CLIC1 Supports Mechanisms Related to Platelet Activation and Thrombosis

Chloride intracellular channel 1 (CLIC1) belongs to a protein family of highly homologous and evolutionary conserved ion channels [1‒4]. Out of these, six CLIC genes (clic1–6) have been identified in humans with expression in most if not all tissues [5]. Depending on cell type and context, the six CLIC proteins have been shown to mediate a diverse array of cell functions, ranging from cell migration, invasion, and proliferation to lumen formation, phagocytosis, and stabilization of the cell membrane [6‒10]. Accordingly, CLICs have been implicated in functions contributing to the development of kidneys, blood vessels, and inner ear as well as to pathologies such as hemorrhage, inflammation, vascular degeneration, and cancer [7, 9‒13]. While all six human CLICs are known to generate ion currents, their exact mechanism of action in cells is still not clearly understood [3, 14‒17]. The Caenorhabditis elegans CLIC orthologue EXC-4 was shown to be necessary for proper lumen formation in the excretory organ of C. elegans, and it has been speculated that EXC-4-mediated ion shunts support vesicle acidification, which is necessary for vesicle fusion and subsequent tube formation [18, 19]. Work with transgenic CLIC1-deficient mice, on the other hand, has indicated that CLIC1-dependent chloride currents promote vesicle acidification during macrophage phagocytosis and that this mechanism is relevant for chronic inflammation in a model of rheumatoid arthritis [8]. In addition, CLIC1-deficient mice also exhibit a platelet function deficit in response to ADP signaling [12]. While it remains uncertain if the resulting bleeding disorder is caused by reduced ion conductivity of CLIC1−/− platelets, it has been shown that extracellular chloride ions support platelet aggregation through P2Y12 [20].

Structurally, CLICs belong to a small group of metamorphic proteins as they can exist in completely different conformational states [1‒4, 21, 22]. As such, they can transition from soluble cytosolic monomers into integral membrane proteins and each conformation appears to be associated with very distinct functions [23]. Soluble CLICs adopt a glutathione S transferase fold, which is characterized by a conserved cysteine residue that conveys oxidoreductase enzymatic activity [24, 25]. Oxidation of this active-site cysteine has also been shown to induce a distinct conformational change that promotes dimerization and subsequent membrane binding, which in turn is thought to be critical for the recruitment of further CLIC molecules to form a membrane spanning oligomer [24, 26]. Based on these data, it has been suggested that CLIC can act as a sensor for reactive oxygen species and that the subsequent membrane insertion serves to equilibrate electrochemical membrane potentials that are generated during cell oxidation [26, 27]. However, additional mechanisms for membrane insertion appear to exist since the C. elegans CLIC orthologue EXC-4 can form membrane pores even though it lacks the active site cysteine [4, 18, 19]. This observation has paved the way for a new concept of CLIC1 membrane insertion, implicating increased intracellular Zn²⁺ ions as crucial mediators of CLIC1 binding to membranes [28].

CLIC1 has regularly been shown to be upregulated in patients with malignant tumors, and in many of these, CLIC1 expression correlates with aggressive disease and poor outcome [29‒34]. CLIC1 has also been shown to be overexpressed in the tumor vasculature, indicating that CLIC1 plays an important role in tumor cells and the angiogenic endothelium [35]. In line with this, we found that CLIC1 promotes tumor and endothelial cell invasion in cooperation with integrin αvβ3, which mediates the recruitment of CLIC1 into invadopodia [9]. Knocking down CLIC1, on the other hand, negatively affects αvβ3-mediated fibronectin matrix formation and, accordingly, inhibits cell invasion, colonization, and tumor metastasis. Taken together, these data support a concept, where CLIC1 is involved in cell-adhesive processes in cooperation with integrin β3. In addition to tumor cell invasion, β3 integrin plays an important role in endothelial cell proliferation and platelet aggregation, suggesting that the specific functional relationship between β3 integrin and CLIC1 is of importance for critical vascular biological processes such as thrombus formation, vascular repair, and angiogenesis [36, 37]. Here, we show that CLIC1 cooperates with integrin β3 and that this interaction affects endothelial cell spreading, platelet activation in vitro, and platelet thrombus formation in vivo.

