Separation of active and inactive forms of recombinant human plasminogen activator inhibitor type 1 (PAI-1) expressed in Chinese hamster ovary cells: comparison with native human PAI-1

Detection of Matrix Metalloproteinases (MMP), Tissue Inhibitor of Metalloproteinase-2, Urokinase and Plasminogen Activator Inhibitor-1 within Matrigel and Growth Factor-Reduced Matrigel Basement Membrane

Aims and background

Matrigel (MG) basement membrane is commonly used for in vitro studies of cellular migration, invasion and angiogenesis. It contains structural molecules such as collagen IV and laminin plus several growth factors which are diminished in growth factor-reduced Matrigel (GFR-MG). A less appreciated but important aspect of MG is the presence of matrix enzymes and their inhibitors. For relevant interpretation of data using MG/GFR-MG models, it may be necessary to know the enzymes or inhibitors contributed by these basement membranes themselves.

Methods

Immunoblot and zymography were used to detect the presence or absence of MMP-1 and 7, tissue inhibitor of metalloproteinase 1 and 2 (TIMP-1, TIMP-2), plasminogen activator activity and plasminogen activator inhibitor-1 (PAI-1). Growth and invasion assays using prostate cancer cells were used to assess the effects of TIMP-2 presence or absence.

Results

We detected MMP-7, urokinase plasminogen activator (uPA) and PAI-1 in both Matrigels, TIMP-2 was detected only in regular Matrigel and no MMP-1 or TIMP-1 was detected in either matrix. Invasion assays comparing regular MG and GFR-MG indicated cell line variability with regard to invasion efficiency as two tested prostate cancer lines were unaffected by the MG type while one was significantly more invasive in regular MG. Growth experiments suggest that the presence of TIMP-2 in regular MG may retard growth but overall proliferation is still greater in regular MG than in GFR-MG.

Conclusions

These data provide a useful reference for interpretation of in vitro Matrigel assays.

Characterization of a small molecule inhibitor of plasminogen activator inhibitor type 1 that accelerates the transition into the latent conformation

A novel class of small molecule inhibitors for plasminogen activator inhibitor type 1 (PAI-1), represented by AZ3976, was identified in a high throughput screening campaign. AZ3976 displayed an IC(50) value of 26 μm in an enzymatic chromogenic assay. In a plasma clot lysis assay, the compound was active with an IC(50) of 16 μm. Surprisingly, AZ3976 did not bind to active PAI-1 but bound to latent PAI-1 with a K(D) of 0.29 μm at 35 °C and a binding stoichiometry of 0.94, as measured by isothermal calorimetry. Reversible binding was confirmed by surface plasmon resonance direct binding experiments. The x-ray structure of AZ3976 in complex with latent PAI-1 was determined at 2.4 Å resolution. The inhibitor was bound in the flexible joint region with the entrance to the cavity located between α-helix D and β-strand 2A. A set of surface plasmon resonance experiments revealed that AZ3976 inhibited PAI-1 by enhancing the latency transition of active PAI-1. Because AZ3976 only had measurable affinity for latent PAI-1, we propose that its mechanism of inhibition is based on binding to a small fraction in equilibrium with active PAI-1, a latent-like prelatent form, from which latent PAI-1 is then generated more rapidly. This mode of action, with induced accelerated latency transition of active PAI-1 may, together with supporting x-ray data, provide improved opportunities for small molecule drug design in the hunt for therapeutically useful PAI-1 inhibitors.

Platelets retain high levels of active plasminogen activator inhibitor 1

The vascular fibrinolytic system is crucial for spontaneous lysis of blood clots. Plasminogen activator inhibitor 1 (PAI-1), the principal inhibitor of the key fibrinolytic enzyme tissue-type plasminogen activator (tPA), is present in platelets at high concentrations. However, the majority of PAI-1 stored in platelets has been considered to be inactive. Our recent finding (Brogren H, et al. Blood 2004) that PAI-1 de novo synthesized in platelets remained active for over 24 h, suggested that PAI-1 stored in the α-granules might be active to a larger extent than previously reported. To re-evaluate this issue, we performed experiments where the fraction of active PAI-1 was estimated by analyzing the tPA-PAI-1 complex formation. In these experiments platelets were lysed with Triton X-100 in the presence of serial dilutions of tPA and subsequently the tPA-PAI-1 complex was evaluated by Western blot. Also, using a non-immunologic assay, tPA was labeled with (125)I, and (125)I-tPA and (125)I-tPA-PAI-1 was quantified by scintigraphy. Interestingly, both methods demonstrated that the majority (>50%) of platelet PAI-1 is active. Further analyses suggested that pre-analytical procedures used in previous studies (sonication or freezing/thawing) may have substantially reduced the activity of platelet PAI-1, which has lead to an underestimation of the proportion of active PAI-1. Our in vitro results are more compatible with the role of PAI-1 in clot stabilization as demonstrated in physiological and pathophysiological studies.

