Figure 8.1 Definition of binding sites for substrates and inhibitors.
(iv) Hydrolysis of the acyl-enzyme by catalytic addition of water produces the N-terminal carboxylic acid fragment of the peptide and the free enzyme which is then ready to repeat the cycle.
Human neutrophil and leucocyte elastases are members of a subfamily of proteases released in response to various inflammatory stimuli and capable of degrading a variety of structural proteins including collagen, elastin and fibronectin. Under normal circumstances, elastolytic activity is controlled by natural proteinaceous inhibitors such as ^-proteinase inhibitor (a1-P1) which is present in plasma (guards the lower airways), secretory leukocyte protease inhibitor (SLP1) secreted by mucosal cells (protects the larger airways) and elafin found mainly in the skin and has also been detected in bronchial secretions. The regulation of elastase activity by these natural proteinaceous inhibitors breaks down in a number of pathophysiological states resulting in unrestrained elastolytic activity associated with diseases such as emphysema, rheumatoid arthritus, chronic bronchitus, cystic fibrosis, adult respiratory distress syndrome and glomerulonephritis. The aim of developing human neutrophil elastase inhibitors has been to identify agents to treat diseases associated with one of the most destructive enzymes in the body.
The primary structure, x-ray crystal structure and gene sequence for human neutrophil elastase (HNE) have been determined. Natural inhibitors have been produced by recombinant technology and formulations have been developed for aerosol and intravenous administration. Attempts have also been made to develop low molecular weight synthetic inhibitors based on the enzyme's mechanism. Design of human leucocyte elastase (HLE) inhibitors has been based on computer modelling studies and proposed enzyme-inhibitor complexes using X-ray crystal structure information from HLE or porcine pancreatic elastase (PPE). Structural information for PPE and PPE-inhibitor complexes has been available for some time and initial efforts were directed towards mapping the active-site of the enzyme with synthetic substrates and irreversible enzyme inhibitors such as peptidyl chloromethyl ketones. Elastases have a relatively small S1 subsite and are more dependent than most other serine proteases on interactions of their extended binding pocket as a means of achieving binding of substrates. Most mapping studies have focused on the N-terminus side of the scissile bond. The optimal tetrapeptide recognition sequence for both synthetic peptide substrates and inhibitors has been identified (MeO-Suc-Ala-Ala-Pro-Val-X). Proline, the optimal residue at P2 for both the substrates and inhibitors, is thought to 'pre-organise' the 'backbone' into a conformation which is complementary to the enzyme. The methoxysuccinyl group increases potency by binding to the S5 subsite and improves the aqueous solubility of the peptide. Various elements of the catalytic machinery, involved in the formation and breakdown of the transition state and reaction intermediates, have been considered for inhibitor design.
Affinity label inhibitors combine the concept of the substrate fragment, an affinity fragment which guides the molecule into the active site, and a chemically reactive group which reacts with the essential catalytic groups, usually by alkylation, to block enzyme activity. This type of inhibitor e.g. peptidyl chloromethyl ketone has been the most studied and used to investigate secondary enzyme-inhibitor interactions.
Transition-state analogues lack the scissile amide linkage. Peptidyl-aldehydes, -boronic acids, -phosphonic acids, and -methyl ketones derivatives all form a complex with the enzyme which resembles the tetrahedral intermediate. Peptidyl aldehydes such as (8.18), which are selective for human leucocyte elastase, borrow the structural features of the substrate elastin and the natural inhibitor a1-P1 to build the appropriate recognition features into the backbone of the inhibitors. Inhibitors of this type lack metabolic stability. Simple aliphatic ketones are poor inhibitors of serine proteases but trifluoromethyl ketones (8.19) are competitive, slow binding, inhibitors of elastase. Introduction of alpha fluorine atoms increases the electrophilicity of the carbonyl group and also stabilises the resulting oxyanion formed with the serine hydroxyl of the tetrahedral complex (now confirmed by X-ray crystallography) which resembles the oxyanion of the enzyme-
substrate reaction intermediate. The slow binding nature of trifluoroketones is due to a rate-limiting conformational change of a highly stabilised enzyme-inhibitor complex which allows for optimum interaction between the enzyme and inhibitor. Optimal inhibitory activity is exhibited by tri- and tetra-peptide analogues (P1-P4) with lipophilic side-chains. However, in vivo activity in the hamster emphysema model does not correlate with the in vitro potency of these compounds when administered directly into the lung. Acylsulphonamide derivatives exhibit a long duration of action and in vivo activity. ICI 200 880 (8.20) (administered intravenously or by aerosol) is undergoing clinical evaluation in adult respiratory distress syndrome. Information from X-ray crystallography studies of some peptidyl inhibitors bound to HLE and PPE has been useful in designing pyridone-based inhibitors (8.21). Crystal structures indicate a pair of hydrogen bonds between the P3 residue of the inhibitors and the Val-216 residue of the enzyme. The carbonyl and amido groups of Val-216 are approximately planar and there is nearly a coplanar arrangement of the reciprocal pair of hydrogen bonding partners (NH, C=O) of the P3 residue of the inhibitor. Maintenance of this coplaner arrangement was considered to be important in the design of non-peptide inhibitors. The S3 subsite of HLE, a relatively shallow area exposed to the solvent, indicated that occupation of this site did not make a significant contribution to the binding affinity of inhibitors. A planar molecular fragment, such as a pyridone ring, is acceptable as the P3 residue of inhibitors. Molecular modelling studies have demonstrated that the pyridone carbonyl and 3-position NH groups could be positioned so that hydrogen bonding interactions could be formed with Val-216. However the pyridone nucleus could not access the enzyme S2 subsite
which normally requires a hydrophobia group to improve inhibitor binding affinity. Pyrimidinone-containing trifluoromethyl ketones (8.22, Ki=0.1 pM) show a good combination of enzyme selectivity, oral bioavailability and reasonable duration of action. Boronic acid inhibitors (8.23, Ki=6.2 nM) show good inhibitory potency in vitro but not in vivo.
