Ibrutinib Inhibits BMX-Dependent Endothelial VCAM-1 Expression In Vitro and Pro-Atherosclerotic Endothelial Activation and Platelet Adhesion In Vivo

Abstract

Introduction

Inflammatory activation of the vascular endothelium leads to overexpression of adhesion molecules such as vascular cell adhesion molecule-1 (VCAM-1), contributing to the pro-thrombotic state underpinning atherogenesis. While the role of TEC family kinases (TFKs) in mediating inflammatory cell and platelet activation is well defined, the role of TFKs in vascular endothelial activation remains unclear. We investigated the role of TFKs in endothelial cell activation in vitro and in a nonhuman primate model of diet-induced atherosclerosis in vivo.

Methods and Results

In vitro, we found that ibrutinib blocked activation of the TFK member, BMX, by vascular endothelial growth factors (VEGF)-A in human aortic endothelial cells (HAECs). Blockade of BMX activation with ibrutinib or pharmacologically distinct BMX inhibitors eliminated the ability of VEGF-A to stimulate VCAM-1 expression in HAECs. We validated that treatment with ibrutinib inhibited TFK-mediated platelet activation and aggregation in both human and primate samples as measured using flow cytometry and light transmission aggregometry. We utilized contrast-enhanced ultrasound molecular imaging to measure platelet GPIbα and endothelial VCAM-1 expression in atherosclerosis-prone carotid arteries of obese nonhuman primates. We observed that the TFK inhibitor, ibrutinib, inhibited platelet deposition and endothelial cell activation in vivo.

Conclusion

Herein we found that VEGF-A signals through BMX to induce VCAM-1 expression in endothelial cells, and that VCAM-1 expression is sensitive to ibrutinib in vitro and in atherosclerosis-prone carotid arteries in vivo. These findings suggest that TFKs may contribute to the pathogenesis of atherosclerosis and could represent a novel therapeutic target.

Introduction

Atherosclerosis is a chronic inflammatory disease characterized by the accumulation of lipoproteins and smooth muscle cells in the arterial intima, endothelial cell activation, and leukocyte recruitment culminating in atheromatous plaque formation.²⁶ During the later stages of atherosclerosis, patients may experience significant complications from plaque rupture and thrombus formation.^12,26 Despite medical advances, atherosclerosis remains the leading cause of mortality in the United States.^1,17

Endothelial activation contributes to the pro-thrombotic and pro-inflammatory milieu underlying atherogenesis. As platelets adhere to the vascular endothelium, pro-inflammatory endothelial activation occurs in concert with an overexpression of adhesion proteins involved in cell-cell interactions, such as intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1).^11,39 Elucidating the mechanisms that govern platelet-endothelial interactions, the release of endothelial-cell derived inflammatory factors, and the expression of cell adhesion molecules may identify novel therapeutic targets to mitigate the progression of atherogenesis.

Due to their role in systemic inflammation and cytokine signaling, TEC family kinases (TFKs) represent promising targets for mitigating inflammatory disorders and autoimmune diseases.^6,25 The TFKs represent the second largest group of non-receptor tyrosine kinases and consist of five members: (1) BTK (Bruton’s tyrosine kinase); (2) RLK (resting lymphocyte kinase; also known as TXK); (3) ITK (interleukin-2-inducible T-cell kinase); (4) TEC (tyrosine kinase expressed in hepatocellular carcinoma); and (5) BMX (bone-marrow tyrosine kinase gene on chromosome X; also known as ETK).³⁴ BTK is mainly expressed in B-cells, whereas RLK/TXK, and ITK are exclusively expressed in myeloid hematopoietic cells. TEC is expressed in certain somatic cells and BMX is specifically expressed by the cardiac endothelium and arterial endothelial cells.^10,29,38 BTK and TEC play important signaling roles in glycoprotein VI (GPVI)-mediated platelet activation, spreading, and aggregation.²

Within the last decade, the TEC family kinase and critical B-cell signaling molecule, BTK, has become a novel and effective therapeutic target in many hematologic malignancies.⁵ Indeed, the first-generation BTK inhibitor, ibrutinib, has demonstrated efficacy in treating several hematological malignancies, such as chronic lymphocytic leukemia, and chronic inflammatory conditions, such as graft-versus-host disease.^5,23 However, ibrutinib also imparts anti-platelet effects as BTK contributes to platelet signaling, with accumulating clinical data demonstrating the effects of TFK inhibition in platelets.³⁶ Additionally, ibrutinib exhibits off-target effects through inhibition of several other TFKs including TEC and BMX.³⁷

Vascular endothelial cells express BMX but not BTK.¹⁰ Both BMX and BTK play a role in systemic inflammation and cytokine signaling, leading to the investigation of several TFK inhibitors for inflammatory and autoimmune diseases.^6,25 In particular, inhibition of inflammatory cytokine activity has been demonstrated to mitigate cardiovascular morbidity in a major clinical trial.³² Although it has been shown that BMX plays a role in signaling via the vascular endothelial growth factor receptors 1 and 2 (VEGFR1, VEGFR2) and that tumor necrosis factor-α (TNFα) induces a reciprocal activation between BMX and VEGFR2, it remains unknown if these mechanisms are sensitive to TFK inhibitors such as ibrutinib.^30,42 Since ibrutinib can also inhibit BMX activity by irreversibly binding a conserved cysteine residue,³⁷ we used ibrutinib as a functional tool to investigate the role of TFKs in the pathogenesis of atherosclerosis.

