inhibitors caused moderate retardation of conversion, but at the same dose range inhibited the surface tension–lowering properties of the LA fraction, which per se explained their inhibitory effect. In contrast, specific dose-dependent inhibition of the LA-to-SA transition was observed for four derivatives of the bis-benzamidine group: full blockage of conversion over 240 min of cycling was noted at doses that did not interfere with the surface activity of the LA fraction. In addition, the prototype of these bis-benzamidines, 1,4-bis-[ß-naphthylsulfonyl-(3-aminophenylalanine)]-piperazide, was found to inhibit the activity of the rabbit liver carboxylesterase ES-2 in two different synthetic substrate assays reflecting the amidase and esterase properties of carboxylesterases. These findings support the hypothesis that the LA-to-SA conversion is an enzymatically-driven process with serine-active carboxyl esterase(s) being centrally involved. Synthetic bis-benzamidine–type serine protease inhibitors may offer specific inhibition of this event.
Abbreviations: minimum surface tension after 5 min film oscillation,
The extracellular pulmonary surfactant pool, mostly assessed by analyzing the bronchoalveolar lavage fluid (BALF), can be separated into large (LA) and small (SA) surfactant aggregates. The large surfactant aggregates mainly consist of lamellar bodies, tubular myelin, and large multilamellar vesicles, and are assumed to represent the biophysically active precursor fraction that provides the interfacial surfactant film (1–3). The small surfactant aggregates mainly consist of small unilamellar vesicles and are suggested to represent the less active metabolic end-product of the interfacial surfactant film (1–3). LA-to-SA transition can be observed in vivo (4) and under conditions of cyclic surface changes in vitro (1, 2, 5). In parallel with the conversion of the LA to the SA fraction, a marked reduction of surface activity (6) and a loss of the hydrophobic surfactant apoprotein-B (SP-B) in the remaining LA fraction (2, 6) were recently observed in rabbit BALF undergoing in vitro cycling. Interestingly, a decrease of the LA fraction in conjunction with a loss of surface activity and SP-B content of this surfactant subpool was also observed in BALF originating from patients with severe adult respiratory distress syndrome (ARDS), suggesting that this type of disturbance of the surfactant homeostasis may contribute to the pathogenesis of respiratory failure (6, 7). At present, the exact molecular mechanisms underlying the surfactant subtype conversion are largely unsettled. Beyond doubt, cyclic surface area changes are a prerequisite for this event (1). In addition, high surface activity of the LA fraction seems to be mandatory, because impairment of the surface tension–lowering properties of the large surfactant aggregates, reflecting less dense packing of saturated phospholipids (in particular dipalmitoylphosphatidylcholine) and hydrophobic apoproteins at the interfacial layer, results in a strong retardation and even blockade of conversion (unpublished data). Finally, an enzymatic activity was proposed to be required for this conversion on the basis of inhibitor studies employing serine protease inhibitors. A diisopropylfluorophosphate (DFP)-binding protein, later named "convertase," was isolated from BALF, purified, and identified by partial amino sequencing as an esterase with high sequence homology to liver esterase (ES-2) (8–10). The substrate of the esterase is presently unknown, and due to the broad range of substrate specifities of carboxylesterases may not be easily deduced from the fact that this type of enzyme is involved in the LA-to-SA conversion. Ester bonds, as well as amide bonds, may be cleaved. In two recent communications, a dramatic loss of the hydrophobic apoprotein SP-B was observed in postcycling LA, without any significant changes within the phospholipid or neutral lipid fraction, suggesting that SP-B might be one possible candidate attacked by the "convertase" under discussion (2, 6). To further characterize the LA-to-SA transition, we now investigated the influence of several serine proteases (plasmin, trypsin, thrombin, tryptase), and a large number of natural and synthetic serine protease inhibitors, on the conversion kinetics in vitro. Admixture of the serine proteases did not substantially affect the conversion rate. Of all natural and synthetic serine protease inhibitors tested, a bis-benzamidine analog and three closely related derivatives exhibited a strong inhibitory effect on the surfactant subtype conversion, at doses far below those interfering with surface activity. In parallel, inhibitory potency of these agents on both the esterase and the amidase activity of rabbit liver esterase was noted when employing synthetic substrate assays. These findings further support the view that the LA-to-SA conversion of natural surfactant requires the presence of a specific esterase and/or amidase activity.
