Pharmacological Targeting of Plasminogen Activator Inhibitor-1 Decreases Vascular Smooth Muscle Cell Migration and Neointima Formation

Abstract

Objective

Plasminogen activator inhibitor-1 (PAI-1), a serine protease inhibitor that promotes and inhibits cell migration, plays a complex and important role in adverse vascular remodeling. Little is known about the effects of pharmacological PAI-1 inhibitors, an emerging drug class, on migration of vascular smooth muscle cells (SMCs) and endothelial cells (ECs), crucial mediators of vascular remodeling. We investigated the effects of PAI-039 (tiplaxtinin), a specific PAI-1 inhibitor, on SMC and EC migration in vitro and vascular remodeling in vivo.

Approach and Results

PAI-039 inhibited SMC migration through collagen gels, including those supplemented with vitronectin and other extracellular matrix proteins, but did not inhibit migration of PAI-1-deficient SMCs, suggesting that its anti-migratory effects were PAI-1-specific and physiologically relevant. However, PAI-039 did not inhibit EC migration. PAI-039 inhibited phosphorylation and nuclear translocation of STAT-1 in SMCs, but had no discernable effect on STAT-1 signaling in ECs. Expression of LDL receptor-related protein 1 (LRP1), a motogenic PAI-1 receptor that activates JAK/STAT-1 signaling, was markedly lower in ECs than in SMCs. Notably, PAI-039 significantly inhibited intimal hyperplasia and inflammation in murine models of adverse vascular remodeling, but did not adversely affect re-endothelialization after endothelium-denuding mechanical vascular injury.

Conclusions

PAI-039 inhibits SMC migration and intimal hyperplasia, while having no inhibitory effect on ECs, which appears to be due to differences in PAI-1-dependent LRP1/JAK/STAT-1 signaling between SMCs and ECs. These findings suggest that PAI-1 may be an important therapeutic target in obstructive vascular diseases characterized by neointimal hyperplasia.

Plasminogen activator inhibitor-1 (PAI-1) is a serine protease inhibitor expressed by vascular endothelial cells, vascular smooth muscle cells (SMCs), and several other cell types.¹ PAI-1 is present in plasma and the extracellular matrix (ECM) of blood vessels, where it rapidly inhibits tissue-type plasminogen activator (t-PA) and urinary-type plasminogen activator (u-PA), which down-regulates plasmin formation and fibrinolysis. Complete deficiency of PAI-1 produces a bleeding disorder,² while elevated plasma levels of PAI-1 are associated with thrombotic disorders.^3–5 Although its primary function appears to be the regulation of fibrinolysis, PAI-1 also regulates cell migration, which plays a critical role in vascular remodeling, including pathological neointimal hyperplasia.⁶ PAI-1 decreases cell migration by inhibiting pericellular plasmin formation and binding to vitronectin (VN) in the ECM, thereby blocking VN’s binding interactions with α_Vβ₃ integrin and the u-PA receptor (uPAR).^7–10 Conversely, PAI-1 exerts a pro-migratory effect by binding to low-density lipoprotein receptor-related protein 1 (LRP1) and activating a Janus kinase/signal transducers and activators of transcription (JAK/STAT) signaling pathway.¹¹ Consistent with its opposing effects on cell migration in vitro, PAI-1 has been shown to either inhibit^12–14 or promote^15–19 neointima formation in vivo. While PAI-1 affects vascular remodeling via multiple pathways, including SMC proliferation and apoptosis,^20–22 ECM synthesis and clearance,²³ and inflammatory signaling pathways,^24–26 its effects on SMC migration are of major importance in vascular remodeling.⁶

Several small-molecular-weight inhibitors of PAI-1 have been developed.²⁷ These compounds inhibit thrombosis, macrophage migration, and vascular senescence.^28–30 However, little is known about their effects on SMC and endothelial cell (EC) migration in vitro and vascular wall remodeling in vivo. Given the potential of PAI-1 to either promote or inhibit SMC migration, and considering the central role of SMC migration in pathological vascular remodeling, including arterial restenosis and vein graft intimal hyperplasia, it is critical to determine the effects of pharmacological PAI-1 inhibition on SMC migration. It is equally important to determine the effects of pharmacological PAI-1 inhibition on endothelial cell migration, as an intact, functional endothelium is essential to maintain vascular health. Therefore, the goals of this study were to examine the effects of PAI-039 (tiplaxtinin), a highly specific and well characterized inhibitor of PAI-1,²⁸ on SMC and EC migration under physiologically relevant conditions in vitro and arterial and venous remodeling in vivo.