Human umbilical venous endothelial cells (HUVECs) were purchased from Lonza, cultured in EBM-2 endothelial growth media, and grown on cell culture-treated dishes at 37°C under a humidified, 5% CO₂ atmosphere according to the manufacturers’ specifications.

On-TARGETplus SMARTpool siRNAs were purchased from Dharmacon and contain a pool of four siRNA sequences directed against the target gene. HUVECs were grown for 24 h prior to transfection with 10 nm CLIC1 (L-009530-00), integrin β1 (L-004506-00), integrin β3 (L-004124-00), or nontargeting control (D-001810-10) siRNA. Cells were transfected in Opti-MEM medium (Thermo Fisher Scientific) using Lipofectamine 2000 reagent (Thermo Fisher Scientific) according to the manufacturers’ instructions. After 6 h, cells were placed in normal culture medium and grown for an additional 42 h.

Microslides were coated with 10 μg/mL fibrinogen or fibronectin (Sigma Aldrich) overnight at 4°C. Collagen I-coated slides (Biocoat) were purchased from Corning. Serum-coated microslides were generated by incubating collagen I-coated slides with EBM-2 endothelial growth media overnight at 4°C. Endothelial cells or platelets were allowed to adhere to culture slides for 10 min to 4 h, then fixed in ice cold 4% paraformaldehyde, permeabilized with 0.2% Triton ×100, and incubated with anti-CLIC1 (Abcam) or isotype control (Millipore), followed by incubation with Alexa Fluor 488-conjugated secondary antibody (Thermo Fisher Scientific) to assess subcellular localization of CLIC1. To visualize the cytoskeleton, slides were incubated with Alexa Fluor 546-phalloidin (Thermo Fisher Scientific) and analyzed by fluorescence microscopy on a Nikon Eclipse Ni fluorescent microscope equipped with NIS Elements software (Nikon). The staining of F-actin with Alexa Fluor 546-phalloidin allowed us to identify spread endothelial cells based on the presence of lamellipodia, filopodia, or actin stress fibers. Spread endothelial cells were scored based on the presence of these features as a percentage of all cells.

Healthy blood donors were recruited from our blood donation service. Approval was received from the Local Ethics Committee (Kenn-Nr. 73/14, Ethik-Kommission Ärztekammer Saarland), and blood was drawn after receiving informed consent. Briefly, venous blood was drawn with a 20-gauge needle into S-Monovette^® Citrate tubes (Sarstedt) and centrifuged at 100 g for 20 min at room temperature. The platelet-rich plasma supernatant was collected and centrifuged at 800 g for 3–8 min, and the resulting platelet pellet was washed with a 5 mm EDTA/0.9% NaCl solution. Following centrifugation, the pellet was washed a second time in HEPES-Tyrode buffer (HTB; 10 mm HEPES, 140 mm NaCl, 2.7 mm KCl, 0.4 mm NaH₂PO₄, 10 mm NaHCO₃, and 5 mm dextrose), pH 6.5, centrifuged, and resuspended in HTB (pH 7.4) containing 1.25 mm CaCl₂ and 6.25 mg/mL bovine serum albumin.

Platelet aggregation was measured using a Born aggregometer (APACT 4SPlus, Rolf Greiner Biochemica) at 37°C under stirring conditions (700–1000 U/min). Washed platelets were preincubated for 5 min with IAA94 (200 µm; Sigma Aldrich) or vehicle (DMSO) as a control. To induce aggregation, platelets were stimulated with 500 µg/mL arachidonic acid (MöLab) or 5 µm ADP (DiaSYS) in the presence of 0.5 mg/mL fibrinogen (Enzyme Research Laboratories), and the change in light absorption was measured for 5 min. Pure HTB (pH 7.4) and unstimulated platelets were used to calibrate the aggregometer to 100% and 0% aggregation, respectively. Platelet spreading was assessed by scoring spread and unspread platelets by phase-contrast microscopy using a Zeiss Primo Vert microscope and Zen light 2012 software.