Heterogeneous glycosylation patterns of human PAI-1 may reveal its cellular origin

The main inhibitor of intravascular fibrinolysis is plasminogen activator inhibitor 1 (PAI-1) which binds to and irreversibly inhibits tissue plasminogen activator (tPA). PAI-1 is present in blood, both in platelets and in plasma, and PAI-1 levels are associated with risk of atherothrombosis. Several tissues express PAI-1 but the source of plasma PAI-1 is not known. We recently found that platelets can de novo synthesize PAI-1 and the amount synthesized in vitro in 24 hours is 35-fold higher than required to maintain normal plasma levels. Recombinant human PAI-1 expressed in different cell types or secreted naturally by human cell lines, exhibit heterogeneous glycosylation patterns. The aim of this study was to investigate the hypothesis that platelets might be the source of plasma PAI-1 and that the cellular source of PAI-1 can be determined by its tissue-specific glycosylation pattern. PAI-1 was isolated from platelets, macrophages, endothelial cells, adipose tissue, as well as plasma from lean and obese subjects. The glycosylation was analyzed by nanoLC-MS/MS. PAI-1 isolated from cell lysates and conditioned media from macrophages, endothelial cells, and adipose tissue expressed heterogeneous glycosylation patterns. By contrast, no glycans were detected on PAI-1 isolated from plasma or platelets from healthy lean individuals. Hence, our data suggest that platelets may be the main source of plasma PAI-1 in lean individuals. Interestingly, plasma PAI-1 from obese subjects had a glycan composition similar to that of adipose tissue suggesting that obese subjects with elevated PAI-1 levels may have a major contribution from other tissues.

Cytochalasin D can improve heterologous protein productivity in adherent Chinese hamster ovary cells

We generated a series of adherent gene-amplified CHO clones expressing human secreted alkaline phosphatase (SEAP) as a model for heterologous protein production. Clones demonstrate a 26- to 52-fold increase in productivity compared to controls after dhfr/methotrexate-mediated gene amplification and clone selection. SEAP is stably expressed in these clones over at least a 6-week period without significant productivity loss. Two-dimensional protein electrophoresis identified 21 proteins that exhibited altered expression in clones of increasing SEAP productivity. Based on MALDI TOF/TOF mass spectrometry of relevant protein spots, changes in translation, energy pathways, chaperones, regulatory proteins, and cytoskeletal proteins were observed, including a 4-fold expression increase in actin capping protein. We hypothesized that an alteration of the actin cytoskeleton using cytochalasin D as a mimic for actin-capping protein could have a beneficial effect on heterologous protein secretion. Treatment with 0.5 mug/mL cytochalasin D increased SEAP productivity 2- to 3-fold compared to an amplified control which resulted in an increase in productivity from 52- to 150-fold compared to a nonamplified parent.

Protein movement during complex-formation between tissue plasminogen activator and plasminogen activator inhibitor-1

Plasminogen activator inhibitor-1 (PAI-1) rapidly inactivates tissue plasminogen activator (tPA). After initial binding and cleavage of the reactive-centre loop of PAI-1, this complex is believed to undergo a major rearrangement. Using surface plasmon resonance and SDS-PAGE, we have studied the influence of a panel of monoclonal antibodies on the reaction leading to the final covalent complex. On the basis of these data, we suggest the mechanisms for the action of different classes of inhibitory antibodies. We propose that the antibodies which convert PAI-1 into a substrate for tPA do this by means of preventing the conversion of the initial PAI-1/tPA complex into the final complex by sterical intervention. Moreover, the localisation of the binding epitopes on free PAI-1, as well as on the PAI-1/tPA complex, suggests that tPA in the final complex cannot be located near helices E and F, as has previously been proposed.

Interfering with the inhibitory mechanism of serpins: crystal structure of a complex formed between cleaved plasminogen activator inhibitor type 1 and a reactive-centre loop peptide

BACKGROUND

Plasminogen activator inhibitor type 1 (PAI-1) is an important endogenous regulator of the fibrinolytic system. Reduction of PAI-1 activity has been shown to enhance dissolution of blood clots. Like other serpins, PAI-1 binds covalently to a target serine protease, thereby irreversibly inactivating the enzyme. During this process the exposed reactive-centre loop of PAI-1 is believed to undergo a conformational change becoming inserted into beta sheet A of the serpin. Incubation with peptides from the reactive-centre loop transform serpins into a substrate for their target protease. It has been hypothesised that these peptides bind to beta sheet A, thereby hindering the conformational rearrangement leading to loop insertion and formation of the stable serpin-protease complex.

