Vascular Endothelial Dysfunction and Immunothrombosis in the Pathogenesis of Atrial Fibrillation

Abstract

Atrial Fibrillation (AF) induces proinflammatory processes which incite vascular endothelial activation and dysfunction. This study seeks to examine the potential relationship between various endothelial, inflammatory, thrombotic, and renin-angiotensin-system (RAS) biomarkers in AF patients.

Blood samples were from AF patients (n = 110) prospectively enrolled in this study prior to their first AF ablation. Control plasma samples (n = 100) were used as reference. All samples were analyzed for endothelial (NO, ICAM-1, VEGF, TF, TFPI, TM, Annexin V), inflammatory (IL-6, TNFα, CRP), thrombotic (vWF, tPA, PAI-1, TAFI, D-dimer), and RAS (Renin, Ang-II) biomarkers using ELISA methods. Biomarker average comparisons and Spearman correlations were performed.

AF patients showed varying levels of biomarker increase compared to controls. We observed a significant decrease of Ang-II in the AF population relative to controls when stratified for the use of angiotensin-converting enzyme inhibitor (ACEI) or angiotensin II receptor blocker (ARB) upon study enrollment. AF patients showed statistically significant correlations between the following biomarkers: TNFα vs IL-6 (r_s= 0.317, p = .004), ICAM-1 vs TNFα (r_s= 0.527, p = .012), Annexin V vs VEGF (r_s= 0.620, p < .001), CRP vs VEGF (r_s= 0.342, p = .031), Ang-II vs tPA (r_s= −0.592, p = .010), and tPA vs PAI-1 (r_s= 0.672, p < .001).

Our study demonstrated significant elevation of endothelial, inflammatory, and thrombotic biomarkers in AF patients compared to controls, with significant correlations between these biomarkers in the AF population. Future investigations are required to better elucidate the mechanistic pathways that lead to endothelial dysfunction and thromboinflammation in AF. This may provide novel therapeutic targets, that in addition to current anticoagulation practices, can best curtail thrombogenicity in AF.

Introduction

Atrial Fibrillation (AF) is the most prevalent cardiac arrhythmia across the world and is one of the main causes of cardiovascular morbidity and mortality. ¹ Currently in the United States, at least 3 to 6 million people have AF and the prevalence is estimated to reach 6 to 16 million by 2050.^2–5 There is a growing role of thromboinflammation in the pathogenesis of AF. ⁶ Approximately one-third of all strokes are caused by AF, and stroke in AF is described to be cardioembolic in nature 52% and 57% of the time in the Stroke Prevention in AF (SPAF) and Asymptomatic AF and Stroke Evaluation in Pacemaker Patients and the AF Reduction Atrial Pacing Trial (ASSERT) trials respectively.^7–9 Anticoagulation is widely known to decrease stroke risk in AF and is the mainstay of medical therapy. However, upon review of landmark clinical trials evaluating the efficacy of direct oral anticoagulants (DOAC) and warfarin among patients with AF, the residual risk of stroke or systemic embolism despite anticoagulation treatment was between 1.11% and 2.40% per year. ¹⁰ Additionally, a large investigation analyzing databases of AF patient health insurance claims illustrated that 12.4% of AF patients had a contraindication to oral anticoagulation and this treatment gap carried high risk of morbid cerebrovascular outcomes. ¹¹ These studies make evident the need for additional therapeutic strategies to curtail the hypercoagulable state of AF.

Currently, the complete pathophysiology of AF is not fully understood. A multifactorial pathogenesis is highly likely and there is an emerging understanding of the mechanistic link between AF and vascular endothelial cell dysfunction that contributes to the thrombogenic state of AF. ¹² Since there are numerous mediators that participate in the hemostatic response in AF, it is important to outline each of their roles as these factors have potential to serve as biomarkers of disease severity and/or therapeutic targets.

