Mini-reviewSmall molecules inhibitors of plasminogen activator inhibitor-1 – An overview
Graphical abstract
Introduction
Plasminogen activator inhibitor type 1 (PAI-1), a glycoprotein of approximately 50 kDa belongs to serine protease inhibitors (serpins). Serpins represent about 10% of human plasma proteins and most of them share the same highly ordered structure which comprises three β-sheets, nine α-helices and a reactive center loop (RCL). The reactive site (P1–P1′) is found in the RCL and mimics the normal substrate of the serine protease. Among serpins, PAI-1 has unique properties as the RCL can insert spontaneously into β-sheet A without requiring any protease [1]. In vivo, PAI-1 is the main inhibitor of tissue-type plasminogen activator (tPA) and urokinase plasminogen activator (uPA).
PAI-1 is secreted as an active protein (Fig. 1(A)) which can form a covalent complex with its target serine proteases uPA and tPA that recognize the RCL of PAI-1 as a substrate and cleave it. Then, PAI-1 undergoes a conformational change called “stressed to relaxed (S to R) transition”. During this transition, the cleaved RCL is inserted into β-sheet A and the target protease is trans-located to the opposite side of PAI-1, irreversibly inhibiting the protease itself. This spontaneous conformational change is required for its inhibitory activity. The active form is not stable and under normal physiological conditions spontaneously converts to a latent form (Fig. 1(B)) which is unable to interact with the target proteases (uPA or tPA). This reactive form has an apparent half-life of 1–2 h at 37 °C and neutral pH. Inactive latent form can be reactivated in vitro by denaturing agents or negatively charged phospholipids [2], [3]. A third non-inhibitory conformation of PAI-1 called cleaved form (Fig. 1(C)) has also been reported. In this case, PAI-1 is cleaved by tPA or uPA without forming a covalent complex [4]. The self-inactivating mechanism of PAI-1 is important in the regulation of its activity.
PAI-1 is synthesized by endothelial cells, platelets, and other mesenchymal cells that surround vasculature. The circulating active free form is relatively unstable. However in the blood circulation and in the extracellular matrix, PAI-1 is stabilized by vitronectin (VN). This abundant glycoprotein plays key roles in cell migration, cell invasion, cell proliferation and tissue remodeling. VN binds to active PAI-1 with a Kd in the subnanomolar range, while affinity to the latent form is at least 200-fold lower [6]. Site-directed mutagenesis [7] and crystallographic data [8] show that PAI-1 binds to the N-terminal somatomedin B domain of VN. However, PAI-1 may contain a second simultaneous vitronectin binding site with a lower affinity [9], [10]. A first study reported by Lawrence et al. showed that VN binds PAI-1 in the flexible joint region, implicating residues Q125 and L118 [10], while, later on, De Prada et al. suggested that residues 116–120 were also involved [11].
PAI-1 plays a critical role in the intra and extracellular fibrinolysis regulation. Binding to plasminogen activators tPA and/or uPA. PAI-1 blocks the activation of plasminogen to plasmin, and therefore the fibrin clot hydrolysis [12]. PAI-1 is also involved in many other physiological processes. It acts on extracellular matrix remodeling by indirect interaction with matrix metalloproteases (MMP). MMP, activated by plasmin, degrade most of the extracellular matrix components. As tPA inhibitor, PAI-1 inhibits the plasmin formation and consequently its activity against the MMP [13], [14], [15]. PAI-1 acts also on cell migration by preventing integrin αvβ3 from binding to vitronectin or by uPA inhibition [16], [17], [18]. Besides, the role of PAI-1 in tumor growth and angiogenesis is not yet fully understood and remains controversial [19], [20]. At physiological concentration, PAI-1 could promote angiogenesis whereas at higher doses, PAI-1 would be antiangiogenic [21], [22].
The PAI-1 activity is tightly regulated at the transcriptional level with TGF-β (transforming growth factor β) as the major regulator [23], [24]. Due to its involvement in various physiological processes, PAI-1 is also implicated in many pathologies such as vascular thrombosis [25], [26], neointimal hyperplasia [27], [28], atherosclerosis [29], [30], tumor progression [20], metastasis [20], Alzheimer disease [31], obesity [32], [33], diabetes [32], [34], [35] and fibrosis [36], [37]. Indeed, it has been shown that high circulating PAI-1 level is associated with the previous disorders. Therefore, many interests have been focused on PAI-1 regulation as a potential therapeutic strategy [38], illustrated by the number of published researches [39].
