The control of neutrophil chemotaxis by inhibitors of cathepsin G and chymotrypsin.

Neutrophil chemotaxis plays an important role in the inflammatory response and when excessive or persistent may augment tissue damage. The effects of inhibitors indicated the involvement of one or more serine proteinases in human neutrophil migration and shape change in response to a chemoattractant. Monospecific antibodies, chloromethylketone inhibitors, and reactive-site mutants of α1-antitrypsin and α1-antichymotrypsin were used to probe the specificity of the proteinases involved in chemotaxis. Antibodies specific for cathepsin G inhibited chemotaxis. Moreover, rapid inhibitors of cathepsin G and α-chymotrypsin suppressed neutrophil chemotaxis to the chemoattractants N-formyl-L-methionyl-L-leucyl-L-phenylalanine (fMLP) and zymosan-activated serum in multiple blind well assays and to fMLP in migration assays under agarose. The concentrations of antichymotrypsin mutants that reduced chemotaxis by 50% would inactivate free cathepsin G with a half-life of 1.5-3 s, whereas the concentrations of chloromethylketones required to produce a similar inhibition of chemotaxis would inactivate cathepsin G with a half-life of 345 s. These data suggest different modes of action for these two classes of inhibitors. Indeed the chloromethylketone inhibitors of cathepsin G (Z-Gly-Leu-Phe-CMK) and to a lesser extent of chymotrypsin (Cbz-Gly-Gly-Phe-CMK) mediated their effect by preventing a shape change in the purified neutrophils exposed to fMLP. Antichymotrypsin did not affect shape change in response to fMLP even at concentrations that were able to reduce neutrophil chemotaxis by 50%. These results support the involvement of cell surface proteinases in the control of cell migration and show that antichymotrypsin and chloromethylketones have differing modes of action. This opens the possibility for the rational design of anti-inflammatory agents targeted at neutrophil membrane enzymes.

The pathogenesis of lung diseases such as adult respiratory distress syndrome and emphysema is thought to result from an imbalance between leukocyte serine proteinases (such as neutrophil elastase, cathepsin G, and proteinase III) and the proteinase inhibitors of the lung (1)(2)(3). There are a variety of these inhibitors, but most important are the serpins, a family of serine proteinase inhibitors, typified by antitrypsin (␣ 1 -proteinase inhibitor) and antichymotrypsin (4). Each of the family members has a unique inhibitory specificity, but they all share a similar overall molecular structure (5). Neutrophil chemotaxis is an important component of the inflammatory response, and the recruitment of neutrophils to the lung may play a pivotal role in the pathogenesis of emphysema (6). If neutrophil migration into the lungs is excessive, there will be enhanced delivery and release of proteolytic enzymes, which may overwhelm the native inhibitors causing lung damage and eventually the development of emphysema.
The physiological mechanisms controlling neutrophil chemotaxis are unclear, although Ward and Becker (7) suggested that this process may be under the control of a surface serine proteinase. They showed that the inhibition of this enzyme by organophosphorus compounds reduced neutrophil chemotaxis (8,9) but were unsure as to the identity of the natural substrate or inhibitor (10). Other workers have reported that chloromethylketone inhibitors of chymotrypsin-like proteinases were able to suppress human neutrophil chemotaxis (11)(12)(13) and superoxide production (14) as well as attenuating membrane potential changes in rat neutrophils (15). Furthermore King and co-workers (16) reported that a monoclonal antibody against human neutrophils inhibited superoxide anion generation in response to the chemotactic peptide fMLP 1 by binding to a surface chymotrypsin-like enzyme. Recent work (17) has demonstrated that serine proteinase inhibitors (particularly antichymotrypsin) are able to reduce neutrophil chemotaxis in response to the same chemoattractant. This has led to the suggestion that these agents exert their effects by binding and inhibiting a cognate surface serine proteinase, possibly cathepsin G, involved in receptor-mediated cell activation. This hypothesis is supported by the demonstration that neutrophil elastase, and probably cathepsin G, can bind to the cell membrane after secretion and therefore play a role in cell migration (18). Moreover, the human neutrophil contains isoforms of elastase and cathepsin G that may have differing roles (19). The current study uses inhibitors to assess the role of serine proteinases in modulating the neutrophil chemotactic response. This was achieved by examining the effects of antibody and chloromethylketone enzyme inhibitors along with P1 mutants of antitrypsin and antichymotrypsin, which have a range of inhibitory profiles and thus could be used to probe the specificity of the serine proteinases controlling chemotaxis.