Washed platelets were activated with arachidonic acid (500 µg/mL), ADP (5 µm), and/or fibrinogen (0.5 mg/mL) for 15 min in the presence of CLIC1 antibody (Abcam) or control IgG (Millipore). Platelets were then fixed for 15 min with room temperature 1% paraformaldehyde, washed with HTB, and centrifuged at 300 g for 5 min. Pellets were incubated with Alexa Fluor 488-conjugated secondary antibody (Thermo Fisher Scientific) for 15 min and washed to remove unbound antibody. Antibody binding was analyzed using a FACSCalibur flow cytometer (Becton Dickinson) equipped with CellQuest software. Where indicated, platelets were pretreated with IAA94 (200 µm), EDTA (1 mm), RGD or RAD (ca. 300 µm; EMD Chemicals), or vehicle control for 15 min prior to activation with ADP. To determine platelet activation, unfixed platelets were incubated with FITC-labeled PAC-1 (BD Biosciences) or FITC IgM Control (BD Biosciences). Integrin β3 and integrin αIIb total expression were monitored by flow cytometry using PE-CD61 (BD Biosciences) and PE-CD41 (Beckman Coulter) antibodies, respectively.

BALB/c mice (body weight 22–25 g) were housed in a temperature-controlled environment under a 12 h/12 h light-dark cycle and fed a standard pellet diet (Altromin, Lage, Germany) and water ad libitum. All experiments were approved by the Local Governmental Animal Welfare Committee (Landesamt für Verbraucherschutz, Abteilung C Lebensmittel- und Veterinärwesen, Saarbrücken, Germany; Permit Number: 15/2014) and were conducted in accordance with the European legislation on protection of animals (Guideline 2010/63/EU) and the NIH Guidelines for the Care and Use of Laboratory Animals (http://oacu.od.nih.gov/regs/index.htm. 8th Edition; 2011).

The dorsal skinfold chamber model was used to investigate the effect of IAA94 on photochemically induced thrombus formation. For this purpose, 12 BALB/c mice were anesthetized with an intraperitoneal injection (i.p.) of 75 mg/kg ketamine and 15 mg/kg xylazine and dorsal skinfold chambers were implanted, as previously described [38]. After 72 h of recovery, mice were treated with 20 mg/kg IAA94 i.p. (n = 6) or vehicle i.p. (DMSO; n = 6) 19 h and 1 h before photochemically induced thrombus formation. For in vivo microscopic analysis, mice were immobilized on a plexiglas plate and the dorsal skinfold chamber was attached to the microscopic stage. Following intravenous injection of 0.05 mL 5% FITC-labeled dextran 150,000 for contrast enhancement of the intravascular blood plasma, intravital epi-illumination fluorescence microscopy was performed using a modified Zeiss Axiotech microscope (Zeiss) with a 100-W mercury lamp attached to a blue (excitation wavelength: 450–490 nm/emission wavelength: >515 nm) and a green (530–560 nm/>585 nm) filter block. The microscopic images were recorded by a charge-coupled device video camera (FK6990; Pieper) and transferred to a monitor (Trinitron) with recording system (DVD-HR775; Samsung) for off-line evaluation. Using a ×20 long distance objective (Achroplan 0.50 W; Zeiss) baseline blood flow was monitored in individual venules (diameter range: 20–40 μm; n = 5 per chamber). Subsequently, thrombus formation was photochemically induced by continuous local exposure of the vessels to filtered light (450–490/>520 nm excitation/emission wavelength) with a ×63 water immersion objective (Achroplan 0.95 W; Zeiss) [39, 40]. Quantitative off-line analysis of the microscopic images was performed using CapImage software (Zeintl). Diameters, centerline red blood cell velocity and wall shear rate were determined in venules prior to thrombus induction. Diameters (d) were measured perpendicularly to the vessel path. Centerline red blood cell velocity (v, given in µm/s) was measured using the line shift method [41], and the wall shear rate (y, given in s^-1) was calculated based on the Newtonian definition: y = 8 *v/d. The kinetics of thrombus formation was analyzed by measuring the time (given in s) until sustained cessation of blood flow due to complete thrombotic vessel occlusion.

Significant differences were determined by Student’s two-tailed t-test or one-way ANOVA followed by the post hoc Tukey’s multiple comparisons test (GraphPad Prism 5). A p < 0.05 was considered to indicate significant differences. Data are given as mean ± SEM.

