Targeting PAI-1 in Cardiovascular Disease: Structural Insights Into PAI-1 Functionality and Inhibition

Abstract

Plasminogen activator inhibitor-1 (PAI-1), a member of the serine protease inhibitor (serpin) superfamily with antiprotease activity, is the main physiological inhibitor of tissue-type (tPA) and urokinase-type (uPA) plasminogen activators (PAs). Apart from being crucially involved in fibrinolysis and wound healing, PAI-1 plays a pivotal role in various acute and chronic pathophysiological processes, including cardiovascular disease, tissue fibrosis, cancer, and age-related diseases. In the prospect of treating the broad range of PAI-1-related pathologies, many efforts have been devoted to developing PAI-1 inhibitors. The use of these inhibitors, including low molecular weight molecules, peptides, antibodies, and antibody fragments, in various animal disease models has provided ample evidence of their beneficial effect in vivo and moved forward some of these inhibitors in clinical trials. However, none of these inhibitors is currently approved for therapeutic use in humans, mainly due to selectivity and toxicity issues. Furthermore, the conformational plasticity of PAI-1, which is unique among serpins, poses a real challenge in the identification and development of PAI-1 inhibitors. This review will provide an overview of the structural insights into PAI-1 functionality and modulation thereof and will highlight diverse approaches to inhibit PAI-1 activity.

Keywords: PAI-1 inhibitors; cardiovascular disease; fibrinolyisis; plasminogen activator inhibitor 1 (PAI-1); serpin (serine proteinase inhibitor).

Figure 1

Cartoon representation of the crystal structure of the active form of plasminogen activator inhibitor-1 (PAI-1) [PDB ID 6ZRV (41)] and the amino acid sequence of native PAI-1. PAI-1 shows the evolutionarily conserved topology of serpins, consisting of three β-sheets (A–C) and nine α-helices (hA-hI). β-sheet A, B, and C are shown in blue, magenta, and yellow, respectively, with numbers labeling the individual strands. The α-helices are indicated in the figure. The reactive center loop (RCL) of PAI-1 connects strand 5 of β-sheet A (s5A) to strand 4 of β-sheet B (s4B) and comprises strand 1 of β-sheet C (s1C). The RCL is shown in red, with the reactive center Arg346 (P1) and Met347 (P1′) represented by a magenta and cyan sphere, respectively. Other important domains that control and modulate PAI-1 conformational changes (the gate, hinge, breach, shutter, and flexible joint regions) are also indicated. Residues missing in the crystal structure are indicated by a dashed line. The amino acid sequence is presented and secondary structures (α-helices and β-strands) are indicated in the colors corresponding to the cartoon representation.

Figure 2

Schematic overview of the PAI-1 (I) conformations as well as its interactions with plasminogen activators (PAs, E) and cofactor vitronectin. Following the formation of a non-covalent PAI-1/PA Michaelis complex (EI), the P1-P1′ bond is cleaved to generate an acyl-enzyme intermediate (E~I). From here on, the reaction proceeds through a branched pathway, resulting in either the formation of an irreversible inhibitory complex (E-I) or the generation of cleaved PAI-1 (I*) due to the hydrolysis of the acyl-enzyme intermediate. PAI-1 is shown in white; the central β-sheet A of the PAI-1 molecule in blue; the flexible reactive center loop (RCL) in red, and Arg346 and Met347 (P1-P1′) of the reactive center are indicated by magenta and cyan spheres, respectively. The PA is shown in green. Vitronectin is shown in orange. PDB structures 1DVN (45), 1DB2 (46), 5BRR (47), 3EOX (48), 1EZX (49), 1H4W (50), and 1OC0 (51) were used to generate this figure. Figure adapted from Sillen et al. (52).

Figure 3

X-ray crystallographic structures of small molecule inhibitors bound to PAI-1. (A) Structure of latent PAI-1 in complex with AZ3976 (128). (B) Superimposition of the structures of active PAI-1 in complex with embelin (129) and CDE-096 (131). PAI-1 is colored white with the central β-sheet A colored blue and the RCL colored red. Secondary structure elements involved in binding to the compounds are indicated. AZ3976 is colored yellow, and embelin is colored green. CDE-096 in the structures obtained by cocrystallization or crystal soaking is colored orange or cyan, respectively. (C) AZ3976 bound to a deep pocket aligned by hD and s2A in latent PAI-1 [PDB ID 4AQH (128)]. (D) CDE-096 bound to active PAI-1 obtained by cocrystallization [PDB ID 4G8O (131)]. (E) Detail of the structure of CDE-096 bound to active PAI-1 obtained by crystal soaking [PDB ID 4G8R (131)]. (F) Detail of the structure of embelin bound a groove aligned by hD, hF, s2A, and the hE-s1A loop in active PAI-1 [PDB ID 3UT3 (129)].

