Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Loss of elastin is a characteristic feature of emphysematous lungs (1). According to the protease-antiprotease hypothesis, the lesion could be caused by excessive elastolysis (2). The three major neutrophil serine proteases, human leukocyte elastase (HLE), cathepsin G (CatG), and proteinase 3 (Pr3), are potential elastases, each of which has been suspected as contributing to emphysemagenesis (2). The postulate that HLE or Pr3 is involved in this disorder originated from observations that genetic deficiency of ₁-proteinase inhibitor (₁-PI), the primary inhibitor for both HLE and Pr3 (3, 4), predisposes to the early onset of emphysema (5) and that instillation of HLE or Pr3 into the lungs of laboratory animals induces emphysema-like lung tissue injury (6, 7). Abnormal turnover of elastin was also detected in patients with other evidences of chronic obstructive pulmonary disease (2).

The degradation of elastin by HLE has been studied extensively (8). One of the significant findings from early studies is that excessive ₁-PI fails to inhibit completely ongoing elastolysis by HLE (8, 13). A similar phenomenon was also observed with CatG (data in this study). Because Pr3 was the last of the three neutrophil serine proteases to be purified and characterized, elastolysis by Pr3 and its inhibition by endogenous inhibitors have not been well studied. This paper compares the degradation of [¹⁴C]elastin by Pr3 and its inhibition by ₁-PI with those of HLE and CatG. It was found that the rate of elastolysis catalyzed by Pr3 was the lowest among the three enzymatic reactions and that elastolysis by Pr3 that had been preincubated with elastin substrate was almost completely and stoichiometrically inhibited by ₁-PI. The mechanism that underlies the highly efficient inhibition of ongoing elastolysis by Pr3 and the incomplete inhibition of ongoing elastolysis by HLE and CatG was explored.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Pr3, HLE, CatG, and ₁-PI were purchased from Athens Research and Technology (Athens, GA). Recombinant eglin c was a gift from Dr. H. P. Schnebli, Novartis (formerly Ciba-Geigy), Basel, Switzerland. The active site concentration of HLE was determined by titration with N-benzyloxycarbonyl-Ala-Ala-Pro-azaAla p-nitrophenyl ester (Enzyme System Products, Livermore, CA) (18). The active site concentrations of ₁-PI and eglin c were measured with titrated HLE, assuming that complete inhibition of 1 mol of HLE consumes 1 mol of ₁-PI or results from binding by 1 mol of eglin c. The active site concentrations of Pr3 and CatG were measured with ₁-PI and eglin c, respectively, also assuming one-to-one inhibition. Throughout, the concentrations of proteases and ₁-PI are expressed as active site concentrations, rather than as total protein. Fibrous elastin, 100-400 mesh, extracted from bovine neck ligament by the method of Partridge and coworkers (19), was purchased from Elastin Products (Owensville, MI). Amino groups of the elastin substrate were labeled by reductive methylation with [¹⁴C]formaldehyde (NEN, Boston, MA) by the procedure of Bielefeld and coworkers (20). Specific activity of the labeled substrate was 0.19 µCi/mg, as assayed after complete hydrolysis with porcine pancreatic elastase.

Assays for Elastolysis

[¹⁴C]Elastin powder was weighed accurately and suspended in Dulbecco's modified phosphate-buffered saline (DPBS; pH 7.2, ionic strength 0.15) containing 0.1% (wt/vol) Triton X-100 to make a stock suspension of 4 mg/ml. In a series of 1.5 ml Eppendorf tubes, 50 µl of the stock suspension (200 µg of [¹⁴C]elastin), increasing amounts of Pr3, and DPBS/0.1% Triton X-100 were mixed to a final volume of 200 µl. The mixtures were incubated at 25°C for 4 h with constant end-over-end rocking. The tubes were centrifuged at 3,000 × g for 5 min. Aliquots of the supernatants were removed to count soluble ¹⁴C-labeled peptide fragments cleaved from [¹⁴C]elastin. The elastolysis by HLE and CatG was measured by the same procedure.