While TFKs appear to be novel, druggable targets to inhibit key pathways that mediate platelet signaling and activity, the effects of inhibiting these pathways on platelet-endothelial interactions and overall cardiovascular risk in vivo remain undescribed. We hypothesized that TFKs play a critical role in atherosclerosis formation via several independent pathways: (1) pro-atherosclerotic vascular endothelial functions; (2) platelet function; and (3) regulation of pro-atherosclerotic cytokine production and systemic inflammation. We therefore aimed to evaluate the role of TFKs on these biochemical pathways.

Materials and Methods

Reagents

Primary human aortic endothelial cells (HAECs) and 0.1% gelatin solution were obtained from American Type Culture Collection (Manassas, VA, USA). VascuLife VEGF Endothelial Medium Complete Kit was obtained from Lifeline Cell Technology (Frederick, MD, USA). The irreversible BTK inhibitor, ibrutinib, was obtained from Selleck Chemicals (Houston, TX, USA). The irreversible BMX inhibitor, BMX-IN-1, was obtained from MedChemExpress (Monmouth Junction, NJ, USA). Recombinant human vascular endothelial growth factor (VEGF) and TNFα were obtained from R&D System (Minneapolis, MN, USA). CRP-XL was obtained from Dr. Richard Farndale (Cambridge University, Cambridge, UK). Protein A/G Sepharose and Protein A/G PLUS-agarose beads were from Santa Cruz Biotechnology (Dallas, TX, USA). Laemmli Sample Buffer was obtained from Bio-Rad Antibodies (Hercules, CA, USA). Accutase and human endothelial serum free media were obtained from Thermo Fisher Scientific (Waltham, MA, USA). Prostacyclin was obtained from Cayman Chemical (Ann Arbor, MI, USA). All other reagents were obtained from Sigma-Aldrich (St. Louis, MO, USA) or previously cited sources.²²

Antibodies

The 15‐amino acid cyclic peptide (CCP‐015b) and the mouse anti‐human monoclonal IgG1 against VCAM‐1 (1.G11B1) used for contrast-enhanced ultrasound (CEU) molecular imaging were obtained from Dr. Gray D. Shaw (Quell Pharma Inc., Plymouth, MA, USA) and Bio-Rad Antibodies (Hercules, CA, USA), respectively. The PE mouse anti-human CD62P (P-selectin) and FITC mouse anti-human PAC-1 antibodies used for flow cytometry were obtained from BD Biosciences (San Jose, CA, USA). The APC anti-human CD62P (P-selectin) antibody used for flow cytometry was obtained from BioLegend (San Diego, CA, USA). The PE-Cyanine7 CD106 (VCAM-1) antibody used for flow cytometry was obtained from ThermoFisher Scientific (Waltham, MA, USA). The anti-p-Tyr antibody used for immunoprecipitation assays was from Santa Cruz Biotechnology (Dallas, TX, USA). The anti-ETK/BMX antibody and anti-p-STAT3 used for immunoprecipitation assays were from Cell Signaling Technology (Danvers, MA, USA). The anti-p-BMX antibody used for immunoprecipitation assays was from Abcam (Cambridge, UK). The anti-VCAM-1 antibody used for Western blotting was from Santa Cruz Biotechnology (Dallas, TX, USA).

Flow Cytometry for VCAM-1 Expression

HAECs were grown to confluence in 0.1% gelatin-coated 6-well plates with VascuLife VEGF Endothelial Medium Complete Kit. HAECs were serum-starved for 4 h and then pre-incubated with vehicle (DMSO) or ibrutinib (10 µM) for 30 min at 37 °C. Cells were then stimulated with VEGF-A (100 ng/mL) or TNFα (5 ng/mL) in serum-free medium with 0.5% bovine serum albumin (BSA) for 6 h at 37 °C. HAECs were washed with PBS, incubated with accutase to detach cells for 2 min at 37 °C, and centrifuged at 1000×g for 5 min. Samples were re-suspended in 0.5% BSA in PBS and incubated with an anti-VCAM-1 PE/Cy7 antibody (1:50) for 30 min at 37 °C at 200 rpm. Samples were then diluted with 0.5% BSA in PBS, centrifuged at 1000×g for 5 min, and re-suspended in 2% PFA. Samples were analyzed by flow cytometry (FACSCanto II, BD Biosciences, Franklin Lakes, NJ, USA) and VCAM-1 expression was determined by flow cytometry analysis using FlowJo software (version 10.8.1).

Immunoprecipitation and Western Blotting

HAECs were grown to confluence in 0.1% gelatin-coated 12-well plates with VascuLife VEGF Endothelial Medium Complete Kit. HAECs were serum-starved for 4 h before being stimulated with vehicle, VEGF-A (100 ng/mL), or TNFα (5 ng/mL) in serum-free medium with 0.5% BSA for 6 h at 37 °C. Prior to stimulation, cells were pre-incubated with vehicle (DMSO) or the TFK inhibitors, JS25 (20, 40, 80 µM), BMX-IN-1 (20, 40 µM), or ibrutinib (10 µM) as indicated for 30 min at 37 °C. Cells were lysed in Laemmli sample buffer containing 200 mM dithiothreitol. Protein sample were separated by SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membranes, and blotted with an anti-VCAM-1 antibody (0.4 µg/mL) and a horseradish peroxidase-conjugated secondary antibody (1:10,000).

