Cell activation by stressors is characterized by a sequence of detectable phenotypic cell changes. A given stimulus, depending on its strength, induces modifications in the activity of membrane phospholipid transporters and calpains, which lead to phosphatidylserine exposure, membrane blebbing and the release of microparticles (nanoscale membrane vesicles). This vesiculation could be considered as a warning signal that may be followed, if the stimulus is maintained, by cell detachment-induced apoptosis. In the present study, plasminogen incubated with adherent cells is converted into plasmin by constitutively expressed tPA (tissue-type plasminogen activator) or uPA (urokinase-type plasminogen activator). Plasmin formed on the cell membrane then induces a unique response characterized by membrane blebbing and vesiculation. Hitherto unknown for plasmin, these membrane changes are similar to those induced by thrombin on platelets. If plasmin formation persists, matrix proteins are then degraded, cells lose their attachments and enter the apoptotic process, characterized by DNA fragmentation and specific ultrastructural features. Since other proteolytic or inflammatory stimuli may evoke similar responses in different types of adherent cells, the proposed experimental procedure can be used to distinguish activated adherent cells from cells entering the apoptotic process. Such a distinction is crucial for evaluating the effects of mediators, inhibitors and potential therapeutic agents.
Cell response to a number of stressors and inflammatory mediators is key to maintenance of tissue homoeostasis. The initial response of activated cells may evolve to apoptosis depending on the type and strength of stimuli . Early manifestations of this cell-activation process are the formation of membrane blebs and the shedding of nanoscale membrane fragments known as MPs (microparticles) (0.1–1 μm) . Since the first discovery of MPs in platelet-free plasma [3,4], the most well-known cellular MPs are those of platelet, leucocyte, erythrocyte and endothelial cell origin found in circulating blood . A number of studies have demonstrated that stimulation of these cells is followed by the characteristic features of cell activation: increased levels of cytoplasmic calcium associated with exposure of phosphatidylserine and activation of calpains (EC 3.4.22) . The increase in intracellular calcium induces a disordered state in the concerted activity of membrane transporter proteins (flippases, floppases and scramblases) that maintain the membrane phospholipid asymmetry of quiescent cells [7,8]. As a result, pro-coagulant phosphatidylserine is translocated from the inner leaflet to the external leaflet of the membrane. The activated calpains cleave cytoskeleton filaments and thereby facilitate membrane blebbing and shedding of MPs. In addition to their pro-coagulant activity, MPs carry at their surface identity antigens that characterize their cellular origin. Since they also convey proteolytic glycoproteins, growth factor or inflammatory mediators, they are currently considered to be a storage pool of bioactive effectors . Proteolytic mediators that have been identified include the plasminogen-activation system and the matrix metalloproteases (EC 3.4.24) [10–13].
The plasminogen-activation system is involved in fibrinolysis and in pericellular proteolysis . Finely regulated cellular plasminogen activation plays an essential role in ECM (extracellular matrix) remodelling , cell migration  and survival . The central mechanism is the transformation of plasminogen into plasmin (EC 184.108.40.206), the active enzyme, at cell membranes, the ECM or the surface of fibrin, by the uPA (urokinase-type plasminogen activator) (EC 220.127.116.11) or the tPA (tissue-type plasminogen activator) (EC 18.104.22.168). MPs function as a catalytic surface for plasmin formation by providing both plasminogen activators and plasminogen-binding sites . Interestingly, the uPA system of MPs also participates in the formation of plasmin at the surface of platelets, fibrin or the ECM via a fibrinolytic/proteolytic cross-talk mechanism described recently . Thrombin (EC 22.214.171.124), another serine protease, is known to induce the release of MPs from platelets, but the potential of plasmin to induce vesiculation is not known as yet.
Beyond activation-dependent release of MPs, the survival of cells within structural–functional units involving tissue-specific components and the microvasculature (e.g. neurovascular unit, glomerulus and pulmonary alveolus), depends on dynamic cell–ECM interactions that ensure their adhesion to the substratum and tissue cohesion. Thus, in the absence of any interaction with the ECM, human endothelial cells rapidly enter apoptosis . Accordingly, excessive proteolysis of the ECM by cells that express a plasminogen-activator system results in loss of cell anchorage and apoptosis , a phenomenon that may be of relevance in pathological situations in vivo . Discerning the steps of this cell-activation and apoptosis process is therefore crucial for evaluating the effects of mediators, inhibitors and potential therapeutic agents.
