THE JOURNAL OF BIOLOGICAL CHEMISTRY (0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No . 23, Issue of August 15, PP. 15356-15362, 1991 Printed in U. S. A.

Heparin Strongly Decreases the Rate of Inhibition of Neutrophil Elastase by al-Proteinase Inhibitor*

(Received for publication, February 20, 1991)

Klaus J. Frommherz, Bernard Faller, and Joseph G. Bieth From the Laboratoire d’Enzymologie, Znstitut National de la Sante et de la Recherche Medicale U 237. Uniuersite Louis Pasteur de Strasbourg, F-67400 Zllkirch, France

Heparin depresses the second-order rate constant k, for the inhibition of neutrophil elastase by cq-protein- ase inhibitor. High molecular mass heparin decreases k. from 1.3 X 10’ M” s” to a limit of 4.6 X lo4 ~ - l s-l. Low molecular mass heparin is about 7-fold less effec- tive. Dermatan sulfate and chondroitin sulfate are less efficient. Heparin preparations used in clinical care also strongly depress k, when tested at concentrations corresponding to their clinical efficacy.

Heparin also decreases the k, for the elastaseleglin c and the cathepsin G/al-proteinase inhibitor systems but not that for the al-proteinase inhibitor/pancreatic elastase or trypsin pairs. These results, together with Sepharose-heparin binding studies, indicate that the k,-depressing effect of the polymer is related to its ability to form a tight complex with elastase but not with Dl-proteinase inhibitor. One mol of high molecular mass heparin binds 3 mol of neutrophil elastase with a K,i of 3.3 nM. Low molecular mass heparin binds elas- tase with a 1:1 stoichiometry and a Kd of 89 nM. For both heparins k, is lowest when elastase is fully satu- rated with heparin. From this we conclude that heparin decreases k,, because the heparin-elastase complex is able to slowly react with a,-proteinase inhibitor and not because the inhibitor slowly dissociates the hepa- rin-elastase complex. These findings may have impor- tant pathophysiological bearing.

The azurophil granules of polymorphonuclear leukocytes contain a number of proteolytic enzymes including NE,’ a 30- kDa serine proteinase. This cationic glycoprotein is formed of 218 amino acid residues with 19 arginines but no lysines (1). The arginine residues are all located in patches on the surface of the protein (2, 3). Two of them (Arg177 and are close to the active center and their chemical modification inactivates the enzyme (4). NE not only solubilizes elastin but also cleaves extracellular matrix macromolecules and plasma proteins such as type IV collagen, proteoglycans, fibronectin, antithrombin 111, and components of the immune system ( 5 ) . It plays a number of physiologic functions includ- ing proteolysis of phagocytosed proteins, neutrophil migra- tion, and tissue remodeling following injury (6).

It has been calculated that the concentration of NE in the

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

‘The abbreviations used are: NE, neutrophil elastase; alPI, al- proteinase inhibitor; SUC, succinyl; MeOSuc, methoxysuccinyl; pNA, pnitroanilide; SDS, sodium dodecyl sulfate; H- or L-heparin, high molecular mass or low molecular mass heparin; Hepes, 4-(2-hydrox- yethy1)-1-piperazineethanesulfonic acid.

azurophil granules is in the millimolar range (7). Th’ 1s enor- mous local concentration requires efficient control mecha- nisms to prevent undesirable extracellular protein degrada- tion during activation or death of neutrophils. The 53-kDa glycoprotein alPI is thought to be the most important neutro- phil elastase inhibitor. It inhibits serine proteinases irrevers- ibly by forming a denaturant-stable complex with them. Among all proteinases tested, NE is inhibited with the highest rate constant ( k , > lo7 M” s-’) (8-10). The high ka value, the elevated plasma and tissue levels of alPI (11, 12), and the irreversible character of the enzyme-inhibitor binding render the inhibition process extremely fast and efficient. For in- stance, the time required to fully inhibit elastase in vivo (13) is only 13 ms in plasma ( 5 ) and 100 ms in the lung epithelial lining fluid (12). The antielastase function of alPI is best illustrated by the well known correlation between hereditary deficiency of the inhibitor and pulmonary emphysema ( 5 ) .

There are a number of circumstances in which the inhibi- tion of NE by alPI may be impaired (i) the antielastase protection may be overcome in the immediate vicinity of an activated or a “dying” neutrophil because of the aforemen- tioned high local concentration of NE; (ii) the close contact between a neutrophil and an extracellular matrix protein may prevent the access of alPI to the cell-matrix interface so that NE-mediated proteolysis may occur (14); (iii) Met3”, the P1 residue of the active center of alPI, may be oxidized to methionine sulfoxide by oxidants secreted by activated phag- ocytes or contained in cigarette smoke; oxidized alPI is unable to prevent NE-mediated tissue injury in vivo and may cause emphysema in smokers (15), because it is a slow-binding elastase inhibitor in vitro ( k , = 8 X lo3 M” s-l, Ref. 16).

Here we show that native alPI also behaves like a slow- binding NE inhibitor if heparin is present in the reaction medium. Heparin is a naturally occurring sulfated polysac- charide that belongs to the family of glycosaminoglycans. Mammalian heparins mostly consist of trisulfated disaccha- ride units formed of L-iduronic acid-2-sulfate linked to D- glucosamine-N,6-sulfate. These regular sequences are var- iously interrupted by irregular sequences containing either undersulfated or oversulfated uronic acid and amino sugar residues. On the average, however, there are three sulfate groups per disaccharide unit. One of the irregular sequences is a sulfated pentasaccharide whose sequence is: glucosamine- N,G-sulfate, iduronic acid, glucosamine-N,3,6-sulfate, idu- ronic acid-2-sul€ate, and glucosamine-N,6-sulfate. This pen- tasaccharide binds antithrombin 111 with resultant accelera- tion of the rate of thrombin inhibition (17). This effect causes heparin to be an anticoagulant. Standard heparin (H-heparin) extracted from porcine mucosa is a pool of chains with differ- ent lengths. This polydisperse material has an average molec- ular mass of 14 f 1 kDa. Recently, heparins with lower molecular masses (L-heparins) have been prepared by chem-

15356

Heparin, Elastase, al-Proteinase Inhibitor Interaction 15357

ical depolymerization of the natural product. These deriva- tives are less polydisperse than the natural compound (17).