RESULTS

We report here the 1.95 A X-ray crystal structure of a complex of a glycosylated mutant of PAI-1, PAI-1-ala335Glu, with two molecules of the inhibitory reactive-centre loop peptide N-Ac-TVASS-NH2. Both bound peptide molecules are located between beta strands 3A and 5A of the serpin. The binding kinetics of the peptide inhibitor to immobilised PAI-1-Ala335Glu, as monitored by surface plasmon resonance, is consistent with there being two different binding sites.

CONCLUSIONS

This is the first reported crystal structure of a complex formed between a serpin and a serpin inhibitor. The localisation of the inhibitory peptide in the complex strongly supports the theory that molecules binding in the space between beta strands 3A and 5A of a serpin are able to prevent insertion of the reactive-centre loop into beta sheet A, thereby abolishing the ability of the serpin to irreversibly inactivate its target enzyme. The characterisation of the two binding sites for the peptide inhibitor provides a solid foundation for computer-aided design of novel, low molecular weight PAI-1 inhibitors.

Epitopes on plasminogen activator inhibitor type-1 important for binding to tissue plasminogen activator

The molecular details of the rapid complex formation between tissue plasminogen activator (tPA, E.C. 3.4.21.68) and plasminogen activator inhibitor type-1 (PAI-1) are still not fully elucidated. We have used surface plasmon resonance (SPR), the BIAcore, to characterize the binding of a large panel of monoclonal antibodies to four forms of recombinant human PAI-1, including active and latent PAI-1 as well as the complex between PAI-1 and recombinant human tc tPA or the protease part of tPA, the B-chain. Antibodies that discriminate between these different forms of PAI-1 have been identified, which is reflected by differences in k(a), k(d) as well as in Kd. In addition, in a chromogenic assay with PAI-1 and tPA we determined the IC50-values for these antibodies, i.e., studied their ability to inhibit the decrease in tPA-activity caused by PAI-1. In a competition assay using SPR, we have also been able to study whether concurrent binding of these antibodies to PAI-1 was possible. We could thereby assign the antibodies to five groups according to their binding areas. Furthermore, by using this technique, we have for the first time been able to identify three distinct epitopes on PAI-1, which are all of importance for the interaction and complex-formation with tPA. Since the antibodies that bind to one of these areas all have very poor affinity for the complex between PAI-1 and tPA, we suggest that this not previously described epitope must be located near the final binding site for tPA in this complex. Altogether, this also supports the theory of a multistep reaction between PAI-1 and tPA, in which tPA interacts with different parts of the PAI-1-molecule.

Plasminogen activator inhibitor-1 represses integrin- and vitronectin-mediated cell migration independently of its function as an inhibitor of plasminogen activation

Cell migration involves the integrins, their extracellular matrix ligands, and pericellular proteolytic enzyme systems. We have studied the role of plasminogen activator inhibitor-1 (PAI-1) in cell migration, using human amnion WISH cells and human epidermoid carcinoma HEp-2 cells in an assay measuring migration from microcarrier beads and a modified Boyden-chamber assay. Active, but not latent or reactive center-cleaved, PAI-1 inhibited migration. A PAI-1 mutant without ability to inhibit plasminogen activation was as active as wild-type PAI-1 as a migration inhibitor, showing that inhibition of plasminogen activation was not involved. PAI-1 specifically interfered with intergrin- and vitronectin-mediated migration: Migration onto vitronectin-coated but not onto fibronectin-coated surfaces was inhibited by PAI-1, a cyclic RGD peptide inhibited migration, and both cell lines expressed vitronectin-binding alpha v-integrins. In addition, active PAI-1, but not latent or reactive center-cleaved PAI-1, inhibited vitronectin binding to integrins in an in vitro binding assay, without affecting binding of fibronectin. Monoclonal antibodies against the urokinase receptor, another vitronectin binding protein, did not affect cell migration in the beads assay, while some inhibitory effect was observed in the Boyden-chamber assay. We conclude that PAI-1, independently of its role as a proteinase inhibitor, inhibits cell migration by competing for vitronectin binding to integrins, while the interference of PAI-1 with binding of vitronectin to the urokinase receptor may play a secondary role. These data define a novel function for the serpin PAI-1, enabling it to regulate cell migration over vitronectin-rich extracellular matrix in the body.

