Activities of wild‐type and variant tissue‐type plasminogen activators retained on vascular endothelial cells

We reported that tissue‐type plasminogen activator (tPA) secreted from vascular endothelial cells (VECs) is retained on the cell surface and effectively evokes both plasminogen activation and fibrin clot dissolution (fibrinolysis) on VECs. Here, to evaluate possibly different behaviors of variants of tPA, we quantitatively assessed these two events separately using green fluorescent protein (GFP)‐conjugated tPA in cultured human VECs. The amount of secreted wild‐type tPA‐GFP correlated well with both the activities of plasminogen activation (r = 0.66) and fibrinolysis (r = −0.93). A variant of tPA‐GFP, with a lower affinity to the surface of VECs but higher affinity to fibrin, showed higher fibrinolysis and lower plasminogen activation activity compared to the wild‐type.

We reported that tissue-type plasminogen activator (tPA) secreted from vascular endothelial cells (VECs) is retained on the cell surface and effectively evokes both plasminogen activation and fibrin clot dissolution (fibrinolysis) on VECs. Here, to evaluate possibly different behaviors of variants of tPA, we quantitatively assessed these two events separately using green fluorescent protein (GFP)-conjugated tPA in cultured human VECs. The amount of secreted wild-type tPA-GFP correlated well with both the activities of plasminogen activation (r = 0.66) and fibrinolysis (r = À0.93). A variant of tPA-GFP, with a lower affinity to the surface of VECs but higher affinity to fibrin, showed higher fibrinolysis and lower plasminogen activation activity compared to the wild-type.
Tissue-type plasminogen activator (tPA), secreted from vascular endothelial cells (VECs) triggers intravascular fibrinolysis by activating plasminogen to plasmin [1]. Using green fluorescent protein (GFP)-conjugated tPA (tPA-GFP) and total internal reflection fluorescence (TIRF) microscopy, we studied the unique exocytotic dynamics of tPA-GFP and observed that tPA-GFP was retained on the VEC surface in a heavy-chaindependent manner after exocytosis [2]. Such a slow discharge of tPA was also reported in adrenal chromaffin cells [3,4], in which tPA showed lower diffusion mobility after the fusion of the granule membrane with the plasma membrane due to a slower fusion-pore expansion by tPA.
Unlike other members of the family of serine proteases, tPA is secreted as an active form [5], and thus its 'retention' on the solid surface is directly related to the facilitated plasmin generation [6], since its substrate of plasminogen also binds to the VEC surface. Our previous findings [6] showed that fluorescentlabeled plasminogen progressively accumulated on tPA-GFP-expressing VEC surface/pericellular areas in a manner that was tPA-GFP activity-dependent and plasmin activity-dependent. Our prior findings further demonstrated that the surface-retained tPA-GFP effectively initiated the dissolution of fibrin clots (fibrinolysis) that had formed over tPA-GFP-expressing cells.
The prerequisite function of the solid phase was demonstrated in both of the phenomena described in the previous paragraph as follows: First, these phenomena were strongly attenuated when a mutant tPA-GFP lacking a heavy chain, which is responsible for the retention on VECs as well as fibrin binding, was expressed. Second, the inhibition of plasminogen binding to either the cell surface or fibrin by a lysine analog completely abolished these phenomena. We also demonstrated that plasminogen activator inhibitor-1 (PAI-1), the primary inhibitor of tPA, formed a complex with retained tPA and facilitated its dissociation from secretory granule membranes to the fluid phase, which resulted in the reduction of the amount and activity of tPA on VECs [2]. The retention of tPA and its modification by PAI-1 thus appears to be important factors in the regulation of PA activity on the VEC surface.
In this study we aimed to quantify VEC-associated plasminogen activator activities by two distinct methods: a modified conventional chromogenic assay for the assessment of plasminogen activation, and a clot lysis assay for the assessment of fibrinolytic activities. A combination of these methods enabled us to define the distinct characteristics of both wild-type and variant tPA (tenecteplase: TNK [T103N, N117Q, K296A, H297A, R298A, R299A]-tPA), the latter of which is shown to be associated with better reperfusion and clinical improvement as a thrombolytic drug for patients with stroke compared to wild-type tPA (alteplase) [7] and with safe and feasible agents for minor stroke with intracranial occlusion [8].