A number of glycosaminoglycans including heparin are able to form tight complexes with NE via electrostatic interactions (18-22). Glycosaminoglycan-bound NE is 2-5 times less ac- tive than the free enzyme on synthetic and natural substrates. Here we demonstrate that heparin-bound NE is inhibited much more slowly by alPI than the free enzyme.

EXPERIMENTAL PROCEDURES

Materials

Human NE and cathepsin G, isolated from purulent sputum (23), were active site-titrated with acetyl-Ala,-azaAla-p-nitrophenylester (24) from Enzyme System Products (Livermore, CA) and N-trans- cinnamoylimidazole (25) from Sigma (Paris, France), respectively. Porcine pancreatic elastase was isolated as described previously (16) and active site-titrated in the same way as NE. Bovine pancreatic trypsin (Choay, Paris, France) was active site-titrated with p-nitro- phenyl-p'-guanidinobenzoate (26) from Sigma. Human plasma a1PI was isolated and titrated as described earlier (27). Recombinant alp1 and eglin c were gifts from Dr. Schnebli (Ciba-Geigy, Basel, Switzer- land). Recombinant eglin c is the N-acetylated form of the 8.1-kDa inhibitor initially isolated from the leech. Suc-Alas-pNA, Suc-Ala,- Pro-Phe-pNA, MeOSuc-Ala,-Pro-Val-pNA, benzoyl-Arg-pNA, and MeOSuc-Ala,-Pro-Val chloromethylketone were from Bachem (Bub- endorf, Switzerland). Stock solutions of the substrates and the chlo- romethylketone were prepared in dimethyl-formamide. The final concentration of the organic solvent in the reaction media was 1% (v/v) throughout.

Reagent-grade heparins were from Calbiochem (Meudon, France). H-heparin, isolated from porcine mucosa, was the sodium salt with a molecular mass of 13.5-15 kDa. It had an anticoagulant activity of 171 units/mg. L-heparin, prepared by depolymerization of H-heparin, was the sodium salt with a molecular mass of 3.7 kDa. It had a biological activity of 51 anticoagulant units/mg and 80 antifactor Xa units/mg. One disaccharide unit of heparin has a molecular mass of 0.6 kDa (21). H-heparin immobilized on Sepharose CL-GB was from Pharmacia-LKB (St-Quentin en Yvelines, France). The standard heparins used in clinical care were from Hoffman-la-Roche, Basel, Switzerland (Liquemin" and Heparin'") and Choay (Calciparin'"). The low molecular mass heparins used in clinical care were from Kabivi- trum, Noisy-le-Grand, France (Tedelparin@) and Pharmuka-Four- nier, Joinville, France (Enoxaparin'"). Dermatan sulfate, chondroitin sulfate, and hyaluronic acid came from Sigma.

Unless otherwise stated, all kinetic experiments were done in 50 mM Hepes, 100 mM NaCl, pH 7.4, 25 "C.

Methods

Rate Constant ka for the Inhibition of NE by a,PI-The rate of the irreversible inhibition of NE by a,PI in the presence of heparin was measured under second-order conditions ([E,] = [I,,] = 0.1 p ~ ) . Ten pl of NE were added to 980 pl of a1PI + heparin. After appropriate intervals, 10 pl of 100 mM MeOSuc-Ala,-Pro-Val-pNA were added and the release of pNA was monitored at 410 nm. The residual enzymatic activities were used to calculate the concentration of free enzyme [E] which is related to the initial enzyme concentration [E,,] and the reaction time t through the following classical second-order equation.

[El = [E01

1 + [Eo].k,.t (1)

The values of k, and their standard errors were calculated by nonlin- ear regression analysis using the Enzfitter software.

In the absence of heparin, the rate of inhibition of NE by alPI was too fast to measure under the above conditions. Attempts made with [En] = [In] = 10 nM were unsuccessful due to the tendency of NE to stick on the walls of the spectrophotometer cuvettes (28). High salt concentrations or detergents prevent this undesirable effect (9, 10). The reaction was therefore done in 50 mM Hepes, 500 mM NaC1, pH 7.4, 25 "C, with [En] = [In] = 10 nM and 1 mM MeOSuc-Ala,-Pro- Val-pNA. Under these conditions the k. was found to be (1.3 & 0.17) lo7 s- I , a value close to that found by Straus et al. (9) (1.1 X lo7 M-l s-') in 100 mM Hepes, 500 mM NaCl, pH 7.5, 25 "C. This figure also agrees with that found by Braun et al. (10) (2.4 X 10' M" s-l) in

physiological phosphate-buffered saline containing 0.05% Triton X- 100 at pH 7.3 and 37 "C. Ionic strength has therefore a negligible effect on ka.

Rate Constant k, for Other Enzyme-Inhibitor Systems-The rate of inhibition of porcine pancreatic elastase and bovine pancreatic trypsin by nlPI in the absence and presence of H-heparin was measured under second-order conditions ([Eo] = [IO] = 0.1 pM) using the same principle as above. Suc-Ala3-pNA and benzoyl-Arg-pNA (final concentrations: 1 mM) were used as substrates of elastase and trypsin, respectively. The k. for the interaction of NE with eglin c in the presence of H-heparin was measured in exactly the same way as the ka for the NE + alPI association. Eglin c is a reversible NE inhibitor (k, = 1.3 X lo' M-' s-', kd = 3.5 X s-', Ref. 10). The time-dependent inhibition data could, however, be fitted to Equation 1, indicating that reversibility of inhibition was undetectable under our conditions (Le. tM (dissociation) >> duration of experiment).