We recently demonstrated that CLIC1 promotes invasion in clot-embedded tumor cells in cooperation with integrin β3 [9]. To assess the role of CLIC1 in endothelial cell adhesion, we analyzed F-actin staining by fluorescence microscopy in siCLIC1-transfected HUVECs 1 h after attachment on serum-coated microslides. This experiment revealed a significant reduction in cell spreading (Fig. 1a, b). To further assess the function of CLIC1 in endothelial cell spreading, we analyzed the subcellular localization of CLIC1 in native HUVECs for up to 4 h after adhesion to serum-coated microslides. Immunohistochemistry and subsequent fluorescence microscopy of endothelial cells attached for 10 min showed strong expression of CLIC1 in the cell margins in close vicinity of F-actin-rich lamellipodia (Fig. 2c, d). The colocalization between CLIC1 and F-actin in the cell periphery was short lived and dissolved over the next few hours with CLIC1 being diffusely expressed in the cytoplasm and F-actin forming large stress fibers as a sign of maturing adhesions. The distribution of CLIC1 to the cell margins during early attachment was significantly reduced, when we treated endothelial cells with the synthetic CLIC1 inhibitor IAA94 or with siRNA to inhibit the expression of integrin β3, whereas knocking down a related cell adhesion receptor, integrin β1, had no effect (Fig. 2e, online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000544115). This process was relevant as inhibiting CLIC1 expression or its relocation to the cell margins significantly diminished the generation of F-actin in endothelial lamellipodia and, therefore, lamellipodia formation itself (Fig. 2f). Together, these results indicate that integrin β3-mediated engagement leads to relocation of CLIC1 to the cell periphery and, this, in turn, promotes assembly of F-actin in the leading edge of endothelial cells.

As endothelial cells and platelets share similar adhesive mechanisms based on the prominent expression of integrin β3 [42], we tested CLIC1 function in platelets. Paralleling our data in endothelial cells, we established that CLIC1 colocalized with cortical F-actin in the leading edge of newly attached platelets (Fig. 2a). CLIC1 expression in the platelet margins reached its maximum after 10 min of adhesion and was strongest after attachment to fibronectin, followed by collagen and fibrinogen (Fig. 2b). To determine CLIC1 membrane expression in nonadherent cells, we performed flow cytometry on fresh, intact platelets in suspension (Fig. 3a, b). In line with our concept that CLIC1 membrane expression is adhesion-dependent, we detected only minimal amounts of CLIC1 on the surface of resting platelets. However, CLIC1 cell surface expression was strongly induced after treatment with platelet agonists known to activate platelet integrin α_IIbβ₃ such as arachidonic acid or ADP and further enhanced when we added ADP in combination with the integrin α_IIbβ₃ ligand fibrinogen (Fig. 3c, d). Inhibiting integrin binding with a small peptide containing the integrin-binding motif RGD, on the other hand, abolished CLIC1 expression on the platelet surface, whereas the inactive RAD peptide had no relevant effect (Fig. 3e). Collectively, these results indicate that stimulation of platelet integrin β3 promotes CLIC1 membrane expression.

To define the role of CLIC1 in platelet adhesion, we utilized the inhibitory function of the synthetic CLIC1 antagonist IAA94 [43]. Pretreating platelets with 200 µm IAA94 significantly reduced the recruitment of CLIC1 into the periphery of platelets attached to collagen (Fig. 4a, b). This negative effect of IAA94 on CLIC1 membrane expression was associated with a diminished capacity of platelets to spread on collagen (Fig. 4c, d). To determine if CLIC1 inhibition also affects platelet aggregation, we performed light transmission aggregometry in the presence of IAA94. Notably, inhibiting CLIC1 with IAA94 caused a significant reduction of platelet aggregation in response to arachidonic acid as well as ADP (Fig. 4e, f). This result was matched by reduced PAC1 binding in flow cytometry as a sign of diminished integrin α_IIbβ₃ activation (Fig. 5a). IAA94 treatment had no effect on cell surface expression of total α_IIbβ₃ integrin or Annexin V binding in activated platelets (Fig. 5b–d). Together, these data indicate that CLIC1 promotes platelet activation.