Inhibition of Elastolysis by ₁-PI

[¹⁴C]Elastin (200 µg) and Pr3 (81 pmol) in 190 µl of the same buffer as described previously were incubated at 25°C for 20 min. After addition of 10 µl of increasing concentrations of ₁-PI into the mixtures, elastolysis was allowed to continue for another 4 h. After centrifugation, [¹⁴C]peptide fragments in the supernatants were measured. Controls were run to account for elastolysis during the 20 min preincubation. Inhibition by ₁-PI of HLE and CatG that had been preincubated with [¹⁴C]elastin was studied using protease concentrations of 16 and 50 pmol, respectively.

Adsorption of Proteases and ₁-PI on [¹⁴C]Elastin

[¹⁴C]Elastin (200 µg) and increasing amounts of protease, Pr3, HLE, or CatG, in 200 µl buffer were incubated for 20 min. Controls contained no [¹⁴C]elastin in the tubes. Extent of adsorption of the proteases by [¹⁴C]elastin was determined from the difference in esterolytic activities of the supernatants from the elastin-containing and elastin-free incubations. Esterolysis was measured at 25°C in DPBS containing 0.1% Triton X-100, 10% (vol/vol) dimethyl sulfoxide and 5,5'-dithio-bis-2-nitrobenzoic acid (100 µM), with the substrates (Bachem, King of Prussia, PA) t-butyloxycarbonyl-Ala-Pro-Nva-p-chlorothiobenzyl ester (200 µM) for Pr3 and HLE, and succinyl-Val-Pro-Phe-thiobenzyl ester (300 µM) for CatG. For a typical assay of Pr3, for instance, 10 µl of a supernatant sample was added to 990 µl of medium containing t-butyloxycarbonyl-Ala-Pro-Nva-p-chlorothiobenzyl ester, and the reaction of p-chlorothiobenzyl group released from the substrate by Pr3 with 5,5'-dithio-bis-2-nitrobenzoic acid was monitored at 410 nm for a period that depended on protease concentration. The amount of Pr3 in the supernatant was calculated from a standard curve covering the entire range of protease concentrations used in the adsorption studies.

To examine the absorption of ₁-PI by [¹⁴C]elastin, 405 pmol ₁-PI in 1 ml buffer with or without 1 mg [¹⁴C]elastin were incubated at 25°C for 20 min. The mixtures were centrifuged to remove elastin. To determine the concentrations of ₁-PI in the supernatants, 10 µl of 2 µM titrated HLE was incubated with increasing amounts of the supernatants in 990 µl buffer for 40 min, and residual activity of HLE was determined by addition of 10 µl of 50 mM methoxysuccinyl-Ala-Ala-Pro-Val p-nitroanilide (Sigma, St. Louis, MO) in dimethyl sulfoxide.

Effects of Dilution and Elastolysis on Adsorption of Proteases

[¹⁴C]Elastin (200 µg) and Pr3 (176 pmol) in increasing volumes (200, 400, 600, and 800 µl) of buffer were incubated at 25°C for 4 h. The concentrations of Pr3 and solubilized [¹⁴C]peptides in the supernatant phase were assayed by esterolysis and radioactive counting, respectively. The amounts of Pr3 absorbed by the remaining insoluble [¹⁴C]elastin were then calculated by difference. Adsorption of HLE or CatG under the same conditions was determined using 25 pmol HLE and 88 pmol CatG.