Some pentafluroethyl derivatives such as MDL 101 146 (8.24, Ki=25 nM) are orally active, whereas the trifluoromethyl derivatives (8.25, Ki=12 nM) show no oral activity possibly due to the difference in the degree of hydration of the respective electrophilic ketones. The E-enol acetate derivative (8.26) of MDL 101 146 acts as an orally active prodrug.
Unlike affinity labels and transition-state analogues, mechanism-based inhibitors of elastase are activated by the catalytic machinery of the target enzyme by two possible mechanisms. With "acyl enzyme" inhibitors the catalytic attack causes the formation of an acyl enzyme which deacylates slowly without irreversibly inactivating the target
enzyme. Studies on the electronic effects of ring substituents on the acylation and deacylation where the transferred group is a benzoyl derivative have shown that, in general, electron donating substituents stabilize the acyl enzyme. Steric factors also contribute to the stabilization process. Inhibitors which act mainly via this mechanism with an appropriate pharmacokinetic profile and selectivity for human elastases have yet to be developed. An alternative approach has been the development of compounds where the catalytic attack causes the formation of chemically reactive intermediates which irreversibly modify the enzyme by forming stable covalent bonds with a different functional group at or near the active site. This type of inhibitor has the advantage of being relatively chemically inert until activation by the target enzyme.
Bacterial penicillin binding proteins, beta-lactamases and serine proteases all hydrolyse an amide bond via formation of a tetrahedral intermediate. The benzyl ester of the beta-lactamase inhibitor clavulanic acid is a weak inhibitor of HLE. Extensive screening of chemical derivatives have identified cephalosporin sulphone (8.27), as a potent beta-lactam elastase inhibitor. Bacterial penicillin binding proteins require a beta configuration at C-7, whereas mammalian elastases prefer the alpha configuration at C-7. Small alpha-orientated substituents such as chloro or methoxy are preferred at C-7 with the sulphone derivatives showing the highest inhibitory potency for elastase. Masking of the free carboxyl group at C-4 of the cephalosporins with ester, amide or ketone substituents also increases inhibitory activity for elastase. Modelling studies show the C-4 substituents to be positioned around the S1'-S2' sites of elastase. The shape and lipophilicity of C-4 substituents also contribute to elastase inhibitory potency by altering the reactivity and structural reorganisation ensuing from beta-lactam cleavage which is essential to the enzyme inactivation mechanism. C-4 tert-Butyl ketones of 7 alpha-chlorocepham series are equal in potency with the ester derivatives and the tert-butyl ketones of 7 alpha-methoxycepham (8.28, Ki= 21 nM, t1/2=75 h) show high inhibitory potency combined with hydrolytic stability which is greater than the ester, thioester or amide (8.29, Ki=75 nM, t1/2=25 h) analogues. Introduction of an acyloxy substituent at the C-2 position further increases potency. X-ray crystallographic data from cephalosporin A bound to the porcine enzyme shows inhibition of elastase to be initially reversible, followed by a time-dependent irreversible inhibition resulting from alkylation of His-57 by the dihydrothiazine ring of the inhibitor. Crystallographic data, biochemical studies and structural characterisation of enzyme-inhibitor complexes and biproducts indicates that the mechanism operating through the enzyme-inhibitor complex involves beta-lactam ring opening by the catalytic Ser-195 residue, expulsion of a leaving group at the 3'-position of the cephem moiety and
binding to the His-57 residue of the enzyme catalytic triad (see Figure 8.3). Cephem 4-ketones, with no adequate leaving group attached to the dihydrothiazide ring, behave as poor substrates rather than inhibitors and are slowly completely hydrolysed by HNE.