Purification of Anti-cathepsin G Antibodies
Cathepsin G (10 mg) was coupled to cyanogen bromide-activated matrix according to the manufacturer's instructions. The gel (3.5 ml) was then packed into a column at 20 ml/h, washed with 0.05 M Tris, pH 8.0, and 1 ml of anti-cathepsin G antibody was applied at the same flow rate. Immunopurified antibody was eluted from the column with 0.2 M glycine, pH 2.5, and fractions of 2 ml were collected into tubes containing Tris to restore neutral pH. The column was reequilibrated with 0.05 M Tris, pH 8.0, and the procedure was repeated. The IgG concentration of the eluate was determined spectrophotometrically (⑀ 280 1% ϭ 14.5), and the specificity of the antibody was confirmed by radial immunodiffusion.

Construction, Expression, and Purification of Antitrypsin Mutants
Active site mutants of antitrypsin were produced by site-directed mutagenesis and expressed in Saccharomyces cerevisiae strain AB 116 as described previously (20). The recombinant protein was extracted and purified to homogeneity according to the method of Travis et al. (21). Purity was assessed by 10 -20% (w/v) SDS-polyacrylamide gel electrophoresis, and total antitrypsin concentration was determined by rocket immunoelectrophoresis (22) with a yeast recombinant methionine (P1) antitrypsin standard that had been calibrated by amino acid analysis. The recombinant P1 antitrypsin mutants migrated as several bands on SDS-polyacrylamide gel electrophoresis; amino-terminal sequencing confirmed these to be due to amino-terminal, but not reactive center loop, cleavage. The specific activities of these mutants against bovine ␣-chymotrypsin were arginine (P1) antitrypsin 77%, methionine (P1) antitrypsin 89%, valine (P1) antitrypsin 67%, and lysine (P1) antitrypsin 81%.

Active Site Determination of Serine Proteinases and Recombinant
Inhibitors-The active site titration of bovine ␣-chymotrypsin and human ␣-thrombin were performed according to the method of Chase and Shaw (26) using the suicide substrates p-nitrophenyl acetate and pnitrophenyl pЈ-guanidino benzoate, respectively. Bovine ␣-chymotrypsin was used to determine the activity of recombinant antitrypsin and antichymotrypsin using the substrate Suc-Ala-Ala-Pro-Phe-pNA as detailed previously (27). These inhibitors were in turn used to determine the activity of cathepsin G and neutrophil elastase using the substrates Suc-Ala-Ala-Pro-Phe-pNA and MeOSuc-Ala-Ala-Pro-Val-pNA, respectively.
Determination of Second Order Association Rate Constants-Equimolar active site concentrations of enzyme and inhibitor were incubated at 20°C as described previously (28). The value of the association rate constant (k ass ) was determined by nonlinear regression analysis according to the following equation (29), where E 0 is the initial enzyme concentration, and t is time. All values were determined on 2-3 separate occasions with the results quoted as weighted mean with standard error.
Determination of Pseudo-first Order Association Rate Constants-7-35-fold active site molar excess of inhibitor was incubated with the enzyme for varying time intervals at 20°C. The quantity of free enzyme was then determined by adding reaction buffer containing substrate, and the association rate constant was calculated as described previously (30,31).
Determination of Association Rate Constants and K i Values Using Progress Curve Analysis-Kinetic parameters for the interaction of recombinant antitrypsin and antichymotrypsin with human neutrophil elastase and cathepsin G were determined at 20°C as detailed previously (27). All buffers contained 0.1% (v/v) Triton X-100, and the substrates MeOSuc-Ala-Ala-Pro-Val-pNA and Suc-Ala-Ala-Pro-Phe-pNA were present at a concentration of 400 M and 6 mM for elastase and cathepsin G, respectively. The K m values for neutrophil elastase with MeOSuc-Ala-Ala-Pro-Val-pNA and cathepsin G with Suc-Ala-Ala-Pro-Phe-pNA were determined to be 0.1 Ϯ 0.01 and 1.9 Ϯ 0.23 mM, respectively.