Effects of Washing with Buffer on Ongoing Elastolysis and its Inhibition by ₁-PI

[¹⁴C]Elastin (400 µg) and Pr3 (158 pmol) in 190 µl of buffer were incubated at 25°C for 20 min. After centrifugation at 3,000 × g for 3 min, 140 µl of each supernatant was removed, and 950 µl of buffer was added to wash the pellets by 3 min of vigorous vortexing. The suspensions were centrifuged for 3 min, and 950 µl of each supernatant was removed. The washed pellets were resuspended in fresh buffer to a final volume of 200 µl, in the absence or presence of ₁-PI (316 pmol), and elastolysis was allowed to proceed for 4 h. Analogous experiments were performed using 20 pmol HLE and 60 pmol CatG, in the absence or presence of 40 and 180 pmol ₁-PI, respectively.

All assays were repeated at least three times. Each data point in the figures that follow represents a mean from three or more determined values. For simplicity, standard deviations, most of which were less than 20% of the means, are not shown in the figures, except in Figure 6, in which the data are expressed in the form of a histogram.

View larger version (21K): [in this window] [in a new window]   Figure 6.   Effects of washing with buffer on ongoing elastolysis and its inhibition by 1-PI. [14C]Elastin (400 µg) and proteases in 190 µl buffer were incubated at 25°C for 20 min. After addition of 10 µl buffer without (black bars) or with (gray bars) 1-PI, elastolysis was continued for another 4 h. Alternatively, protease-[14C]elastin complexes were isolated by centrifugation, washed with 1 ml buffer, and resuspended in 200 µl buffer without (white bars) or with (bars with lines) 1-PI, and elastolysis was continued for another 4 h. Initial protease amounts: HLE, 20 pmol; CatG, 80 pmol; and Pr3, 158 pmol. Molar ratio of 1-PI:protease: HLE and Pr3, 2:1, and CatG, 3:1.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Elastolysis of Three Neutrophil Serine Proteases

Figure 1 illustrates the degradation of [¹⁴C]elastin by HLE, CatG, and Pr3. For all the three proteases, quantities of soluble [¹⁴C]peptides released from the substrate increased linearly with increasing amounts of proteases added in the reaction mixtures. In the case of CatG-catalyzed elastolysis, where as much as 18% of the [¹⁴C]elastin was consumed, the curve still maintained good linearity. From the slopes of the linear plots, the rates at which soluble peptides were liberated from the [¹⁴C]elastin substrate were calculated: 0.04 µg/(pmol Pr3 h), 0.08 µg/(pmol CatG h) and 0.32 µg/(pmol HLE h). Thus, the catalytic efficiency of Pr3 to degrade [¹⁴C]elastin is half that of CatG, and one-eighth that of HLE.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Elastolysis by neutrophil proteases. [14C]Elastin (200 µg) and proteases in 200 µl buffer, pH 7.2, were incubated at 25°C for 4 h. After centrifugation, soluble [14C]peptides in the supernatants were measured by scintillation counting. Open circles, Pr3; closed circles, HLE; and squares, CatG.

