Protein C Inhibitor (PCI) Binds to Phosphatidylserine Exposing Cells with Implications in the Phagocytosis of Apoptotic Cells and Activated Platelets

Protein C Inhibitor (PCI) is a secreted serine protease inhibitor, belonging to the family of serpins. In addition to activated protein C PCI inactivates several other proteases of the coagulation and fibrinolytic systems, suggesting a regulatory role in hemostasis. Glycosaminoglycans and certain negatively charged phospholipids, like phosphatidylserine, bind to PCI and modulate its activity. Phosphatidylerine (PS) is exposed on the surface of apoptotic cells and known as a phagocytosis marker. We hypothesized that PCI might bind to PS exposed on apoptotic cells and thereby influence their removal by phagocytosis. Using Jurkat T-lymphocytes and U937 myeloid cells, we show here that PCI binds to apoptotic cells to a similar extent at the same sites as Annexin V, but in a different manner as compared to live cells (defined spots on ∼10–30% of cells). PCI dose dependently decreased phagocytosis of apoptotic Jurkat cells by U937 macrophages. Moreover, the phagocytosis of PS exposing, activated platelets by human blood derived monocytes declined in the presence of PCI. In U937 cells the expression of PCI as well as the surface binding of PCI increased with time of phorbol ester treatment/macrophage differentiation. The results of this study suggest a role of PCI not only for the function and/or maturation of macrophages, but also as a negative regulator of apoptotic cell and activated platelets removal.


Introduction
Protein C Inhibitor (PCI) is a serine protease inhibitor belonging to the class of serpins [1]. Initially PCI has been described as an inhibitor of the vitamin K dependent anticoagulant protease activated protein C (APC) [2]. In the course of further research PCI revealed broad protease reactivity and was acknowledged as inhibitor of different coagulation factors, like thrombin, factor Xa and XIa, but also of fibrinolytic enzymes, uncovering PCI as a balancing factor in hemostasis [3,4]. Synthesized mainly in the liver, human PCI circulates in plasma at a concentration of about 5mg/ml (approx. 100 nM). Besides that the serpin is synthesized by many cell types and tissues and is present in almost every body fluid. In order to define new markers for myocardial or thrombotic diseases, PCI and the PCI-APC complex have been frequent targets of interest. Indeed elevated blood plasma levels of PCI were detected in male survivors of myocardial infarction [5] and high APC-PCI levels are associated with higher early death rates after aortic surgery [6]. Watanabe et al. [7] found higher APC-PCI complex levels in patients suffering from disseminated intravascular coagulation, thrombotic thrombocytopenia, acute myocardial infarction, pulmonary embolism, and deep vein thrombosis. Decreased PCI levels were found in disseminated intravascular coagulation and thrombotic thrombocytopenia, showing the importance but also the complexity of the APC-PCI balance in coagulation and several thrombotic disorders [7].
Binding to certain glycosaminoglycans, like heparin stimulates the inhibitory activity of PCI for certain target proteases [8,9]. Heparin binding increases the rate of APC inhibition by PCI about 400 fold [10] and thrombin inhibition 30 fold [11,12]. PCI also binds oxidized and unoxidized negatively charged phospholipids like phosphatidylethanolamine (PE) and phosphatidylserine (PS), which also modulate its inhibitory activity [13].
These phospholipids are normally present on the inner leaflet of the plasma membrane. PS becomes surface exposed during apoptosis and cell activation. In apoptosis PS serves as an important marker and phagocytosis signal [14,15]. The phagocyte interacts either directly with PS via a receptor [16,17], or via a linking molecule, opsonizing PS before the docking of the phagocyte. Milk fat globule -EGF-factor-8 and protein S are well known opsonins, stimulating phagocytosis of apoptotic cells (efferocytosis) acting as a bridge for the exposed PS [18,19]. Current literature also reveals PS-binding proteins that inhibit phagocytosis of apoptotic cells, like the well known apoptosis marker Annexin V [20], or high mobility group protein-1 [21]. Data obtained in the study by Malleier et al. [13], revealing PS as a binding partner of PCI, were performed in a purified system, not taking into account the influence of plasma membranes. It was therefore our aim to analyse binding of PCI to phospholipds on cell membranes, and to determine whether there is a difference in PCI binding to apoptotic and activated cells, which expose a higher percentage of PCI-binding phospholipids on their surface in comparison to quiescent cells.