To determine if the inhibition of CLIC1 is antithrombotic in vivo, we assessed thrombus formation in response to a photochemically induced vascular injury in a mouse dorsal skin fold chamber model (Fig. 6a). Monitoring postcapillary venules with intravital microscopy revealed that vaso-occlusion time was significantly delayed in mice treated with two subsequent doses of IAA94 (20 mg/kg i.p.) 19 and 1 h before inducing vascular injury (Fig. 6b). These data demonstrate that inhibiting CLIC1 in vivo has significant antithrombotic effects.

We previously showed that CLIC1 promotes tumor cell invasion [9]. Here, we demonstrate that CLIC1 supports endothelial spreading and platelet aggregation. Mechanistically, we demonstrate that interactions with integrin β3 lead to the recruitment of CLIC1 into lamellipodia of newly attached endothelial cells, while RGD-dependent interactions with platelet integrins cause cell surface expression and recruitment of CLIC1 into the leading edge of newly attached platelets. In addition, we identified a role of CLIC1 in the activation of integrin β3 on platelets, suggesting that the integrin-mediated recruitment of CLIC1 into the cell membrane generates an adhesive feed forward loop. This mechanism appears to be important for prothrombotic functions in vitro as well as in vivo and could have practical implications for measuring platelet activation.

CLIC1 expression is particularly pronounced in tumor cells and the angiogenic endothelium [9, 25, 32, 35, 44]. In line with this pattern of overexpression, CLIC1 has been shown to be associated with cell functions typically associated with tumor progression and angiogenesis such as cell invasion, survival, and proliferation [9, 32, 44]. Moreover, targeting CLIC1 on endothelial cells with the CLIC1-binding peptide CLT1 had a pronounced antiangiogenic effect in vivo, as CLIC1 mediates the uptake of CLT1 into proliferating endothelial cells [44]. CLIC1-mediated endocytosis of the CLT1 peptide depended on complexation of CLT1 with fibronectin, whereby interaction of the fibronectin part of the complex with integrin β3 promoted cell surface expression of CLIC1, binding of CLT1, and subsequent internalization of the CLT1-fibronectin complex. Together, these data suggest a close coordination between cell adhesion and CLIC1 function, which we confirm now by demonstrating that CLIC1 is involved in endothelial cell spreading and platelet aggregation.

Cooperation with integrins is a defining feature of CLIC family members as they promote adhesive functions in a diverse array of cells [9, 45, 46]. As such, CLIC3 and CLIC4 have been shown to promote tumor cell adhesion, invasion, and survival by recycling activated integrins to the cell surface [7, 46]. This mechanism is associated with an aggressive tumor phenotype and reduced overall survival in ovarian, pancreatic, and breast cancer [7, 46]. While CLIC3 is particularly important in cancer types that depend on β1 integrin activation, we identified a close association between CLIC1 and integrin β3 in cancer types that benefit from clotting during lung metastasis [9, 47]. The coordination of CLIC1 with β3 integrin also became evident in endothelial cells and platelets, where inhibiting CLIC1 expression with siRNA or CLIC1 function with the synthetic inhibitor IAA94 leads to an impairment of integrin β3-mediated functions such as endothelial spreading and platelet aggregation. Moreover, targeting CLIC1 on platelets has significant antithrombotic effects in vivo as demonstrated by the delay in venous occlusion time in the mouse dorsal skin fold chamber model in response to photochemically induced vascular injury after treatment with IAA94.

CLIC1 was originally discovered as a nuclear chloride channel but has since been shown to be expressed in mainly two forms, a cytosolic monomer and a membrane-bound oligomer that inserts in the plasma membrane as well as the membrane of phagosomes [8, 24, 32]. The membrane form can be induced through the oxidation of the active site cysteine at position 24, which leads to a cascade of conformational changes that precede the oligomerization of the protein and its membrane insertion [24]. This process seems to take place in glia cells after exposure to amyloid β, which upon binding to the cell membrane induces reactive oxygen species production and promotes CLIC1 membrane insertion [27]. Alternatively, it has been shown that conformational change and membrane insertion of CLIC1 can be stimulated through the release of intracellular Zn²⁺ ions, which causes activating effects in platelets similar to those of CLIC1 [12, 28, 48]. Notably, intracellular Zn²⁺ ions are typically bound to reduced thiols and, therefore, are released into the cytoplasm upon oxidation, suggesting that the generation of reactive oxygen species can drive CLIC1 membrane insertion directly as well as indirectly [49].