Isolation of Neutrophils
Venous blood (10 ml) was collected from healthy volunteers into lithium heparin tubes, and the neutrophils were isolated on percoll gradients according to the method of Jepsen and Skottun (32). The cells (Ͼ95% pure, Ͼ95% viable by trypan blue exclusion) were washed, counted, and resuspended at the required concentrations in Hepesbuffered RPMI 1640 medium/bovine serum albumin (2 mg/ml). All reagents were endotoxin-free (Ͻ20 pg/ml) as determined by a Limulus assay. Following isolation, the cells were spherical, indicating that they had not been activated during harvesting (33).

Neutrophil Chemotaxis in the Multiple Blind Well Assay System
The chloromethylketones and diazomethylketone were dissolved in a minimum quantity of dimethylsulfoxide and then diluted to the required concentration with Hepes-buffered RPMI 1640 medium. The effect of these inhibitors on chemotaxis was assessed by incubating them with the isolated neutrophils for 30 min at room temperature prior to placing the cells in the chemotactic chamber. Similarly the effect of the antitrypsin and antichymotrypsin mutants was assessed after incubating with the neutrophils for 30 min, and that of the antibodies was assessed after incubating for 60 min.
The chemotactic response was based on the multiple blind well assay system (34). The lower well contained 27 l of 10 Ϫ8 M fMLP or 7% (v/v) zymosan-activated serum as the chemoattractant (35), and the upper chamber was filled with 50 l of the neutrophil suspension (final concentration, 3 ϫ 10 6 cells per ml) with or without antibody or inhibitor. The two chambers were separated by a 2.0-m pore size polycarbonate filter. Negative control wells contained medium but no chemoattractant in the lower chamber. The assay plates were incubated at 37°C for 20 min, and the membrane was then washed, fixed, and stained. The cells were counted from 5 areas of each membrane at 400ϫ magnification, and the mean value was obtained. Each experiment was performed in triplicate, and the mean value of the three membranes was taken as the result. All experiments were repeated on six different cell preparations unless otherwise stated. In the event of the negative control exceeding 5% of the positive control, the membrane was discarded. All values are expressed as mean cells/high power field (cpf) Ϯ S.E., and differences between groups of six subjects were assessed by the Wilcoxon-Signed Rank test. Under these conditions, the chemotactic assay had a control within-batch co-efficient of variation of 4.0% (n ϭ 5).

Neutrophil Chemotaxis under Agarose
Human neutrophils were isolated as described previously and incubated at 5 ϫ 10 7 cells/ml in Hepes-buffered RPMI 1640 medium with 10% (v/v) fetal calf serum with increasing concentrations of the chloromethylketone or antichymotrypsin at room temperature. The cells were then allowed to migrate under agarose to a chemoattractant (10 Ϫ7 M fMLP) or buffer (36), which allows the measurement of both chemotaxis and chemokinesis. The value for chemokinesis was then subtracted from that of chemotaxis to give a final result. The coefficient of variation was 9% (n ϭ 6).

Neutrophil Polarization to fMLP
Neutrophils were isolated on percoll gradients, washed twice with sterile calcium-free buffer (normal saline with 0.1% (w/v) human albumin, 8 mM glucose, and 10 mM Hepes), counted, and resuspended at 1 ϫ 10 7 cells/ml in Hanks' buffered saline solution containing 0.1% (w/v) albumin and 8 mM glucose. Cells were incubated with chloromethylketones (0 -100 M) or antichymotrypsin (0 -0.5 M active site) at room temperature for 30 min prior to the addition of fMLP to a final concentration of 10 Ϫ8 M. The suspensions were then left at room temperature for a further hour before fixing with 10 volumes 2% (v/v) formaldehyde. Polarization was measured by flow cytometry on a Becton Dickenson FACScan with Lysis II software after gating on the neutrophil population.