Cell culture and differentiation
Human myeloid U937 cells, human monocytic THP-1 cells (collections of the Department of Vascular Biology and Thrombosis Research, Medical University of Vienna, purchased from ATCC, American Type Culture Collection, Manassas, VA, USA) [22,23] and human Jurkat cells (a kind gift of Prof. Veronika Sexl, purchased from ATCC, American Type Culture Collection, Manassas, VA, USA) [24], a T-lymphocyte cell line, were cultured in RPMI medium supplemented with 10% FBS (Sigma-Aldrich, St.Luis, MO, USA), 1% Penicillin/Streptomycin/Fungizone, 1% L-glutamine and 2% HEPES (Lonza, Basel, Switzerland), defined as full medium. All cells were maintained in a humidified atmosphere of 5% CO 2 and a temperature of 37uC. For differentiation U937 cells were seeded at 5610 5 cells/ml into 24-well plates (Nunc, Thermo Fisher Scientific, Waltham, MA, USA) and treated for 24 h or 72 h with 20 nM PMA (Sigma-Aldrich, St.Luis, MO, USA) [25,26]. THP-1 cells were differentiated with 1mM PMA for 24 h at 37uC and 5% CO 2 . Jurkat cells were driven into apoptosis by treatment with 6mM Camptothecin (Sigma-Aldrich, St.Luis, MO, USA) at a density of 1610 6 cells/ml in full medium for 24 h (Ct cells). Apoptosis was determined by Annexin V-EGFP or -FITC/propidium iodide staining (Apoptosis kit, Biovision, Milpitas, CA, USA) according to the manufacturer's protocol and analysed by FACS (FACS Calibur, Becton Dickinson, Franklin Lakes, NJ, USA).
Isolation and maintenance of platelets and monocytes from human blood was performed as described by Badrnya et al. [27]. In brief, healthy blood donors gave their informed consent and experiments were approved by the Human Ethics Committee of the Medical University of Vienna, project number EK112/2009. Platelets isolation was performed with citrated blood by two centrifugation steps, 1256g for 20 minutes and 30006g for 2 minutes on the platelet rich fraction with the addition of prostacyclin (PGI2) (Sigma-Aldrich, St.Luis, MO, USA) at a concentration of 1mM. Monocytes were isolated from blood samples anticoagulated with ethylenediaminetetraacetic acid (EDTA). Peripheral mononuclear blood cells were obtained by density centrifugation with Histopaque 1077 and Histopaque 1119 (Sigma-Aldrich, St.Luis, MO, USA). On the whole, experiments were performed with blood samples, drawn from 15 healthy individuals.

PCI binding to cells
Plasma PCI was isolated as described by Malleier et al. [13] and coupled with Cy3 (Cyanine dye with an excitation at ,550 nm) fluorophor (Cy-3 Ab labelling kit Amersham, GE Healthcare, Uppsala, Sweden) according to the product manual. Live and apoptotic Jurkat cells (5610 5 /tube) were washed with cold PBS (pH = 7.4) and incubated with 75 nM Annexin V-EGFP and/or 250 nM PCI-Cy3 for 1 h at room temperature (RT) in 500ml Annexin V binding buffer containing 1.8 mM Ca 2+ (apoptosis kit, Biovision, Milpitas, CA, USA) or PBS without Ca 2+ (pH = 7.4). After washing with the respective incubation buffer (26500ml), cells were fixed and stored in 1% paraformaldehyde (PFA) at 4uC in the dark until FACS analysis (excitation: 488 nm, detection: Annexin V-EGFP FL-1 channel, PCI-Cy3 FL-2 channel).