Cell culture and plasmids
Vascular endothelial cells, that is, human umbilical vein endothelial cell-derived EA.hy926 cells-which were established to retain endothelial cell-specific functions including their fibrinolytic characteristics-were kindly provided by Dr C.J. Edgell of the University of North Carolina at Chapel Hill [9,10] and cultured as described [2]. Using tPA-GFP plasmid vector [2] as a template, we generated TNK-tPA-GFP and K-tPA (K296A, H297A, R298A, R299A)-GFP using the QuickChange II site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA), repeating the reaction several times to achieve the substitutions. The primer information is shown in the Table 1.

Analysis of tPA-GFP retention on cell surface
Cells cultured on 35-mm glass-bottom dishes (Asahi Techno Glass, Tokyo, Japan) were transiently transfected with tPAs-GFP by lipofection using a commercially available transfection reagent (TransIT-LT1, MIR2300; Mirus, Madison, WI, USA). After overnight incubation, the cells were washed with HEPES-buffered solution (HBS; 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2.5 mM CaCl 2 , 10 mM glucose, and 10 mM HEPES-NaOH, pH 7.3) and then set on the TIRF microscope stage. Details of the microscope system and method of image capture are explained below.
The retained ratio (RR) of tPA-GFP in individual secretory granules after secretion was determined by recording 3-min movies captured at 3.2 frames per s under the nonstimulated condition. We fit the region of interest (ROI) fitting to a single granule and calculated the averages of the fluorescence intensity in the ROI during the exocytosis procedure. The fluorescence intensity at 2 and 5 s after the opening of the secretory granules was divided by the maximum intensity and is shown as a percentage (RR-2s and RR-5s).

TIRF microscope system
The optical instrument for visualizing the exocytosis process of tPA-GFP secretory granules consisted of an Olympus inverted microscope (IX81; Olympus Corp., Tokyo, Japan, 609/1.45 numeric aperture oil-immersion objective) equipped with a TIRF unit for generating an evanescent field which enabled us to excite only the fluorophore present in the immediate vicinity of the plasma membrane; an automatic focus device (ZDC2); a digital complementary metal oxide semiconductor (CMOS) camera (ORCA-Flash4.0, C11440; Hamamatsu Photonics, Hamamatsu, Japan), and a temperature controller to keep the cells at 37°C on the microscope stage (INUG2-ONID-BE; Tokai Hit, Shizuoka, Japan). tPAs-GFP was excited by a 488nm laser with a 1.3 neutral density filter (Edmond Optics, Tokyo, Japan), and emissions were collected through a 520/35-nm band pass filter (Semrock, Rochester, NY, USA). HCIMAGE software (Hamamatsu Photonics) was used to capture the fluorescence images. The fluorescence intensity of an ROI in individual secretory granules was measured and analyzed on an Aquacosmos imaging station (Hamamatsu Photonics).

