Role of the NO-GC/cGMP signaling pathway in platelet biomechanics

Abstract

Cyclic guanosine monophosphate (cGMP) is a second messenger produced by the NO-sensitive guanylyl cyclase (NO-GC). The NO-GC/cGMP pathway in platelets has been extensively studied. However, its role in regulating the biomechanical properties of platelets has not yet been addressed and remains unknown. We therefore investigated the stiffness of living platelets after treatment with the NO-GC stimulator riociguat or the NO-GC activator cinaciguat using scanning ion conductance microscopy (SICM). Stimulation of human and murine platelets with cGMP-modulating drugs decreased cellular stiffness and downregulated P-selectin, a marker for platelet activation. We also quantified changes in platelet shape using deep learning-based platelet morphometry, finding that platelets become more circular upon treatment with cGMP-modulating drugs. To test for clinical applicability of NO-GC stimulators in the context of increased thrombogenicity risk, we investigated the effect of riociguat on platelets from human immunodeficiency virus (HIV)-positive patients taking abacavir sulfate (ABC)-containing regimens. Our results corroborate a functional role of the NO-GC/cGMP pathway in platelet biomechanics, indicating that biomechanical properties such as stiffness or shape could be used as novel biomarkers in clinical research.

Plain Language Summary

Increased platelet activation and development of thrombosis has been linked to a dysfunctional NO-GC/cGMP signaling pathway. How this pathway affects platelet stiffness, however, has not been studied yet. For the first time, we used novel microscopy techniques to investigate stiffness and shape of platelets in human and murine blood samples treated with cGMP modifying drugs. Stiffness contains information about biomechanical properties of the cytoskeleton, and shape quantifies the spreading behavior of platelets. We showed that the NO-GC/cGMP signaling pathway affects platelet stiffness, shape, and activation in human and murine blood. HIV-positive patients are often treated with medication that may disrupt the NO-GC/cGMP signaling pathway, leading to increased cardiovascular risk. We showed that treatment with cGMP-modifying drugs altered platelet shape and aggregation in blood from HIV-negative volunteers but not from HIV-positive patients treated with medication. Our study suggests that platelet stiffness and shape can be biomarkers for estimating cardiovascular risk. 

Keywords:

• Cellular stiffness
• cyclic guanosine monophosphate (cGMP)
• inhibition of platelet activation
• NO-sensitive guanylyl cyclase (NO-GC) enzyme
• platelet shape

Introduction

Platelets are small anucleate blood cells responsible for critical roles in hemostasis and thrombosis.¹ Upon activation, platelets release inflammatory cytokines and become highly adhesive with a drastic change in shape.² The cytoskeleton generates contractile forces and essentially contributes to mechanotransducive and haptotactic mechanisms and platelet motility.³ Therefore, measuring cellular stiffness (the amount of resistance of an object against deformation in response to an applied force) could help in the assessment of the bleeding risk in patients with cytoskeletal disorders.⁴ Inherited platelet disorders (IPDs) could potentially be diagnosed based on their cytoskeleton changes and platelet stiffness.⁴ The present study shows, for the first time, the pharmacological effects of cGMP-modulating drugs upregulating the NO-GC/cGMP pathway in platelets on platelet stiffness, suggesting that a decrease in platelet stiffness could be beneficial from a therapeutic point of view.

Nitric oxide (NO) is an endogenous platelet inhibitor that binds to NO-sensitive guanylyl cyclase (NO-GC). In platelets, NO-GC has been shown to be the only NO receptor. NO-GC catalyzes the conversion of guanosine 5’-triphosphate (GTP) to cyclic guanosine 3’,5’- monophosphate (cGMP). An increase in cGMP activates the cGMP-dependent protein kinase (PKG), which phosphorylates, among other proteins, vasodilator-stimulated phosphoprotein (VASP) in platelets and other cell types. NO is involved in signaling cascades that support a healthy cardiovascular system including inhibition of platelet aggregation in vitro⁵ and in vivo,⁶ making NO-GC an attractive target for cardiovascular pathologies including thrombosis⁷; however, continuous nitrate administration can lead to drug resistance because of the desensitization of NO-GC.⁸ In a cardiovascular pathology such as ischemia, heme, the essential NO-GC co-factor becomes oxidized, preventing NO from binding NO-GC and impairing cGMP generation.^9,10 Riociguat is a NO-GC stimulator with a dual effect on NO-GC: it stimulates the enzyme itself and decreases the dissociation of NO from NO-GC.¹¹ Cinaciguat is a NO-GC activator that can activate NO-GC independently of its heme redox status (heme-oxidized or heme-free NO-GC) and irrespective of impaired NO signaling.¹²