The influence of heparin on the rate of inhibition of cathepsin G by alPI was measured under pseudo-first-order conditions ([a1PI] >> [cathepsin GI) by adding enzyme to inhibitor + substrate + H-heparin mixtures and recording the progress curves. The final concentrations of reactants were: cathepsin G, 0.1 pM, a,PI, 5 p ~ , Suc-Ala,-Pro- Phe-pNA, 1.2 mM, H-heparin-variable. First-order progress curves were obtained as predicted by theory (29). These curves were analyzed using the following relationships (29):

p = U. (1 - e-'*) k (2)

where LJ, is the rate at t = 0 and k is an apparent first-order rate constant given by:

The constant k and its confidence interval were calculated by nonlin- ear regression analysis. Control experiments done without alPI showed that heparin decreased the activity of cathepsin G on Suc- Ala,-Pro-Phe-pNA up to a limiting value representing 55% that of the activity of the free enzyme, a result that agrees with that of Redini et al. (21).

Stoichiometries and Equilibrium Dissociation Constants Kd of the Heparin-NE Complexes-Binding stoichiometries and Kd values were measured by adding increasing amounts of heparin to constant amounts of NE and measuring the residual enzyme activities with an appropriate substrate. For the determinations of stoichiometry, 0.5 p M NE was used and the substrate was Suc-Alaa-pNA (1.4 mM for L- heparin and 1.8 mM for H-heparin). For the measurement of Kd, the concentration of NE was either 20 nM (L-heparin) or 40 nM (H- heparin) and the substrate was MeO-Suc-Ala2-Pro-Val-pNA (1 mM for L-heparin and 1.5 mM for H-heparin).

Kd was calculated using the following equation (30),

where A = + [I0] = [heparin] X number of binding 01 + ISnl/Km [Ed '

sites; 01 = K, in the presence of a saturating concentration of heparin/ K,,, in the absence of heparin; ui/un = rate in the presence of heparin/ rate in the absence of heparin; LJ,, limit of vi for saturating concen- trations of heparin.

Equation 4 takes into account both the tight-binding behavior of glycosaminoglycans (i.e. significant depletion of inhibitor by enzyme) and the residual enzymic activity of the NE-glycosaminoglycans complexes (18, 21, 31). To calculate a , we measured the K,,, for the NE/MeO-Suc-Ala,-Pro-Val-pNA system in the absence and presence of variable concentrations of heparin using conventional means. K, did not vary with H-heparin (K,,, = 0.085 mM, a = l), whereas 01 was equal to 1.4 for L-heparin. Kd and v, were calculated by nonlinear regression analysis using either Equation 4 for L-heparin or its simplified form ( a = 1) for H-heparin.

Sepharose-heparin Affinity Chromatography-The affinity support was equilibrated with 20 mM Hepes buffer, pH 7.4, and 8 ml of gel slurry were poured into a 1-cm diameter HR 10/10 Pharmacia-LKB column mounted on a Pharmacia fast protein liquid chromatography

15358 Heparin, Elastase, cq-Proteinase Inhibitor Interaction

apparatus. The various enzymes and inhibitors were dissolved in the above buffer. The equimolar complexes of NE with aIPI or eglin C were prepared just before chromatography. One-hundred pl of 100 p~ trypsin, 20 p~ porcine pancreatic elastase, 30 PM nlPI, 100 p~ eglin c, 10 p~ aIPI .NE complex, 50 p~ eglin-NE complex, and 300 pl of 30 p~ cathepsin G were chromatographed a t a flow rate of 0.5 ml/min. After elution of the void volume of the column, a linear NaCl gradient was applied.

Electrophoresis-SDS-polyacrylamide gel electrophoresis and sil- ver staining were done using the Phastsystem apparatus from Phar- macia as described (16). NE (0.1 p ~ ) was added to a IP I (0.12 p ~ ) in the absence or presence of 0.2 p~ H heparin. Following 1 or 60 min of incubation, 3 mM MeO-Suc-Ala2-Pro-Val-chloromethyl ketone was added to block free enzyme. After a further 1 min of incubation, the solutions were denatured and electrophoresed.

Elastolytic Activity of NE-The rate of solubilization of human lung elastin by NE was monitored a t pH 7.4 and 37 "C as described previously (32). Variable amounts of H-heparin were mixed with constant amounts of NE (final concentration: 1 PM) before addition to the elastin suspension.

RESULTS

Effect of Heparin on the Rate of Inhibition of NE by alPI- Fig. 1 shows that the rate constant k,, for the inhibition of NE by aIPI decreases with increasing heparin concentrations and then reaches a plateau. H-heparin, formed of 22-25 disaccha- ride units, was much more effective than L-heparin composed of about 6 disaccharide units. With the former, the maximal decrease of k, (290-fold) was observed with a 2-fold molar excess of heparin over NE. In contrast, a 54-fold excess of L- heparin was required to get the largest (40-fold) decrease in k,. The results were the same whether NE was added to a mixture of heparin + alPI (condition of Fig. 1) or whether a lPI was mixed with heparin + NE.

Heparin was somewhat less efficient on recombinant alPI than on wild-type alPI although in the absence of heparin the two inhibitors reacted with NE with about identical rates. For instance, with a 2.5-fold molar excess of H-heparin over NE, the decrease of k, for recombinant aIPI was 59% that determined with wild-type alPI. Similarly, with an 8-fold excess of L-heparin, the decrease of k,, for recombinant a lPI was 34% that observed with the wild-type inhibitor (data not shown).