For immunoprecipitation experiments, cells were seeded in 12-well plates and serum-starved for 4 h before being stimulated with vehicle or VEGF-A (100 ng/mL) in serum-free medium with 0.5% BSA for 5 to 60 min at 37 °C. Cells were lysed with cold lysis buffer (10 mM Tris/HCl, pH 7.4; 150 mM NaCl; 2 mM EDTA; 1% Triton X-100) containing PMSF (1:100) and phosphatase inhibitor (1:100) for 30 min at 4 °C. HAEC lysates were then pre-cleaned with Protein A/G Sepharose and incubated with 2 µg of an anti-p-Tyr antibody or nonspecific IgGs overnight at 4 °C. Antibody-protein complexes were then captured with Protein A/G PLUS-agarose beads at 4 °C for 2 h. At the end of the incubation period, the beads were washed three times in a lysis buffer. Precipitates were then eluted with Laemmli sample buffer supplemented with 200 mM dithiothreitol. Protein sample were separated by SDS-PAGE, and transferred to PVDF membranes and blotted with an anti-BMX antibody (0.1 µg/mL), an anti-p-BMX Y566 antibody (0.03 µg/mL), or an anti-p-STAT3 Y705 antibody (0.2 µg/mL), and horseradish peroxidase-conjugated secondary antibodies. Films were scanned and blots representative of 3-4 independent experiments are shown.

Human Platelet Isolation

Human venous blood was drawn by venipuncture from healthy subjects into 3.8% sodium citrate (1:10) and acid citrate dextrose (1:10) as previously described.^27,43 Written informed consent from volunteers was obtained in accordance with an approved protocol from the Institutional Review Board of Oregon Health & Science University. Platelet rich plasma (PRP) was isolated from whole blood by centrifugation at 200×g for 20 min. PRP was then centrifuged at 1000×g for 10 min in the presence of prostacyclin (0.1 µg/mL) and the platelet poor plasma (PPP) was decanted to obtain platelets. The platelet pellet was re-suspended in modified HEPES-Tyrode buffer (129 mM NaCl, 0.34 mM Na₂HPO₄, 2.9 mM KCl, 12 mM NaHCO₃, 20 mM HEPES, 5 mM glucose, 120 mM MgCl₂, pH 7.3), and washed via centrifugation at 1000×g for 10 min in the presence of prostacyclin (0.1 µg/mL) and acid citrate dextrose (1:10). The washed platelets were re-suspended in modified HEPES-Tyrode buffer to the desired concentration.

Platelet Aggregation

Platelet aggregation was measured in response to GPVI-agonist, cross-linked collagen-related peptide (CRP-XL). Platelet aggregation studies were performed by pre-incubating 300 µL of human washed platelets (3×10⁸/mL) with inhibitors for 10 min at room temperature. Platelets were then stimulated with CRP-XL (1 µg/mL) and monitored under continuous stirring at 37 °C. Platelet aggregation was measured by changes in light transmission using an aggregometer (Chrono-Log Corporation, Havertown, PA, USA). Data were presented as representative aggregation tracings. The averages of maximal platelet aggregation for 3 different experiments were analyzed for statistical significance by one-way ANOVA using GraphPad PRISM 8 software. For the NHP studies, 300 µL of PRP was warmed to 37 °C and then stimulated with CRP-XL (1 µg/mL) under continuous stirring. Platelet aggregation was measured by changes in light transmission using an aggregometer (PAP-4, Bio/DataCorporation, Horsham, PA, USA). Data were presented as representative aggregation tracings.

Flow Cytometry for Platelet Activation

Platelet integrin activation and granule secretion in response to CRP-XL were measured using flow cytometry. Flow cytometry studies were performed with human whole blood anticoagulated with 3.8% sodium citrate and pre-incubated with either vehicle (DMSO) or ibrutinib (10 µM) for 10 min at 37 °C. Human whole blood samples were diluted in modified HEPES-Tyrode buffer (1:4) and incubated with anti-CD62P PE (1:20) or anti-PAC-1 FITC (1:20) antibody for 20 min at 37 °C in the presence of CRP-XL (5 µg/mL). Whole blood from rhesus macaques was also diluted in modified HEPES-Tyrode buffer (1:4) and incubated with anti-CD62P PE (1:10) or anti-PAC-1 FITC (1:10) antibody for 20 min at 37 °C in the presence of CRP-XL (5 µg/mL). The reactions were stopped by adding 2% PFA and analyzed by flow cytometry (BD FACSCanto II, BD Biosciences, Franklin Lakes, NJ). Platelet activation was determined by flow cytometry analysis using FlowJo software (version 10.8.1) and presented as mean fluorescence intensity.