We examined the effects of riociguat and cinaciguat on platelet biomechanics using scanning ion conductance microscopy (SICM), a noninvasive imaging technique that allows simultaneous measurement of morphological and mechanical properties in living cells.^13,14 SICM has already been used to investigate the stiffness of living platelets¹⁵ and to show that reduced platelet stiffness is related to increased bleeding in MYH9-related disease.¹⁶ Measuring stiffness brings additional biomechanical information about the functional structure and crosslinking of the platelet cytoskeleton,^15,17 which cannot be obtained by just quantifying cytoskeletal components like actin. We correlated SICM stiffness measurements with platelet shape (circularity) investigated by deep learning platelet morphometry¹⁸ in human and wild-type (C57Bl6/J) or megakaryocyte/platelet-specific NO-GC knockout (KO) murine platelets.

Finally, we studied the effect of riociguat on platelet circularity in human immunodeficiency virus (HIV)-positive patients taking the anti-viral drug abacavir-sulfate (ABC). Carbovir triphosphate (CBV-TP), the active ABC anabolite, is a guanosine derivative shown to compete and antagonize cGMP in the NO-GC/cGMP signaling pathway in platelets.¹⁹ Therefore, the potential blockage of the NO-GC/cGMP pathway by CBV-TP could increase the risk of thrombogenic events and lead to an increased development of cardiovascular pathologies in HIV patients taking ABC.¹⁹

Materials and methods

Animals

Platelets from 3 to 12 months old wild-type (C57Bl6/J) and megakaryocyte/platelet specific NO-GC KO mice with a Pf4-Cre (wt/Cre); NO-GC (flox/flox) genotype were used.²⁰ NO-GC KO mice were generated by crossing the Pf4-Cre line (B6-Tg(Cxcl4-cre)Q3Rsko/J)²¹ to the NO-GC (flox/flox) line (B6.129-Gucy1b3tm1.2Frb).²² Homologous recombination in bacteria was used to generate the Pf4-Cre transgene construct (for detailed experimental procedures²¹). The Cre/loxP system was used to generate KO mice, where the Exon 10 of the β1 subunit was floxed after embryonic stem cell transfer.²² The null mutation of the β1 subunit, which leads to the complete loss of NO-GC, was confirmed by PCR²² and Western blot analysis (Figure 1i). Mouse genotyping was done for all animals before all experiments to confirm their genotype. Experiments have been approved by the local authority (Regierungspräsidium Tubingen, IB 02/20 M) and are reported in accordance with the ARRIVE guidelines.²³

Figure 1. Cellular stiffness is decreased in human platelets and wild-type (C5BI6/J) murine platelets when treated with 8-Br-cGMP, riociguat or cinaciguat. (a,d,g) topography and cellular stiffness images of washed human, wild-type murine, and platelet-specific NO-GC KO murine platelets. (b,e,h) cellular stiffness quantification. (c,f,i) Western blot analysis of VASP phosphorylation (p-VASP) and NO-GC. GAPDH was used as loading control. DMSO (1:1000), 8-Br-cGMP (1 mM), riociguat (10 μM), cinaciguat (10 μM). The significance level of P-values is indicated by asterisks (* p < .05; ** p < .01; *** p < .001; ns, not significant; Tukey’s test). Scale bars: 5 μm. SICM experiments were performed with 4–24 platelets per donor from n = 2 (human), n = 2 (wild-type mice), n = 2 (NO-GC KO mice) donors. Western blots were performed for n = 3 (human), n = 3 (wild-type mice), n = 4 (NO-GC KO mice) donors.