AF causes vascular endothelial cell injury and inflammation due to turbulent blood flow in the left atrium and/or systemically secondary to a lack of regular pulsatile laminar flow.^13,14 This shear stress on endothelial cells (EC) leads to local reactive vasoconstriction via contraction of vascular smooth muscle and can be initiated by the factors endothelin (ET-1) and angiotensin II (Ang-II). ¹⁵ ET-1 levels are induced by Ang-II, which along with nitric oxide (NO) are the primary regulators of vascular tone. ¹⁶ During primary hemostasis, EC injury leads to both EC secretion of von Willebrand factor (vWF) and damage induced exposure of the subendothelial extracellular matrix (ECM), which together promote platelet adhesion and activation to form a hemostatic plug. This is followed by secondary hemostasis where the coagulation system works together to form an insoluble clot composed of fibrin along with erythrocytes and leukocytes such as neutrophils. One such mediator in this process is endothelial derived tissue factor (TF) whose cell surface localization activates the coagulation system. ¹⁷ After the formation of a stable thrombus and transition towards wound healing, EC's secrete tissue plasminogen activator (tPA) which initiates the fibrinolytic system through plasmin activation and results in the enzymatic degradation of the clot.^18,19 This leads to thrombus breakdown and the potential dissemination of thrombotic emboli. Plasmin also contributes to wound healing processes both individually and through the activation of matrix metalloproteinases (MMPs) which leads to subendothelial ECM degradation and collagen remodeling.^20,21 These mediators have also been illustrated to play a role in the pathogenesis of AF.^22,23 It is well described that the vascular endothelium plays two distinct roles in each phase of the hemostasis-coagulation system. These roles are the endothelium's antithrombotic mechanisms during normal resting state, and its prothrombotic mechanisms during response to vascular injury. ²⁴

The objectives of our study were (1) to quantify the plasma concentration of endothelial, inflammatory, and thrombotic biomarkers in AF patients compared to a human control population; and (2) to determine potential correlations between these various biomarkers in the AF cohort. We hypothesized that AF exacerbates the vascular endothelium's immunothrombotic mechanisms illustrated by elevated levels of endothelial, inflammatory, and thrombotic biomarkers in AF patients relative to healthy controls, and that there may be an array of correlations between these biomarkers in AF patients that reinforce the role of endothelial dysfunction in the pathogenesis of AF.

Material and Methods

Patients with symptomatic AF undergoing first AF ablation (n = 110) were consecutively enrolled in this prospective study which was approved by the institutional review board of Loyola University Medical Center. Patients with untreated structural heart disease, severe left ventricular dysfunction (ejection fraction <30%), prior cardiac surgery or ablation procedures, uncontrolled hypertension, severe valvular heart disease, pregnancy, bleeding diathesis and coagulopathy, and those unable or unwilling to provide informed consent were excluded from the study. Blood samples were collected from patients upon study enrollment prior to AF ablation. Patient demographic information for the AF cohort is presented in Table 1. Blood samples were obtained in tubes containing 3.2% sodium citrate and subsequently centrifuged at 3000 x g for 15 min. We separated the plasma supernatant, aliquoted, and froze the samples at −80 °C prior to analysis. The plasma samples were labeled using each patient's study identification number without any patient identifying information. We then blindly analyzed the samples in the Hemostasis and Thrombosis Research Laboratories at Loyola University Medical Center. The control group in our study consisted of 100 normal human plasma samples (50% female) obtained from a commercial vendor, George King Biomedical (Overland Park, Kansas).

Table 1.

Baseline Characteristics of the Atrial Fibrillation Cohort.