We unveil in the present paper our findings that plasmin provokes the release of cell-derived MPs and propose a sequence of events initiated by plasminogen activation on adherent cells and spanning from early blebbing and vesiculation to subsequent cell detachment and apoptosis/survival. Because other proteolytic or inflammatory stimuli may evoke similar responses in different types of adherent cells, the proposed mechanistic procedure can be used to distinguish activated adherent cells from cells entering the apoptotic process.
Reagents and proteins
Human Glu-plasminogen was purified and characterized as described previously  and was over 99% pure as assessed by SDS/PAGE and by N-terminal sequence analysis. Plasmin was prepared by activation of Glu-plasminogen with immobilized uPA using the method of Wiman and Wallen  with modifications. Rabbit anti-(mouse laminin) polyclonal antibody was kindly provided by H.P. Erickson (Duke University Medical Center, Durham, NC, U.S.A.). The HRP (horseradish peroxidase)-labelled monoclonal antibody directed against plasminogen kringle 1 (CPL15-PO) was prepared as described in . Active-site blocked plasmin (D-Val-Phe-Lys-chloromethyl ketone dihydrochloride, Pn-VFK) was prepared as described previously . The plasminogen activators uPA (140000 IU/mg) and tPA (578000 IU/mg) both over 99% single-chain form were obtained from Biopool AB and Abbot respectively. Aprotinin was a gift from Bayer HealthCare AG. Glutamine, fetal bovine serum and DMEM (Dulbecco's modified Eagle's medium)/Ham's F12 were obtained from Invitrogen. The chromogenic substrate selective for plasmin (methylmalonyl)-hydroxyprolylarginine-p-nitroaniline (CBS0065) was from Stago. The plasmin inhibitor VFK was from Calbiochem. The lysine analogue ϵ-ACA (ϵ-aminocaproic acid), the tetrazolium salt MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide], DMSO, CpB (carboxypeptidase B) (porcine pancreas) and DAPI (4′,6-diamidino-2-phenylindole) were from Sigma. The TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling) reagents were from Roche Applied Science.
Identification of cell plasminogen activators
CHO (Chinese-hamster ovary)-K1 cells (A.T.C.C. CCL-61, fibroblast phenotype), HEK (human embryonic kidney)-293 cells  and HMEC-1s (human microvascular endothelial cells-1)  were used as prototype adherent cells. Cells were grown in a 37 °C humidified atmosphere of 5% CO2 using DMEM/Ham's F12 (CHO-K1 cells) or RPMI 1640 medium (HEK-293 cells and HMEC-1s) supplemented with 2 mM L-glutamine and 10% fetal bovine serum. Cells were then seeded in 96- or 24-well plates, depending on experiments. Fibrin autography following SDS/PAGE was performed as described previously . Briefly, cells were lysed in 100 mM Tris/HCl buffer (pH 8.1), containing 1% Triton X-100. Proteins in cell lysates (100 μg) and reference proteins [10 μl of tPA (1 pmol), uPA (0.2 pmol) and plasmin (100 pmol)] were subjected to 8% PAGE under non-reducing conditions. SDS was then exchanged with 2.5% Triton X-100. After washing-off excess Triton X-100 with distilled water, the gel was carefully overlaid on a 1% agarose gel containing 1 mg/ml bovine fibrinogen, 100 nM plasminogen and 0.2 NIH unit/ml of bovine thrombin. Zymograms were allowed to develop at 37 °C for 12 h and photographed at regular intervals using dark-ground illumination. Active proteins in cell lysates were identified by reference to the migration of known markers (uPA, tPA and plasmin). To verify the activator identity, cell lysates were incubated with a polyclonal antibody directed against tPA or uPA and their residual activities were detected by radial diffusion on fibrin–agarose gels.
Cell activation by plasmin
Cells were incubated with Glu-plasminogen (1 μM) at 37 °C, in the presence of 0.75 mM of CBS0065, a plasmin-selective chromogenic substrate. In parallel experiments, the effect of various inhibitors (ϵ-ACA, a lysine analogue that blocks plasminogen binding to cells; CpB, an exopeptidase that cleaves C-terminal lysine residues on cells; aprotinin, a serine protease inhibitor with restricted specificity) on plasminogen activation (1 μM) was determined in the presence of 0.75 mM CBS0065.
The kinetics of plasmin formation were monitored by measuring the release of p-nitroaniline from the chromogenic substrate, detected as a change in absorbance (ΔA405/min), using a multiwell plate reader (Biotek ELx808) thermostatically controlled at 37 °C. Rates of plasmin formation were calculated from slopes of A405 against time. Initial rates were computer-calculated at the inflexion point of the curve by non-linear regression analysis, and were fitted to the Michaelis–Menten equation considering a non-specific component (k[Pg]) proportional to plasminogen concentration ([Pg]).