Nonhuman Primate Model of Early Atherosclerosis

This study was approved by the Institutional Animal Care and Use Committee of the Oregon Health & Science University (Approval Number: IP00002332) prior to initiation. Rhesus macaques (Macaca mulatta), were cared for and housed at the Oregon National Primate Research Center (ONPRC) at Oregon Health & Science University.

Adult male rhesus macaques (n = 2; ages 14 and 18 years) were fed a high fat diet (45% carbohydrates, 18% protein, and 36% fat by caloric content LabDiet 5L0P, Purina Mills) for > 2 years and were selected on the basis of having carotid intimal-medial thickening as a marker of early atherosclerotic changes. Animals were administered ibrutinib orally daily for 7 days at a dose of 10 mg/kg/day and studied at baseline and days 1 and 7 after drug administration. The dose used in this study was selected to test the maximal response of ibrutinib within the dose range of 1.25–12.5 mg/kg/day used in clinical studies ibrutinib.²¹ Anesthesia was induced with ketamine (10 mg/kg intramuscularly) and maintained with isoflurane (1.0–1.5%) at two discrete time points during the study to facilitate CEU molecular imaging.

At designated study intervals, animals underwent phlebotomy for analysis of lipid levels and inflammatory biomarkers, platelet aggregation studies, and whole blood flow cytometry. Hematological analyses were also performed on adult male rhesus macaques (n = 3; ages 8, 12, 15 years) fed a standard diet (58.52% carbohydrates, 26.82% protein, and 14.65% fat by caloric content LabDiet 5L0P, Purina Mills) as lean controls.

Nonhuman Primate Blood Collection

For the NHP studies, venous blood samples were collected from rhesus macaques by venipuncture into 3.2% sodium citrate or plastic vacutainers at baseline and days 1 and 7 following ibrutinib treatment. PRP was isolated from blood samples anticoagulated with 3.2% sodium citrate by centrifugation of whole blood at 200×g for 8 min. The supernatant was removed, and PPP was obtained by further centrifugation of the remaining blood at 1000×g for 3 min. To isolate serum, blood was drawn into a plastic vacutainer and allowed to sit for 30 min at room temperature. The blood clot was then removed by centrifuging the sample at 1000×g for 10 min.

Hematological Analysis for Lipid Profiles and Inflammatory Biomarkers

Venous blood samples collected from rhesus macaques by venipuncture were used to measure lipids and inflammatory biomarker profiles. Serum samples were analyzed for lipid levels by the Clinical Pathology Laboratory at the ONPRC using an ABX Pentra 400 Clinical Chemistry System (Horiba Medical, Irvine, CA, USA). C-reactive protein (CRP) concentrations were assessed by an enzyme-linked immunosorbent assay (ELISA) kit, following the manufacturer’s instructions (30-9710S, ALPCO, Salem, NH, USA). The assay range was 0.95-150 ng/mL and intra-assay CV was 3.6%.

Serum samples were analyzed for cytokines using a custom 9-plex monkey cytokine Luminex panel (ThermoFisher, Waltham, MA, USA), which included IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12/IL-23p40, IFN-γ, and IFN-α. The analysis was performed by the Endocrine Technologies Core at the ONPRC. Briefly, 25 µL of each serum sample was diluted in assay diluent and incubated overnight with antibody-coated, fluorescent-dyed capture microspheres specific for each analyte, followed by detection antibodies and streptavidin-phycoerythrin. Washed microspheres with bound analytes were resuspended in reading buffer and analyzed on a Milliplex LX-200 Analyzer (EMD Millipore, Billerica, MA, USA) bead sorter with XPonent Software version 3.1 (Luminex, Austin, TX, USA). Data were calculated using Milliplex Analyst software version 5.1 (EMD Millipore). An in-house generated rhesus macaque serum pool was run in quadruplicate as a quality control. Intra-assay CVs were as follows: IL-1β, 4.7%; IL-2, 12.7%; IL-8, 1.6%; IL-10, 4.3%; IL-12/IL-23p40, 6.6%; IFN-γ, 12.8%; IFN-α, 3.2%. Intra-assay CVs for IL-4 and IL-6 were not calculated, as they were undetectable in the QC. Since all samples were analyzed in a single assay, no inter-assay variation was calculated.

Targeted Molecular Imaging Agent Preparation

For CEU molecular imaging, microbubbles targeted to platelet glycoprotein-Ibα (GPIbα), endothelial vascular cell adhesion molecule-1 (VCAM-1), or control microbubbles were prepared as previously described.^4,7,24 In brief, biotinylated lipid‐shelled decafluorobutane microbubbles were prepared via sonication of a gas-saturated aqueous suspension of distearoylphosphatidylcholine (2 mg/mL), polyoxyethylene‐40‐stearate (1 mg/mL), and distearoylphosphatidylethanolamine‐PEG (2000) biotin (0.4 mg/mL). Microbubbles targeted to platelet GPIbα or endothelial VCAM-1 were prepared by conjugating biotinylated ligands to the microbubble surface using either a 15-amino acid cyclic peptide (CCP-015b) biotinylated at an added C-terminal lysine residue or a monoclonal antibody against VCAM-1 (1.G11B1), respectively. Microbubbles targeted to platelet GPIbα and endothelial VCAM-1 were previously validated using in vitro flow chamber assays and immunohistochemistry of spleen and carotid artery from rhesus macaques that were fed a high fat diet for 2 years, respectively.^4,7 Control microbubbles were unconjugated with no targeting ligand. Electrozone sensing was used to measure microbubble concentration and to ensure similar size distribution between agents for each experiment (Multisizer III, Beckman-Coulter, Brea, CA, USA).