Materials Human plasmin (specific activity 8 U/mg) and bovine thrombin (specific acitivity 1,250 U/mg) were purchased from Boehringer Mannheim (Mannheim, Germany). Trypsin,  1-antitrypsin, 4-nitrophenylacetat, bis-(p-nitrophenyl)phosphate (BNPP), and porcine liver esterase were obtained from Sigma (Munich, Germany). Butanilicainephosphate (Hostacain) was kindly provided by Hoechst Roussel Vet (Unterschleissheim, Germany). Purified human mast cell tryptase and leech-derived tryptase inhibitor were generous gifts from C. Sommerhoff (Department of Clinical Chemistry and Biochemistry, University of Munich, Germany). Aprotinin (Trasylol) was obtained from Bayer (Leverkusen, Germany). Twenty-four synthetic serine protease inhibitors, some of them depicted in Figure 1
, were provided by J. Stürzebecher (Center for Vascular Biology and Medicine, FSU Jena, Germany). They included regular benzamidines (group A), 3-amidinophenylalanine derivatives (group B), bis-benzamidines (group C), and analogs of naphthylsulfonyl-glycyl-4-amidinophenylalanine piperidide (group D), and are listed in Table 1. One inhibitor (Inhibitor 19, pentamidine) was obtained from Rhone Poulenc Pharma (Cologne, Germany).
Preparation of a Rabbit BALF Pool Rabbits of either sex were killed by intravenous application of a lethal dose of pentobarbital/ketamine. A catheter was immediately placed into the trachea and the lungs were lavaged three times with 50 ml saline. After filtration through sterile gauze and sedimentation of cells (300 x g, 15 min, 4°C), supernatants were pooled and aliquoted in 5-ml fractions in polypropylene tubes (15–ml Falcon blue cups; Becton Dickinson, Meylan, Cedex, France) while being continuously stirred. The aliquots were then stored at -85°C until further processing. The phospholipid content was 81 µg/ml.
Addition of Serine Proteases and Protease Inhibitors to BALF
In Vitro Cycling of Rabbit BALF
Surfactant Subtype Separation
Analysis of Phospholipid Content
Assessment of Surface Activity
Synthetic Substrate Esterase Assays
Statistics
As shown in previous studies, in vitro cycling of rabbit BALF resulted in a time-dependent decrease in the relative concentration of large surfactant aggregates. After 120 min of cycling, the relative amount of LA was reduced to  23% (control: 86%). By means of a synthetic substrate assay we could detect esterase activity (13.0 mU/ml) as well as amidase activity (27.4 mU/ml) in the rabbit BALF preparation used for cycling experiments. Addition of increasing amounts of the natural serine proteases plasmin, trypsin, thrombin, or tryptase, all tested at a dose range equivalent to serum concentrations, did not accelerate the conversion rate of rabbit BALF (Table 2); instead, in the case of trypsin, a slight retardation of the conversion process was encountered. Various natural and synthetic serine protease inhibitors were tested for their ability to inhibit the LA-to-SA conversion rate. Aprotinin was entirely ineffective, as was benzamidine (Table 3). Among the synthetic serine protease inhibitors tested, some turned out to be completely ineffective in the currently used assay system (Inhibitors 1–3, 6–8, 10–12, 17–24; see Table 3), whereas others showed some moderate capacity to retard the conversion kinetics at the highest dosage used (2 x 10-4 M) (Inhibitors 4, 13–15). As shown in Figures 2 and 3 , four of the tested inhibitors were highly effective in retarding the LA-to-SA conversion process, namely Inhibitors 5, 9, 16, and 25. Almost complete inhibition of the conversion (LA content of > 80% after 120 min cycling) was achieved at molar ratios of inhibitor to phospholipids of 0.2–0.4 and at absolute concentrations of 10-5 M.
In view of the role of the surface tension for the kinetics of the LA to SA conversion, we additionally analyzed the impact of some of these inhibitors on the surface activity of the LA fraction prepared from the currently employed cycling pool. In the case of the only partially effective inhibitors, such as Inhibitor 4, some increase in minimum surface tension occurred in parallel to the retardation of conversion kinetics (Figure 3). In contrast, the highly effective Inhibitor 5 fully blocked the LA-to-SA conversion at a molar range that did not influence the surface activity at all.