Phagocytosis assays
Apoptotic cells. Differentiated, adherent U937 cells were stained with 1mM Dodecyldimethylamine oxide Succinimidyl-Ester (DDAO-SE) far red (excitation 633 nm, detection FL-4 channel) and apoptotic Jurkat cells, in suspension, with 0.5mM Carboxyfluorescein Diacetat Succinimidyl Ester (CFDA-SE) cell tracker dyes (excitation 488 nm, detection FL-1 channel) according to the manufacturer's protocols (Life Technologies, Carlsbad, CA, USA). Either Jurkat cells or macrophages were incubated with 100 or 300 nM human recombinant PCI (expressed in HEK293 cell line; Novoprotein Scientific, Summit, NJ, USA), heat-treated PCI (10 Min. 100uC), human plasma purified antithrombin III (Sigma-Aldrich, St.Luis, MO, USA), or PBS buffer control, respectively, in Hepes buffer pH = 7.4 (3.1610 6 cells/125ml) for 1 h at 37uC. Thereafter cells were either washed twice with Hepes buffer to remove unbound protein or seeded immediately onto adherent macrophages (2.5610 5 macrophages in 200ml Hepes buffer containing 0.5% FBS +1.25610 6 / 50ml Ct Jurkat cells). Phagocytosis assays were conducted for 24 h at 37uC or at 4uC. To stop phagocytosis, unengulfed apoptotic cells were washed away with cold PBS and after Accutase treatment (30 min. 37uC) dislodged U937 macrophages were fixed and stored in 1% PFA at 4uC in the dark until FACS analysis. Polysterene microspheres. Phagocytosis experiments with fluorescent polystyrene microspheres were conducted in a similar manner. U937 cells were differentiated with 100 nM PMA for 24 h. Yellow green fluorescent microspheres with a diameter of 1mm (carboxyl modified, Sigma-Aldrich, St.Luis, MO, USA) were washed with 1 ml PBS/2610 7 particles and by centrifugation at 11,000 g for 15 minutes. Beads (1610 7 microspheres/100ml) or macrophages (250ml/well) were preincubated with buffer, 100 nM human plasma derived PCI or 100 nM AT3 in serum free medium for 1 h at 37uC. Unbound protein was washed away and 10 beads/macrophage were seeded onto differentiated macrophages and phagocytosis was conducted in cell culture medium, containing 10% FBS, for 0.5, 1, 2, 3, 4 h at 37uC or 4uC. Detachment, fixation, FACS and analysis were performed as described for phagocytosis of apoptotic cells.
Platelets. Phagocytosis of platelets was conducted similarly as previously described by Badrnya et al. [27]. Platelets were activated by the addition of ADP (Sigma-Aldrich, St.Luis, MO, USA) at a final concentration of 20mM for 10 minutes with or without 300 nM PCI, and dyed with CMFDA Cell Tracker green (0.1mg/ml, Life Technologies, Carlsbad, CA, USA). Excess dye was washed away with PBS and phagocytosis was carried out for 60 min with freshly isolated human blood derived mononuclear cells, isolated from the same blood donor as the platelets, at a ratio 1:100 monocytes:platelets. Phagocytosis was stopped by fixation with 1% PFA. Surface bound platelets were stained with anti CD61-Alexa647 (BioLegend, San Diego, CA, USA) and monocytes stained with anti-CD14 PerCP (Becton Dickinson, Austria). Flow cytometric analysis was performed on an Accuri Flow Cytometer (Becton Dickinson, Austria). The amount of surface bound Cell tracker green and Alexa 647 double positive platelets on monocytes was substracted from all Cell tracker green positive platelets associated with monocytes resulting in the percentage of phagocytosed particles.

Determination of activation markers on platelets
Isolated platelets were incubated with or without 300 nM PCI for 15 minutes, and activated by the addition of ADP (20mM) for 5 minutes. Platelet surface expression of PS and P-selectin and glycoprotein IIb/IIIa (GPIIb/IIIa) activation were detected by flow cytometry after staining with Annexin V-FITC or anti-CD62P-Alexa647 (Biolegend, San Diego, CA, USA) and PAC-1 (Becton Dickinson, Austria) for 20 minutes in binding buffer followed by fixation in 1% formaldehyd, as previously described [28].

Statistics
Statistics were calculated in GraphPad PRISM Software (GraphPad Software Inc., La Jolla, CA, USA) using One-way ANOVA and the Dunnett's multiple comparison, or a two tailed paired Student's T-test where indicated, P,0.05 was considered as significant.