Isolation of human platelets

All procedures were approved by the institutional ethics committees (Medical Faculty and University Clinics at the University of Tübingen, 273/2018BO2 and 064/2022BO2 and Imperial College London Research Ethics Committee) in accordance with the declaration of Helsinki. Informed consent was obtained from all participants. Venous blood was collected from the antecubital vein of healthy volunteers in acid-citrate-dextrose (ACD) anticoagulant at 1:4 ratio (ACD:blood) at room temperature (RT). The time between blood collection and isolation, the storage conditions, and the isolation procedure are pre-analytical variables that can potentially affect results. To keep these variables as constant as possible the following procedure was followed. Blood was stored at 37°C under continuous rotation. Platelets were isolated within 4 h after collection. Monovettes containing 4.8 mL of blood were centrifuged at 200× g for 20 min at RT. Tyrode-HEPES buffer (6 mM HEPES, 136.9 mM NaCl, 12.1 mM NaHCO₃, 2.6 mM KCl, 5.5 mM D-glucose), pH 6.5 (adjusted with HCl) was added to the platelet-rich plasma (PRP) at a ratio of 1:3 and centrifuged at 920× g for 10 min, without brake, at RT. The supernatant was discarded, and the platelet pellet was carefully re-suspended in 1 mL of Tyrode-HEPES buffer, pH 7.4 (adjusted with HCl) and immediately used for experiments described below.

In addition, platelets from HIV-positive patients taking ABC-containing regimens registered at Chelsea and Westminster Hospital NHS Trust were obtained in accordance with research ethics permits 294 707 21/NW/0148 approved by the NHS Health Research Authority and Chelsea and Westminster Hospital NHS Trust. All experiments were performed in accordance with the Declaration of Helsinki. Briefly, blood was collected from consented HIV-negative volunteers and HIV-positive patients taking ABC-containing regimens by venipuncture in VACUETTE tubes (Greiner Bio-One Ltd, United Kingdom) containing 3.2% (v/v) sodium citrate. Platelets from HIV-positive patients were isolated immediately after blood collection. The whole blood was centrifuged at 175× g for 15 min to obtain PRP. Washed platelets were obtained by the addition of 150 µL of citrate-dextrose solution (C3821, Sigma-Aldrich, Dorset, UK) and 5 µL of prostaglandin E1 (PGE₁) (P5515; Sigma-Aldrich, Dorset, UK) to the PRP, mixing by inversion and centrifuging at 1400× g for 10 min. Platelet poor plasma was discarded, and the platelet pellet was re-suspended in a total volume of 20 mL of Tyrode-HEPES buffer (10 mM HEPES, 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 5 mM glucose, 0.42 mM NaH₂PO₄, 12 mM NaHCO₃, pH 7.4). Then, 3 mL of ACD and 5 µL of PGE₁ were added and the tube was centrifuged again at 1400× g for 10 min. Washed platelets were obtained by re-suspending the pellet in 1 mL of Tyrode-HEPES buffer and immediately used for microscopy and aggregometry experiments.

Isolation of murine platelets

Murine blood was withdrawn from the retro-orbital sinus of wild-type and megakaryocyte/platelet-specific NO-GC KO mice under isoflurane anesthesia into a capillary with ACD buffer (85 mM sodium citrate, 72.9 mM citric acid, 110 mM D-glucose). A total of 300 μL of platelet wash buffer (PWB) solution (4.3 mM K₂HPO₄, 4.3 mM Na₂HPO₄, 24.3 mM NaH₂PO₄, 5.5 mM D-glucose, 113 mM NaCl, pH 6.5), supplemented with 0.1% (w/v) bovine serum albumin (BSA), were added into a plastic test tube where the blood was collected and then centrifuged at 200× g for 2 min, without brake, at RT. Collected PRP was transferred into a fresh tube, while 600 μL of PWB solution was added to the remaining blood and centrifuged for a second time, with the same settings. Platelet supernatants from the two centrifugations were combined, and the PRP obtained after centrifugation at 2000 × g for 1 min, without brake, at RT. Murine platelets were re-suspended in 600 μL Tyrode-HEPES buffer (10 mM HEPES, 137 mM NaCl, 12 mM NaHCO₃, 2.7 mM KCl, 5.5 mM D-glucose, pH 7.4; supplemented with 0.1% (w/v) BSA) and immediately used for experiments described below.