Materials and Methods

Materials and Methods are available in the online-only Data Supplement.

Results

PAI-039 inhibits SMC migration

We studied the migration of SMCs through 3-dimensional collagen matrices in the presence or absence or PAI-039. PAI-039 significantly inhibited migration of wild-type murine SMCs when present in concentrations ≥17.5 µM (Fig. 1A, C). PAI-039 also inhibited the migration of human coronary artery SMCs through collagen (Supplementary Fig. IA in online data supplement). However, PAI-039 did not inhibit migration of SMCs isolated from pai1−/− mice (Fig. 1B, C), suggesting that the anti-migratory effect of PAI-039 was due specifically to PAI-1 inhibition. PAI-039 had no significant effect on proliferation of wild-type SMCs (Supplementary Fig. IB in online data supplement), suggesting that the reduced migration of PAI-039-treated cells was not mediated by an anti-mitogenic effect. Addition of a mixture of ECM proteins (MaxGel ECM) or purified VN to collagen gels increased SMC migration (Supplementary Fig. II in online data supplement), but did not blunt the anti-migratory effect of PAI-039 (Fig. 1 D–E). These results suggest that PAI-039 inhibits SMC migration under physiological conditions, including in the presence of VN, which binds PAI-1 and reduces its susceptibility to inhibition by PAI-039.³¹

Figure 1.

PAI-039 inhibits SMC migration. (A) PAI-039 inhibits migration of wild-type (WT) murine venous SMCs through collagen. (B) PAI-039 does not inhibit migration of PAI-1-deficienct SMCs. (C) Representative images of SMCs that migrated through collagen gel to lower-chamber side of transwell membranes. (D) PAI-039 inhibits SMC migration through more complex collagen gels containing a mixture of extracellular matrix proteins (MaxGel ECM). (E) PAI-039 inhibits SMC migration through collagen gels supplemented with purified vitronectin (10 µg/mL). Data represent mean of 3 independent experiments. *P<0.05 vs. control group lacking PAI-039.

PAI-039 does not inhibit EC migration

We next studied the effect of PAI-039 on EC migration in vitro, utilizing a 2-dimensional scratch assay. Contrary to its inhibitory effect on SMCs, PAI-039 had no significant effect on human aortic EC (AEC) migration (Supplementary Fig. III in online data supplement).