The Effect of Anti-cathepsin G, Anti-elastase, and Anti-cathepsin B Antibodies on Neutrophil Chemotaxis
The anti-cathepsin G antibodies reduced the chemotactic response of neutrophils to 10 Ϫ8 M fMLP in a dose-dependent manner (Fig. 1) from a control value of 30.5 (Ϯ3.1) to 11.0 (Ϯ2.2) cpf at 5.7 g/ml (p Ͻ 0.025). The antibody did not affect cell viability (as measured by trypan blue exclusion) or cause neutrophil agglutination (as visualized under light microscopy at 400ϫ magnification), and the inhibition could not be prevented by 5 g/ml polymyxin B sulfate, a sequester of endotoxin (37). Furthermore, the antibody was able to exert its effect over a range of fMLP concentrations (10 Ϫ10 -10 Ϫ5 M) and was not itself chemotactic when placed in the lower chamber (positive control, 10 Ϫ8 M fMLP 44.5 cpf; negative control, 1.3 cpf; 5.7 g/ml anti-cathepsin G antibody, 1.1 cpf: n ϭ 2).
It was possible that the inhibition was nonspecific and resulted from the interaction of the antibody Fc fragment with neutrophil receptors (38). To assess this possibility, the cells were incubated with an immunopurified antibody to the macrophage protein cathepsin B, which is not present in human neutrophils (39). The antibody had no effect on neutrophil chemotaxis (control 24.2 Ϯ 2.3 cpf; anti-cathepsin B antibody 23.3 Ϯ 1.9 cpf at 5.7 g/ml: n ϭ 4), whereas similar concentrations of anti-cathepsin G antibody inhibited cell migration by 64%.
Furthermore, polyclonal anti-elastase antibodies also had no effect on the chemotactic response of neutrophils to fMLP, suggesting that the effect was specific to cathepsin G.

The Effect of Chloromethylketones on Neutrophil Chemotaxis
The specific inhibitor of cathepsin G, Z-Gly-Leu-Phe-CMK (40), was highly effective at reducing neutrophil chemotaxis to fMLP. Cell migration was reduced from a control value of 39.5 Ϯ 4.4 to 3.9 Ϯ 2.6 cpf (90.2% inhibition; p Ͻ 0.025) at 55 M (Fig. 2a). Similarly, Cbz-Gly-Gly-Phe-CMK, an inhibitor of chymotrypsin (41) and to a lesser extent of cathepsin G (40), reduced chemotaxis from a control value of 36.6 Ϯ 2.7 to 1.5 Ϯ 1.1 cpf (97.0% inhibition; p Ͻ 0.025) at 55 M (Fig. 2a). MeO-Suc-Ala-Ala-Pro-Val-CMK, a specific inhibitor of neutrophil elastase that has no effect on cathepsin G (40,42), also inhibited neutrophil chemotaxis although less effectively. At 65 M this chloromethylketone reduced the chemotactic response to fMLP from a control value of 31.8 Ϯ 6.6 to 14.5 Ϯ 5.5 cpf (54% inhibition; p Ͻ 0.025), although lower concentrations (55 M and below) had no effect (Fig. 2b). None of these agents at concentrations of up to 100 M had any effect on cell viability as determined by trypan blue exclusion or by lactate dehydrogenase release. Furthermore, none of the chloromethylketones was chemotactic when placed in the lower chamber.
Z-Gly-Leu-Phe-CMK (45 M) suppressed the neutrophil chemotactic response over a range of fMLP concentrations (Fig.  3a). In contrast, similar concentrations of both Cbz-Gly-Gly-Phe-CMK and MeOSuc-Ala-Ala-Pro-Val-CMK produced a shift in the peak chemotactic response to fMLP. Cbz-Gly-Gly-Phe- CMK, at a concentration of 45 M, produced a peak cell migration of 46.6 Ϯ 4.1 cpf in response to 10 Ϫ7 M fMLP. This was significantly greater (p Ͻ 0.025) than the value obtained at 10 Ϫ8 M fMLP (29.1 Ϯ 6.1 cpf) but not significantly less than the peak response in the control curve (53.0 Ϯ 2.7 cpf) that occurred, as expected, at 10 Ϫ8 M fMLP (Fig. 3b).