Drug treatments for human and murine platelets

Thirty-five mm cell culture dishes with a glass bottom (81218, ibidi, Gräfelfing, Germany) were coated with 0.1 mg/mL fibrinogen (F3979, Sigma Aldrich, St. Louis, MO, USA) for 30 min at 37°C. Tyrode-HEPES buffer was supplemented with CaCl₂ (1 mM) (10043-52-4, Sigma Aldrich) and MgCl₂ (1 mM) (7786-30-3, Sigma Aldrich) to wash the platelets after spreading. NO-GC stimulator riociguat (10 μM) (9000554, Cayman, Ann Arbor, MI, USA), and NO-GC activator cinaciguat hydrochloride (10 μM) (SML1532, Sigma Aldrich), [1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) (10 μM) (495320, Sigma-Aldrich, Dorset, UK), carbovir-5'-triphosphate triethylammonium salt (CBV-TP) (50 μM) (443952, US Biologicals, Salem, MA, USA) were dissolved in dimethyl sulfoxide (DMSO) (472301, Sigma Aldrich). 8-Br-cGMP (1 mM) (B1381, Sigma Aldrich) and adenosine 5”-diphosphate sodium salt (ADP) (3 μM) (20398-34-9, Sigma Aldrich) were dissolved in phosphate buffered saline (PBS) before addition to the platelets for 10 min at 37°C.

For SICM measurements and morphometric analysis, 50 μL of pre-treated platelets were added to the fibrinogen-coated cell culture dishes containing human or murine supplemented Tyrode-HEPES buffer and used as controls. Platelet adhesion and spreading were allowed to occur for 15 min at 37°C, followed by three washes with human or murine supplemented Tyrode-HEPES buffer to remove non-adherent platelets before imaging.

SICM imaging

Cell culture dishes were installed in two custom-built SICM setups.²⁴ Live imaging of platelets was done within 1 h from the collection because of their short life span. Individual isolations were performed for each drug treatment for each SICM measurement. Borosilicate nanopipettes with an inner radius of about 90 nm were fabricated using a CO₂-laser-based micropipette puller (p-2000; Sutter Instrument, Novato, CA, USA). Topography and cellular stiffness images were obtained with a constant pressure of 10 kPa applied to the upper end of the pipette to allow for quantitative cellular stiffness measurements. The local cellular stiffness (Young’s modulus) was calculated from the slope of the ion current vs. distance curve (IZ-curve) between 99% and 98% of the saturation current (where the saturation current is the constant current at large tip-sample distances) for each pixel, using a model based on finite element calculation.²⁵ Imaging was done at a 25 Hz pixel rate, with 30 × 30 or 65 × 65 pixels at a scan size of 15 × 15 µm². The cellular stiffness value of a single platelet was obtained by calculating the median value of each cellular stiffness map within the area of the platelet.

Western blot analysis

Blood was collected from healthy humans and wild-type or genetically modified mice. The platelets were isolated and then re-suspended in human or murine Tyrode-HEPES buffer after isolation. DMSO (1:1000), 8-Br-cGMP (1 mM), riociguat (10 µM), cinaciguat (10 µM) were added to 100 µL of platelet suspension and pre-incubated for 10 min at 37°C. Pre-treated platelets with the different drugs were centrifuged at 2000 × g for 1 min and then the supernatant was removed. Platelets were re-suspended in a lysis buffer (21 mM Tris pH 8,3; 0.67% (w/v) SDS, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mL H₂O), followed by freezing in liquid nitrogen and thawing. Protein lysates were separated by SDS-PAGE and proteins of interests detected with the following antibodies: phosphorylated (p-VASP) Ser239 (rabbit, 1:1000, Cell Signaling #3114S), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (rabbit, 1:1000, 14C10, Cell Signaling #2118S), NO-GC (sGC β1, rabbit, 1:100, Abcam 24 824).

Microplate aggregation assay

PRP samples from HIV-negative volunteers and HIV-positive patients taking ABC-containing regimens were pre-incubated with riociguat (10 µM), ODQ (20 µM), DMSO (1:1000), or the different drug combinations for 30 min at 37°C. PRP samples from HIV-negative volunteers were pre-incubated with CBV-TP (50 µM). An absorbance microplate reader (Sunrise, Tecan UK Ltd., UK) was used to read the light absorbance values in a 96-well plate (VWR; Leicestershire; UK) containing ADP (3 µM) and the pre-incubated PRP. Platelet aggregometry readings were performed at 20 s intervals for 16 min at 37°C, shaking for 7 s before each reading. The maximum aggregation values at minute 16 were obtained by calculating the percentage change in comparison to the baseline.