Measurement of clot lysis time, plasmin generation, and GFP concentration
Confluent cells in 96-well plates were used for transfection with plasmids. After the 24-h cultured supernatant was collected, the cells were washed with HBS and supplemented with 0.5 lM plasminogen (purified from fresh-frozen human plasma), 1 UÁmL thrombin (Benesis, Osaka, Japan), and 3 lM fibrinogen (Enzyme Research Laboratories, South Bend, IN, USA) to form fibrin clots (final volume 200 lL).
The clots were overlaid with liquid paraffin, and the absorbance at 405 nm (A405) of each well was measured every 20 min at 37°C to monitor the turbidity. The clot lysis time was determined as the time when A405 declined to one-half of its original value.
Plasmin activity on the cell surface was assessed by a chromogenic assay [11]. First, the 24-h cultured supernatant after plasmid transfection to confluent cells in 96-well plates was collected, and then the cells were washed with HBS and supplemented with 1 lM plasminogen and 0.4 mM S-2251 (Chromogenix, M€ olndal, Sweden) (final volume 100 lL). The reaction mixture was overlaid with liquid paraffin, and A405 was measured every 10 min at 37°C. The rate of plasmin generation was determined by the following calculation: A405 [m optical density] divided by the square of the time [h]. The values, normalized to estimate the rate of plasmin generation per 1 ng of GFP (which was calculated based on the GFP concentration in individual supernatants) in the variants are shown as the ratio compared to that in the wild-type.
The concentrations of tPA-GFP in the supernatants obtained after 24 h of incubation from the individual wells described above were determined using the GFP ELISA Kit (ab117992, Abcam, Cambridge UK).

Detection of tPAs-GFP in supernatant by western blotting of GFP
Culture medium was collected after 24 h of incubation from untransfected or transfected cells with tPAs-GFP expression vectors using a 6-well culture plate. Under the nonreduced condition, 15 lL of sample mixed with the sample buffer were separated by 7% SDS/PAGE. Protein bands transferred to a nitrocellulose membrane were incubated with anti-GFP antibody (Molecular Probes, Eugene, OR, USA), and horseradish peroxidase (HRP)-conjugated secondary antibodies. Signals were visualized by enhanced chemiluminescence (ECL) western blotting detector reagents (GE Healthcare, Little Chalfont, UK).

Accumulation of plasminogen on VECs
A trace amount of Alexa Fluor 568-labeled plasminogen (plg-568) with 0.5 lM nonlabeled plasminogen in HBS was supplemented to tPAs-GFP-expressing cells in 35-mm glass-bottom dishes and incubated at 37°C on the stage of a confocal laser scanning microscope (CLSM; FV1000, Olympus). The accumulation of plg-568 on and around the surface of individual tPAs-GFP-expressing cells was determined as the increase in values in plg-568 fluorescence intensity for 30 min (dF 30 min ; the ratio of fluorescence intensity at 50 min compared to that at 20 min) using the most basal focal images in which the cells were just adhering to the glass.

Statistical analysis
Correlation coefficients between the clot lysis time/the rate of plasmin generation and the amounts of GFP secretion were evaluated with Pearson's methods for normally distributed data. The significance of the difference in each parameter was evaluated with either the Williams multiple comparison test, Student's t-test (for normal distribution), or Mann-Whitney's U-test (for non-normal distribution) as needed.

Quantitative evaluation of PA activities
Plots of the GFP concentration values in the supernatant from tPA-GFP-expressing VECs versus the corresponding clot lysis time (CLT) showed a strong negative correlation ( Fig. 1A; open circles, r = À0.93, P < 0.001). The fibrin clots that formed on the cells not expressing tPA-GFP showed very little or no dissolution in the 40-h observation period after clot formation (n = 5), indicating that the activity of endogenous plasminogen activators, that is, tPA as well as urokinase-type plasminogen activator, were not enough to change the fibrin clot turbidity. This inhibition was mainly due to the endogenously expressed PAI-1, because much of the endogenous tPA existed as tPA-PAI-1 complex in the supernatant [2].
The real-time visualization of fibrin clot lysis using labeled plasminogen and fibrinogen showed that lysis started only where tPA-GFP-expressing cells existed [6], suggesting that tPA-GFP secreted from its expressing cells overcome the inhibitory effect of endogenous PAI-1 and expresses its plasminogen activator activity at a local area. Therefore, the clot lysis time strongly correlated with the amounts of secreted tPA-GFP.
Similarly, the plots of GFP versus the surface-associated plasmin generation showed a positive correlation (Fig. 1A, closed circles; r = 0.66, P = 0.0014). Plasmin generation in cells not expressing tPA-GFP after 1 lM plasminogen supplementation was detected but was negligible (1.48 AE 0.42 mODÁh À2 [mean AE SD, n = 5]) even though the cells expressed endogenous plasminogen activators. Plasminogen activation activity, represented as the rate of plasmin generation (RPG)/ ng.GFP, normalizing the secretion levels of GFP, was substrate concentration-sensitive and increased dosedependently with increasing concentrations of supplemented plasminogen (Fig. 1B). Although the dissociation rates of tPA-GFP from the membranous secretory granules after exocytosis were heterogeneous (shown as the retention rate in Fig. 2B, C; open circles), the GFP concentration in the supernatant (mostly tPA-GFP-PAI-1 complex, and a negligibly small amount in free form; Fig. 3, wild-type) seems to correlate well with the amounts and activities of surface-retained tPA-GFP.