Recombinant P1 Antitrypsin Mutants
Association Rate Constants of P1 Antitrypsin Mutants-The association rate constants of the antitrypsin mutants with bovine ␣-chymotrypsin, cathepsin G, and neutrophil elastase are shown in Table I. Methionine (P1) antitrypsin has been reported to be labile, losing inhibitory activity upon standing at 4°C (21). To avoid this, the protein was harvested in the presence of 0.01 M ␤-mercaptoethanol, and full inhibitory activity was confirmed by dialysis against this agent (0.05 M Tris, 0.01 M ␤-mercaptoethanol, pH 8.0), which had no effect on the association rate constant with bovine chymotrypsin (2.5 Ϯ 0.1 ϫ 10 6 M Ϫ1 s Ϫ1 prior to dialysis; 2.4 Ϯ 0.2 ϫ 10 6 M Ϫ1 s Ϫ1 following dialysis). The progress curve kinetic analysis allowed the determination of the K i value for methionine and arginine (P1) antitrypsin with cathepsin G as 5.8 Ϯ 0.5 and 143 Ϯ 5 nM, respectively.
The Effect of Antitrypsin Mutants on Neutrophil Chemotaxis-None of the recombinant antitrypsin preparations inhibited neutrophil chemotaxis at concentrations of up to 0.48 M active site (data not shown). However when the concentration of methionine (P1) antitrypsin was increased to 0.95 M, there was a 30% reduction in chemotactic response to fMLP (p Ͻ 0.05).
There was no effect on cell viability, and the proteins were not themselves chemoattractants. Furthermore the effects were not reduced by the endotoxin sequester polymyxin B, and the inhibition was apparent over a range of fMLP concentrations (data not shown). Arginine (P1) antichymotrypsin had no effect on the chemotactic response at active site concentrations (against human ␣-thrombin) of up to 1.5 M when a slight but significant (p Ͻ 0.025) fall was observed (control 32 Ϯ 3.6 cpf; arginine (P1) antichymotrypsin 1.5 M active site 26.6 Ϯ 3.4 cpf: n ϭ 6).
The inhibitory effect of antichymotrypsin was also apparent with a second chemoattractant, zymosan-activated serum. Using this agent, the peak chemoattractant response was observed at a dilution of 7% (v/v). Leucine (P1) antichymotrypsin reduced the control value from 38.6 (Ϯ6.0) to 30 The Effect of Inhibitors of Cathepsin G on Neutrophil Migration under Agarose-There has been some debate about the ability of the multiple blind well assay system to differentiate between neutrophil chemotaxis and random movement, chemokinesis. Inhibitors of cathepsin G were therefore assessed, by a different observer, for their effect on neutrophil migration to fMLP under agarose. The peak leading front response was apparent at 10 Ϫ7 M fMLP and was reduced in a dose-related manner by inhibitors of cathepsin G and chymotrypsin. Z-Gly-Leu-Phe-CMK reduced the leading front from a control value of The Effect of Inhibitors of Cathepsin G on Neutrophil Polarization-Z-Gly-Leu-Phe-CMK abolished neutrophil polarization at 100 M in a dose-dependent manner (Fig. 5). Cbz-Gly-Gly-Phe-CMK was less effective at inhibiting neutrophil shape change and at higher concentrations activated the cells inducing polarization above the value of the negative control. Antichymotrypsin had no effect on polarization at concentrations of 0.5 M (Fig. 5). DISCUSSION Despite the obvious importance of neutrophil migration, little is known about the mechanisms available to control or reduce this response at a site of inflammation after the initiating insult has been removed. We have shown that inhibitors of cathepsin G (antibodies, chloromethylketones, and active site mutants of antichymotrypsin) are able to attenuate neutrophil migration in vitro. The chloromethylketones and antichymotrypsin that retard neutrophil chemotaxis were also rapid inhibitors of bovine ␣-chymotrypsin.
Immunopurified antibodies to cathepsin G were able to reduce neutrophil migration to the chemoattractant fMLP, presumably by binding to this enzyme on the surface of the neu-