Flow cytometry assay

PRP samples were incubated for 20 min with different drug treatments (DMSO, 8-Br-cGMP, riociguat, or cinaciguat). PRP was centrifuged at 1000× g for 10 min at RT, and then the cells for each donor were counted with Sysmex to verify that the platelet concentration was about 300∙10³/μl. 10 μl PRP for all treatments, and 10 μl of CD62p-FITC (Beckman Coulter) per well were incubated for 25 min at RT. Then, 150 μl of PBS (Gibco 15 460 624) was added and centrifuged at 2000× g for 7 min at RT. The supernatants were removed, and the cells were resuspended in 80 μl freshly diluted 0.2% (v/v) paraformaldehyde in PBS and incubated for 10 min. A total of 100 μl of PBS was added, mixed gently, and transferred into flow cytometry tubes. At least 100.000 platelets were acquired at the Becton Dickinson Biosciences FACSCanto^TM II. Unstained controls for each donor were included. Data analysis was performed with FlowJo 10.8.1.

Deep learning platelet morphometry

Platelet circularity and area were analyzed using a convolutional neural network.¹⁸ Phase contrast images for platelet morphometry were recorded with an optical microscope (TiE, Nikon, Tokyo, Japan) on fixed platelets using a 100×/1.45 NA immersion oil objective for human and murine platelets. Phase contrast images of fixed and washed platelets from HIV-negative volunteers and HIV-positive patients taking ABC-containing regimens were taken at the FILM (Facility for Imaging by Light Microscopy) at Imperial College London. WF3 Zeiss Axio Observer, with a Zen Blue software with a 100×/1.4 Oil Ph3 Plan Apochromat and phase illumination for the transmitted light channel, and a Hamamatsu Flash 4.0 camera with a pixel size of 65 nm were used to record the images.

Statistics

Data were analyzed and processed in Igor Pro (WaveMetrics, Lake Oswego, Oregon, USA). Data are presented as median unless stated otherwise. All results were tested using Tukey’s test for parametric multiple comparisons, except for circularity and area data where Dunn’s test for non-parametric multiple comparisons was applied. The results are considered to be significantly different for p-values < .05.

Results

Platelet activation, stiffness, and shape changes were investigated in platelets isolated from humans and from wild-type and megakaryocyte/platelet-specific NO-GC KO mice.

Platelet stiffness is decreased by riociguat and cinaciguat

SICM was used to investigate the role of NO-GC in platelet stiffness. Topography and stiffness images of platelets were recorded with high spatial resolution (Figure 1). Human (Figure 1b) and wild-type murine (Figure 1e) platelets treated with cGMP-modulating drugs had a significantly decreased average cellular stiffness (≈2 kilopascal, kPa) in comparison to DMSO-treated platelets (≈5 kPa), indicating a decrease of ≈ 50% in cellular stiffness.

In contrast, NO-GC KO platelets (Figure 1h) showed no decrease in cellular stiffness after the treatment with riociguat or cinaciguat (≈3 kPa) in comparison with DMSO (≈3 kPa). A change in the cellular stiffness by ≈ 30% was only observed when NO-GC KO platelets were treated with 8-Br-cGMP (≈2 kPa). Furthermore, we found increased levels of phosphorylated VASP (p-VASP), a downstream effector of cGMP and regulator of F-actin polymerization,²⁶ in wild-type and human platelets for all the treatments (Figure 1c,f). In contrast, VASP phosphorylation was almost undetectable in NO-GC KO platelets treated with riociguat or cinaciguat, but is still detectable when NO-GC KO platelets are stimulated with 8-Br-cGMP (Figure 1i).

Efficient NO-GC deletion in platelets of megakaryocyte/platelet-specific NO-GC KO mice was confirmed via Western blot (Figure 1i). Taken together, these results suggest that NO-GC-mediated VASP phosphorylation may be linked to the regulation of cellular stiffness in platelets (Figure 1). Specific deletion of NO-GC in platelets corroborated the role of NO-GC in platelet stiffness treated with riociguat or cinaciguat, suggesting that stiffness can be modulated through the NO-GC/cGMP pathway.