Secretory characteristics of TNK-/K-tPA-GFP
We visualized the secretory dynamics of wild-type and TNK-/K-tPA-GFP expressed in EA.hy926 cells ( Fig. 2A) by using TIRF microscopy. The supplemental movies (Movie S1; wild-type tPA-GFP and S2; K-tPA-GFP) showed more rapid disperse of exocytosed K-tPA-GFP than that of wild-type tPA-GFP. We determined the retained ratio (RR) of tPAs-GFP on the membrane surface after the secretory granular opening, which reflected the residual amounts of tPAs-GFP on the membrane surface facing the extracellular fluid. The results obtained for TNK-tPA-GFP and K-tPA-GFP were significantly less than that for wild-type tPA (Fig. 2B, C), indicating that positively charged amino acid sequences, KHRR(296-299) residues are important for the retention of tPA-GFP. In light of our detection of the free forms of TNK-/K-tPA-GFP in 24-h cultured supernatant, which clearly contrasts with the detection of only the complex form with PAI-1 in wild-type tPA-GFP (Fig. 3), the replacement of KHRR to AAAA seems to have altered the charge-dependent interaction with the membrane and facilitated the diffusion of TNK-/K-tPA-GFP from the site of the opened granule into the liquid phase. Our trial to confirm the existence of free TNK-/K-tPA-GFP in the supernatant by western blotting using anti-tPA antibody, however, was not successful due to the fact that the signals derived from the endogenous tPA-PAI-1 complex and the expressed free form of tPAs-GFP were too close and difficult to be distinguished (Data not shown).
Interestingly, the accumulation of plg-568 at the cell-surface/pericellular area where cells were adhering to the glass was significantly decreased in K-tPA-GFP (decreased tendency in TNK-tPA-GFP; P = 0.06)expressing cells compared to that in the wild-type (Fig. 5A,B). This result coincided with the plasminogen activation activity represented as the values of normalized plasmin generation (Fig. 5C)   retained tPA was primarily responsible for the increase in plasmin generation assessed by this method.