Detection of cell blebbing and release of MPs by activated cells
Supernatants of activated cells, collected at different time points during plasminogen activation, were supplemented with 2% glutaraldehyde before cytospinning detached cells at 800 g for 5 min (Cytospin® Thermo Fisher Scientific) on small Thermanox® (Nunc) coverslips. These coverslips supported cells all along the preparation until the end of polymerization and embedding of the resin for electron microscopy. Immediately after cytospinning, the Thermanox® coverslips were transferred into a cap containing phosphate buffer. Cells were then rinsed three times in 0.1 M phosphate buffer (pH 7.4) before and after post-fixation with 0.5% osmium tetroxide, dehydrated through a graded ethanol series (70% 2×, 95% 3×, 100% 4×) and finally embedded in epon resin. After polymerization, the coverslips were removed and ultrathin sections were stained with 2% uranyl acetate and lead citrate. All observations were performed with a JEOL 1011 transmission electron microscope. Images were acquired with a megaview III camera (SIS).
MPs were isolated from culture medium conditioned by subconfluent adherent cells and stimulated for 24 h with 1 μM plasminogen. Culture supernatants were collected and cleared from detached cells or large cell fragments by centrifugation at 1500 g for 15 min. The supernatants were then centrifuged at 20000 g for 90 min at 4 °C. Pelleted MPs were washed twice using the same conditions and were suspended in 0.1 M phosphate buffer.
Determination of cell survival using the MTT method
At the indicated time intervals during plasminogen activation, the conditioned medium was carefully removed from a 96-well plate without disturbing the cell monolayer and was replaced with 100 μl of 0.5 mg/ml MTT. The tetrazolium salt was transformed into formazan by the mitochondrial succinate dehydrogenase of living cells. After 1 h at 37 °C, excess MTT was removed and the formazan crystals formed were dissolved in 100 μl of DMSO and colorimetrically detected at 550 nm. Absorbance readings are proportional to the number of living cells. The results are expressed as a percentage compared with control cells.
TUNEL and DAPI staining
Cells in supernatants collected at different times after plasminogen activation were cytospun on to glass slides. TUNEL was used to visualize DNA fragmentation. The cells were then counterstained with DAPI to visualize all nuclei. After washing, the slides were mounted and observed under an epifluorescence microscope. The TUNEL index was calculated as the percentage of TUNEL-positive nuclei relative to total DAPI-stained nuclei.
Plasmin-activated cells or MPs were lysed in 100 mM Tris/HCl buffer (pH 8.1) containing 1% Triton X-100. Proteins in lysates (10 μg) were subjected to 8% PAGE under reducing conditions. Proteins were transferred on to PVDF membranes and revealed with primary specific antibodies (a sheep antibody directed against uPA, a rabbit antibody directed against laminin and a HRP-labelled monoclonal antibody directed against plasminogen kringle 1) at appropriate dilutions and with HRP-coupled secondary antibodies when indicated.
Results are expressed as means±S.E.M. The statistics were performed using Kruskall–Wallis and Mann–Whitney non-parametric tests (MYSTAT12 software). Statistical significance was set at P<0.05.
RESULTS AND DISCUSSION
Cell membrane stability and functioning depend on coupling of the transbilayer membrane phospholipid asymmetry to the cytoskeleton. In adherent cells, stability is enhanced further by coupling integral membrane protein receptors (integrins) to ECM glycoproteins such as fibronectin, laminin, collagen, thrombospondins and tenascins that act as a scaffold surface for cell adhesion. They function as a mediator of cytoskeleton reorganization leading to formation of focal adhesions and intracellular signalling [26,27] that direct cell growth, morphology, migration and anchorage-dependent survival . Disruption of the transbilayer phospholipid asymmetry, a feature of cell activation, or of ECM–integrin interactions by proteolysis may have consequences for membrane cytoskeleton reorganization and cell adhesion during tissue remodelling or in pathological proteolysis [20,29,30]. We therefore explored the possibility that plasmin formation on cells may induce these phenotypic cell changes. The present study provides the first demonstration that plasmin generated at the surface of adherent cells induces an early response characterized by membrane blebbing and vesiculation, i.e. release of MPs. In addition, we show that plasmin proteolyses ECM proteins and thereby disrupts interactions with cells that result in phenotypic changes characteristic of cell detachment and apoptosis. Thus, in agreement with previous reports [8,31], localized decoupling of the cytoskeleton to the membrane results in formation of a bleb exposing phosphatidylserine , and loss of anchoring to the ECM results in reorganization of focal adhesions and cell detachment . Membrane blebs, MPs and cells activated or entering apoptosis express phosphatidylserine at the outer leaflet . Fluorescent annexin A5 has been used to detect the exposed phosphatidylserine by flow cytometric analysis . Since MPs may contaminate cell preparations , the annexin A5-binding assay cannot discriminate between cells or MPs expressing phosphatidylserine . To better identify early morphological changes (membrane blebbing and vesiculation) before anchorage-dependent survival of cells is compromised, we first determined the ability of cells to express plasminogen activators and then monitored plasmin formation and its effects on cell blebbing and survival.