Platelet shape is influenced by riociguat and cinaciguat

A convolutional neural network (CNN) was used to generate binary prediction images from optical-phase contrast images (Figure 2a,d,g) for the analysis of circularity (circ) and area of human, murine wild-type, and megakaryocyte/platelet-specific NO-GC KO murine platelets. Circularity is a unitless morphological parameter ranging between 0 and 1; a lower circularity indicates platelet shape irregularity, while a higher circularity represents a rounder shape.¹⁸ Platelets can become circular (circ ≈1) in a resting state or when they are fully spread and activated.¹⁸

Figure 2. Platelet shape is altered in human platelets and wild-type (C57Bl6/J) murine platelets when treated with 8-Br-cGMP, riociguat, or cinaciguat. (a,d,g) phase contrast and binary prediction images obtained via deep learning morphometry in washed human, wild-type murine, and megakaryocyte/platelet-specific NO-GC KO murine platelets. (b,e,h). platelet circularity. (c,f,i) platelet area. DMSO (1:1000), 8-Br-cGMP (1 mM), riociguat (10 μM), cinaciguat (10 μM). The significance level of P-values is indicated by asterisks (*p < .05; **p < .01; ***p < .001; ns, not significant; Dunn’s test). Scale bars: 5 μm. Experiments were performed with ≈ 15 platelets per donor from n=4 (human), n=4 (wild-type mice), and n=4 (NO-GC KO mice) donors.

When human platelets and wild-type murine platelets were treated with cGMP-modulating drugs, a significant increase in platelet circularity was observed (Figure 2b,e). However, the circularity of platelets from megakaryocyte/platelet-specific NO-GC KO mice did not change when treated with riociguat or cinaciguat, compared to DMSO (Figure 2h). This demonstrates that the platelet shape is likely modulated via NO-GC-mediated signaling.

Human platelets had a median area of about 25 µm², while the wild-type murine platelets and megakaryocyte/platelet-specific NO-GC KO murine platelets had a smaller median area of about 10 µm² (Figure 2c,f,i). However, no significant differences in area after the drug treatments were observed, except for riociguat-treated human platelets, where a significant decrease was observed (Figure 2 c,f,i).

Circularity and aggregation are decreased in riociguat-treated platelets from HIV-negative volunteers but not in riociguat-treated platelets from HIV-positive patients treated with ABC

The circularity of platelets from HIV-negative volunteers was significantly increased by ≈50% when treated with riociguat (circ ≈0.6) compared to DMSO (circ ≈0.3) (Figure 3a), consistent with our findings of riociguat-treated platelets from healthy human volunteers (Figure 2b). However, the circularity of platelets from HIV-negative volunteers did not change when treated with the active ABC anabolite CBV-TP (circ ≈0.3), compared to DMSO control (circ ≈0.3) (Figure 3a). Interestingly, the circularity of riociguat-treated platelets from HIV-positive patients taking ABC-containing regimens (circ ≈0.35) did not change compared to DMSO control (circ ≈0.35) (Figure 3b). Additionally, the circularity of the ADP-treated platelets from HIV-negative volunteers (circ ≈0.2) and HIV-positive patients taking ABC-containing regimens (circ ≈0.25) was ≈ 30% smaller than for DMSO, albeit not significantly different (Figure 3a, b). These findings could indicate a disruption of the thrombo/cardio-protective NO-GC/cGMP signaling pathway in HIV-positive patients treated with ABC-containing regimens.

Figure 3. Riociguat increases platelet circularity and decreases platelet aggregation in HIV-negative volunteers, but not in HIV-positive patients taking ABC-containing regimens. (a, b) platelet circularity. (c, d) maximum of ADP-induced aggregation in response to indicated drug treatments. ADP (3 µM), DMSO (1:1000), CBV-TP (50 µM), riociguat (10 µM), ODQ (20 µM). Platelet numbers for circularity ≈ 15 per donor from n=2 donors. Donors for aggregation experiments: n=3. The significance level of p-values is indicated by asterisks (* p < .05; ** p < .01; *** p < .001; ns, not significant; Tukey’s test).