The influence of heparin on the rate of NE-aIPI complex formation could also be studied by SDS-polyacrylamide gel electrophoresis since the complex migrates as a 79-kDa band whereas free aIPI migrates as a 53-kDa protein (11). For the sake of comparison, NE, aIPI, and H-heparin were reacted at concentrations close to those used to measure k. (shown by an asterisk in Fig. 1). Fig. 2 shows that in the absence of heparin the bulk of SDS-stable complex is formed within 1 min, in agreement with the low half-life of the reaction (tnh (association) = 1/k. [Eo] = 0.8 s ) . In contrast, virtually no

PERCENT S A T U R A T I O N OF NE BY H E P A R I N

- 0 86 9 3 965 90 99 0 90 9f 99.6 99.6

10' - - H - H E P A R I N L - H E P A R I N

0 01 0 1 21 35 0 68 135 2 1 51 108 MOL H E P A R I N / MOL NE

FIG. 1. Effect of heparin on the rate constant k., for the inhibition of NE by alPI at pH 7.4 and 25 "C. The heparin concentration was expressed either as the molar ratio of heparin to NE or as the degree of saturation of NE by heparin calculated using the constants given in Table 11. The asterisk indicates the concentra- tion conditions used for demonstrating complex formation by electro-

C 1' 60'1' 6 0 " -

FIG. 2. SDS-polyacrylamide gel electrophoresis of NE + alPI mixtures incubated for 1 or 60 min in the absence (-) or presence (+) of H-heparin. The concentrations were about the same as those shown by an asterisk in Fig. 1: NE = 0.1 pM alPI = 0.12, H-heparin = 0.2 p ~ . Lane C, nIPI control. The 49-kDa band corresponds to post-complex nlPI (11, 16).

complex forms during this time interval if H-heparin is pres- ent, in agreement with the high half-life of the reaction under these conditions ( tH (association) = 220 s). However, after 1 h ( 4 6 half-lives) of incubation in the presence of heparin, the 79-kDa complex is clearly visible. This experiment con- firms in a nonenzymatic way the rate-depressing effect of heparin and indicates that heparin does not prevent the formation of a denaturant-stable NE. a lPI complex but sim- ply decreases its rate of formation.

Specificity of the Heparin Effect-Table I shows that H- heparin does not impair the interaction of alPI with porcine pancreatic elastase, an enzyme closely related to NE (5), but decreases the k,, for the inhibition of NE by eglin c, an inhibitor unrelated to alPI (33). On the other hand, H-heparin does not change the k, for the reaction of alPI with bovine pancreatic trypsin but strongly decreases the ko for the reac- tion of a lPI with neutrophil cathepsin G, another cationic serine proteinase. The k,-depressing effect of heparin is there- fore not specific for the NE + aIPI association.

The above data rather suggest that the ability of heparin to decrease k, is related to its proteinase-binding capacity. To confirm this, we chromatographed the various proteinases and inhibitors on a Sepharose-heparin column at a very low ionic strength (20 mM Hepes, pH 7.4) and eluted them with a linear salt gradient. Fig. 3 shows that neither wild-type aIPI nor recombinant eglin c bind to the column. The same was true for recombinant a lPI (curve not shown). In contrast, NE, cathepsin G, and pancreatic elastase and trypsin all bound to the affinity column. NE interacted so strongly with the matrix that even a 0-4 M NaCl gradient could not elute it. Elution occurred, however, when the affinity chromatog- raphy was done in the presence of 10% (v/v) acetonitrile. In that case, active NE eluted at 1.65 M NaCl (data not shown). Cathepsin G was also tightly bound to heparin-Sepharose since an NaCl concentration of 1.45 M was required to elute it. The strong interaction of the leukoproteinases contrasts with the weak binding of pancreatic elastase and trypsin which eluted at 0.35 and 0.5 M NaCl, respectively. The shoul- der seen on the elution profile of trypsin is probably due to the inert proteins present in the commercial preparation of this enzyme which was only 60% active by active site titration. Sepharose-heparin affinity chromatography thus entirely confirms the aforementioned hypothesis: NE and cathepsin G, whose k, with alPI is strongly decreased by heparin, tightly interact with this polymer while pancreatic elastase and tryp- sin, whose inhibition by alPI is not impaired by heparin, do not form tight complexes with it.

Sepharose-heparin chromatography was also used to study the effect of alPI and eglin c on the affinity of NE for heparin. The 1:l enzyme-inhibitor complexes were chromatographed in the same way as free NE. Fig. 3 shows that the complexes

phoresis (Fig. 2). of NE with alPI and eglin c elute from the column at 0.6 and

15359 Heparin, Elastase, al-Proteinase Inhibitor Interaction TABLE I

Effect of H-heparin on proteinase-inhibitor association The estimates for the association rate constants 12, are given together with their standard errors obtained from

the nonlinear regression analyses. Mole heparin

Proteinase Inhibitor k,

decrease in k. Mole proteinase Without heparin With heparin

” s-1 n-fold Neutrophil elastase (YIP1 2.1 (1.3 k 0.17) lo7 (4.5 f 0.4) 104 290 Neutrophil elastase Eglin c 2.1 (7.7 f 2.1) lo6 (4.7 f 0.7) 104 164 Porcine pancreatic elas- a,PI 2.1 (2.3 f 0.4) lo5 (2.7 f 0.4) lo5 None

Bovine pancreatic trypsin a,PI 3.5 (6.5 f 1.8) lo5 (8.3 f 2.1) lo5 None Neutrophil cathepsin G ff,PI 2.4 (3.9 f 0.2) lo5 130 f 3.4 3000

tase

‘ From Braun et al. (10).