Carotid Molecular Imaging

We utilized CEU molecular imaging at the carotid bifurcation to measure endothelial VCAM-1 and platelet GPIbα before and after treatment with the TFK inhibitor ibrutinib as previously described.^4,7,24 In brief, longitudinal-axis imaging at the carotid bifurcation was performed using multi-pulse, contrast-specific imaging at 7 MHz, a mechanical index of 1.9, a dynamic range of 55 dB, and a frame rate of 1 Hz (Sequoia, Siemens Medical Imaging, Mountain View, CA, USA). Intravenous injections of microbubbles targeted to endothelial VCAM-1 or platelet GPIbα (1×10⁸) were performed in a random order. Following injection of targeted microbubbles, the ultrasound was paused for 1 min before locating the carotid artery using two‐dimensional ultrasound at low power (mechanical index <0.10) and activating contrast‐specific imaging for several frames. The signal for the retained microbubbles was quantified by digitally averaging the first two frames acquired and then subtracting several averaged frames acquired after >5 destructive pulse sequences (mechanical index 1.3) to mitigate signal from freely circulating microbubbles. Data from regions-of-interest at the near and far walls of the common carotid artery were averaged.

Statistical Analysis

Data were tested for normality using a Shapiro–Wilk test. Two group data presented in the study were not normally distribution and were analyzed by a Mann–Whitney test. For three or more groups, data were analyzed by one-way ANOVA with a Dunnett’s post-hoc test. Statistical significance was considered for P < 0.05. For all the experiments, n indicates the number of independent experiments performed. Individual data from the NHP experiments were not evaluated statistically to avoid overpowering the analysis given the small size of the animal cohort. Statistical analyses were performed using GraphPad PRISM 9 (San Diego, CA, USA).

Results

Effects of VEGF-A on Endothelial Cell VCAM-1 Expression

Tyrosine kinases play a central role in endothelial cell signal transduction by phosphorylating cellular proteins to facilitate growth and differentiation signal cascades, as well as to regulate vessel tone, expression of adhesion molecules, and chemoattractants.⁶ Endocardial and arterial endothelial cells express the TEC family kinase (TFK), BMX. BMX is known to be involved in signaling downstream of VEGFR1 and VEGFR2.^30,42 Since VEGF-A, the ligand for the aforementioned receptors, is upregulated in atherosclerotic plaques and enhance atherosclerosis, we studied whether VEGF-A-mediated VCAM-1 expression by endothelial cells was mediated by BMX.

First, we validated that VCAM-1 expression was increased in response to VEGF-A using flow cytometry. As seen in Fig. 1a, VCAM-1 expression was upregulated in response to VEGF-A (100 ng/mL) compared to vehicle; VCAM-1 expression was confirmed to be upregulated by the cytokine TNFα (5 ng/mL). We next confirmed that VEGF-A-induced phosphorylation of BMX (Fig. 1b) Previous studies have shown that BMX mediates STAT3 activation,^13,37 and that STAT3 contributes to a number of cardiovascular diseases, including atherosclerosis, cardiac hypertrophy, and heart failure.⁸ Indeed, as shown in Fig. 1b, VEGF-A stimulation of HAECs upregulated phosphorylation of both BMX and STAT3 as a function of time, providing a mechanistic link between the initiator involvement of BMX as well as the pathological roles of STAT3 in atherosclerosis.

Figure 1.

VCAM-1 expression in human aortic endothelial cells (HAECs) following exposure to VEGF-A or TNFα. (a) Representative flow cytometry histograms from HAECs stimulated with vehicle, VEGF-A (100 ng/mL), or TNFα (5 ng/mL). In select experiments, HAECs were pre-incubated with ibrutinib (20 µM). (b) Phosphorylation of BMX (p-BMX) and STAT3 (p-STAT3) immunoprecipitated from HAECs stimulated with VEGF-A (100 ng/mL) for 0, 15, 30 or 60 min. Representative blots for at least n = 3. IP immunoprecipitation.

Effects of BMX Inhibition on Endothelial Cell VCAM-1 Expression

When the endothelium is activated, adhesion molecules (e.g. E-selectin, ICAM-1, and VCAM-1) are expressed on the surface. VCAM-1 represents the most prevalent adhesion molecule in atherosclerosis and is present in early atherosclerotic lesions.³⁹ Since VEGF-A stimulates VCAM-1 expression in HAECs (Fig. 1a) and VEGF-A signals through BMX and STAT3 (Fig. 1b), we next tested if the kinase pathway downstream of VEGF-A stimulation is sensitive to BMX inhibition. Our results show that pharmacological inhibition of BMX with the BMX inhibitors, JS25 and BMX-IN-1, reduced VCAM-1 expression in response to VEGF (Fig. 2a). Along these lines, VEGF-A-induced VCAM-1 expression was also sensitive to ibrutinib (Fig. 2a). With the exception of the higher concentration of JS25 (40µM), VCAM-1 expression in response to TNFα was not affected by any of the aforementioned inhibitors (Fig. 2b). Taken together, our data demonstrate that VEGF-A-mediated endothelial VCAM-1 expression is sensitive to pharmacological inhibition of BMX.