Demographic Categories
Age (years ± SD)	61.3 ± 10.9
Sex; N (%)	110
Male	77 (70.0%)
Female	33 (30.0%)
BMI ± SD	30.62 ± 5.10
Smoking status; n (%)
No smoking history	59 (53.6%)
Former Smoker	48 (43.6%)
Current smoker	3 (2.7%)
Comorbidities; n (%)
Hypertension	68 (61.8%)
Hyperlipidemia	64 (58.2%)
Diabetes Mellitus	14 (12.7%)
Coronary Artery Disease	26 (23.6%)
Acute Coronary Syndrome	16 (14.5%)
Heart Failure	14 (12.7%)
Stroke/Transient Ischemic Attack	5 (4.5%)
Thromboembolism	9 (8.2%)
Peripheral Vascular Disease	5 (4.5%)
Chronic Kidney Disease	4 (3.6%)
Anticoagulation Therapy; n (%)
On anticoagulation	83 (75.5%)
- Rivaroxaban	36 (43.4%)
- Apixaban	32 (38.6%)
- Warfarin	11 (13.3%)
- Dabigatran	3 (3.6%)
- Edoxaban	1 (1.2%)
Antiplatelet Therapy; n (%)
On antiplatelet	38 (34.5%)
- Aspirin only	29 (76.3%)
- Clopidogrel only	7 (18.4%)
- DAPT	2 (5.3%)
Rate Control Therapy; n (%)
On rate control agent	76 (69.1%)
Beta Blockade	63 (82.9%)
- Metoprolol	53 (84.1%)
- Carvedilol	4 (6.3%)
- Bisoprolol	4 (6.3%)
- Atenolol	1 (1.6%)
- Nebivolol	1 (1.6%)
Nondihydropyridine CCB	17 (22.4%)
- Diltiazem	14 (82.4%)
- Verapamil	3 (17.6%)
- BB and CCB	4 (5.3%)
Antiarrhythmic Therapy; n (%)
On AAD agent	52 (47.3%)
- Dronedarone	15 (28.8%)
- Amiodarone	13 (25.0%)
- Flecainide	12 (23.1%)
- Sotalol	7 (13.5%)
- Propafenone	4 (7.7%)
- Dofetilide	1 (1.9%)
Other Pharmacotherapy; n (%)
- ACEI	24 (21.8%)
- ARB	12 (10.9%)
- Diuretic	22 (20.0%)
- Statin	48 (43.6%)
- Digoxin	2 (1.8%)
Transthoracic Echocardiogram prior to study enrollment; n (%)	70 (63.6%)
Valvular Heart Disease; n (%)
No VHD	18 (25.7%)
Mild VHD	35 (50.0%)
Significant VHD	17 (24.3%)
2D ECHO LA Size Assessment; n (%)	56 (80.0%)
LA Volume (mL ± SD)	87.9 ± 29.9

Abbreviations: SD = standard deviation, BMI = body mass index, DAPT = dual antiplatelet therapy, BB = beta blocker, CCB = calcium channel blocker, AAD = antiarrhythmic drug, ACEI = angiotensin-converting enzyme inhibitor, ARB = angiotensin II receptor blocker, VHD = valvular heart disease, LA = left atrium

To quantify the endothelial, inflammatory, and thrombotic biomarkers, we used commercially available sandwich enzyme-linked immunosorbent assay (ELISA) kits. The endothelial biomarkers screened were Nitric Oxide (NO), Intercellular Cell Adhesion Molecule-1 (ICAM-1), Vascular Endothelial Growth Factor (VEGF), Tissue Factor (TF), Tissue Factor Pathway Inhibitor (TFPI), Thrombomodulin (TM), and Annexin V. The inflammatory biomarkers studied were Interleukin-6 (IL-6), Tumor Necrosis Factor-Alpha (TNFα), and C-reactive Protein (CRP). The thrombotic biomarkers analyzed were von Willebrand Factor (vWF), Tissue Plasminogen Activator (tPA), Plasminogen Activator Inhibitor-1 (PAI-1), Thrombin Activatable Fibrinolysis Inhibitor (TAFI), and D-dimer. We also examined patients for levels of Renin and Angiotensin-II (Ang-II) which highlights the Renin-Angiotensin-System (RAS) involvement in regulation of vascular tone and resistance. If there was inadequate supply of plasma remaining for any of the healthy controls or AF patients, they were excluded from future enzyme linked immunosorbent assay (ELISA) analyses. The ELISA kits were obtained from various commercial vendors such as R&D Systems (Minneapolis, Minnesota), BioMedica Diagnostics (Windsor, NS Canada), Hyphen BioMed (Neuville-sur-Oise, France), and Abcam (Cambridge, United Kingdom).