I I 0 LO 8 0 120

T I M E I m i n u t e s I

FIG. 3. Sepharose-heparin chromatography of al-protein- ase-inhibitor (al), eglin c (Ec), pancreatic elastase (PE) , pan- creatic trypsin (PT), cathepsin G (CG), eglin c-neutrophil elastase complex (Ec-NE), and al-proteinase inhibitor-neu- trophil elastase complex (al-NE). The proteins were chromato- graphed separately and eluted with a linear NaCl gradient (----I.

0.7 M NaCl, respectively and therefore interact much weaker with heparin than does free NE. The inhibitors thus strongly decrease the affinity of NE for heparin.

Stoichiometries and Dissociation Constants of the Heparin- NE Complexes-Heparin inhibits partially the activity of NE on synthetic substrates (21,22). Use was made of this property to measure the binding stoichiometry and the equilibrium dissociation constant Kd of the heparin-NE complexes. The former was measured with high reactant concentrations to ensure pseudo-irreversible binding of NE to heparin. The inset of Fig. 4 shows the effect of increasing concentrations of H-heparin on the activity of constant concentrations of NE (0.5 p ~ ) . It can be seen that the polysaccharide titrates the enzyme with an equivalence point corresponding to the binding of about three molecules of NE/molecule of H-hepa- rin composed of 22-25 disaccharide units. With L-heparin, formed of six disaccharide units, the binding stoichiometry was 1:l (Table 11). Thus, 1 mol of NE binds 6 to 8 disaccharide units. Arteparon@, another sulfated glycosaminoglycan, was found to bind NE with a stoichiometry of about 5 disaccharide units/mol of enzyme (18).

With lower NE concentrations, the inhibition curves were concave (e.g. Fig. 4) so that the data could be used to calculate K d by nonlinear regression analysis. A good fit was obtained between the experimental points and the inhibition curve calculated according to Equation 4 (Fig. 4). Table I1 shows that the Kd for the H-heparin-NE complex is in the nanomolar range whereas that for L-heDarin-NE is about 30-fold higher.

1 1.0

0.11

0.0

0.6 > \

’ 0.4 .-

I I

H E P A R I N [ M x l O ’ ) 0 0.1 0.3 0.5

FIG. 4. Determination of the equilibrium dissociation con- stant and the stoichiometry of the H-heparin-NE complex at pH 7.4 and 25 “C. Increasing concentrations of H-heparin were added to constant concentrations of NE and the residual enzymatic activity was measured with a synthetic substrate. Closed circles, NE = 40 nM, substrate = 1.5 mM MeO-Suc-Ala2-Pro-Val-pNA. Open circles, NE = 0.5 PM, substrate = 1.8 mM Suc-Ala3-pNA. ui/vo = rate in the presence of heparin/rate in the absence of heparin.

TABLE I1 Binding stoichiometry, equilibrium dissociation constant 0 , and

fractional activity at infinite heparin concentration (v&d on MeOSuc-Alap-Pro-Val-pNA of the complexes of NE

with H- and L-heparin

Constants Complex of NE with:

H-heuarin L-heuarin

Binding stoichiometry 3 1 Kd (nM)‘ 3.3 k 1.0 89 f 12 V J V r I 0.28 0.31

Best estimates and standard errors obtained by nonlinear regres- sion analysis.

These K d values were used to calculate the degree of satu- ration of NE by heparin in the experiments describing the effect of heparin concentration on the k, for the inhibition of NE by alPI. Fig. 1 shows that for both heparins the leveling off of ka corresponds to a full saturation of NE by heparin.

The above data analysis also provides urn, the enzymic activity at infinite heparin concentration (see “Experimental Procedures”). As can be seen from the u,/uo ratio reported in Table 11, the NE-heparin complexes are about 30% active on MeO-Suc-Ala:,-Pro-Val-pNA.

When increasing concentrations of H-heparin were added to constant concentrations of NE and the mixtures tested for elastolvtic activitv on insoluble elastin. a titration curve sim-

Using a similar method, Redini et al. (21) found a Kd of 40 ilar to tha t shown in Fig. 4 was recorded (data not shown). nM for the complex of NE with a 4.8-kDa heparin preparation. The elastolytic activity of the NE-heparin complex was 20%

15360 Heparin, Elastase, al-Proteinase Inhibitor Interaction

that of the free enzyme. This is in agreement with the data obtained by Baici (18) with the sulfated polysaccharide Ar- teparona whose complex with NE was found to have a residual elastolytic activity of 28%.

Effect of Other Glycosaminoglycans-To investigate the im- portance of sulfate groups, we tested the effect of dermatan sulfate and chondroitin sulfate which have only one sulfate group per disaccharide unit as well as that of hyaluronic acid which lacks sulfate groups (34). The rate of inhibition of NE by aIPI decreased with increasing concentrations of these glycosaminoglycans and then reached a plateau as in the case of heparin. The lowest values of ka are reported in Table 111. It can be seen that these poorly sulfated compounds have a much smaller k,-depressing effect than heparin which bears three sulfate groups/disaccharide unit. Also, the amounts of these compounds required to get the maximal decrease in k, are much higher than those of heparin, suggesting a weaker affinity. These results suggest that the efficacy of glycosami- noglycans is, at least in part, related to their degree of sulfa- tion. Consistent with this assumption, the effect of heparin was lost at high ionic strength (data not shown).