PCI binds to Camptothecin-treated cells
In order to determine the binding of PCI, untreated and Ct Jurkat cells were incubated with PCI-Cy3 and Annexin V-EGFP, used as apoptosis marker and analysed by flow cytometry (Figure 1). By treatment of Jurkat cells with Camptothecin 50% -80% Annexin V positive Jurkat cells were obtained (data not shown). Calcium-free conditions served as control for specific Annexin V binding. Fluorophor-linked PCI bound to a high percentage of Ct Jurkat cells, and only to a small percentage to untreated cells. Binding of PCI to Ct cells was not dependent on calcium, but the binding to untreated cells partially was ( Figure 1B).
Confocal microscopy data of PCI binding to untreated and Ct Jurkat cells confirmed an association of PCI with Ct cells (Figure 2A). PCI was probably internalized into the cells in vesicular structures. About 10-30% of untreated Jurkat cells bound PCI. Here the binding was limited to specific spots probably on the cell surface. To prove that these results were not limited to T-lymphocytes a myeloid cell line (U937 cells) was investigated for PCI binding as well ( Figure 2B). The binding to U937 cells showed a similar binding pattern, binding to Ct cells was intracellular while for the untreated cells, PCI was localised only on few cells and at distinct spots, most likely on the cell surface. PCI binding to U937 cells increases with macrophage differentiation U937 cells were differentiated to macrophage like cells by the incubation with PMA for 24 h or 72 h, respectively, and thereafter incubated with PCI-Cy3. Macrophage differentiation was confirmed by the upregulation of the macrophage markers CD14 and CD11b ( Figure 3A). The amount of PCI-Cy3 positive cells increased with time of PMA treatment, as was assessed by flow cytometry. Moreover, also the amount of Annexin V-FITC positive cells increased ( Figure 3B), which is in line with current literature [29]. According to confocal microscopy in Figure 3C, PCI (red) was obviously binding to the cell surface and showed also accumulation at Annexin V binding regions (green).

PCI colocalizes with PS in close proximity on Ct cell membranes
In order to prove that PCI does not only bind PS in purified in vitro studies we used fluorescence resonance electron transfer microscopy (FRET) to identify a close proximity of PCI to PS on apoptotic cells. Ct Jurkat cells, which were considered as late apoptotic according to Annexin V/propidium iodide binding, were treated with Annexin V-FITC as the donor and PCI-Cy3 as the acceptor molecule of energy transfer. Corrected pictures revealed significant FRET signals (Figure 4) suggesting that PCI and Annexin V bind in close proximity to apoptotic cells.

PCI inhibits phagocytosis of apoptotic cells
In our phagocytosis model differentiated U937 macrophages and Ct Jurkat cells were stained with different fluorophores and phagocytosis was assessed by flow cytometry ( Figure 5A). PCI inhibited phagocytosis in a dose dependent manner in an assay system where apoptotic Jurkat cells were preincubated with PCI and where PCI was present during phagocytosis ( Figure 5B). At 100 nM, corresponding to the plasma concentration, PCI showed a mild, but not significant decrease, while 300 nM PCI resulted in a reduction of the phagocytosis index to half. In this model PCI influenced mainly the quantity of phagocytosing macrophages (average buffer control 17.56%68.33), without appreciable impact on the amount of ingested material per macrophage. The heat-treated form of PCI, or antithrombin III (AT3) did not display inhibitory effects on phagocytosis under the same conditions. Preincubation of either the apoptotic Jurkat cells or the macrophages with PCI and subsequent washing ( Figure 5C) did not result in a significant effect on efferocytosis but in comparison to AT3 or heat-treated PCI, active PCI showed a mild inhibitory effect. However the tendency of phagocytosis inhibition by PCI was more prominent for preincubated Jurkat cells, than preincubated U937 macrophages, while the stage of macrophage differentiation seemed to be irrelevant.
To visualize cellular binding of PCI in the phagocytosis model, Jurkat cells (green) alone (untreated or Ct-treated) or together with THP-1 macrophages (green) were incubated with fluorophore coupled PCI (red) and analysed by fluorescence microscopy. THP-1 macrophages were used for microscopy since these cells show less clustering upon differentiation than U937 cells. The pictures revealed binding of PCI to the apoptotic Jurkat cells during phagocytosis, shown by intensely red stained small apoptotic bodies inside of larger macrophages, indicated by small arrows ( Figure 5D).