Figure 2.

Effect of ibrutinib on VEGF-A-induced VCAM-1 expression in human aortic endothelial cells (HAECs). (a) HAECs were stimulated with either VEGF-A (100 ng/mL) or (b) with TNFα (5 ng/mL) for 6 h in the presence of different concentrations of BMX inhibitors (JS25 or BMX-IN-1) or ibrutinib (10 µM). Representative blots for at least n = 3.

Effects of TFK Inhibition on Platelet Aggregation

To advance toward in vivo testing in a nonhuman primate (NHP) model, we first validated the effect of ibrutinib on known platelet pathways in response to ITAM-mediated agonists, including CRP-XL. CRP-XL mediates platelet activation in a receptor tyrosine kinase-dependent manner.¹⁹ We found that pre-incubation of human washed platelets with ibrutinib significantly abrogated platelet aggregation in response to CRP-XL (1 µg/mL) (Fig. 3a) (P = 0.029). Next, we examined the effect of TFK inhibition on NHP platelet aggregation in response to CRP-XL. Our data show robust aggregation of NHP platelets in response to CRP-XL (1 µg/mL) prior to administration of ibrutinib (Fig. 3b, Baseline). Platelet aggregation in response to CRP-XL (1 µg/mL) was abrogated following one week of treatment with the BTK inhibitor, ibrutinib (Fig. 3b, Ibrutinib). These data validate that ibrutinib inhibits GPVI-mediated platelet aggregation.

Figure 3.

Platelet aggregation and activation in response to the GPVI-agonist, CRP-XL. (a) Human washed platelets (3x10⁸/mL) were pre-treated with ibrutinib for 10 min at room temperature and stimulated with CRP-XL (1 μg/mL). Representative tracings from human washed platelets pre-incubated with ibrutinib (n = 3) prior to stimulation with CRP-XL (arrow). Maximal aggregation from human washed platelets pre-incubated with ibrutinib (n = 3). (b) Platelet rich plasma from obese NHPs (n = 2) was stimulated with CRP-XL (1 µg/mL). Traces are shown for primates at baseline and after one week of ibrutinib treatment following the addition of CRP-XL (arrow). Flow cytometry analysis of P-selectin (c, d) and PAC-1 (e, f) expression from human (n = 3) or NHP (n = 2) whole blood samples stimulated with CRP-XL (5 μg/mL). MFI mean fluorescent intensity. *indicates P < 0.05. **Indicates P < 0.01. n indicates the number of independent experiments using blood from different donors. Error bars indicate SEM.

Effects of TFK Inhibition on Platelet P-selectin and PAC-1 Expression

Upon activation, platelets release α-granules containing numerous substances including P-selectin (CD62P).^33,43 Therefore, we used flow cytometry to quantify P-selectin expression as a marker of GPVI-mediated platelet activation. We first pre-incubated human whole blood with ibrutinib (10 µM) before stimulating with CRP-XL and staining with an anti-P-selectin antibody. Flow cytometry analyses confirmed that ibrutinib significantly decreased P-selectin expression in response to CRP-XL relative to control (P = 0.012) (Fig. 3c). Stimulation of NHP whole blood with CRP-XL upregulated P-selectin surface expression at baseline, and dramatic inhibition of CRP-XL-induced P-selectin expression was observed as early as one day following the initial dose of ibrutinib (Fig. 3d).

Downstream of the initial GPVI-mediated platelet activation, intracellular signaling events promote “inside-out” activation of platelet integrins such as αIIbβ3 that bind to fibrinogen, vWF, and matrix proteins with RGD-like sequences.¹⁶ These processes mediate stable platelet adhesion and aggregation, as well as thrombus formation.⁴³ To monitor platelet surface integrin αIIbβ3 activation, whole blood samples were treated as previously described and an anti-PAC-1 antibody was used to recognize the active conformation of platelet integrin αIIbβ3. We confirmed that PAC-1 expression in human samples treated with ibrutinib decreased relative to control (P = 0.042) (Fig. 3e). Indeed, PAC-1 expression in the NHP samples was elevated at baseline and dramatically inhibited as early as one day following the initial dose of ibrutinib (Fig. 3f). Taken together, these results confirm that inhibiting TFK blocks platelet GPVI-mediated α-granule secretion and integrin activation.

Effects of TFK Inhibition on Body Weight, Lipid Profiles, and Inflammatory Biomarkers

Although TFKs represent an attractive target to reduce platelet activation, the effects of inhibiting these pathways on platelet-endothelial interactions and overall cardiovascular risk in vivo remains unknown. Having validated the efficacy of ibrutinib with respect to platelet activity in vitro, we next sought to perform a pilot experiment examining the systemic effects of BTK inhibition using a model of obese NHPs exhibiting an atherosclerotic phenotype (Fig. 4a). The two rhesus macaques in this pilot study were fed a high fat diet for a minimum of two years, resulting in the development of detectable atherosclerotic lesions in the aorta, coronary, carotid and renal arteries.⁷

Figure 4.