Effects of Heparins Used in Clinical Care-Two types of heparins are used as anticoagulant drugs: the standard (high molecular mass) heparins and the low molecular mass ones. Both preparations are ill-defined mixtures of polysaccharide chains whose concentration is not given but whose range of therapeutic concentration (units/ml blood) is provided by the manufacturer. We have tested three standard heparins and two low molecular mass heparins (see "Materials") for their ability to depress the ka for the NE + alPI association. The solutions were diluted in such a way that the final concentra- tions of heparin in the reaction medium (units/ml reaction medium) was close to the range of the therapeutic concentra- tion (units/ml blood) given by the manufacturer. All prepa- rations depressed ka, the standard heparins being significantly more efficient (-220-fold decrease in ka) than the low molec- ular mass heparins (-14-fold decrease in k,J.

DISCUSSION

Heparin modulates the activity of numerous biological sys- tems. For instance, it accelerates the inhibition of thrombin and factor Xa by antithrombin I11 (35), promotes the binding of thrombin to fibrin (36), stimulates the activity of plasmin- ogen activator (37), and accelerates the inactivation of anti- thrombin I11 by NE (38). We report here for the first time that it also decreases the rate of inhibition of NE by alPI, a member of the serpin superfamily of proteinase inhibitors.

TABLE 111 Compared effects of various glycosaminoglycans on the rate constant

k. for the inhibition of NE by alPZ Glycosaminoglycan Lowest value Maximal decrease

Nature Concentration of '0" in k.

" s" n-fold None (1.3 f 0.17) lo7 H-heparinb 0.21 p M (4.5 k 0.40) io4 289 L-heparinb 5.4 p M (3.2 * 0.30) io5 40 Dermatan sulfate' 0.2 mg/ml (1.9 f 0.22) 10' 6.8 Chondroitrin 4-sulfate' 0.2 mg/ml (2.4 f 0.40) 10' 5.4 Chondroitrin 6-sulfate' 0.4 mg/ml (2.2 f 0.19) 10' 5.9 Hyaluronic acidd 1 mg/ml (6.5 * 0.05) 10' 2.0

sion analysis. ' Best estimates and standard errors obtained by nonlinear regres-

Three sulfate groups per disaccharide unit. One sulfate group per disaccharide unit, no molecular mass avail-

No sulfate, no molecular mass available. able.

Unlike some members of this family (antithrombin 111, hep- arin cofactor, protease nexin, and protein C, Ref. 39), alPI does not bind heparin as suggested previously (40) and con- firmed here. The rate-depressing effect of heparin must there- fore be related to the ability of the sulfated polymer to form ion pairs with some of the 19 surface-exposed (2, 3) arginine residues of NE. Two sets of observations strengthen this view: (i) k,, reaches its lowest value once NE is fully saturated by the sulfated polysaccharide (Fig. 1); (ii) heparin does not decrease the k, for the reaction of alPI with porcine pancreatic elastase and bovine trypsin, two enzymes that do not form tight complexes with it (Fig. 3), whereas the rate of association of alPI with neutrophil cathepsin G, a proteinase that tightly binds heparin, is strongly depressed by the sulfated polysac- charide (Table I).

The rate of inhibition of NE by alPI decreases with the heparin concentration and then reaches a plateau. On the other hand, the levelling off of k, is paralleled by saturation of NE by the polymer (Fig. 1). These results suggest the following reaction scheme:

k,,o E + I + E 1 + + H H 41 K d B K d

E H + I + E H I k , m

where E, I, and H stand for NE, alPI, and heparin, respec- tively. In this scheme, ka,rn represents the value of ka measured with saturating concentrations of heparin. With lower heparin concentrations the inhibition process is governed by both k,,o, and k , , so that plots of [E] uersus t (Equation 1) should be biphasic. This was however not observed because, in our experiments, NE was saturated with heparin to the extent of a t least 86% so that the k,,o pathway was barely detectable. With lower heparin concentrations the inhibition reaction was too fast to follow. The dimensionless number fi is probably much larger than unity at least in the case of H-heparin because the NE-alPI complex has much less affinity for heparin than free NE (Fig. 3). This suggests that alPI desta- bilizes the NE-heparin complex after reacting with it, an effect that could be due to an alPI-induced change in the conformation of NE. Scheme I also suggests that NE reacts much faster with heparin than with alPI, since the effect of heparin on ka was found to be the same whether NE was added to a mixture of heparin + alPI or whether alPI was mixed with heparin + NE.

The mechanism of action of alPI involves a limited prote- olysis of the inhibitor followed by an irreversible trapping and inhibition of the proteinase (11, 41). Proteolysis of alp1 is probably the rate-limiting step for proteinase inhibition since chymotrypsinogen, whose catalytic activity on the pseudo- substrate p-nitrophenyl p'-guanidinobenzoate is lo6 to lo7 times lower than that of the active enzyme (42), reacts about 105-fold slower with alPI than does the active enzyme (43). At first sight, the rate-depressing effect of heparin may there- fore be directly related to its ability to decrease the catalytic activity of NE. The data compiled in Table IV, however, contradict this hypothesis. First, the complexes of H- and L- heparin with NE have about identical residual activities (urn/ uo) on a synthetic substrate but react with alPI with quite different rates (7.3-fold difference in k0,.Jka,d. Second, the urn/ uo ratio is higher for the heparin-cathepsin G complex than for the heparin-NE complex while ka,rn/k,,o is 10-fold lower for the former than for the latter association. Third, H-heparin decreases the rate of solubilization of the protein substrate

Heparin, Elastase, cY1-Proteinase Inhibitor Interaction 15361 TABLE IV

Compared effects of heparin on the action of NE and cathepsin G on substrates and on the ka for their association with a1PZ

The ratios vo/u, and k.,o/ko,, have been defined in Table 111 and in Scheme I, respectively.