PCI inhibits the phagocytosis of activated platelets
As activated platelets expose PS on their surface [30], we investigated if PCI has an influence on phagocytosis of platelets. Therefore untreated and ADP activated platelets were coincubated with human blood derived monocytes in the presence or absence of PCI. The activation of platelets after ADP treatment was confirmed by an increase in activation markers like PS exposure (2.361.2 fold), glycoprotein IIb/IIIa activation (2.260.7 fold) and surface P-selectin expression (3.760.9 fold).
In general, the phagocytosis of platelets by monocytes isolated from the same blood donor, increased with ADP activation 3.7 times compared to untreated controls. PCI showed no effect on the phagocytosis of untreated platelets, while the phagocytosis of activated, PS exposing platelets was significantly decreased with PCI incubation compared to the control ( Figure 6A). Binding of platelets to monocytes was not affected by PCI ( Figure 6B). In order to rule out a difference in platelet activation in the presence of PCI, activation markers of untreated and ADP treated platelets were compared with or without PCI incubation. Annexin V binding decreased on activated platelets in comparison to the control, while the surface expression of P-selectin and GPIIb/IIIa did not show any difference (Figure 7), which suggests that PCI does not influence platelet activation itself but interacts with exposed PS.

PCI preincubation inhibits phagocytosis of polystyrene microspheres
To further evaluate the interaction site of PCI in the phagocytosis process phagocytosis was conducted with carboxyl modified polystyrene microspheres as target material to be engulfed by macrophages. Preincubation with proteins was conducted on macrophages or the beads. Pretreatment of the beads results in PCI binding (data not shown), which decreases the amount of phagocytosing macrophages significantly, while the ingested amount of microspheres per macrophage does not alter in relation to the control. Preincubation of macrophages with PCI or AT3 shows no effect on phagocytosis (Figure 8).

PCI expression is upregulated during U937 macrophage differentiation
PCI expression was upregulated during differentiation of U937 cells with PMA into macrophage like cells. After 72 hour treatment, PCI mRNA expression reached about three fold the expression rate of the undifferentiated U937 monocytes ( Figure 9A). Also PCI-antibody binding to the cell surface of U937 cells increased with differentiation ( Figure 9B).