This review focuses on synthetic organic compounds, natural products or small peptides used as PAI-1 inhibitors. Studies on monoclonal antibodies [40] or RNA aptamers [41] will be excluded. Several possible mechanisms for inactivating PAI-1 are plausible [40]: (a) sterically blocking the initial formation of the Michaelis complex between PAI-1 and its target proteases (b) preventing the conformational change of PAI-1 associated with the complex formation, resulting in the cleavage of PAI-1 as a substrate or (c) conversion of PAI-1 into an inactive latent or otherwise inert form. The structural complexity of PAI-1 has made the identification and development of inhibitors very challenging [42]. For more details on PAI-1 biochemistry see Ref. [43].
Section snippets
Peptides or pseudo-peptide inhibitors
PAI-1 is unique among the serpin superfamily because of its structural flexibility. It is secreted in its active form which is not particularly stable. Under normal physiological conditions, it spontaneously converts to a latent state. This irreversible conversion (in vivo) is due to the spontaneous insertion of its reactive center loop into β-sheet A. Therefore, as it has been demonstrated for others serpins, a peptide mimicking a portion of the RCL could be able to inhibit PAI-1. Based on
Diketopiperazines and analogs
The first low molecular weight molecules targeting PAI-1 were described in the mid-90s. In 1996, research scientists from Xenovia Limited [52] isolated two molecules bearing a diketopiperazine moiety (XR334 (8) and XR330 (9), Fig. 6) from the lyophilized biomass of an unknown Streptomyces sp. These compounds inhibited PAI-1/tPA interactions in vitro (IC50 = 51.7 μM and 30 μM respectively in chromogenic assays) and exhibited a good ex vivo activity on fibrinolysis. When administered in rats at
Flufenamic acid derivative
In 1998, Bjöquist and coll. [60] from Astra Hässle had developed from the known fibrinolytic activator flufenamic acid (17), the low molecular weight inhibitor AR-HO29953XX (18) (Fig. 7).
AR-HO29953XX (18) inhibits PAI-1 by preventing complex formation with tPA. It was found that the acid function is essential for inhibitory activity. In a chromogenic assay with tPA, the IC50 value for this compound was 12 μM. Construction of several PAI-1 mutants suggests that this inhibitor interacts with
Benzofuran derivative
In 2003, Crandall et al. [62], [63] from Wyeth Research described a new PAI-1 inhibitor, the benzofuran derivative WAY-140312 (19) (Fig. 8). Immunofunctional and chromogenic assays with tPA or uPA showed that it inhibits PAI-1 activity with an IC50 of approximately 10 μM. It was the first molecule to exhibit efficacy in rat models of vascular diseases following oral administration. A 10 mg/kg oral dose improved time to occlusion, reduced thrombus weight and maintained artery blood flow with no
Menthol based derivatives
Some of the most potent PAI-1 inhibitors described to date are small compounds with a menthol scaffold. HTS of a compound library allowed Ye et al. [80] to identify compound 25 (Fig. 10) as a hit (IC50 = 1.4 μM in chromogenic assay using uPA). SAR study aimed at optimizing segments A and B showed that the first one could be replaced by various aromatic (but not aliphatic) diamines without loss of potency and that segment B could be simplified by diphenyl ether carboxylic acids moieties. This
Piperazines and hybrides
In 2004, Ye et al. [82] identified a novel piperazine-based scaffold as a PAI-1 inhibitor using HTS. Compound 27 (Fig. 11) was identified as hit displaying moderate potency (IC50 = 1.8 μM in chromogenic assay) but poor oral bioavailability. Systematic SAR studies were conducted to optimize consecutively the A, B and C segments of 27. The synthesis and biological valuation of about thirty derivatives showed that no improvement was observed by modifying segment A, while aminomethylphosphonic acid
Oxadiazolidinediones
In 2004, Gopalsamy et al. [84] from Wyeth Research identified a new series of PAI-1 inhibitors containing an oxadiazolidinedione moiety as a result of HTS. After identification of the lead compound 32 (Scheme 17) which inhibited PAI-1 with an IC50 of 5.29 μM, the authors undertook a SAR study by modifying the benzylether substituents on the central phenyl core and the nature of the linker between the phenyl and oxadiazolidinedione moieties. From this study it appears that a highly lipophilic
Oxalamide derivatives
In 2009, other HTS and SAR studies conducted by Jain et al. [85] allowed the identification of oxalamide derivatives as PAI-1 inhibitors. Few compounds such as derivatives 34, 35, 36 and 37 (Scheme 19), containing a carboxylic acid function, trifluoromethyl groups on phenyl ring and a sulfonamide function as a spacer showed significant inhibitory activity. IC50 evaluation in chromogenic assays with tPA gave values of 4.5 μM, 5.4 μM, 5.0 μM and 8.4 μM respectively.