To further test whether inhibition of conversion by serine protease inhibitors also influences surface activity after cycling, we analyzed the surface tension–lowering properties (
All four highly effective inhibitory agents of the LA-to-SA conversion (5, 9, 16, and 25) belong to the bis-benzamidine type of inhibitor, with differences in the side chains as compared with Inhibitor 5 1,4-bis-[ß-naphthylsulfonyl-(3-amidinophenylalanine)]-piperidide (see Figure 1). Inhibitor 5 was originally designed as a mast cell tryptase inhibitor; however, the structurally related Inhibitors 9, 16, and 25, which also highly effectively block the LA-to-SA conversion, are characterized by a different spectrum of preferential inhibition (see Table 1). In addition, a natural inhibitor against mast cell tryptase (leech-derived tryptase inhibitor) was ineffective in retarding the LA-to-SA conversion (data not given in detail). To probe further whether a liver carboxyl esterase, closely related to the lung convertase proposed to promote surfactant conversion, is inhibited by the presently tested serine protease inhibitors, we used synthetic substrate assays based on the cleavage of p-nitrophenylacetate and butanilicaine, in the absence of surfactant material. As shown in Figure 5 , Inhibitor 5 was indeed capable of inhibiting the enzymatic activity of the rabbit liver esterase in both assays, thus blocking the esterase activity (Figure 5A) and the amidase activity (Figure 5B) of this enzyme. In contrast, inhibitors 1, 2, and 7 did not interfere with the esterase activity in these assay systems when tested in the same dose range as Inhibitor 5 (Figures 5A and 5B).
In an effort to further characterize the mechanisms underlying the conversion of large to small surfactant aggregates, we analyzed the impact of various serine proteases and serine protease inhibitors on the conversion rate of rabbit BALF in vitro. Preceding studies had demonstrated that, next to cyclic surface area changes, an enzymatic activity is required for such a conversion (8). Initially, it had been suggested that the enzyme involved in conversion would be a "classical" serine protease (8, 9) because it was possible (i) to inhibit conversion by use of DFP, a toxic prototype serine protease inhibitor; (ii) to isolate a protein by use of DFP, which promoted surfactant conversion of convertase-depleted BALF; and finally (iii) to alter the in vitro and in vivo conversion rate by admixture of -1-antitrypsin to the surfactant system (9). However, partial sequence analysis of the isolated enzyme indicated that it belongs to the multigene family of carboxyl esterases (10), also representing serine-active enzymes. Meanwhile, the enzyme, entitled "convertase," has been completely sequenced (14) and was found to possess far-reaching sequence homology with the liver carboxyl esterase ES-2 (10). The fact that the enzyme responsible for the LA-to-SA conversion is not a "classical" serine protease is further supported by the present finding that the additional admixture of various such serine proteases (plasmin, tryptase, trypsin, and thrombin) to rabbit BALF did not result in any acceleration of the conversion rate. Instead, trypsin even caused some decrease of the LA-to-SA transition when added to rabbit BALF. The reasons for this effect are presently unclear; however, trypsin attack on the convertase itself might even be imaginable. Another explanation might be that addition of trypsin interfered with the surface activity of LA, as suggested by previous work in the field (15). By doing so, it may also cause a retardation of LA-to-SA conversion, as recently suggested by our group (6). The main focus of the present study was to rigorously test whether well-defined synthetic serine protease inhibitors would interfere with the LA-to-SA conversion rate, and to seek additional information as to the nature of the convertase based on the profile of inhibitory and noninhibitory agents. Particular emphasis was placed on testing the specificity of the inhibitors under investigation. This is based on the recent finding of our group that all agents interfering with the surface activity of the large surfactant aggregates cause a retardation or even blockade of the LA-to-SA transition (unpublished data). A dense packing of the surfactant material, or an adequate presentation of the relevant compounds at the air–liquid interface, is thus obviously a prerequisite for the convertase to be active. Thus, putative specific convertase inhibitors are to be tested for their impact on surface tension properties. Indeed, in the present study, the Inhibitors 4 and 14 effected a moderate retardation of the LA-to-SA conversion rate, but at a dose range that caused at the same time a marked increase in minimal surface tension, sufficient to explain the inhibitory effect on the conversion rate. Thus, these agents may not be regarded as specific