Discussion
PCI has been extensively studied for interactions with glycosaminoglycans and phospholipids in purified systems. Malleier et al. [13] showed the binding of PCI to negatively charged phospholipids (oxidized phosphatidylethanolamine, unoxidized and oxidized PS) and modulation of the inhibitory activity upon phospholipid binding.
In order to investigate a physiological function of this interaction, it was the intention to study the binding characteristics of PCI to cell membranes, as PS and PE are exposed on the surface of apoptotic cells. By comparing the binding of exogenously added PCI it was observed that the majority of Ct cells, considered as apoptotic cells, bound fluorophor coupled PCI, but only a small extent of untreated cells. From previous in vitro data it is known that PCI binding to phospholipids is not dependent on calcium [13], as it is the case for Annexin V [31], frequently used as apoptosis marker. PCI binding to apoptotic cells was not dependent on calcium. However, on untreated cells PCI binding was at least partially dependent on calcium, suggesting the involvement of additional or other binding partners on intact cells. From confocal microscopy it was further obvious that not only the quantity of cells binding PCI differs between Ct-and untreated cells, but also the localisation pattern. Pictures confirm that the fluorophor coupled PCI bound to almost all apoptotic cells and seemed to be internalized into the cells in vesicular structures. Baumgä rtner et al. [32] propose that PCI is internalized into cells upon phosphatidylethanolamine binding. As the loss of lipid asymmetry on apoptotic cell membranes leads to an unusually high exposure of PS and PE, the internalization of PCI, visible by confocal microscopy of apoptotic cells, may result from binding to phospholipids. A similar effect was described by Kenis, showing an internalization of PS upon Annexin V binding, which consequently reduces the phagocytosis of apoptotic cells by phagocytes [20]. Live cells showed binding of PCI only at spotted sites probably located on the cell surface, although cytosolic staining cannot be excluded. Further studies are needed to investigate the binding partners at these specific punctuated sites. Our group also described the internalization and targeting of PCI to the nucleus of HL60 cells [32]. The difference between Baumgä rtner's study and our present study may be explained on the one hand by different assay conditions (serum free, 37uC) and on the other hand by using a PCI directly coupled to a fluorophor, which was linked to the protein via free amino groups. This protein modification might have altered the charge and the size of the protein, or modified the accessibility of the nuclear localisation signal, resulting in impaired nuclear transport. However, the PCI binding pattern on each live and apoptotic cells seemed to be similar in T-lymphocytes (Jurkat) and in cells of the myeloid linage (U937).
PS is known as signal for the removal of apoptotic cells by phagocytes [15]. As PCI binding to and internalization into apoptotic cells was observed by FACS analysis as well as by confocal microscopy, it was intended to elucidate the role of this  interaction in an in vitro model of phagocytosis. As phospholipid composition is also changing upon cell activation and PS exposure increases during U937 macrophage differentiation, binding of PCI was examined on U937 macrophages. Treatment with phorbol esters induced U937 monocytes to differentiate into the macrophage linage. PCI binding during differentiation was studied on U937 cells up to 72 hours and revealed an increase in PCI binding cells with macrophage maturation. According to literature also the amount of Annexin V binding cells increases with differentiation [29]. Strikingly, binding of PCI to differentiated live U937 was not dependent on calcium. As this was the same for apoptotic but not for live Jurkat cells, this might point out in an increased binding to phospholipids on U937 macrophages. The question is, why there were much more Annexin V binding cells than PCI binding cells within the differentiated macrophage population, while on apoptotic cells the amount was very similar. We can show that endogenous PCI expression and surface exposure increases with macrophage differentiation. It is therefore also possible that PS is partially occupied by endogenous PCI and no longer available for PCI-Cy3. Binding and PCI localization was studied by confocal microscopy and clearly showed PCI on the surface of differentiated macrophages. Even an accumulation at Annexin V positive sites and colocalization was obvious, suggesting binding of PCI to PS.
Apart from PCI binding to PS exposing cells, further proof was needed to identify an interaction of PCI with PS on cell membranes. FRET analysis using FITC coupled Annexin V as donor molecule and PCI linked to Cy3 as acceptor, revealed clear FRET signals, suggesting a very close proximity between Annexin V and PCI. Therefore the binding partners of each protein must be very close. Competition experiments (data not shown) with Annexin V and PCI on apoptotic cells showed that the two proteins do not compete for the same binding site, therefore the positive FRET signals suggest binding to different epitopes. PS is known to be expressed as ''patch''-like structures at the surface, probably together with a panel of other apoptosis signals, meaning that a dense area of PS lipids might also bind PCI and Annexin V simultaneously but to separate PS molecules, bringing both proteins in close connection [33]. PCI was probably covered all over with the Cy3 fluorophor as the dye was coupled to all free amino-groups on the protein. Hence, it seems likely that the close proximity of the molecules, even if not specifically interacting witch each other, resulted in a positive FRET signal. Confocal microscopy showed only little colocalization of PCI and Annexin V, which might be explained by higher affinity of PCI for PE and internalization, of PCI. By accumulation of PCI at PE rich sites, fainter stainings at Annexin V binding sites might be overseen according to acquisition and resolution parameter adjustments which were adapted to areas of highest signals. Therefore FRET signals prove an overlap of PCI and Annexin V at various sites of apoptotic cells, which is not easily recognized by two color confocal microscopy.