Statistical analyses were carried out using Microsoft Excel, GraphPad Prism, and QI Macros Statistical Software. The mean values ± standard error of the mean (SEM) and standard deviations (SD) were calculated for each biomarker in the healthy control and AF groups. We also performed a percent change analysis between the healthy control and AF mean values for each respective biomarker. For the biomarkers Renin and Ang-II, patients were screened for use of either an angiotensin-converting enzyme inhibitor (ACEI) or angiotensin II receptor blocker (ARB) at time of study enrollment and these patients were excluded from analysis. To determine statistical significance between control and experimental groups, each biomarker data set was screened for normality and equal variance. Data sets that passed for normality and equal variance were subsequently analyzed for significance using a parametric two-tailed t test. The only biomarker was TAFI. Data sets that passed for normality, but had unequal variance were analyzed using an unpaired two-tailed t test with Welch's correction. The only biomarker was vWF. If a data set was identified to not be of Gaussian distribution, we performed statistical analysis using the nonparametric Mann-Whitney U test. These biomarkers included: NO, ICAM-1, VEGF, TF, TFPI, TM, Annexin V, IL-6, TNFα, CRP, tPA, PAI-1, D-dimer, Renin, and Ang-II. For all analyses, we set a p-value of < .05 as the threshold for statistical significance. Additionally, median values and interquartile ranges (IQR) were also calculated for each biomarker. The Mood's median test was used to assess for significance between the median values of the healthy control and AF groups. We also set the level of statistical significance for the Mood's median test to be a p-value < .05. Lastly, we completed Spearman correlation analyses to assess for potential relationships between biomarkers in the AF cohort.

A post-hoc power analysis revealed that the study achieved a statistical power of at least 80% across all biomarkers, which meets commonly accepted criterion for sufficient power in biomedical research. Of the majority of the biomarkers analyzed; the statistical power exceeded 90%.

Results

Endothelial Biomarkers

The mean comparisons of endothelial biomarker results in the normal human plasma control group (NHP) (n = 17-76) and AF populations (n = 33-58) are depicted in Table 2. There was varying degree of biomarker elevation in the AF group compared to NHP. The most pronounced increases were noted in VEGF, TF, and Annexin V which were statistically significant. The AF cohort showed a wider range in each of the biomarkers compared to healthy controls. The medians and IQRs for each endothelial biomarker are represented in Table 3. All of the endothelial biomarkers in the AF cohort, except for TFPI, displayed higher medians and IQRs compared to NHP. There was a statistically significant median elevation for VEGF, TF, and TM in the AF group relative to NHP.

Table 2.

Comparison of Endothelial, Inflammatory, Thrombotic, and RAS Biomarker Mean Concentrations Between Healthy Control and Atrial Fibrillation Groups.