PAI-039 inhibits STAT-1 signaling in SMCs

PAI-1 induces cell migration via an LRP1-dependent pathway involving STAT-1 nuclear translocation.¹¹ However, the effects of PAI-1 and PAI-1 inhibition on STAT-1 phosphorylation have not been investigated previously. Therefore, human coronary artery SMCs were incubated in the presence or absence of recombinant wild-type PAI-1 (PAI-1-WT) and STAT-1 phosphorylation was analyzed by Western blotting. PAI-1-WT significantly increased STAT-1 phosphorylation at Tyr⁷⁰¹ (Fig. 2A). We also studied the effect of PAI-1 inhibition on STAT-1 phosphorylation by incubating human coronary artery SMCs in the presence or absence of PAI-039. PAI-039 significantly decreased the level of constitutive STAT-1 phosphorylation in SMCs (Fig. 2B). To explore effects of PAI-039 on STAT-1 intracellular trafficking we treated murine arterial SMCs and AECs with PAI-1-WT and analyzed STAT-1 nuclear translocation by fluorescence immuno-histochemistry/confocal scanning laser microscopy. PAI-1-WT induced nuclear translocation of STAT-1 in SMCs, but not in AECs (Fig. 3). PAI-039 blocked the capacity of PAI-1-WT to induce STAT-1 nuclear translocation in SMCs, but had no discernable effect on STAT-1 signaling in AECs. LRP1 gene and protein expression were markedly lower in AECs than in SMCs (Fig. IV in online Data Supplement), suggesting that the resistance of ECs to the anti-migratory effect of PAI-039 may be explained by a very low level of expression of LRP1. PAI-039 blocked PAI-1-WT-induced STAT-1 nuclear translocation in wild-type mouse embryonic fibroblasts (MEFs), but there was no discernable effect of PAI-1-WT or PAI-039 on STAT-1 nuclear translocation in LRP1-deficient MEFs (Fig. 3). Furthermore, recombinant PAI-1-AK, a recombinant PAI-1 mutant that promotes cell migration by binding to LRP1, but not VN,³² significantly increased migration of wild-type MEFs, but not LRP1-deficient MEFs (Supplementary Fig. V in online Data Supplement). Together, these results suggest that the differential effects of pharmacological PAI-1 inhibition on SMC vs. EC migration are mediated by a distinct difference in the expression of LRP1 and downstream JAK/STAT-1 signaling between these two cell types.

Figure 2.

PAI-1 and PAI-039 regulate STAT-1 phosphorylation in SMCs. (A) Recombinant PAI-1 stimulates STAT-1 phosphorylation. Human coronary artery SMCs were incubated with vehicle control or PAI-1-WT (0.1 µg/mL) for 20 min, after which cells were lysed and STAT-1 phosphorylation was assessed by Western blotting and quantified by band intensity analysis. Data shown are from 12 experiments. *P<0.05 vs. control. Representative images from an individual experiment shown are. (B) PAI-039 inhibits STAT-1 phosphorylation. Human coronary artery SMCs were incubated with PAI-039 (10 µM) or vehicle control for 24 hrs, after which cell lysates were prepared and analyzed by Western blotting. Quantitative analysis of STAT-1 phosphorylation is shown (n=4 experiments/group; *P<0.01 vs. control), along with representative images from an individual experiment.

Figure 3.

Effects of PAI-1 and PAI-039 on STAT-1 trafficking, assessed by fluorescence immuno-histochemistry. PAI-1-WT (0.1 µg/mL) induces translocation of STAT-1 (green) from cell cytoplasm to nuclei (blue) in murine arterial SMCs and wild-type mouse embryonic fibroblasts (arrows), which is inhibited by PAI-039 (25 µM). In contrast, PAI-1 does not induce nuclear translocation of STAT-1 in murine endothelial cells or LRP1-deficient mouse embryonic fibroblasts. Scale bar equals 14 µm. All images shown are representative of data from at least 2 independent experiments.

PAI-039 inhibits intimal hyperplasia

To examine the effects of PAI-039 in vivo, we subjected mice to carotid artery ligation, which induces formation of a SMC-rich neointima proximal to the site of vascular occlusion.³³ We used pluronic gel to deliver PAI-039 (n=5) or vehicle control (n=5) to the surface of ligated carotid arteries. In addition, mice received PAI-039 or vehicle control by intraperitoneal (IP) injection for 4 weeks after surgery. PAI-039 significantly inhibited neointima formation and percent lumen stenosis, but not vessel total cross-sectional area (Fig. 4). Immunohistochemical analysis revealed that there was negligible macrophage (Mac-3-positive cell) invasion into the neointima of either experimental group at this time point (data not shown), consistent with a prior study.³³ We also studied the effects of pharmacological PAI-1 inhibition on the intimal hyperplasia that develops in segments of inferior vena cava (IVC) after they are grafted into the transected carotid arteries of recipient mice.³⁴ RT-PCR analysis revealed that PAI-1 gene expression was increased greater than 2-fold in carotid artery vein grafts harvested 4 weeks after surgery as compared to naïve IVC segments (Supplementary Fig. VI in online Data Supplement). Similar to our arterial study, PAI-039 significantly inhibited vein graft intimal hyperplasia and lumen stenosis, while having no significant effect on total vessel cross-sectional area (Fig. 5). In addition, PAI-039 significantly decreased macrophage invasion into vein grafts, as well as SMC content within vein graft neointima (Supplementary Fig. VII in online Data Supplement).