Figure 4

Schematic representation of an antibody structure and different antibody fragment formats. (A) The left panel shows the cartoon and biological surface representation of the full-size crystal structure of Pembrolizumab, a human anti-PD1 immunoglobulin G4 (IgG4) antibody [PDB ID 5DK3 (226)]. A conventional antibody is a Y-shaped heterotetrameric glycoprotein consisting of two identical heavy (cyan) and two identical light chains (green). The heavy chain comprises one variable domain (V_H) and three constant domains (C_H1, C_H2, and C_H3), whereas the light chain comprises one variable (V_L) and one constant (C_L) domain. The panel on the right shows a simplified schematic representation of a conventional antibody and a heavy-chain-only antibody (HCAb). Each arm of the conventional antibody represents the antigen-binding fragment (Fab) that comprises the constant region (C_H1 and C_L domains) and the variable fragment (Fv) containing V_L and V_H. The stem of the antibody comprises two copies of the C_H2-C_H3 domains that form the crystallizable fragment (Fc). The HCAb comprises two heavy chains, each combining one variable V_HH domain, referred to as nanobody, and two constant domains (CH2 and CH3). (B) Schematic representation of a selection of antibody fragments, including Fab, diabody, single-chain variable fragment (scFv), and nanobody. Figure adapted from Rodrigo et al. (227).

Figure 5

Localization of different epitopes in the structure of active PAI-1. (A) Cartoon representation of the crystal structure of PAI-1 in complex with the Fab fragment of MEDI-579 (PDB ID 6I8S) (132). The heavy and the light chain of the Fab fragment are colored cyan and green, respectively. The constant and variable heavy (C_H1 and V_H) and light (C_L and V_L) domains are indicated in the figure. Residues closer than 4 Å to MEDI-579 are indicated as spheres. Residues located in the RCL are colored red, residues residing in the exosite regions for the 37-loop and 60-loop of PAs are colored yellow and orange, respectively. (B) Localization of different epitopes of monoclonal antibodies (mAbs) as determined by mutagenesis studies. The epitopes of mAbs that prevent the interaction between PAI-1 and PAs are indicated as red (MA-42A2F6, MA-56A7C10, and MA-44E4) (235) and yellow (MA-124K1) (236) spheres. The epitopes of switching mAbs that bind to hF or the hF-s3A loop (MA-33H1F7, MA-55F4C12, and Mab2) (237, 238) are indicated as magenta spheres. The epitope of switching mAb MA-8H9D4 (239) that binds to the hI-s5A loop is indicated as green spheres. The epitopes of latency-inducing antibody MA-33B8 (240, 241), MA-H4B3 (90), and MA-159M12 (242) are indicated as cyan, orange, and brown spheres, respectively. (C) Localization of different epitopes of nanobodies as determined by mutagenesis studies (35). The epitope of substrate-inducing nanobody Nb98 is indicated by green and magenta spheres, whereas only the magenta spheres indicate the epitope of Nb64. The epitope of Nb93 that interferes with PAI-1/PA complex formation is indicated as red spheres. (D) Cartoon representation of the superimposed crystal structures of PAI-1 in complex with Nb42 [PDB ID 6GWN, 6GWP, and 6GWQ (52)], Nb64 [PDB ID 6GWN and 6GWP (52)], and Nb93 [PDB ID 6ZRV (41)]. Nb42, Nb64, and Nb93 are colored yellow, green, and magenta, respectively. Residues closer than 4 Å to Nb42 are indicated as yellow spheres. Residues closer than 4 Å to Nb64 are indicated as green spheres. Residues closer than 4 Å to Nb93 are indicated as red and magenta spheres. Red spheres represent residues located in the RCL of PAI-1, whereas magenta spheres represent residues located in the exosite binding regions for PAs.