Inhibition of Ongoing Elastolysis by ₁-PI

Figure 2 gives the inhibitory curves of ₁-PI against [¹⁴C]elastin degradation by the three proteases. The experiments were carried out under conditions in which proteases were preincubated with [¹⁴C]elastin (these conditions are referred to as ongoing elastolysis). As reported by others (8, 13), Figure 2 shows that ₁-PI in large molar excess failed to inhibit completely ongoing elastolysis by HLE. In this set of experiments, the highest molar ratio of ₁-PI to HLE used was 6.3. With this inhibitor:protease ratio, the residual rate of degradation of [¹⁴C]elastin by HLE was apparently equal to 12% of the elastolytic activity of the uninhibited enzyme. With a molar ratio of inhibitor:protease of 6.5, the apparent residual elastolytic activity of CatG in the presence of ₁-PI was 33% of that of the uninhibited enzyme. At this point, it is useful to consider the stoichiometry of inhibition of ₁-PI against HLE or CatG. The stoichiometry of inhibition in an isotropic reaction system, SI, is defined as the number of moles of an inhibitor required for the complete inactivation of 1 mol of an enzyme. The SI for ₁-PI against HLE is commonly accepted to be 1, though this parameter has never been verified experimentally. The SI for ₁-PI/CatG is somewhat controversial. Some investigators have reported this stoichiometry to be 1.1 (21), but others have reported a value of 2.4 (22). Because of this uncertainty in the value of SI for ₁-PI/CatG, we did not use ₁-PI, but rather eglin c, to measure the active site concentration of CatG, which resulted in values of the active site concentration close to those obtained by direct titration with N-trans-cinnamoylimidazole. The difference in SI of ₁-PI against HLE and CatG is reflected in the profiles of inhibitory curves for HLE and CatG in Figure 2: the absolute value of initial slope of the curve for CatG is significantly larger than that for HLE. Clearly different from the inhibition of HLE- and CatG-catalyzed elastolysis, Figure 2 shows that ongoing elastolysis catalyzed by Pr3 was completely inhibited by ₁-PI. To achieve complete inhibition of Pr3, an inhibitor:protease ratio not greater than 1.5 was required.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Inhibition of ongoing elastolysis by 1-PI. [14C]Elastin (200 µg) and proteases in 190 µl buffer were incubated at 25°C for 20 min. After the addition of 10 µl of 1-PI, the incubation was continued for another 4 h. Initial protease amounts: Pr3 (open circles), 81 pmol; HLE (closed circles), 16 pmol; and CatG (open squares), 50 pmol.

Absorption of Proteases and ₁-PI by [¹⁴C]Elastin

To address the significant differences in inhibitory efficacy of ₁-PI reported in Figure 2, we examined the adsorption of the three proteases by [¹⁴C]elastin. Figure 3 shows that the absorption of CatG by [¹⁴C]elastin increased linearly with increasing amounts of the protease added in the reaction mixtures. The absorption curve of HLE began to level off slightly when the amounts of HLE added were larger than 0.4 pmol/µg [¹⁴C]elastin. In comparison with CatG and HLE, the amounts of Pr3 absorbed by [¹⁴C]elastin were the least. At a ratio of 0.3 pmol total Pr3/µg [¹⁴C]elastin, it appeared that sites on the substrate for this protease had been fully saturated. It was estimated that each microgram of [¹⁴C]elastin adsorbed a maximum of 18 ± 3 fmol of Pr3, i.e., the site number for Pr3 on [¹⁴C]elastin was 18 nmol/g. Up to the highest protease concentrations tested, absorption sites for HLE or CatG were not saturated. From the relevant absorption curves, site numbers for HLE or CatG on [¹⁴C]elastin were estimated to be > 300 or 400 nmol/g, respectively. In the inhibitory experiments reported in Figure 2, 81 pmol of Pr3 was added to the reaction mixtures. According to the absorption curve for Pr3 in Figure 3, 92% of the protease would not be absorbed by the [¹⁴C]elastin under these conditions. It is expected that this portion of the enzyme, which still remained in the fluid phase, would be inactivated rapidly by ₁-PI. The relatively small number of sites on [¹⁴C]elastin that can be occupied by Pr3 during preincubation may be one of the causes for the high efficacy of ₁-PI against ongoing elastolysis by Pr3.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Adsorption of proteases by [14C]elastin during preincubation. [14C]Elastin (200 µg) and proteases in 200 µl buffer were incubated at 25°C for 20 min. Free protease concentrations in the fluid phase were measured with soluble synthetic substrates. Open circles, Pr3; closed circles, HLE; and squares, CatG.

We also examined the absorption of ₁-PI by [¹⁴C]elastin. Figure 4 shows that after exposure to [¹⁴C]elastin for 20 min, the antiprotease capability of ₁-PI was slightly increased compared with that of the controls. The increment, however, was within the region of experimental errors. Under the conditions used, absorption of ₁-PI on the elastin substrate was negligible.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Adsorption of 1-PI by [14C]elastin. 1-PI (405 pmol) in 1 ml buffer with or without 1 mg [14C]elastin were incubated at 25°C for 20 min. To determine the concentrations of 1-PI in the liquid phase, 10 µl of 2 µM titrated HLE and increasing amounts of the supernatants in 990 µl buffer were incubated for 40 min, and the residual amidolytic activity of HLE was assayed by addition of 10 µl 50 mM methoxysuccinyl-Ala-Ala-Pro-Val p-nitroanilide. Closed circles, 1-PI exposed to [14C]elastin; and open circles, controls.