Having shown binding of PCI to both interaction partners (CT Jurkat cells as well as differentiated U937 macrophages), it seemed important to define the impact of PCI binding on the process of apoptotic cell phagocytosis. We can show that PCI inhibits phagocytosis in a dose dependent manner, which is in line with another member of the Serpin family, plasminogen activator inhibitor-1, which has been shown to inhibit efferocytosis of neutrophils [34]. Furthermore the effect of PCI was not dependent on the degree of macrophage differentiation. To show that inhibition of phagocytosis was not caused by a general protein effect, heat-treated PCI and another protein of the serpin family were studied. Heat denatured PCI did not reveal such an effect, which suggests either the importance of protease inhibitory activity for phagocytosis inhibition or altered phospholipid binding of the denatured protein. Antithrombin, another serpin and member of the same phylogenetically related class as PCI, did not show any effect on phagocytosis, which is consistent with previous in vitro data showing no binding of AT3 to PS [13]. The exact mode of interference of PCI in this system still needs to be elucidated, but the fact that efferocytosis inhibition seemed to be dependent on binding of PCI to apoptotic cells, the hypothesis of covering phagocytosis signals like PS, as it was already published for High mobility group protein-1 [21], seems very likely. Also Annexin V binding to and internalization of PS on apoptotic cell surfaces was reported to inhibit efferocytosis. As PCI also shows accumulation in endocytic vesicles in apoptotic cells, according to confocal microscopy ( Figure 2), a similar mechanism might be considered for PCI [20].
Experiments, investigating the influence of PCI on the phagocytosis of untreated and activated human blood derived platelets by monocytes showed similar effects. PCI decreased phagocytosis of PS exposing, activated platelets, but does not influence phagocytosis of untreated platelets. Furthermore, our results clearly show no difference in platelet binding to monocytes. Therefore it seems very likely, that PCI is not affecting the binding step, but probably influences the internalization step during the course of phagocytosis. In order to rule out a difference in platelet activation in the presence of PCI, different activation markers were measured on the surface of the platelets with and without PCI. Previous studies on neutrophil-mediated phagocytosis of platelets revealed that while P-selectin is important for platelet adhesion to neutrophils, surface expression of PS is essential for phagocytosis of platelets [35]. In our experiments we did not observe any effect of PCI on surface expression of platelet Pselectin, which might explain the unchanged (rather slightly upregulated) formation of platelet-monocyte aggregates. However, PCI decreased the surface expression of PS on the platelet surface thereby potentially interfering with platelet uptake by monocytes. One hypothesis to explain this phenomenon, in line with our observations on apoptotic cells, is internalization of PS-PCI resulting in decreased amounts of surface exposed PS, available for Annexin V binding. Although the exact interaction mechanism of PCI in phagocytosis needs to be clarified, the internalization of PS upon PCI binding could well explain similar monocyte binding rates but decreased phagocytosis.
The physiological relevance of PCI in inhibiting efferocytosis might be described as balancing factor. PS serves as a signal for phagocytosis and PS exposure occurs also on activated platelets, on activated lymphocytes and on mature macrophages [29,30,36]. Therefore it seems reasonable that PS alone is not sufficient as a phagocytosis signal but that the interaction with other expressed surface markers and linking molecules defines the fate of a cell. As PCI is present in almost every body fluid, the inhibition of such an important process as apoptotic cell removal by PCI seems pointless. The serpin may rather act as a balancing regulator, preventing the phagocytosis of functional cells, exposing PS for other reasons than apoptosis. In addition, PCI binding to surfaces without PS, also inhibited phagocytosis, as it was shown in experiments with polystyrene microspheres. Here the presence of PCI on the beads causes a decline in the amount of phagocytosing macrophages. This might indicate that PCI alone acts as a phagocytosis inhibitor and the lipid surface composition of activated and apoptotic cells together with local PCI expression levels contribute to balancing immunemodulatory processes.
Real Time PCR data on the differentiation of U937 cells revealed an increase in PCI expression with degree of differentiation. The reason for this increase is not yet clear and needs further study but might indicate locally and tissue specific higher PCI levels compared to plasma concentration. Therefore the application of more than 100 nM (plasma concentration) PCI is assumed as reasonable and physiological relevant amount. According to previous publications, pointing at the stimulation of pathogen removal by PCI, this upregulation of PCI in immune cells might not only relate to phagocytosis of apoptotic cells but, indicates a much more expansive, and as yet undefined, role of PCI in immunity and inflammation [32,37].