Biomarker	Healthy control mean ± SEM (SD)	Atrial fibrillation mean ± SEM (SD)	P-value	AF vs control % Change
NO (µmol/L)	13.80 ± 0.791 (5.596)	23.51 ± 3.368 (19.35)	.0564	70%
ICAM-1 (ng/mL)	128.9 ± 7.46 (34.19)	155.4 ± 7.13 (45.07)	.0206	21%
VEGF (pg/mL)	18.54 ± 2.368 (14.60)	60.97 ± 7.346 (46.46)	<.0001	229%
TF (pg/mL)	0.9045 ± 0.441 (1.820)	86.33 ± 44.65 (334.1)	.0003	9445%
TFPI (ng/mL)	130.2 ± 10.03 (45.97)	140.45 ± 14.58 (111.1)	.5855	8%
TM (ng/mL)	3.809 ± 0.146 (0.6519)	4.848 ± 0.2877 (2.191)	.0606	27%
Annexin V (ng/mL)	5.451 ± 0.857 (7.468)	11.62 ± 2.352 (17.91)	.0207	113%
IL-6 (pg/mL)	2.940 ± 2.438 (19.20)	18.04 ± 5.396 (52.60)	<.0001	514%
TNFα (pg/mL)	0.7933 ± 0.087 (0.3796)	3.977 ± 0.480 (4.499)	<.0001	401%
CRP (µg/mL)	0.6282 ± 0.116 (0.9832)	3.257 ± 0.327 (3.235)	<.0001	418%
vWF (%)	98.24 ± 3.115 (22.03)	4796 ± 289.4 (1941)	<.0001	4782%
tPA (ng/mL)	4.233 ± 0.4529 (3.202)	5.010 ± 0.9205 (6.377)	.9972	18%
PAI-1 (ng/mL)	5.577 ± 0.9676 (5.120)	5.931 ± 0.7009 (3.839)	.3716	6%
TAFI (%)	69.79 ± 1.645 (7.890)	75.43 ± 1.244 (9.472)	.0135	8%
D-dimer (ng/mL)	192.8 ± 31.39 (212.9)	502.1 ± 107.7 (628.1)	.0001	160%
Renin (ng/mL)	401.1 ± 26.88 (137.1)	970.4 ± 234.7 (1197)	.0006	142%
Ang-II (pg/mL)	1127 ± 79.32 (469.3)	932.4 ± 113.5 (786.6)	<.0001	−17%

Table 3.

Comparison of the Medians and Interquartile Range (IQR) of Endothelial, Inflammatory, Thrombotic, and RAS Biomarker Concentrations Between Healthy Control and Atrial Fibrillation Groups.

Biomarker	Healthy control median (IQR)	Atrial fibrillation median (IQR)	P-value
NO (µmol/L)	12.70 (9.98–16.28)	18.45 (10.30–33.19)	.0529
ICAM-1 (ng/mL)	118.43 (103.55–141.76)	153.66 (119.81–186.52)	.0200
VEGF (pg/mL)	14.59 (9.42–20.76)	49.94 (30.25–78.05)	<.0001
TF (pg/mL)	0 (0–1.11)	5.64 (0–17.97)	.0029
TFPI (ng/mL)	127.53 (97.10–156.32)	120.07 (92.59–137.38)	.7471
TM (ng/mL)	3.87 (3.43–4.15)	4.15 (3.50–5.01)	.0380
Annexin V (ng/mL)	1.14 (0–8.80)	4.47 (1.19–10.63)	.0812
IL-6 (pg/mL)	0 (0–0.26)	1.31 (0–5.77)	<.0001
TNFα (pg/mL)	0.74 (0.42–0.93)	2.31 (1.67–5.02)	<.0001
CRP (µg/mL)	0.12 (0–1.06)	2.08 (0.86–5.46)	<.0001
vWF (%)	102.53 (78.74–114.98)	4456 (3479–5774)	<.0001
tPA (ng/mL)	4.17 (1.42–6.35)	3.63 (3.00–4.68)	.2253
PAI-1 (ng/mL)	3.87 (1.56–8.50)	5.19 (2.79–8.55)	.2932
TAFI (%)	70.11 (65.81–73.46)	75.86 (70.00–82.62)	.0083
D-dimer (ng/mL)	126.17 (4.65–368.41)	356.01 (198.55–546.90)	.0015
Renin (ng/mL)	381.15 (320.15–488.53)	730.89 (403.93–1018)	.0055
Ang-II (pg/mL)	1011 (934–1126)	811 (422–941)	.0001

Inflammatory Biomarkers

The mean inflammatory biomarker results in the NHP (n = 19-72) and AF groups (n = 88-98) are illustrated in Table 2. All of the inflammatory biomarkers, IL-6, TNFα, and CRP, showed statistically significant elevation in the AF cohort relative to NHP, with all three exhibiting four to five-fold increases. The medians and IQRs for each inflammatory biomarker illustrated in Table 3 were increased in the AF group compared to NHP with wider ranges seen in the AF cohort. There was a statistically significant elevation in the median concentration of IL-6, TNFα, and CRP in the AF cohort compared to NHP.