Similarly, a microplate reader-based analysis of platelet aggregation showed decreased aggregation of platelets isolated from HIV-negative volunteers when treated with riociguat (≈20% for riociguat, ≈50% for DMSO; p = .028) (Figure 3c), but not in platelets isolated from HIV-positive patients taking ABC-containing regimens (≈55% for both riociguat and DMSO; p = .58) (Figure 3d). Platelets isolated from HIV-negative volunteers treated with CBV-TP showed a similar percentage of aggregation as platelets from HIV-negative volunteers treated with ODQ (≈40% for both CBV-TP and ODQ; p = 1) (Figure 3c). No pharmacological interaction of CBV-TP with ODQ was observed (Figure 3c). Likewise, ODQ showed no effect in platelet aggregation when pre-incubated together with riociguat in HIV-positive patients taking ABC-containing regimens (Figure 3d). Additionally, when comparing the different drug combinations (ODQ-CBV-TP for HIV-negative volunteers, ODQ-riociguat for both HIV-negative volunteers and HIV-positive patients taking ABC-containing regimens), the maximum ADP-induced aggregation was identical (≈40%), and no significant difference compared to aggregation values for DMSO (≈50% for HIV-negative volunteers, ≈55% for HIV-positive patients taking ABC-containing regimens) was detected.

Platelet activation is dampened by riociguat and cinaciguat

We examined the effect of the NO-GC stimulator riociguat and the NO-GC activator cinaciguat on platelet activation for human platelets. We found that the cGMP-modulating drugs 8-Br-cGMP, riociguat, or cinaciguat decreased the P-selectin signal measured by flow cytometry in isolated human platelets (Figure 4). Downregulation of P-selectin suggests an inhibitory role of the NO-GC/cGMP pathway in platelet activation.

Figure 4. P-selectin signal measured by flow cytometry is decreased in human platelets, when treated with 8-Br-cGMP, riociguat, or cinaciguat. DMSO (1:1000), 8-Br-cGMP (1 mM), riociguat (10 μM), cinaciguat (10 μM). The significance level of P-values is indicated by asterisks (*p < .05; **p < .01; ***p < .001; ns, not significant; Tukey’s test). Experiments were performed with n=3 donors, and each individual point is the average of three samples from the same donor.

Overall, these results indicate that P-selectin is modulated by the NO-GC enzyme and that an increase in cGMP leads to the downregulation of P-selectin.

Taken together, our findings show that activation of the NO-GC/cGMP pathway in platelets results in inhibition of platelet activation and interestingly correlates with a decrease in cellular stiffness and a decrease in platelet circularity (Figure 5). Riociguat did not exert any changes in platelet shape in platelets isolated from HIV-positive patients currently taking ABC (CBV-TP).

Figure 5. Role of the NO-GC/cGMP pathway in platelet biomechanics. addition of riociguat or cinaciguat leads to a decrease in platelet stiffness, circularity, and activation via the NO-GC/cGMP pathway.

Discussion

The goal of this study was to investigate mechanisms that mediate platelet biomechanics, which could lead to novel strategies for the treatment or diagnosis of thrombosis and related cardiovascular pathologies in a variety of clinical settings. Specifically, we sought to elucidate the role of the NO-GC/cGMP signaling pathway in human and murine platelets, with a focus on biomechanics.

Platelets store P-selectin in their alpha granules and upon platelet activation, the external membrane of the platelet exposes P-selectin.²⁷ We found that P-selectin was decreased in human platelets after treatment with different cGMP-modulating drugs (Figure 4), indicating that upregulation of the NO-GC/cGMP signaling pathway led to the inhibition of platelet activation. This change correlated with decreased platelet cellular stiffness (Figure 1a). The specificity of our findings was confirmed with platelets from platelet-specific NO-GC KO mice. Some studies have suggested that increased activity of cAMP-dependent kinase was related with inhibition of platelet activation and decreased thrombus formation.²⁸ Considering the interaction and resemblance between cAMP and cGMP signaling, the inhibition of platelet activation observed in both pathways is congruent.²⁸

Platelet activation induces changes in the platelet cytoskeleton and shape.²⁹ Circularity is mediated by actin-binding proteins,³⁰ and upon platelet activation, filopodia and lamellipodia are formed. Semi-automated analysis of platelet spreading in untreated human platelets spread on fibrinogen (area ≈20 µm²; circ ≈ 0.38) had similar values compared to the values presented in Figure 2b,c.³¹ Platelet morphological features such as circularity could define morphological subtypes such as not-spread, partially, or fully spread.³² Increased platelet circularity and decreased P-selectin expression were observed in platelets treated with cGMP-modulating drugs, which may show an intermediate state of platelet activation and spreading. Furthermore, circularity could be a potential indicator for platelet activation in an initial high-throughput screening. NO-GC has a basal, NO-independent activity, which would add to the NO-induced, endothelium-derived inhibition and probably influence platelet shape.