A novel mechanism of enzyme inhibition has been suggested for benzisothiazolone inhibitors. The inhibitors inactivate the enzyme by a suicide mechanism but the enzyme is then able to regain its full activity. The lead compound for benziothiazolone inhibitors, KAN 400 473, (8.30, Ki=15 nM), could not be detected in human blood after an incubation time of less than 1 minute. An isopropyl substituent at the 4-position
Figure 8.3 p-Lactam inhibition of elastase (after
Figure 8.3 p-Lactam inhibition of elastase (after
significantly improves inhibitory potency and metabolic stability in human blood (Ki= 0.3 nM, t1/2=45 min). Introduction of a methoxy group in the 6-position further enhanced blood stability (t1/2=260 minutes) probably due to the increased reactivity of the benzisothiazolone carbonyl by the electron-donating 6-methoxy group. The 2,6-dichlorobenzoate leaving group was optimum for potency and the compound retained stability in human blood. These inhibitors, however, have poor in vivo activity due to low hydrophilicity. Compounds such as WIN 64733 (8.31, Ki=0.014 nM) and WIN 63759 (8.32, Ki=0.013 nM), with aqueous solubilizing substituents show good pharmacokinetic properties and specificity for HNE. Replacement of mercaptotetrazole (see 8.30) with a diethylphosphate as leaving group significantly increases in vivo activity (8.33, Ki=0.035 nM).
Thrombin plays a central role within the coagulation cascade initiating not only fibrin clotting but exerting several cellular effects, too. The serine proteinase thrombin is a member of the trypsin family which attack peptide bonds following Arg or Lys residues. Therefore, inhibitors occupying the active site must possess or imitate the basic aminoor guanidinoalky side chain of Lys and Arg. As will be described later, in extensive biochemical and pharmacological studies thrombin inhibitors were shown effective as anticoagulants and antithrombotics. The main criteria for a low-molecular weight
thrombin inhibitor to be an ideal anticoagulant are high selectivity and systemic bioavailability after oral application.
Thrombin is not present in an active form in blood but is formed from prothrombin after activation of the coagulation cascade, whereas its substrates (fibrinogen, thrombin-activatable clotting factors) are permanently present. Consequently, inhibitors to be of therapeutic value must be present in the plasma at adequate concentrations to immediately neutralize the thrombin generated upon massive activation after vascular injury. It has been calculated that a pulse of thrombin is formed reaching a peak of about 200 nmol/l. However, immediately after initiation of the coagulation cascade the concentration of thrombin will be lowered by endogenous inhibitors. To be effective in anticoagulation the plasma concentration of a potent inhibitor should be at least 100 nmol/l.
X-ray crystal structures of complexes between thrombin and several inhibitors and substrate analogues have been solved providing the basis for rational drug design (Figures 8.4 and 8.5). Besides the primary specificity binding site to which the basic P1 amino acid of substrates is bound, there are two further important binding sites: the hydrophobic aryl-binding site and the anion-binding exosite, also called fibrinogen recognition site. The aryl-binding site located close to the active site is occupied by Phe at P9 of the fibrinopeptide A sequence, it is important in the binding of inhibitors of small size. The anion-binding exosite was discovered first from the crystal structure of the complex between thrombin and the naturally occurring thrombin inhibitor hirudin isolated from the medicinal leech Hirudo medicinalis. Four Arg and five Lys residues but also hydrophobic amino acids contribute to this positively charged region, involved in both the recognition of the substrates fibrinogen and thrombin receptor but also in the binding of thrombomodulin and some natural inhibitors.
Three main types of inhibitors have been developed which potently inhibit thrombin. These include peptide inhibitors based on natural substrates, arginine analogues, and benzamidine-derived compounds. Peptide chloromethyl ketones, aldehydes, esters or amides which possess the thrombin-sensitive Gly-Val-Arg sequence of the natural substrate fibrinogen and those resembling the Pro-Arg cleavage sites of factor XIII and prothrombin, are less effective inhibitors. However, extending of the Pro-Arg sequence with a D-Phe at P3 position gives effective inhibitors such as the chloromethylketone H-D-Phe-Pro-Arg-CH2Cl (PPACK, (8.34)), the aldehyde H-D-Phe-Pro-Arg-H (8.35) and the boronic acid derivative Ac-D-Phe-Pro-Arg-B(OH)2 (DuP 714, (8.36)). The D-Phe at P3 resembling Phe at P9 of the fibrinopeptide A sequence occupies the aryl binding site.