Characterisation of the complex of plasminogen activator inhibitor type 1 with tissue-type plasminogen activator by mass spectrometry and size-exclusion chromatography

Glycosylated human plasminogen activator inhibitor type 1 (PAI-1), produced in Chinese hamster ovary (CHO) cells, showed a variety of compounds with different molecular weights when subjected to electrospray mass spectrometry (ES-MS), owing to the heterogeneity of the carbohydrate chains. However, non-glycosylated human PAI-1, produced in E. coli, gave rise to a prominent species with a molecular weight of 42,774, consistent with the amino-acid sequence. A non-glycosylated mutant of the proteinase domain (B-chain) of tissue-type plasminogen activator (tPA) produced in C 127 cells, had a molecular weight of 28,168. Full-length, glycosylated, tPA showed a large heterogeneity in molecular mass. For a mass study, a tPA-PAI-1 complex was formed, composed of non-glycosylated PAI-1 and non-glycosylated B-chain. This complex was remarkably stable at room temperature in buffer with a neutral pH. The mass spectrum of the complex provided two main species, a peptide with a mass of 3803 and a dominating species of 67,133. These masses are consistent with a complex where PAI-1 is cleaved at the P1-P1' position. A trace of a species with a molecular mass of 70,942 was also found, corresponding to the complete, non-dissociated complex with PAI-1. Separation of the cleaved peptide, corresponding to the hydrophobic C-terminal 33 amino-acid residues of PAI-1, from the complex, was achieved by size-exclusion chromatography in the presence of 30% acetonitrile. Thus, in the complex between tPA and PAI-1, the proteins are held together by a tight covalent bond, but the C-terminal cleaved peptide of PAI-1 is only bound to the complex by hydrophobic forces. To assess whether this is specific to the tPA B-chain alone, experiments with the complex of full-length, glycosylated tPA and glycosylated PAI-1 were also performed, and it was possible to demonstrate the release of the C-terminal PAI-1 peptide by chromatography, mass spectrometry, as well as by SDS-PAGE.

The inhibition mechanism of serpins. Evidence that the mobile reactive center loop is cleaved in the native protease-inhibitor complex

Inhibitors that belong to the serine protease inhibitor or serpin family have reactive centers that constitute a mobile loop with P1-P1' residues acting as a bait for cognate protease. Current hypotheses are conflicting as to whether the native serpin-protease complex is a tetrahedral intermediate with an intact inhibitor or an acyl-enzyme complex with a cleaved inhibitor P1-P1' peptide bond. Here we show that the P1' residue of the plasminogen activator inhibitor type 1 mutant (P1' Cys) became more accessible to radiolabeling in complex with urokinase-type plasminogen activator (uPA) compared with its complex with catalytically inactive anhydro-uPA, indicating that complex formation with cognate protease leads to a conformational change whereby the P1' residue becomes more accessible. Analysis of chemically blocked NH2 termini of serpin-protease complexes revealed that the P1-P1' peptide bonds of three different serpins are cleaved in the native complex with their cognate protease. Complex formation and reactive center cleavage were found to be rapid and coordinated events suggesting that cleavage of the reactive center loop and the subsequent loop insertion induce the conformational changes required to lock the serpin-protease complex.

Plasminogen activator inhibitor type-1 interacts exclusively with the proteinase domain of tissue plasminogen activator

Two different techniques have been used to study the complex formation of recombinant human plasminogen activator inhibitor type-1, PAI-1, with either recombinant human two-chain tissue plasminogen activator, tc tPA (EC 3.4.21.68), or the tPA deletion variants tc K2P, containing the kringle 2 domain and the proteinase domain, and P, containing only the proteinase domain. The same value for Kon, 2.10(7) M-1s-1 for binding of PAI-1 was found for the three tPA forms by direct detection of the complex formation in real time by surface plasmon resonance, BIAcore, or indirectly by monitoring the time course of the inhibition of tPA using the chromogenic substrate N-methylsulfonyl-D-Phe-Gly-Arg-4-pNA-acetate. Apparently, no conformational change is involved in the rate-limiting step, since the kon value was found to be independent of the temperature from 20 to 35 degrees C. By the BIAcore technique, it was found that the complex between PAI-1 and tPA covalently coupled to the surface, was stable at 25 degrees C, since no dissociation was seen in buffer. However, extended treatment with 1 M NH4OH destroyed the complex with t 1/2 = 5 h. The same kon values and complex composition were found by measuring either the binding of tPA to PAI-1 captured on the monoclonal antibody MAI-11 or the binding of PAI-1 to tPA captured on the monoclonal antibody 2:2 B10. Quantification of the complex composition between PAI-1 captured on the monoclonal antibody MAI-11 with either tPA, K2P or P gave a one-to-one ratio with the fraction of active PAI-1, consistent with the results from SDS-PAGE and the specific activity of PAI-1. The complexes of the three tPA forms with PAI-1 captured on a large surface of MAI-11 dissociated more rapidly from MAI-11, with the same apparent koff, kdis, = 2.10(-3) s-1, compared with 0.7-10(-3) s-1 for the dissociation of PAI-1 alone. In consistance, the Kd, calculated from the direct determination of the kon and koff for the association of different form of PAI-1 to a small surface of MAI-11, was found to be higher for PAI-1 in complex with tPA than for free active PAI-1. Apparently, upon complex formation, a change is induced in PAI-1 at the binding epitope for MAI-11.(ABSTRACT TRUNCATED AT 400 WORDS)