Mobility of Absorbed Proteases

To estimate how tightly Pr3 can be absorbed onto insoluble elastin, we studied the partition of Pr3 between the solid and liquid phases in an elastin suspension. Fixed amounts of [¹⁴C]elastin and Pr3 were mixed in increasing volumes of buffer. The incubation time was extended to 4 h. Thus, the data obtained reflect the distribution of the enzyme between solid and liquid phases after a significant period of ongoing elastolysis. Because a substantial fraction of the substrate has been degraded, we express the results as amounts of Pr3 absorbed per microgram of residual insoluble [¹⁴C]elastin. Figure 5 shows that in a total 200 µl volume of reaction mixture, each microgram of [¹⁴C]elastin absorbed 302 fmol of Pr3 (normalized to 100% absorption in the figure) after 4 h, as compared with only 18 fmol/µg absorbed after a 20 min preincubation period, as reported in Figure 3. These results indicate that a large number of new sites for Pr3 were created when elastolysis proceeded. As the volume of the fluid phase in the suspension was increased, the amount of Pr3 absorbed to the remaining insoluble elastin after 4 h decreased sharply. When the total volume of the reaction mixture was increased to 800 µl, the amount of Pr3 absorbed decreased to 14 fmol/µg. Under the same conditions, the amounts of HLE and CatG absorbed on [¹⁴C]elastin were diminished only slightly as a result of increasing the volume of the fluid phase (Figure 5). Though the data are complicated by the large increases in the amounts of proteases absorbed during elastolysis, the results are consistent with a model in which Pr3 can dissociate more easily than HLE and CatG from the relatively small number of sites on [¹⁴C]elastin to which it can bind.

View larger version (20K): [in this window] [in a new window]   Figure 5.   Effects of reaction mixture volume on absorption of proteases by [14C]elastin. Fixed amounts of [14C]elastin (200 µg) and proteases in increasing volumes of buffer were incubated at 25°C for 4 h. Free proteases and [14C]elastin solubilized in the liquid phase were measured by esterolysis with synthetic substrates and radioactivity counting, respectively. Initial protease amounts: Pr3 (open circles), 176 pmol; HLE (closed circles), 25 pmol; and CatG (squares), 88 pmol. Data shown in the figure were normalized by arbitrarily setting the amounts of the different proteases absorbed per microgram of residual [14C]elastin in the 200 µl reaction mixtures to 100% (the actual absorbed amounts were: Pr3, 0.30 pmol/µg; HLE, 0.12 pmol/µg; and CatG, 0.51 pmol/µg).

Effects of Washing with Buffer

To explore the mobility of absorbed proteases further, [¹⁴C]elastin-protease complexes were washed with buffer, and the rates of elastolysis by the remaining bound enzymes were determined. Figure 6 shows that after washing the insoluble protease-bound elastin once with 1 ml buffer, the rates of elastolysis catalyzed by HLE and CatG were diminished by 39% and 25%, respectively, whereas elastolysis catalyzed by Pr3 was diminished by 91%. Together, results in Figures 5 and 6 demonstrate that unlike bound HLE and CatG, the majority of the Pr3 bound to [¹⁴C]elastin was highly mobile and easily dissociated from the insoluble substrate. The high mobility of absorbed Pr3 should be another cause for the high efficacy of ₁-PI against ongoing elastolysis by Pr3.