Figure 4.

PAI-039 inhibits intimal hyperplasia in murine carotid arteries. Wild-type mice were subjected to carotid ligation surgery and treated with PAI-039 (n=5) or vehicle control (n=5). Four weeks after surgery arterial (A) intima/media ratio, (B) % lumen stenosis, and (C) total area were measured. *P<0.01 vs. vehicle control; ^†P<0.05 vs. vehicle control; N.S. = difference between groups was not statistically significant (P>0.05). (D) Representative images of arteries from control and PAI-039-treated mice. Scale bar = 100 µm.

Figure 5.

PAI-039 inhibits intimal hyperplasia in murine vein grafts. Wild-type mice were subjected to vein graft surgery and treated with PAI-039 (n=8) or vehicle control (n=8), as described in Methods. Four weeks after surgery vein graft (A) intimal thickness, (B) % lumen stenosis, and (C) total area were measured. *P<0.03 vs. vehicle control; N.S. = difference between groups was not statistically significant (P>0.05). (D) Representative images of vein grafts from control and PAI-039-treated mice. Scale bar = 50 µm. L = lumen. Arrow denotes internal elastic lamina.

PAI-039 does not inhibit endothelial repopulation in vivo

To determine if the lack of an inhibitory effect of PAI-039 on EC migration in vitro was also observed in vivo, we denuded the endothelium in mouse carotid arteries by mechanical injury and used the Evans Blue dye method to quantify the effects of PAI-039 on re-endothelialization, which is mediated by EC migration from the adjacent segments of normal artery into the denuded zone.³⁵ PAI-039 had no significant inhibitory effect on EC migration in vivo (Fig. 6). In contrast, paciltaxel, which inhibits intimal hyperplasia,³⁶ significantly inhibited EC recovery.

Figure 6.

PAI-039 does not inhibit re-endothelialization after mechanical carotid artery injury. Murine carotid arteries were injured and treated with PAI-039, vehicle control, or paclitaxel. At 1, 3, and 7 days after injury, absence of an intact endothelium was determined with Evans blue dye, which stains denuded endothelium. (A) Representative images of carotid arteries. (B) Quantitative analysis of denuded endothelium. The number of mice in each treatment group at each time point is indicated within each bar (a total of 36 mice were used in the experiment). Statistical analyses are shown (one-way analysis of variance with pairwise multiple comparison procedures).

Discussion

SMC migration from the tunica media to the intima results in neointima formation and is a key process in atherosclerosis, restenosis after coronary angioplasty/stent implantation, and vein graft disease. PAI-1 plays an important role in regulating SMC migration, with both pro- and anti-migratory effects being reported.^{8, 11} In this study we sought to determine the effects of pharmacological inhibition of PAI-1 on SMC migration under conditions that are relevant to human cardiovascular disease. We showed that pharmacological inhibition of PAI-1 with PAI-039 attenuates SMC, but not EC, migration in vitro. Our data suggest that PAI-039 inhibits SMC migration by down-regulating LRP1-dependent STAT-1 activation. Furthermore, PAI-039 inhibits arterial and venous intimal hyperplasia and vascular inflammation without adversely affecting re-endothelialization after vascular injury.