TABLE I
The association rate constants (M Ϫ1 s Ϫ1 ) for P1 antitrypsin mutants with bovine ␣-chymotrypsin, neutrophil elastase, and cathepsin G M, V, K, and R represent methionine, valine, lysine, and arginine (P1) antitrypsin, respectively. Active site values were determined against bovine ␣-chymotrypsin. The association rate constants were determined under second order conditions apart from those highlighted ‡, which were determined using progress curves, and those marked *, which were determined under pseudo-first order conditions using 7-28 fold excess of inhibitor over enzyme. The enzyme concentration is given in brackets, and the values for second order kinetics and progress curves are quoted as weighted mean with standard error. Pseudo-first order kinetics are quoted with standard deviation. All assays were performed at 20°C. The correlation coefficients of those values determined under pseudo-first order conditions were K(P1)AT and neutrophil elastase, 0.98, and R(P1)AT and neutrophil elastase, 0.99.  King et al. (16) also support the hypothesis that cathepsin G plays a role in neutrophil activation; antibodies against a chymotrypsin-like enzyme on the surface of neutrophils were able to inhibit superoxide anion production induced by fMLP. The absence of an effect with anti-elastase and anti-cathepsin B antibodies suggested a specific role for cathepsin G in mediating the cellular response to fMLP. In order to confirm this role for cathepsin G synthetic chloromethylketone inhibitors of neutrophil elastase, cathepsin G, chymotrypsin and trypsin were assessed for their ability to attenuate neutrophil chemotaxis. Those that inhibited predominantly cathepsin G (Z-Gly-Leu-Phe-CMK) and chymotrypsin (Cbz-Gly-Gly-Phe-CMK) were potent inhibitors of chemotaxis, whilst the specific inhibitor of elastase (MeOSuc-Ala-Ala-Pro-Val-CMK) exerted its only effect at a higher concentration. This is despite Z-Gly-Leu-Phe-CMK and Cbz-Gly-Gly-Phe-CMK having significantly lower association rate constants with cathepsin G (51.2 M Ϫ1 s Ϫ1 and 4.1 M Ϫ1 s Ϫ1 , respectively (40)) than MeOSuc-Ala-Ala-Pro-Val-CMK has with elastase (922 M Ϫ1 s Ϫ1 (40)). The concentrations of chloromethylketones used here were similar to those used by other workers to inhibit chemotaxis (11)(12)(13) and attenuate antibody-dependent cellular cytotoxicity (44). Strikingly, the reduction in chemotaxis by both Cbz-Gly-Gly-Phe-CMK and MeOSuc-Ala-Ala-Pro-Val-CMK was mitigated by increasing the concentration of chemoattractant. This suggests that these agents were not irreversible inhibitors of chemotaxis. In contrast, the effect of Z-Gly-Leu-Phe-CMK, which is significantly more efficient in inactivating cathepsin G than chymotrypsin (51.2 M Ϫ1 s Ϫ1 and 3.0 M Ϫ1 s Ϫ1 , respectively (40)), was apparent over a range of fMLP concentrations. The results with synthetic inhibitors led to the assessment of variants of the naturally occurring proteinase inhibitors, antitrypsin, and antichymotrypsin, on neutrophil chemotaxis.
Active site mutants were used to probe the specificity of the neutrophil surface enzyme involved in the chemotactic response. The results were assessed using active concentrations of proteins as this allowed a more useful comparison between proteins and avoided inaccuracies and distortions when using preparations with different specific activities. Recombinant antitrypsin preparations had no effect on neutrophil chemotaxis at active site concentrations of up to 0.48 M (20 g/ml). Indeed the concentration of methionine (P1) antitrypsin had to be raised to 0.98 M (40 g/ml) before there was a significant fall in the chemotactic response. Nevertheless this response still occurred at physiological concentrations of antitrypsin found within the plasma or at sites of inflammation (45). Clearly the inhibition of neutrophil elastase alone (which was efficiently obtained with valine (P1) antitrypsin) was insufficient to inhibit the chemotactic response at concentrations up to 0.48 M.