We demonstrate that CHO-K1 and HEK-293 cells express active tPA, whereas HMEC-1s express uPA as indicated by the position of a fibrinolytic band similar to the migration of purified uPA or tPA detected by fibrin zymography (Figure 1A, upper panel). Confirmation was obtained by inhibition of the fibrinolytic activity with specific antibodies as detected by radial diffusion on fibrin–agarose gels (Figure 1A, lower panel). Unrelated antibodies had no effect on plasminogen activation (Figure 1A, lower panel). Kinetics of plasmin formation (Figure 1B) indicated that plasminogen incubated with cells were assembled at their surface and transformed into plasmin in a time-, lysine- and dose-dependent manner until saturation (Figure 1C). Kinetic parameters were determined by fitting data to the Michaelis–Menten equation including a non-specific component of uptake (; Table 1). The Km value calculated is comparable with values obtained for tPA expressed by other cell types including endothelial cells  and neurons . It is also comparable with values determined for the activation of plasminogen by tPA on fibrin (40–160 nM) [36,37], suggesting that the cell surface provides a similar plasminogen-activation enhancement. Lysine-dependent specific plasminogen binding was indicated by inhibition of plasmin formation both with the lysine analogue ϵ-ACA that blocks the LBS of plasminogen, and with CpB that cleaves C-terminal lysine residues from the cell surface. The activity of plasmin formed in situ was efficiently inhibited by aprotinin (Figure 1D).
Identification of plasminogen activators and cellular activation of plasminogen
|Cell line||Treatment||Activator||Km (nM)||TUNEL index at 48 h (%)||Viability at 48 h, MTT assay (%)||Nanovesicles (A280)|
|Cell line||Treatment||Activator||Km (nM)||TUNEL index at 48 h (%)||Viability at 48 h, MTT assay (%)||Nanovesicles (A280)|
Since thrombin, which is a serine protease like plasmin, induces membrane blebbing and platelet vesiculation, we explored the possibility that plasmin may induce a similar response on adherent cells. Membrane blebbing is a short-lived phenomenon observed at the very initial phase of cell stimulation that can only be visualized by electron microscopy. For that purpose, the early response of cells to plasmin formation was interrupted by glutaraldehyde fixation, thus allowing capture of multiple blebbing of the membrane and formation of numerous MPs that were about to be released from cells (Figure 2). A small number of MPs could be detected in control culture conditions (Figure 2A and Table 1). After incubation of cells with plasminogen, the number of MPs detected in the culture medium was dramatically increased in all adherent cell types (Figure 2A). These MPs are approx. 300 nm in size and contain substructures having an electron density similar to that of cell cytoplasm, which is surrounded by a well-defined membrane (Figures 2B–2D). This phenomenon was dependent on the activity of plasmin as indicated by its prevention in the presence of aprotinin (results not shown). Since plasmin was the only aprotinin-sensitive serine protease present in the culture supernatant (Figure 1B), these results indicate that membrane blebbing is one of the early responses of cells to plasmin. MPs, which are released, carry the plasminogen activator of their parental cell (e.g. uPA for HMEC-1, Figure 3A) and plasmin(ogen) bound at the membrane (Figure 3B). These results confirm that MPs could be considered as potential messengers of proteolytic activity [9,10].
Early response of cells to plasmin formation: membrane blebbing and vesiculation
Western blot analysis of MPs and cell lysate
As indicated above, the first and immediate consequence of cellular stimulation by plasmin is membrane blebbing followed by the release of MPs. The continuous formation of plasmin has other consequences on cell phenotype and behaviour, which are related to the proteolytic activity of plasmin on the substratum. Active plasmin proteolyses and degrades components of the ECM including laminin (Figure 3C) and fibronectin . In the absence of a substratum for anchorage, the cells became round (Figure 4A) and gradually detached from their support as shown by the amount of residual living cells (MTT assay, Figure 4B). After 48 h of incubation with plasminogen, the detached CHO-K1 cells entered the apoptotic process as suggested by DNA fragmentation observable by TUNEL (Figures 4C and 4D).