The chloromethylketone (PPACK, 8.34) is the most powerful and most selective irreversible inhibitor of thrombin known, with a second order rate constant three to five
Figure 8.4 View of the active site cleft of thrombin, displayed with its Connolly
Figure 8.4 View of the active site cleft of thrombin, displayed with its Connolly dot surface; molecular surfaces are colored dark blue, red, and light blue if donated by basic, acidic, or other residues, respectively. The bound PPACK molecule (8.34) is shown in red, with its Arg side chain disappearing into the primary specificity binding site. The prominent cleft, running from left to right, is where the substrate polypeptide chain would bind. The insertion loop which partially occludes the active site gives the thrombin molecule its selectivity. The aryl-binding site is located to the left close to the insertion loop; the anion binding exosite is to the right. (The Figure was courteously provided by Drs. W.Bode and M.Stubbs).
Figure 8.5 Active-site region of bovine NAPAP-thrombin (light green) superimposed with the experimentally determined inhibitors NAPAP (blue, (8.45)), argatroban (yellow, (8.41)), and PPACK (red, (8.34)). In contrast to the extended peptide-like PPACK molecule, the nonpeptidic inhibitors bind in compact, U-shaped
Figure 8.5 Active-site region of bovine NAPAP-thrombin (light green) superimposed with the experimentally determined inhibitors NAPAP (blue, (8.45)), argatroban (yellow, (8.41)), and PPACK (red, (8.34)). In contrast to the extended peptide-like PPACK molecule, the nonpeptidic inhibitors bind in compact, U-shaped conformation. (The Figure was courteously provided by Drs. W.Bode and M.Stubbs).
orders of magnitude higher than that for the inhibition of other trypsin-like proteases, such as factor Xa, plasmin, urokinase, plasma and glandular kallikrein. Binding of (8.34) in the thrombin-inhibitor complex is shown in Figure 8.4. After i.v. application, the D-Phe-Pro-Arg-derived inhibitors (8.34, 8.35, 8.36) exhibited anticoagulant effects in various animal experiments, however, oral bioavailability is low (<10 %). Low selectivity with respect to the fibrinolytic system would limit the therapeutic use of the aldehyde
(8.35) and the boronic acid derivative (8.36). Methylation of the terminal nitrogen of the parent compound gives Me-D-Phe-Pro-Arg-H (efegatran; (8.37)) with improved stability and selectivity. Exchange of boroArg at P1 by methoxypropylboroGly enhanced both oral bioavailability and selectivity of inhibition. The derivative Z-D-Phe-Pro-boro-Mpg-OPin (8.38), lacking the basic guanidino function of Arg which was thought essential for binding to the primary specificity site, is a potent thrombin inhibitor with nanomolar Ki.
Transformations of the Arg carboxyl in D-Phe-Pro-Arg peptides, such as introduction of nitrile, fluoroalkylketones, phosphonic acid or a-keto carbonyl residues do not improve selectivity and oral bioavailability. Remarkable improvement of the pharmacokinetic properties—both absorption and half life in circulation—was obtained with inogatran (8.39), containing agmatine at P1 and an N-terminal acetic acid residue.
Another type of inhibitor has been developed from synthetic Na-arylsulfonylated arginine ester-type substrates. Potent reversible inhibitors of thrombin are DAPA (8.40) and argatroban (OM 805, MQPA; (8.41)) with Ki values in the 20-40 nM range.
Argatroban is the only direct thrombin inhibitor used in clinical practice so far. The drug has been successfully used in man as an antithrombotic agent instead of heparin.
Argatroban is not sufficiently absorbed after oral administration, however, replacement of the guanidino group of Arg by amino-substituted heterocycles enhanced cell permeation. Benzamidine derivatives are effective competitive reversible inhibitors of thrombin. 4-Amidinophenylpyruvic acid (APPA; (8.42)) is an outstanding inhibitor with respect to oral bioavailability (up to 80%). Despite its low selectivity and affinity (Ki for thrombin 6.5 pM) anticoagulant and antithrombotic effects could be demonstrated in vivo.
Selective inhibitors of thrombin were found among derivatives of amino acids containing a benzamidine moiety at the side chain. The Na-tosylated piperidides of 3-amidinophenylalanine (3-TAPAP; (8.43)) and of 4-amidinophenylalanine (4-TAPAP; (8.44)) had potencies and binding properties similar to those of the arginine derivative
argatroban (8.41) (Figure 8.5). Na-Naphthylsulfonyl-glycyl-4-amidinophenylalanine piperidide (NAPAP; (8.45)), with a glycine spacer at the a-amino substituent, binds tightly to thrombin. It was the first synthetic thrombin inhibitor with nanomolar Ki. Both anticoagulant and antithrombotic effects correspond to its pronounced antithrombin
Was this article helpful?