Their synthesis, presented in
Pyrrolin-2-one derivatives
In 2009, Miyazaki et al. [86] from Mitsubishi Tanabe Pharma synthesized and evaluated a new series of pyrrolin-2-one derivatives from which T-1776Na (38) (Scheme 20) emerged as a soluble selective potent PAI-1 inhibitor (IC50 = 7.9 μM in a chromogenic assay with tPA). It also presented high antithrombotic activity in rat models when used in intravenous infusions. A 43.4% decrease in thrombus weight was observed when administrated at 0.5 mg/kg/min during electrical stimulation. T-1776Na (38) was
Polyphenolic derivatives
Polyphenolic compounds, mostly from natural origins, constitute another potentially appealing family. In 1999, following a large screening of extracts of plant and marine organisms, Astra Hässle AB company identified racemic mixtures of phloroglucinol dimers sideroxylonal A–C (41–43) (Fig. 13), isolated from extracts of the of Eucalyptus albens Benth. (Myrtaceae) flowers, as PAI-1 inhibitors in the presence of tPA [88]. These compounds presented IC50 values of 3.3 μM, 5.3 μM and 4.7 μM
Azetidine derivative AZ3976 (56)
AZ3976 (56) (Fig. 19) is a small PAI-1 inhibitor discovered in 2012 through HTS which exhibits an IC50 value of 26 μM in enzymatic chromogenic assays (tPA/PAI-1), an IC50 value of 16 μM in a plasma clot lysis assay and a Kd value of 0.29 μM for the latent PAI-1.
This molecule has a unique way of irreversibly inhibiting binding to the latent form of PAI-1. The X-ray structure (Fig. 20) shows that AZ3976 (56) binds to a deep pocket between helices D and F and strand 1 of β-sheet A. This area is
Embelin (57)
Recently (2013–4) a third crystallized structure of PAI-1 in complex with Embelin, a small molecule antagonist, has been published (PDB 3UT3) [98]. Fruit of the Embelia ribes Burm. plant (Myrsinaceae) has been used to treat fever, inflammatory diseases, and a variety of gastrointestinal ailments for thousands of years [99]. In 1940, the active component from this plant was isolated, named Embelin (57) [100] and chemically synthesized in 1948 [101] by causing the dilauroyl peroxide to decompose
Conclusion
In addition to its role in the fibrinolytic system, PAI-1 is involved in a large number of cardiovascular diseases, mostly because it inhibits tPA and uPA. It is also involved in cell migration and tumor development by inhibiting uPA and interacting with vitronectin [20]. Several studies using transgenic mice lacking [103] or over expressing PAI-1 [104] have demonstrated the utility of inhibiting PAI-1 in vivo. For these reasons, PAI-1 has been intensively studied and significant efforts have
Acknowledgments
The authors acknowledge the financial support from IdRS, CNRS and University Paul Sabatier.
References (104)
- et al.
Endothelial cells produce a latent inhibitor of plasminogen activators that can be activated by denaturants
J. Biol. Chem.
(1985) - et al.
Identification of a conformationally distinct form of plasminogen activator inhibitor-1, acting as a non-inhibitory substrate for tissue-type plasminogen activator
J. Biol. Chem.
(1992) - et al.
High quality structure of cleaved PAI-1-stab
J. Struct. Biol.
(2009) - et al.
Characterization of the binding of different conformational forms of plasminogen activator inhibitor-1 to vitronectin. Implications for the regulation of pericellular proteolysis
J. Biol. Chem.
(1997) - et al.
Evidence that type 1 plasminogen activator inhibitor binds to the somatomedin B domain of vitronectin
J. Biol. Chem.
(1991) - et al.
A deletion mutant of vitronectin lacking the somatomedin B domain exhibits residual plasminogen activator inhibitor-1-binding activity
J. Biol. Chem.
(2008) - et al.
Localization of vitronectin binding domain in plasminogen activator inhibitor-1
J. Biol. Chem.
(1994) - et al.
Structure of plasminogen activator inhibitor 1 (PAI-1) and its function in fibrinolysis: an update
Fibrinolysis
(1995) - et al.
Development and disease in proteinase-deficient mice: role of the plasminogen, matrix metalloproteinase and coagulation system
Thromb. Res.
(1998) - et al.
PAI-1 promotes extracellular matrix deposition in the airways of a murine asthma model
Biochem. Biophys. Res. Commun.
(2002)