inhibitors of the surfactant convertase. In contrast, 1,4-bis-[ß-naphthylsulfonyl-(3-amidinophenylalanine)]-piperazide (Inhibitor 5) completely blocked surfactant conversion in concentrations far below those for which an increase in surface tension was observed. The same was true for the three other substances of the bis-benzamidine group, Inhibitors 9, 16, and 25, each differing in the side chains from Inhibitor 5. Additionally, inhibition of surfactant conversion also prevented deterioration in surface activity to a great extent. The following conclusions may be drawn from this finding: First, the LA-to-SA conversion of surfactant may be blocked by a family of well-defined serine protease inhibitors, the bis-benzamidine group, independent of any impact of these agents on surface tension regulation. This observation further supports the hypothesis that an enzymatic activity is involved in the transition of large to small surfactant aggregates. Second, the four bis-benzamidine inhibitors markedly differ in their preferential inhibition with respect to the "classical" serine proteases. As based on synthetic substrate assays, Inhibitor 5 displays a preferential inhibition of tryptase > trypsin > thrombin, and in fact, this inhibitor had been "designed" as a tryptase inhibitor. As displayed in Table 1, different profiles have been disclosed for each of the other members of the bis-benzamidine group. This observation again strongly argues against a role of a "classical" serine protease in surfactant conversion. Third, the employment of a completely different test procedure, the use of synthetic substrates assaying the esterase and the amidase function of the liver carboxyl esterase ES-2 (13), corroborated the presence of a carboxylesterase activity in the currently BALF pool, a pronounced inhibitory potency of the bis-benzamidine agents toward this type of enzyme, and, finally, no inhibition of esterase and amidase activity upon use of those inhibitors being also ineffective in retardation of LA-to-SA conversion. Thus, these data add further evidence to the hypothesis that a carboxyl esterase is centrally enrolled in the surfactant LA-to-SA conversion. Unfortunately, the fact that both the esterase and the amidase function of the ES-2 carboxyl esterase was inhibited at almost equivalent doses does not allow further conclusions to be drawn as to the natural substrate attacked by this type of enzyme in the surfactant LA fraction. On the basis of the present results, lipids or proteins could be the target. In view of the recently published data (2, 6), SP-B offers as candidate for the convertase attack; however, this assumption and the putative additional role of SP-C remain to be directly addressed in future studies. In summary, the present study shows that four synthetic bis-benzamidine–type serine protease inhibitors dose-dependently inhibit the LA-to-SA conversion of natural surfactant occurring upon cycling, at doses not yet interfering with surface tension regulation. In addition, inhibition of the esterase and the amidase function of the carboxyl esterase ES-2 was demonstrated. These findings are compatible with the hypothesis that a carboxyl esterase is operative in the large to small aggregate transition in natural surfactant. Inhibitors such as those presently characterized, devoid of the toxicity of DFP, may thus represent a useful tool for studying surfactant conversion also under in vivo conditions, thus overcoming possible limitations in the direct translation of in vitro cycling experiments to the clinical siutation (6). Moreover, and of most importance, these inhibitors may be considered for pharmacologic intervention, when targeting the enhanced conversion of LA to the biophysically less active or even inactive SA known to occur under conditions of acute respiratory failure (6, 7, 16, 17).
This study was supported by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 547 "Kardiopulmonales Gefäßsystem" and GU 405/3-1). Portions of this investigation were performed in partial requirement for the doctoral thesis of C.P. Received in original form April 10, 2001 Received in final form June 13, 2002
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
min • acute respiratory distress syndrome, ARDS • bronchoalveolar lavage fluid, BALF • bis-(p-nitrophenyl)phosphate, BNPP • diisopropylfluorophosphate, DFP • large surfactant aggregates, LA • phospholipids, PL • small surfactant aggregates, SA • surfactant protein, SP
= 405 nm in the absence or presence of Inhibitor 5 (10-4 M) or in the presence of the specific inhibitor BNPP (5 x 10-4 M) was recorded at 25°C, pH 7.5, for 3 min in 10-s intervals. The amidase activity of the rabbit liver esterase (used at 1 U/ml = 1.35 x 10-7 M) was demonstrated using butanilicaine as substrate (8 x 10-3 M), which is split into 2-chloro-6-methylaniline and N-butyl-glycine. The absorption kinetics at 
-amidinophenyl-