Experimental design and targeted contrast-enhanced ultrasound molecular imaging. (a) Experimental design for the in vivo study using an NHP model of early atherosclerosis. (b) Examples obtained at baseline and after one week of ibrutinib showing non-contrast, two-dimensional image of the carotid artery at the bifurcation (left) and a background-subtracted color-coded image (right) obtained after injecting VCAM-1-targeted microbubbles. (c) Background subtracted video intensity for VCAM-1 and GPIbα-targeted microbubbles in obese rhesus macaques (n = 2) treated with TEC family kinase inhibitor, ibrutinib, compared to vehicle. Floating bar plots illustrating the median (bar), and the maxima and minimum (box) for CEU molecular imaging. NHP nonhuman primate, CEU contrast-enhanced ultrasound.

At baseline, animals on the high fat diet had a greater body mass compared to lean control animals. The obese animals had elevated levels of C-reactive protein (CRP) and interleukin (IL)-4 compared to the lean control animals, while IL-8 levels were not elevated compared to lean controls. Treatment of obese animals with ibrutinib for one week did not consistently alter the body mass, lipid levels, or levels of various inflammatory biomarkers (Table 1).

Table 1.

Body weight, lipid profiles, and inflammatory biomarkers in obese NHPs (n = 2) at baseline and after one week of ibrutinib treatment.

	Lean	Obese
		Animal A		Animal B
		Baseline	Ibrutinib	Baseline	Ibrutinib
Weight, kg	11.0 ± 1.9*	20.5	20.5	23.8	23.9
Total cholesterol, mg/dL	134 ± 21*	137	138	162	187
LDL cholesterol, mg/dL	58 ± 11*	67	75	74	98
HDL cholesterol, mg/dL	65 ± 15*	30	34	72	82
CRP, ng/mL	0.381 ± 0.150	0.891	0.111	0.925	0.922
IL-4, pg/mL	1.14 ± 0.13	1.53	2.06	5.47	5.18
IL-8, pg/mL	196.9 ± 95.1	91.4	74.4	175	135

Data are shown as mean ± SEM. *Historic data from Chadderdon et al⁷

LDL low-density lipoprotein, HDL high-density lipoprotein, CRP C-reactive protein, IL interleukin

Effects of TFK Inhibition on Carotid Endothelial Activation and Platelet Adhesion

Previous studies have demonstrated that obese NHPs exhibit a significant increase in platelet GPIbα and endothelial VCAM-1 compared to lean control animals, indicating local platelet adhesion and endothelial inflammation.⁴ It has also been shown that endothelial expression of VCAM-1 increases over time in NHPs fed a high fat diet.⁷ We performed a pilot study to determine the effects of TFK inhibition on local platelet activation and endothelial inflammation. We used a non-invasive molecular imaging technique to measure platelet GPIbα and endothelial VCAM-1 at the carotid bifurcation before and after treatment with ibrutinib in a NHP model of diet-induced obesity. On visual inspection, a decrease in VCAM-1 expression was observed in the presence of ibrutinib (Fig. 4b). Quantification of intensity per unit area showed a decrease from intensity levels at or above 1.8 V/U to below 1.6 V/U in the presence of ibrutinib. The quantification of platelet deposition at baseline indicated a variable response with a high of 19.5 V/U to a low of − 2.5 V/U. The maximum intensity recorded for GPIbα was near 2.3 V/U in the presence of ibrutinib V/U (Fig. 4c).

Discussion

This study examined the effects of TFK inhibition on endothelial activity in vitro and makers of endothelial cell activation in atherosclerosis-prone arteries in NHPs in vivo. Our findings suggest that the TFK inhibitor, ibrutinib, inhibited VEGF-A-mediated endothelial VCAM-1 expression, as well as GPVI-mediated platelet aggregation, integrin activation, and granule secretion in vitro. In a pilot study using an NHP model of diet-induced atherosclerosis, ibrutinib also decreased markers of vascular endothelial cell activation and platelet adhesion compared to baseline in vivo. While the systemic and global platelet inhibition demonstrated in this work can be a major cause of bleeding risks in vivo, previous studies have shown that inhibition of BTK with ibrutinib analogs in an NHP model did not result in bleeding, capillary abnormalities, or thrombocytopenia.³³ Similarly, clinical trials for ibrutinib reported predominantly low-grade bleeding events in humans.^5,40 The sample size for this pilot experiment was small (n = 2) due to the limited resources available to study NHP models of early atherosclerosis. This is a caveat to the findings from these preliminary studies that generate a hypothesis that TFKs contribute to the pathogenesis of atherosclerosis and that BTK represents an efficacious and safe target to prevent atherothrombosis.