H-heparin L-heparin Proteinase Substrate or inhibitor

U J U O k,-lka,o u - 1 ~ 0 k a , d k o

NE MeOSuc-Ala2-Pro-Val-pNA 0.31 0.28 NE Elastin 0.20 NE a1PI 0.0034 0.025 Cathepsin G Suc-Alaz-Pro-Phe-pNA 0.55 Cathepsin G alPI 0.00033

elastin by a factor of 5, whereas it decreases the k, for the protein inhibitor alPI by a factor of 290. Further investiga- tions are therefore needed to get a better insight into the mechanism by which heparin decreases the ka for the NE + alPI association.

We believe our finding has important pathophysiological implications. For instance, NE and other positively charged enzymes are stored in the azurophil granules of neutrophils together with negatively charged proteoglycans. The latter serve as a scaffold for the condensation of the cationic en- zymes whose self-destruction is thus avoided (discussed by Navia et al. (3)). Neutrophil activation releases the granule content in the extracellular milieu where alPI should rapidly inhibit NE. Our results suggest that the above proteoglycans may lower the rate of this inhibition process so that part of the liberated NE may attack surrounding biological sub- strates. On the other hand, the lung parenchyma contains proteoglycans (44) whose glycosaminoglycan part may bind NE and hinder its inhibition by alPI. The rate of inhibition of NE in vivo is therefore likely to be lower in the lung interstitium than in the alveolar epithelial lining fluid where glycosaminoglycans are absent.

On the other hand, NE has been reported to cleave plasma proteins such as antithrombin I11 and other coagulation fac- tors in vitro (5). Neutrophil activation with release of NE occurs in the plasma of patients with septicemia (45, 46) or during hemodialysis (47). Standard heparin is widely used as an anticoagulant drug in these two clinical situations. We have shown that standard heparins decrease k, more than 200-fold when tested at concentrations corresponding to their clinical efficacy. The time required for full inhibition of NE in plasma (13) will therefore be increased by a factor of >200 in the above patients. Hence, NE-mediated intravascular proteolysis is likely to occur. This might account for the decreased activity of a variety of coagulation factors observed in patients with severe septicemia (45,46). Moreover, Jordan and co-workers (38) recently reported that heparin consider- ably increases the rate of inactivation of antithrombin I11 by NE. Their finding, together with our data, suggest that if neutrophil activation takes place in patients under heparin therapy, massive degradation of antithrombin I11 may occur. The 50% antithrombin I11 inactivation found by Duswald et al. (46) in patients with severe septicemia might be accounted for by these observations.

Furthermore, NE is able to fragment lung elastin, a process that leads to pulmonary emphysema (5). Rao et al. (48) recently showed that sulfated glycosaminoglycans known un- der the trade name Arteparona are able to prevent NE- induced emphysema in the hamster. They suggested that sulfated polysaccharides might be useful drugs for the human disease. From the current study it may be inferred that such drugs could rather be harmful. The hamster's emphysema was induced with a single and massive intratracheal dose of NE.

Arteparona tightly binds NE and inhibits its elastolytic activ- ity to the extent of 70% (18). :It therefore reduces by two thirds the amount of active NE instiled in the animal's lung. This, together with the observation that the severity of ham- ster emphysema depends upon the amount of NE used to induce it (5), easily accounts for the observed prevention of emphysema by Arteparon@. Human emphysema does not result from an acute NE-mediated lung injury but develops as a result of a long-term inflammation of the lung. Active alPI is present in the lower respiratory tract of alPI-sufficient patients (49) and of most of the so-called alPI-deficients patients (12). Administration of a glycosaminoglycan to such patients would dramatically impair the physiologic function of alPI because the glycosaminoglycan-NE complexes would react so slowly with alPI that most of them would escape inhibition and attack elastin fibers with an efficiency equal to 30% that of free NE. Hence, the beneficial effect of the glycosaminoglycan (70% reduction of elastase activity) would be annihilated by its detrimental effect on alPI function. Our contention is supported by the observation that the lowest k, value found with H-heparin (4.5 X lo4 M-' s-l) is of the same order of magnitude as that reported for the inhibition of NE by oxidized alPI (16), a derivative thought to be quite ineffi- cient in preventing NE-mediated proteolysis in vivo (5, 11).

Acknowledgments-We thank Dr. A. Baici, University Hospital, Zurich, Switzerland for helpful suggestions and Dr. H. P. Schnebli, Ciba-Geigy, Basel, Switzerland for the gift of recombinant eglin c and %PI.

1.

2.

3.

4.

5.

6.

7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

20.

REFERENCES

Sinha, S., Watorek, W., Karr, S., Giles, J., Bode, W., and Travis,

Bode, W., Meyer, E., Jr., and Powers, J. C. (1989) Biochemistry

Navia, M. A., McKeever, B. M., Springer, J. P., Lin, T. Y., Williams, H. R., Fluder, E. M., Dorn, C. P., and Hoogsteen, K. (1989) Proc. Natl. Acad. Sci. U. S. A. 8 6 , 7-11

Tyagi, S. C., and Simon, S. R. (1990) Biochemistry 29, 9970- 9977

Bieth, J. G. (1986) in Regulation of Matrix Accumulation (Me- cham, R. P., ed) pp. 217-320, Academic Press, New York

Taylor, J. C., and Mittman, C. (1987) Pulmonary Emphysema and Proteolysis, Academic Press, New York

Campbell, E. J. (1986) Am. Rev. Respir. Dis. 134,984-986 Beatty, K., Bieth, J., and Travis, J. (1980) J. Biol. Chem. 255 ,

Straus, S. D., Fells, G. A., Wewers, M. D., Courtney, M., Tessier, L. H., Tolstoshev, P., Lecoq, J. P., and Crystal, R. G. (1985) Biochem. Biophys. Res. Commun. 130,1177-1184

Braun, N. J., Bodmer, J. M., Virca, D. G., Metz-Virca, G., Maschler, R., Bieth, J. G., and Schnebli, H. P. (1987) Biol. Chem. Hoppe-Seyler 3 6 8 , 299-308

Travis, J., and Salvesen, G. S. (1983) Annu. Rev. Biochem. 5 2 ,

Ogushi, F., Hubbard, R. C., Fells, G. A., Casolaro, M. A., Curiel, D. T., Brantly, M. L., and Crystal, R. G. (1988) Am. Rev. Respir. Dis. 137 , 364-370

Bieth, J. G. (1980) Bull. Eur. Physiopathol. Respir. 16 , (suppl.) 183-195

Campbell, E. J., and Campbell, M. A. (1988) J. Cell Biol. 106,

Janoff, A., Carp, H., Laurent, P., and Raju, L. (1983) Am. Rev.