The coupling of mechanotransduction with redox signaling plays a significant role in endothelial cells and platelets, which are continuously exposed to blood flow-dependent shear stress [50]. In line with this, we found significant coordination between integrin β3 ligation and CLIC1 membrane recruitment that was particularly pronounced in nascent lamellipodia during the early phase of endothelial attachment. In contrast, colocalization of CLIC1 with F-actin in the membrane was lost at later time points, when endothelial attachment was less dynamic and more dominated by large stress fibers. Recruitment of CLIC1 into lamellipodia was also evident in platelets, where it was strongest after attachment to fibronectin, which mediates platelet spreading exclusively through ligation of β3 integrins [51]. Flow cytometry revealed that relocation of CLIC1 to the cell membrane of platelets was contingent on platelet activation and further enhanced by subsequent RGD-dependent adhesion to specific ligands of integrin β3 such as fibrinogen [52, 53]. Resting platelets, on the other hand, were negative for transmembrane CLIC1. Therefore, our results confirm previous confocal microscopy data, demonstrating that platelets only express CLIC1 in the cell membrane after conversion from a resting to an activated state [12]. Moreover, they suggest that the activation-dependent interaction of β3 integrins on platelets with soluble fibrinogen is an important inducer of CLIC1 membrane relocation.

Recruitment of CLIC1 into the platelet membrane is functionally significant as we found in platelets pretreated with the synthetic CLIC1 inhibitor IAA94 reduced binding of PAC1, an antibody that specifically binds to the high affinity conformation of platelet integrin α_IIbβ₃. These data indicate that CLIC1 contributes to the activation of integrin α_IIbβ₃, which in turn is a prerequisite of platelet spreading and aggregation [51, 54]. Accordingly, we detected a significant inhibition of platelet spreading as well as platelet aggregation following incubation with the CLIC1 inhibitor IAA94, which conforms with previous data showing reduced spreading and aggregation in platelets isolated from CLIC1 knockout mice [12]. The work derived from CLIC1 knockout platelets suggests a role of CLIC1 in ADP signaling, which is in line with our data as ADP is a known inducer of integrin α_IIbβ₃ activation [55]. However, platelet adhesion to immobilized ligands such as fibronectin or fibrin can be initiated through resting integrins and, therefore, does not require additional soluble agonists, suggesting that CLIC1 acts downstream of ADP on the level of β3 integrin-mediated adhesion or cohesion [51].

The original concept of CLIC1 function was deduced from its ability to generate chloride currents, which are thought to be associated with vesicle acidification and the generation of intracellular reactive oxygen species [8, 27, 56]. This mechanism could be relevant for platelet function as extracellular chloride ions have been shown to affect platelet activation [20]. In addition, there is strong evidence that CLIC1 supports integrin function through either activation of the integrin adaptor protein talin or Rho GTPase signaling [57, 58]. While the exact mechanism of CLIC1 function remains unknown, it has become apparent that CLIC1 fulfills an important purpose in endothelial cells and platelets by modulating integrin β3-related functions. Accordingly, CLIC1 was involved in platelet aggregation in vitro and thrombus formation in vivo. Therefore, our data suggest that further research into the exact function of CLIC1 as an effector of platelet function is warranted.

Study approval was received from the Local Ethics Committee (Kenn-Nr. 73/14, Ethik-Kommission Ärztekammer Saarland). Blood was collected after receiving informed consent.

The authors have no potential conflict of interest.

This work was supported by a HOMFOR Pilot Project Grant of the Saarland University Medical School (JP).

L.M. Knowles, E. Ampofo, M.D. Menger, H. Eichler, and J. Pilch designed the research; L.M. Knowles, E. Ampofo, and A. Drawz performed the experiments; L.M. Knowles, E. Ampofo, and J. Pilch analyzed the data; and L.M. Knowles and J. Pilch wrote the manuscript.

All data generated or analyzed during this study are included. Further inquiries can be directed to the corresponding author.