During the progression of atherosclerosis, a number of pro-inflammatory signaling cascades are implicated through the release of cytokines and growth factors (e.g. interleukins, TNF, VEGF), and the expression of adhesion proteins (e.g. E-selectin, ICAM-1, VCAM-1).^31,39 For example, VEGF can signal through the PLCγ-sphingosine kinase-PKC cascade or the PI3K/AKT/MAPK cascade to activate NF-kB and induce VCAM-1 expression.^3,15 Furthermore, VEGF signals through the TEC family kinase member, BMX, via VEGFR1 and VEGFR2.^30,42 Indeed, our in vitro studies showed that VEGF-A-mediated VCAM-1 expression is sensitive to pharmacological inhibition of BMX with selective BMX inhibitors or ibrutinib. The fact that ibrutinib inhibited VCAM-1 expression induced by VEGF-A may explain the effect of ibrutinib on the endothelium in vivo. Alternatively, TNFα can induce activation of BMX via TNFR2, and the formation of a reciprocally activated complex between BMX and VEGFR2; this in turn elicits the common PI3K/AKT pathway.⁴² Of note, TNFR2 is tightly regulated compared to its ubiquitous counterpart, TNFR1. TNFR1-mediated pathways have been extensively studied as the main receptor for TNFα on endothelial cells; nevertheless, induction of VCAM-1 expression by TNFα-TNFR1 seems to be independent of BMX.^20,42

Although it has been suggested that both TNFα and VEGF increase phosphorylation of BMX on Y566, our studies indicate that TNFα-mediated VCAM-1 expression is not affected by pharmacological inhibition of BMX. One possible explanation is that, compared to VEGF, TNFα induces phosphorylation at fewer tyrosine residues. More specifically, TNFα induces Tyr(P)-1054/1059 on VEGFR2, but neither TNFα nor BMX induce phosphorylation at Tyr(P)-1175, a site that is critical for VEGF-mediated activation of PI3K/AKT and PLCγ pathways upstream of VCAM-1.³⁵ Taken together, this suggests that while BMX is dispensable in TNFα-mediated VCAM-1 expression, BMX plays a role in VEGF-mediated VCAM-1 expression on aortic endothelial cells, and as such is sensitive to ibrutinib. Furthermore, the observed decrease in VCAM-1 expression following one week of treatment with ibrutinib suggests that ibrutinib disrupts an inflammatory signaling pathway driving endothelial cell activation, or at least markers of activation, in vivo.

Although ibrutinib is often referred to as a first generation BTK inhibitor, previous studies showed that ibrutinib inhibits the kinase activities of numerous TFKs including BMX by irreversibly binding a conserved cysteine residue.³⁷ The TFKs share a common domain organization that consists of: (1) an N-terminal pleckstrin-homology domain; (2) a TEC-homology domain with a BTK motif and one to two proline-rich regions; (3) SRC homology 3 and 2 domains; and (4) a carboxy-terminal kinase domain.^34,41 Within the kinase domain, the residues in the ATP binding site share 40–65% identity and 60–80% similarity.³⁵ Ibrutinib reacts with a cysteine residue (C481) within the ATP binding site, thus blocking the catalytic activity of BTK by forcing the BTK kinase domain to adopt an inactive conformation.¹⁴ Similarly, the covalent BMX inhibitors, BMX-IN-1 and JS25, bind a cysteine residue (C496) in the ATP binding site. This residue is a unique occurrence found in the ATP binding pocket and is present in all five members of the TFKs (BTK, ITK, RLK/TXK and TEC), as well as members from the EGFR family (EGFR, HER2, HER4), JAK3, BLK and MAP2K7.³⁵ By virtue of structural homology, BMX-IN-1 and JS25 could also be covalent inhibitors of the other kinases in the TEC family.³⁵

Beyond compromised endothelial function, atherosclerosis is characterized by dysregulated platelet-endothelial interactions, which encourages aberrant leukocyte recruitment.^18,22 In this setting, platelets facilitate monocyte adhesion to the vascular endothelium and subsequent transmigration through the endothelial cell layer to facilitate the shift of lipid-laden macrophages to foam cells. Additionally, neutrophils as well as neutrophil-platelet interactions are implicated in the progression of atherogenesis.^9,28 Although the pro-atherosclerotic phenotype imparted by endothelial cell activation and the secretion of growth factors/chemokines/cytokines by platelets to facilitate leukocyte recruitment has been extensively studied, the role of TFKs in this system as a whole remains poorly defined. Since TFKs are not only expressed by hematopoietic cells, but also by other somatic cells, and ibrutinib promiscuously inhibits TFKs, the utility of ibrutinib could extend beyond its use in treating hematological malignancies and quelling endothelial dysfunction to more global effects within the blood microenvironment.

Abbreviations

BMX

Bone marrow tyrosine kinase gene in chromosome X

BTK

Bruton’s tyrosine kinase

CEU

Contrast-enhanced ultrasound

CRP-XL

Cross-linked collagen-related peptide

GPIbα

Glycoprotein-Ibα

GPVI

Glycoprotein VI

HAEC

Human aortic endothelial cell

IL

Interleukin

ICAM-1

Intercellular adhesion molecule 1

NHP

Nonhuman primate

ONPRC

Oregon National Primate Research Center

PPP

Platelet poor plasma

PRP

Platelet rich plasma

STAT3

Signal transducer and activator of transcription 3

TEC

Tyrosine kinase expressed in hepatocellular carcinoma

TFKs

TEC family kinases

TNF

Tumor necrosis factor

TNFR

Tumor necrosis factor receptor

VCAM-1

Vascular cell adhesion molecule 1

VEGF

Vascular endothelial growth factor

VEGFR

Vascular endothelial growth factor receptor

vWF

von Willebrand factor