Discussion
Vascular endothelial cells, as well as fibrin clots, provide a solid surface where fibrinolytic components can assemble and interact with each other for effective plasminogen activation [5,12]. We previously demonstrated that the heavy chain of tPA-GFP is essential for the cell-surface retention by showing that the heavy-chain-deleted mutant form of tPA-GFP is not retained on the membrane surface after exocytosis [2]. Since the retention of tPA-GFP on a VEC is essential for both cell surface-associated plasminogen activation and effective fibrin clot lysis [6], we attempted to quantify the expressed tPA-GFP-triggered plasminogen activator activities by employing two different methods to analyze fibrinolysis and plasminogen activation. Our present findings demonstrated that the amounts of secreted tPA-GFP strongly correlated with retained tPA-GFP-triggered plasminogen activator activities.
Since most of the tPA-GFP in the fluid phase exists as a complex form with PAI-1 in EA.hy926 cells, plasmin generation analysed by a chromogenic assay mainly reflects plasminogen activation initiated by VEC surface-retained tPA-GFP. In this facilitated plasminogen activation, augmented plasminogen bind-ing to VEC surface, shown as plg-568 accumulation in the previous study [6], is indeed deeply involved in the newly exposed C-terminal lysine residues by the generated plasmin-dependent degradation of proteins at the cell-surface/pericellular area.
According to observations of the secretory dynamics of tPA-GFP by TIRF microscopy, exocytosed tPA-GFP was retained as fluorescent spots on the membrane surface without any internalization in VECs [2], which is different from that in excitable cells [13,14]. In the present study, the retained fluorescent spots of tPA-GFP were gradually dimming, suggesting that tPA-GFP was gradually released. We thus suspected that this slow release of tPA-GFP consisted of two factors: one is a diffusion of tPA-GFP from the site of the opened granule into the liquid phase, and the other is a diffusion of tPA-GFP into the surrounding plasma membrane (lateral diffusion). Although the latter seems more important for the expression of its enzymatic activity, we have not identified the binding molecule of tPA-GFP on the cell surface, despite our efforts to assess the possible binding to reported tPA receptors [15] such as annexin A2 and galectin 1 (Y. Suzuki, unpubl. data). The fluorescence of tPA-GFP diffused into the plasma membrane-as well as that on the fibrin clot lytic edge where a large amount of labeled plasminogen accumulated-was less than the detection limit and was hardly visualized [6], but the amount was enough to express plasminogen activator activity. In any case, the GFP concentration in the supernatant, which correlates with the plasminogen activator activity triggered by secreted tPA-GFP, might be useful as a substitute for the active form of tPA-GFP. Further, measuring the GFP concentration instead of the tPA concentration to determine the amounts of tPA-GFP is more practical for assessing variant tPA-GFP concentrations, since anti-tPA antibody may react differently with different kinds of variant tPA.
Since PAI-1 dissociates retained tPA from the VEC surface by forming high molecular weight complexes [2], we expected TNK/K-tPA-GFP to be retained longer due to its impaired charge-dependent interaction with PAI-1 through its negatively charged residues 'EEIIMD (350-355)' [16]. Contrary to our expectations, TNK/K-tPA-GFP dissociated faster than wild-type tPA-GFP after its secretory granular opening. Our present findings demonstrated for the first time that tPA interacts with the plasma membrane surface in a charge-dependent manner. Further studies are needed to clarify which negatively charged factor is responsible for tPA interactions on the membrane surface. The reduced binding affinity of TNK/K-tPA-GFP toward the membrane developed interesting results in the present novel quantitative methods. In addition to their important characteristics of both higher resistance to PAI-1 and higher affinity to fibrin clot [17,18], faster release from VEC surface as a free tPA seems to have also contributed to the more effective fibrin clot lysis. These newly identified characteristics in the present study also contributed toward the lesseffective plasmin generation on the surface of VEC. Thus, the characteristics of tenecteplase having lower affinity to the plasma membrane compared to alteplase might be beneficial as a thrombolytic agent due to the following two reasons. First, higher effective drug concentration in the circulating blood could be expected after intravenous infusion. Second, tPA-mediated deleterious effects on neurovascular units in the ischemic brain [19], which were partially evoked by the overexpression of VEC surface-associated plasmin activity [20,21], could be reduced.
In conclusion, the combination of our novel quantitative methods demonstrated that fibrinolysis and plasmin generation on VECs are distinctly regulated, and the methods were shown to be useful for evaluations of the function of variant tPA molecules with different characteristics. The elucidation of such activity on the surface of VEC, which seems different from those in a purified system, may contribute to the development of a more beneficial tPA variant, including tenecteplase, in which KHRR residues are modified.