Figure 6 also illustrates the effects of washing on the resistance of ongoing elastolysis catalyzed by the three proteases to inhibition by ₁-PI. Under this set of conditions, ongoing elastolysis by Pr3 was somewhat resistant to inhibition by the antiprotease. However, elastolysis catalyzed by Pr3 (6.0 ± 1.0%) was significantly less than that catalyzed by HLE (23.9 ± 3.0%) or CatG (31.8 ± 2.0%). It is interesting to note that after a single wash, the remaining ongoing elastolysis catalyzed by the three proteases appeared to become progressively more resistant to inhibition by ₁-PI. The reason for this enhanced resistance is unclear. In our experiments, a single washing operation took about 15 min. It has been thought that during the various steps in the incubations, new sites might be created on the insoluble elastin, to which the proteases might be more tightly absorbed and resistant to inhibition by ₁-PI than was the case at the start of the incubations. This hypothesis is not supported by the data obtained from experiments with multiple washing. In the experiments, effects of multiple washings on the inhibition of ongoing elastolysis catalyzed by HLE by ₁-PI were compared. No significant difference can be found between the magnitude of the antiprotease-resistant fraction of ongoing elastolysis catalyzed by HLE in the presence of ₁-PI that was added after a single wash and after three washes (data not shown). These results suggest that the resistance of ongoing elastolysis to ₁-PI may be attributed to a fraction of HLE that was so tightly absorbed to [¹⁴C]elastin that repeated washings were unable to liberate these molecules from the solid substrate. Based on this and other absorption properties of proteases on elastin, we have proposed a mechanism to address the incomplete inhibition of ongoing elastolysis (see DISCUSSION).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The Elastolytic Rates

The rates of elastolysis catalyzed by Pr3, CatG, and HLE measured in this study are consistent with a ratio of 0.125:0.25:1. The data were obtained with purified and ¹⁴C-labeled bovine neck ligament elastin, which differs structurally and physico-chemically from that in human lung elastic fibers. The assays were performed under conditions of a single pH and a single buffer composition, which may also differ from the local environment of the inflamed lung in vivo. For these reasons, the results may be extended to the human lung only with caution.

Elastin is an extremely insoluble substrate. This property makes elastolysis a multiphase reaction, in which the first step is mass transport of proteases from surrounding fluid phase to the solid substrate. Results in Figure 3 indicate that in the early stages of elastolysis, only a small number of sites on [¹⁴C]elastin were occupied by bound Pr3. The amounts of Pr3 absorbed by the substrate during these early stages in its digestion were much less than those of HLE and CatG. Nevertheless, the effective concentrations of Pr3 absorbed on the surface of the substrate may be estimated to reach much higher concentrations than the bulk fluid phase concentrations of the protease. The apparent surface area of an elastin preparation that was isolated from the same source and purified by the same procedure as the elastin preparation used in the present study has been determined to be 1.0 m²/g (23). Pr3 crystallizes with unit cell dimensions a, b, c of 85.6, 54.1, and 113.5 Å, respectively (24). It is reasonable to assume that absorbed Pr3 is present in a monolayer with thickness about 120 Å. For a reaction mixture containing 200 µg [¹⁴C]elastin and a total of 60 pmol Pr3, in which 3.5 pmol of the protease is absorbed by the substrate, as deduced from the absorption curve in Figure 3, the effective local concentration of Pr3 absorbed on the surface of [¹⁴C]elastin can be calculated to be 1.5 mM. Under the same conditions (200 µg [¹⁴C]elastin and 60 pmol protease), the local concentrations of HLE and CatG can be estimated to be 17 and 21 mM, respectively. Because elastolysis also obeys the mass action law, the low effective concentration of Pr3 absorbed on the surface of elastin would partly account for the low elastolytic rate. However, local concentration of enzyme is not a dominant parameter for defining elastolytic rate, as judged from the data for HLE- and CatG-catalyzed elastolysis. [¹⁴C]elastin absorbed more CatG than HLE (Figure 3), whereas the elastolytic rate catalyzed by HLE was higher than that catalyzed by CatG (Figure 1). Obviously, steps other than absorption, which have not been identified in the work reported here, contribute significantly to the elastolytic rate.