VN is present in the vascular wall and its expression is increased during neointima formation.^37–39 VN binds PAI-1 and regulates its function.³² Previous studies have shown that PAI-039 effectively inhibits free, but not VN-bound, PAI-1.^{28, 31, 40} However, we have shown in this study that PAI-039 retains its anti-migratory effects on SMCs in vitro in collagen matrices supplemented with VN, and inhibits neointima formation in vivo. In a dynamic system involving living cells or in vivo it is likely that microenvironments exist in which free (i.e. PAI-039-susceptible) PAI-1 is present, due to either molar excess over VN or transient dissociation from VN, as occurs in an equilibrium binding interaction. The capacity of PAI-039 to selectively inhibit free PAI-1, while “bypassing” active, VN-bound PAI-1, is intriguing and could be advantageous in the context of adverse vascular remodeling, as VN-bound PAI-1 inhibits cell migration by blocking VN binding interactions with its cellular receptors.^{8, 9} VN-bound PAI-1 is also inhibited from binding to LRP1, a motogenic PAI-1 receptor.⁴¹ Therefore, PAI-039 does not inhibit the anti-migratory function of PAI-1, but does inhibit its pro-migratory effect by preventing its functional interaction with LRP1.

JAK/STAT-1 signaling plays an important role in regulating SMC activation and intimal hyperplasia.^42–44 PAI-1 promotes cell migration by binding LRP1 and inducing nuclear translocation of STAT-1.¹¹ In this study we have shown that PAI-1 induces, while PAI-039 decreases, STAT-1 phosphorylation in SMCs. We also have demonstrated that PAI-039 inhibits PAI-1-induced STAT-1 nuclear translocation in a LRP1-dependent manner. Together with our in vitro cell migration and in vivo vascular remodeling data, our results support the hypothesis that PAI-039 exerts important inhibitory effects on LRP1/JAK/STAT-1 activation in SMCs that down-regulate intimal hyperplasia. Our results, however, do not exclude the possibility that PAI-039 also targets other molecular pathways to inhibit neointima formation.

EC migration plays an essential role in restoring an intact vascular endothelium after vascular injury. In contrast to its inhibitory effect on SMC migration, PAI-039 did not significantly affect EC migration. This differential effect could be clinically significant – i.e. while it is desirable to inhibit SMC migration to reduce intimal hyperplasia and vascular stenosis, concomitant inhibition of EC recovery, as occurs with currently employed anti-restenosis drugs, promotes thrombosis and other adverse vascular effects. Our data suggest that the non-responsiveness of ECs to the anti-migratory effect of PAI-039 may be due to their very low expression level of LRP1, a finding that is consistent with published reports.^{11, 45}

In a previous study we demonstrated that mice with complete PAI-1 deficiency exhibit increased vein graft intimal hyperplasia compared to WT controls.⁴⁶ We hypothesized that the enhanced neointima formation in the absence of PAI-1 resulted from 1) enhanced availability of ECM VN to its cellular receptors (due to loss of binding of PAI-1 to VN), which would favor cell migration, and 2) enhanced activity of proteases that are susceptible to inhibition by PAI-1, such as plasmin and thrombin, which promote intimal hyperplasia.^{7, 46} Thus, it would seem paradoxical that pharmacological inhibition of PAI-1 would decrease intimal hyperplasia in the same model. Several factors may explain our results. It is known that PAI-1 has motogenic and anti-motogenic effects,^{8, 9, 11} both of which would be affected by genetic PAI-1 deficiency. However, it is likely that PAI-039 has a significantly greater inhibitory effect on the motogenic pool of free PAI-1 than on the anti-motogenic pool of VN-bound PAI-1.³¹ Such a differential effect of PAI-039 could potentially produce a different net effect on intimal hyperplasia than complete genetic PAI-1 deficiency. Consistent with this hypothesis, we showed previously that PAI-1-deficient SMCs exhibit enhanced migration compared to WT SMCs,³² yet, as we demonstrate in the current study, PAI-039 inhibits SMC migration. Other factors may also contribute to the differential effects of genetic- vs. pharmacologically-induced reductions in PAI-1 activity on vascular remodeling. First, it is possible that genetic deficiency of PAI-1 produces compensatory changes in the expression of other proteins that affect vein graft remodeling. Second, PAI-1 regulates enzyme activity not only in the extracellular space, but also within the intracellular environment.²⁰ While genetic deficiency would affect both compartments of PAI-1, it is not known if PAI-039 inhibits intracellular PAI-1. Lastly, it is possible that complete genetic deficiency of PAI-1 might produce different phenotypic effects than partial or near-complete inhibition of PAI-1, which would be expected from a pharmacological inhibitor. Of note, genetic deficiency of soluble epoxide hydrolase produced an opposite effect on cardiac fibrosis than that produced by a pharmacological inhibitor.⁴⁷ Thus, there is a precedent for genetic deficiency and pharmacological inhibition of a factor producing opposing phenotypes.