For platelets treated with riociguat, increased circularity values were observed in healthy volunteers but not in HIV-positive patients taking ABC (Figure 3), implying that the NO-GC/cGMP pathway may be downregulated or blocked by off-target effects of antiretrovirals such as ABC in these patients. This is supported by the fact that platelet circularity was not altered by any NO-GC-modulating treatment, suggesting platelets are hyperactivated in these patients. Consistent with this finding, riociguat decreased platelet aggregation in HIV-negative volunteers but not in HIV-positive patients currently taking ABC (Figure 3c,d). However, further studies on the interaction between cGMP and ABC are warranted. Our studies suggest that treatment of HIV-positive patients with ABC-containing regimens blocks NO-GC/cGMP-mediated thrombo-protection (readout circularity), potentially putting these patients at higher risk of thrombotic events. This is supported by observational studies, indicating an increased risk of myocardial infarction in patients taking ABC.³³

In summary, our study indicates that the NO-GC/cGMP signaling pathway is involved in the inhibition of P-selectin expression. Moreover, when cGMP/PKG is stimulated via a NO-GC stimulator (riociguat) or a NO-GC activator (cinaciguat), a decrease in platelet cellular stiffness and platelet circularity occurs via NO-GC. As cellular stiffness indicates the mechanical state of the platelet, measuring cellular stiffness could become an important feature to characterize and evaluate patients’ cardiovascular risk, considering that platelet hyperactivation is associated with multiple cardiovascular conditions such as increased arterial stiffness,³⁴ detrimental blood flow,³⁵ thrombosis, or atherosclerosis. Also, platelet circularity could help to distinguish platelet spreading stages and activation state, which could allow to identify patients with a tendency of developing a cardiovascular pathology. However, the precise mechanism of how the NO-GC/cGMP signaling pathway affects stiffness and shape needs to be examined. In particular, the involved downstream phosphorylation targets of the PKG need to be identified.

Author contribution

Conceptualization: A.B., J.R., J.S., D.P.Q., M.E., F.S., R.F., S.F., and T.E.S.; Investigation: A.B., J.R., J.S., D.P.Q., and A.K.; Resources: A.F., M.E., M.B., S.F., R.F., and T.E.S.; Writing – original draft preparation: J.R.; Writing – review and editing: A.B., J.R., J.S., D.P.Q., A.K., M.B., M.F., A.F., M.E., F.S., R.F., S.F., and T.E.S.; Visualization: A.B., J.R., J.S., D.P.Q., A.K., M.B., M.F., A.F., M.E., F.S., R.F., S.F., and T.E.S.; Supervision: M.E., F.S., R.F., S.F., and T.E.S.; Project administration: A.B., J.R., and T.E.S.; Funding acquisition: T.E.S. All authors have read and agreed to the published version of the manuscript.

Informed consent statement

Informed consent was obtained from all the subjects involved in the study in Tübingen and London.

Acknowledgments

The authors acknowledge support by the High Performance and Cloud Computing Group at the Zentrum für Datenverarbeitung of the University of Tübingen and the state of Baden-Württemberg through bwHPC . The authors acknowledge support from the Open Access Publication Fund of the University of Tübingen. Flow cytometry sample acquisition and data analysis were done on shared instruments of the Flow Cytometry Core Facility Tübingen with the help of Dr. Kristin Bieber. A.B. is supported by the Add-on Fellowship of the Joachim Herz Foundation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Projektnummer [335549539/GRK2381], Projektnummer [374031971 – TRR 240], and Projektnummer [INST 37/935-1 FUGG]. The Facility for Imaging by Light Microscopy (FILM) at Imperial College London is partly supported by the Wellcome Trust (grant 104931/Z/14/Z) and BBSRC (grant BB/L015129/19).