Resistance of Ongoing Elastolysis to ₁-PI

This study demonstrates that elastolysis by Pr3 could be almost stoichiometrically inhibited by ₁-PI even after the protease had been preincubated with the substrate. We have attempted to explain this highly efficient inhibition by ₁-PI by considering the absorption properties of Pr3 on [¹⁴C]elastin (data in Figures 3, 5, and 6). In contrast, even large molar excess concentrations of ₁-PI failed to inhibit completely elastolysis catalyzed by HLE and CatG if the proteases had been preincubated with the same substrate. It has been two decades since Reilly and Travis reported on the phenomenon of "antiprotease-resistant" ongoing elastolysis by HLE (8), but the mechanism for this observation has not been established definitively. It might be hypothesized that steric hindrance prevents penetration of ₁-PI into the network of elastin fibers, so that a fraction of the absorbed proteases inside the network is inaccessible to ₁-PI. However, no experimental evidence exists in the literature to support this hypothesis. On the contrary, ongoing elastolysis catalyzed by HLE or porcine pancreatic elastase was reported to be similarly resistant to inhibition by small natural inhibitors and low-molecular-weight synthetic inhibitors (16, 25). Data obtained in this work now allow us to suggest a mechanism that may address the resistance of ongoing elastolysis to antiproteases.

Let us consider the process of inhibition of ongoing HLE-catalyzed elastolysis by ₁-PI as described in Figure 2. In this set of experiments, mixtures containing 200 µg [¹⁴C]elastin and 16 pmol HLE were preincubated. According to the absorption curve in Figure 3, 85% of the total HLE added should have become absorbed to [¹⁴C]elastin during preincubation. By calculations similar to that described in the preceding section, the effective local concentration of HLE on the surface of [¹⁴C]elastin could be estimated to be 5.7 mM. After ₁-PI was added in a molar ratio of inhibitor:protease of, for example, 6.3 (the far right data point on the inhibitory curve of HLE in Figure 2), a nonisotropic reaction system was established in which HLE was condensed on the surface of [¹⁴C]elastin with a local concentration of 5.7 mM, and ₁-PI was distributed in the fluid phase with a concentration of 0.5 µM. At the time of addition of the antiprotease, a fraction of HLE, which was not absorbed by the substrate, and which could dissociate from the substrate over time, should be rapidly inactivated by ₁-PI. According to the data in Figure 2, this fraction appears to account for three-fourths of the protease. After inhibition of this fraction of elastase, the local concentration of HLE on the surface of [¹⁴C]elastin declined to 1.4 mM. Inhibition of this tightly bound fraction of HLE appears to take place only at the interface between the solid and the fluid phases. Because the molar ratio of HLE to ₁-PI on this interface was 2.8 × 10³ (= 1.4 mM/0.5 µM), the reaction might be assumed to obey first-order kinetics, determined by the concentration of ₁-PI. The time required for nearly complete inhibition of a stoichiometric concentration of HLE in the interface, d(t), can be calculated by using Bieth's equation (26), d(t)  5/k_a[I], where [I] is the concentration of ₁-PI, 0.5 µM, and k_a is the association rate constant for HLE and ₁-PI, 6.5 × 10⁷ M¹ s¹ (3). The calculation yielded a d(t) value of 0.15 s. For complete inhibition of the elastase absorbed on the surface of the substrate, a period of as long as 7 min ( 0.15 s × 2.8 × 10³) might be required. However, a more appropriate estimate of the time required for inhibition of absorbed elastase by ₁-PI on the surface of elastin must accommodate some other considerations. The concentration of ₁-PI was assumed in the calculation to be constant throughout the inhibitory process. Because the total molar ratio of ₁-PI to HLE in this special case was 6.3, this assumption may be applicable during inhibition of the major portion of the tightly bound enzyme, but would lead to an underestimation of the time required for complete inhibition. The substrate was assumed not to compete with ₁-PI for absorbed HLE in the calculation, so that the total absorbed HLE might be accessible to the inhibitor. This assumption is clearly problematic. Because both [¹⁴C]elastin and bound HLE achieve very high local concentrations at their interface, at least a portion of the absorbed HLE was "productively" bound to the substrate, which would thereby strongly compete with ₁-PI for the enzyme. These considerations should result in a significant extension of the time required for complete inhibition beyond the calculated value of 7 min. Due to the nonisotropic characteristics of the inhibitory reaction on the interface, which are likely to delay complete inhibition, there should be time for HLE to degrade substantial amounts of elastin before its complete inhibition by ₁-PI. We suggest that this is one of the major mechanisms that contribute to the incomplete inhibition of ongoing elastolysis. Recently, Campbell and coworkers proposed a nonisotropic HLE-₁-PI interaction mechanism, by which they explained how a portion of the HLE in azurophilic granules released by neutrophils escapes inhibition by ₁-PI in the pericellular milieu (27). The strong absorbance of enzymes to the surface of solid natural substrates, such as elastin, seems to represent another nonisotropic interaction mechanism by which neutrophil proteases resist rapid inhibition by endogenous inhibitors.