Recent studies have suggested that pharmacological PAI-1 inhibition has considerable potential as a therapeutic strategy to prevent and treat vascular diseases by pathways that are independent of fibrinolysis. TM5441, a small molecule PAI-1 inhibitor, inhibits vascular senescence and other organ dysfunction associated with aging.^{30, 48} TM5275, another PAI-1 inhibitor, inhibits macrophage migration.²⁹ Consistent with this study, we found that PAI-039 decreased macrophage invasion into vein grafts. In contrast to our findings, Leik et al showed that PAI-039 promoted SMC migration in vitro.⁴⁰ In this study purified PAI-1 was pre-incubated with PAI-039, then added to SMCs migrating on a 2-dimensional, VN-coated surface. These experimental conditions, while informative regarding the potential of PAI-039 to inhibit PAI-1, probably did not adequately reproduce the physiological conditions under which SMCs migrate in vivo. A recent study showed that PAI-039 induces apoptosis of SMCs, thereby producing an anti-proliferative effect,⁴⁹, which is consistent with the known anti-apoptotic properties of PAI-1.^{20, 22} In addition, another recent study demonstrated that PAI-039 improved dermal wound closure in diabetic mice,⁵⁰ which is intriguing because vascular remodeling has been considered as a form of wound repair.⁵¹

In conclusion, we have shown that pharmacological inhibition of PAI-1 with PAI-039 inhibits SMC migration and intimal hyperplasia without adversely affecting EC migration or vascular re-endothelialization. These differential effects appear to be due to the marked difference in LRP1 expression between SMCs and ECs, with the low level of LRP1 expression in ECs shielding them from the effects of PAI-1 and PAI-039 on LRP1-induced JAK/STAT-1 signaling and cell migration.¹¹ Our study suggests that pharmacological targeting of PAI-1 with PAI-039, and perhaps other PAI-1 inhibitors, may be an effective strategy to treat and prevent vascular disease without disrupting EC function, which is a major limitation of currently employed anti-restenosis drugs. Additional studies that assess the effects of pharmacological PAI-1 inhibitors on adverse vascular remodeling in the setting of diabetes mellitus and obesity, which are associated with increased PAI-1 expression,^{52, 53} and preexisting atherosclerosis are warranted. Such studies could lead to clinical trials to determine if pharmacological PAI-1 inhibition exerts beneficial effects on human atherosclerosis, restenosis, vein graft disease, and other obstructive vascular diseases characterized by increased PAI-1 expression.^{54, 55}

Highlights.

• PAI-039, a small molecule, specific inhibitor of plasminogen activator inhibitor-1 (PAI-1), decreases vascular SMC migration.

• PAI-039 does not inhibit endothelial cell (EC) migration.

• The differential effect of PAI-039 on SMCs vs. ECs appears to be mediated by a marked difference in the expression level of LDL receptor-related protein-1 (LRP1) between cell types, and consequently, in PAI-1-dependent activation of the LRP1/JAK/STAT-1 signaling pathway.

• In mice, PAI-039 inhibits intimal hyperplasia, but not re-endothelialization after vascular injury, suggesting that PAI-1 may be an important drug target for preventing and treating obstructive vascular diseases.

Nonstandard Abbreviations and Acronyms

AEC

aortic endothelial cell

ECM

extracellular matrix

JAK

Janus kinase

LRP1

low density lipoprotein receptor-related protein 1

PAI-1

plasminogen activator inhibitor-1

SMC

smooth muscle cell

STAT

signal transducers and activators of transcription

VN

vitronectin