Padrines, M., Schneider-Pozzer, M., and Bieth, J. G. (1989) Am.

Casu, B. (1990) Haemostasis 20, Suppl. 162-73 Baici, A., Salgam, P., Fehr, K., and Boni, A. (1980) Biochem.

Baici, A., and Bradamante, P. (1984) Chem. Biol. Interact. 5 1 ,

Baici, A., Salgam, P., Fehr, K., and Boni, A. (1981) Biochem.

J. (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 2228-2232

28,1951-1963

3931-3934

655-709

667-672

Respir. Dis. 127, 531-538

Rev. Respir. Dis. 139,783-790

Pharmacol. 29, 1723-1727

1-11

Pharmacol. 30, 703-708

15362 Heparin, Elastase, a*-Proteinase Inhibitor Interaction 21. Redini, F., Tixier, J.-M., Petitou, M., Choay, J., Robert, L., and 36. Hogg, P. J., and Jackson, C. M. (1990) J. Biol. Chem. 265, 241-

22. Marossy, K. (1981) Biochim. Biophys. Acta 659,351-361 37. Stein, P. L., van Zonneveld, A. J., Pannekoek, H., and Strickland, 23. Martodam, R. P., Baugh, R. J., Twumasi, D. Y., and Liener, I. E. 5. (1989) J. Biol. Chem. 264, 15441-15444

(1979) Prep. Biochem. 9, 15-31 38. Jordan, R. E., Kilpatrick, J., and Nelson, R. M. (1987) Science 24. Powers, J. C., Boone, R., Carroll, D. L., Gupton, B. F., Kam, C.- 237,777-779

M., Nishino, N., Sakamoto, M., and Tuhy, P. M. (1984) J. Bioi. 39. Huber, R., and Carrell, R. W. (1989) Biochemistry 28,8951-8966 Chem. 259,4288-4294 40. Danishefsky, I., and Pixley, R. (1979) Biochem. Biophys. 'Res.

25. Schonbaum, G. R., Zerner, B., and Bender, M. L. (1961) J. BWL Commun. 91,862-870 Chem. 236,2930-2935 41. Carrell, R. W., and Travis, J. (1985) Trends Biochem. Sci. 10,

26. Chase, T., Jr., and Shaw, E. (1967) Biochem. Biophys. Res. 20-24 Commun. 29,508-514 42. Gertler, A., Walsh, K. A,, and Neurath, H. (1974) Biochemistry

27. Bruch, M., and Bieth, J. G. (1986) Bwchem. J. 238,269-273 13,1302-1310 28. Gauthier, F., Frysmark, U., Olhsson, K., and J. G. Bieth. (1982) 43- Brodrick* J* w** Glaser? c* B.9 Larman* c., Geokas* M. c.,

Bwchim. Biophys. Acta 700,178-183 Graceffo, M., Fassett, M., and Maeda, H. (1980) Biochemistry 29. Tian, W.-X., and Tsou, C.-L. (1982) Biochemistry 21,1028-1032 19,4865-4870 30. Szedlacsek, S. E., Ostafe, V., Serban, M., and Vlad, M. 0. (1988) 44. Clark, J. G., Kuhn, C., McDonald, J. A., and Mecham, R. P.

(1983) in International Review of Connective Tissue Research 31. Baici, A. (1987) Biochem. J. 244, 793-796 (Hall, D. A,, and Jackson, D. S. e&) Vol. 10, pp. 249-331,

32. Boudier, c., Holle, c., and Bieth, J. G. (1981) J. Uzern. 266, 45. Egbring, R., Schmidt, w., Fuchs, G., and H ~ ~ ~ ~ ~ ~ ~ , K. (1977) Academic Press, New York

33. Schnebli, H. p.9 and Liersch, M. H. (1989) in Elastin and Elastases 46. Duswald, K. H., Jochum, M., Schramm, w., and Fritz, H. (1985) (Robert, L., and Hornebeck, W. eds) Vol. 11, pp. 137-143, CRC Press, Boca Raton, FL

Surgery 98,892-899 47. Horl, W. H., Steinhauer, H. B., and Schollmeyer, R. P. (1985)

Biochemistry (Piez, K. A., and Red&, A. H., eds) pp. 277-328, 48. b o , N. V., Kennedy, T. P., Rao, G., Ky, N., and Hoidal, J. R. Elsevier Science Publishers B. V., Amsterdam (1990) Am. Rev. Respir. Dis. 142,407-412

35. Danielsson, A., h u b , E., Lindahl, U., and Bjork, I. (1986) J. Biol. 49. Gast, A., Dieteman-Molard, A., Pelletier, A., Pauli, G., and Bieth, Chem. 261, 15467-15473 J. G. (1990) Am. Rev. Respir. Dis. 141,880-883

Hornebeck, W. (1988) Biochem. J. 262,515-519 247

Biochem. J. 254,311-312

10256-10258 Blood 49,219-231

34. Heinegard, D., and Paulason, M. (1984) in Extracellular Matrix Kidney Int. 28, 791-796