Active site mutants of antichymotrypsin, the cognate inhibitor of cathepsin G, were then assessed for their ability to inhibit chemotaxis. The kinetic analysis confirmed that those mutants, which were efficient inhibitors of cathepsin G and chymotrypsin (leucine, methionine, and phenylalanine (P1) antichymotrypsin) were also able to attenuate neutrophil chemotaxis to the peptide fMLP. Once again, this occurred at concentrations well below those obtained in the plasma or at sites of inflammation (46) The importance of inhibitor specificity was confirmed by arginine (P1) antichymotrypsin, which has a 10fold lower association rate constant and forms a 10-fold less stable complex (as determined by the K i value) with cathepsin G than the other active site mutants. This protein had only a small effect on chemotaxis at concentrations over 3-fold greater than those required by the other mutants to reduce the chemotactic response by over 50% (Fig. 4). The effect of inhibitors of cathepsin G on neutrophil migration was confirmed in assays under agarose. Such assay systems also take into account random migration or chemokinesis.
Previous work has shown that the neutrophil surface serine proteinase has a K m value for a synthetic substrate closer to chymotrypsin than cathepsin G (24), suggesting that the enzyme, although cathepsin G-like, is not cathepsin G. The present data, however, suggest that the enzyme involved in the control of chemotaxis is unlikely to have a specificity identical to chymotrypsin, as methionine (P1) antitrypsin is able to inhibit chymotrypsin more efficiently than the antichymotrypsin mutants and yet has little effect on chemotaxis until higher concentrations of inhibitor are used. The association rate constants reported here allow the determination of half time for the inhibition of cathepsin G at a given inhibitor concentration. For the antichymotrypsin mutants tested, there was a correlation between a short half-life for the inhibition of cathepsin G and a reduction in the chemotactic response. The concentrations of antichymotrypsin mutants that reduced chemotaxis by 50% would inactivate cathepsin G with a half-life of approximately 1.5-3 s. The correlation between a short half-life for inactivation of cathepsin G and the inhibition of chemotaxis could not be extended across different classes of inhibitors. Z-Gly-Leu-Phe-CMK reduced chemotaxis by 50% at a concentration that would inactivate cathepsin G with a half-life of 345 s. This is significantly longer for the same inhibition of chemotaxis by the antichymotrypsin mutants and suggests that the enzyme may not be cathepsin G or that, by virtue of their size, chloromethylketones are more able to inactivate membrane bound cathepsin G than larger proteins. A third possibility is that the uncharged chloromethylketones are able to cross the cell membrane and inhibit an intracellular cathepsin G that is released from intracellular granules during chemotaxis. The differential effects of antichymotrypsin and the chloromethylketone inhibitors is underscored by their effect on neutrophil shape change in response to fMLP. The cathepsin G inhibitor Z-Gly-Leu-Phe-CMK abolished neutrophil shape change in a dose-dependent manner, Cbz-Gly-Gly-Phe-CMK, a less potent inhibitor, had less effect, whereas antichymotrypsin had no effect at all. Thus chloromethylketone inhibitors mediate their effect, at least partly, by blocking cell shape change in response to a chemoattractant, but the point of action of antichymotrypsin occurs subsequent to polarization.
The results suggest that the serpins antitrypsin and antichymotrypsin can play an important role in modulating neutrophil migration by interacting with cathepsin G or a chymotrypsin-like enzyme on the surface of the neutrophil. This contrasts with the effect of antichymotrypsin on neutrophil superoxide anion production, which appears to be independent of the active site (24) and may be mediated by enzyme-inhibitor complex formation (47). Interestingly, both complexed (48,49) and cleaved (50,51) serpins are able to stimulate neutrophil chemotaxis. Thus native and cleaved proteins may interact to modulate neutrophil migration at sites of inflammation, the former reducing chemotaxis during periods of health, and the latter promoting cell migration to sites of inflammation.