Advanced response to plasmin formation: detachment and cell death
Electron microscopy allowed detailed observation of cell morphology and characteristic apoptotic changes (Figure 5). In control cells, the nuclear membrane is clearly delimitated and the characteristics of normal cytoplasm are maintained, including endoplasmic reticulum and mitochondria with their internal crests (lower panel, Figure 5A). In plasminogen-treated cells (Figure 5B; 48 h), the nucleus shows chromatin condensation, the cytoplasm is disorganized, contains lysis vesicles and mitochondria became electron-dense (Figure 5B, lower panel). After 72 h of stimulation, the consequences of plasmin formation were dramatic: higher condensation and fragmentation of the chromatin (Figure 5C).
Advanced response to plasmin formation: electron microscopy of apoptosis
To investigate the relation between plasmin formation and cell detachment, we used active-site blocked plasmin, Pn-VFK (Figure 6A). Pn-VFK was unable to develop a proteolytic activity on ECM components, e.g. laminin (Figure 3C). In the absence of ECM degradation, cells did not detach from their support and remained viable (Figure 6B).
Plasmin-induced cell detachment upon proteolysis of the extracellular matrix
Altogether, these results indicate that, upon plasmin formation, the proteolytic activity of this enzyme induces (i) at the cellular level, membrane blebbing and vesiculation, and (ii) on ECM proteins, cell detachment-induced apoptosis.
Conditioned medium from cell culture experiments and animal or human biological fluids (e.g. urine) or exudates (e.g. ascites, cyst fluid and pleural effusion) contain detached cells that can be cytospun and processed for visualization of structural and morphological changes (Figure 7). This procedure allows identification of the initial state of cell activation before irreversible cell damage occurs, and further explores a more advanced state of cell injury characterized by nuclear and cytoplasmic abnormalities typical of apoptosis.
A sequential procedure to analyse membrane blebbing, vesiculation and cell apoptosis
The advantage of the proposed procedure is that early morphological changes (membrane blebbing and vesiculation) can be easily identified before anchorage-dependent survival of cells is compromised [19,32,39–42]. In the absence of other morphological or structural modifications, these initial membrane changes indicate a reversible response to stimulation by stressors and can be used as indicators of cell stimulation/suffering. Another advantage is that, from the same culture supernatant or biological fluid, it could be possible to isolate (Figure 2) and to analyse MPs (Figures 3A and 3B).
Finally, in vivo proteolytic-dependent degradation of ECM and apoptosis [43–45] may also be studied. In the case of human pathologies or animal experiments, this mechanistic procedure has potential application for diagnostic purposes and in the follow up of disease progression and treatment.
In conclusion, the present study has demonstrated that plasmin formed at the surface of adherent cells induces two types of response (Figure 7): (i) membrane blebbing and vesiculation (grey arrow), the early response of adherent cells to proteolytic activation; and (ii) a delayed response characterized by cell detachment and apoptosis (black arrow). Investigation and analysis of this succession of events may help to identify the outcome of adherent cells in response to insults that may potentially shift from minimal structural changes (membrane blebbing and vesiculation) to apoptosis. This sequence may be operative in inflammatory processes such as atherosclerosis  and in adherent cells derived from the central nervous system .
We thank Dr Alan Young and Dr Mary Osborne-Pellegrin for proofreading and editing the paper. Access to Cytospin® facilities was a courtesy of Dr Laurent Poulain.
Dulbecco's modified Eagle's medium
human embryonic kidney
human microvascular endothelial cell-1
tissue-type plasminogen activator
terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling
urokinase-type plasminogen activator
D-Val-Phe-Lys-chloromethyl ketone dihydrochloride
Loïc Doeuvre performed the research, analysed data and wrote the paper. Laurent Plawinski provided experimental support and participated in drafting the paper. Didier Goux contributed electron microscopy support. Denis Vivien, the director of the research unit, secured funding. Eduardo Anglés-Cano designed the research, analysed data and wrote the paper.
These studies were supported by funds from the Inserm (Institut National de la Santé et de la Recherche Médicale) and the Lower-Normandy Regional Council. L.D., L.P., D.V. and E.A.C. are members of the European Community's Seventh Framework Programme (FP7/2007–2013) [grant agreement 201024]. L.D. is a recipient of a Ph.D. fellowship from the French Ministry of Education and Research.