Pathologic Implications

Pr3 is the most abundant serine protease identified so far in the neutrophil. Each neutrophil was estimated to store 3, 1.1, and 0.85 pg of Pr3, HLE, and CatG, respectively (30). Another study also reported that neutrophils released much more Pr3 than HLE during phagocytosis (31). On the other hand, in comparison with HLE, the antiprotease screen against Pr3 in the lungs seems to be relatively weak. Pr3 and HLE share the same inhibitors, ₁-PI and elafin. The association rate constants of Pr3 with ₁-PI and elafin were measured to be 8 × 10⁶ and 4 × 10⁶ M¹ s¹ (4, 32), respectively, as compared with the comparable constants for HLE, 6 × 10⁷ and > 2 × 10⁷ M¹ s¹ (3, 32), respectively. In situations where the concentrations of ₁-PI and elafin are inadequate to inhibit both proteases, which frequently occurs at inflamed sites, both inhibitors would be expected to associate preferentially with HLE. Moreover, airway secretions contain high concentrations of a third inhibitor of HLE, secretory leukoprotease inhibitor, which does not bind Pr3. The findings reported here, that the elastolytic rate of Pr3 is relatively low and that ongoing elastolysis catalyzed by Pr3 is inhibited nearly completely by ₁-PI, may reflect an intrinsic balance mechanism that moderates the elastolytic potential of Pr3. The resistance of ongoing elastolysis to elastase inhibitors was not only found in vitro, but also in vivo. Studies with animal models of emphysema revealed that the efficacy of synthetic elastase inhibitors depended critically upon whether elastase inhibitors or elastases were administered first (33, 34). To protect animals' lungs from proteolytic destruction, inhibitors must be administrated "prophylactically" before challenge with elastases. Otherwise, no protection could be achieved or less protection was afforded when inhibitors were administered subsequently, in a "therapeutic" mode. A most likely explanation for these observations is that elastase inhibitors failed to control ongoing elastolysis. Resistance of ongoing elastolysis by HLE to endogenous inhibitors is probably one of the major factors that lead to elastin loss at inflamed sites. Conversely, the contribution of Pr3, which is engaged in ongoing elastolysis to lesions marked by elastin loss, might be expected to be relatively small. However, all of these in vitro findings for Pr3 remain to be verified in vivo. Studies on inhibition of ongoing elastolysis in vivo might be helpful in explaining the vexing observation that supplemental therapy with ₁-PI does not terminate rapidly the accelerated turnover of elastin in patients with emphysema with genetic deficiency of ₁-PI (35).