Thrombotic Biomarkers

The mean thrombotic biomarker results in the NHP (n = 23-50) and AF groups (n = 30-58) are represented in Table 2. There was varying magnitude of biomarker increase in the AF group compared to NHP. The most pronounced increases were observed in vWF and D-dimer. There was a statistically significant elevation of vWF, TAFI, and D-dimer in the AF group relative to NHP. The medians and IQRs for each thrombotic biomarker are depicted in Table 3. There were increased median values and IQRs for the biomarkers vWF, PAI-1, TAFI, and D-dimer in the AF group compared to NHP. For the biomarker tPA, the AF population had a lower median relative to the NHP cohort though was not statistically significant. There was a statistically significant median concentration elevation of vWF, TAFI, and D-dimer in the AF group compared to NHP.

RAS Biomarkers

The mean RAS biomarker results in the NHP (n = 26-35) and AF groups (n = 26-48) are shown in Table 2. There was a statistically significant elevation of Renin and a statistically significant decrease of Ang-II in the AF cohort relative to NHP. These significant findings were also observed upon analysis of the median values between the AF and NHP populations. For the biomarker Renin, there was a larger IQR in the AF group. Although the Ang-II median was higher in the NHP group, the IQR in the AF group was observed to be larger.

Spearman Correlation Analysis

We performed Spearman correlation analyses between all studied biomarkers in our AF cohort. Several statistically significant correlations are observed between the endothelial, inflammatory, thrombotic, and RAS biomarkers. These correlations graphically visualized in Figure 1 include: TNFα vs IL-6 (r_s= 0.317, p = .004), ICAM-1 vs TNFα (r_s= 0.527, p = .012), Annexin V vs VEGF (r_s= 0.620, p < .001), CRP vs VEGF (r_s= 0.342, p = .031), Ang-II vs tPA (r_s= −0.592, p = .010), and tPA vs PAI-1 (r_s= 0.672, p < .001).

Figure 1.

Spearman Correlation Analyses of Biomarker Concentrations in the Atrial Fibrillation Cohort.

Discussion

This was a prospective study evaluating endothelial, inflammatory, thrombotic, and RAS biomarkers in AF patients in comparison to normal human plasma controls. We observed varying levels of increase of each endothelial biomarker in the AF population relative to NHP. There were statistically significant increases in the endothelial biomarkers VEGF, TF, and Annexin V in AF who exhibited two, ninety, and one-fold increases respectively. These findings align with previous AF blood plasma investigations which also demonstrated elevation of VEGF and TF mediators and endothelial cell (EC) apoptosis represented by Annexin V.^25–27 Interestingly, in our studies the biomarker TM exhibited a statistically significantly higher median concentration in AF relative to NHP. TM when localized to the membrane of ECs serves as a receptor for thrombin, fulfilling anticoagulant properties through their joint activation of Protein C and prevention of thrombin induced formation of fibrin. ²⁸ However, increased plasma levels of TM are considered to be a marker of endothelial dysfunction and have been shown in previous studies to be upregulated in AF patient plasma.^28,29 These findings together reaffirm the role of endothelial dysfunction in the pathogenesis of AF.

Additionally, there was statistically significant plasma elevation of the inflammatory biomarkers IL-6, TNFα, and CRP and the thrombotic biomarkers vWF, TAFI, and D-dimer in the AF population compared to healthy controls. The inflammatory biomarkers together exhibited four to five-fold increases in AF patients while vWF illustrated a forty-fold increase. Several studies have described upregulation of these same biomarkers in patients with AF.^30–33 To the best of our knowledge, it appears that we may be the first study illustrating increased levels of TAFI in AF relative to a normal control population, though it has been described that TAFI elevation in AF is associated with prior history of stroke, increased thromboembolic event risk, and hypofibrinolysis.^34,35 These results support the mechanistic role of thromboinflammation in the development and maintenance of atrial fibrillation and their interrelationship will be discussed further below.

Furthermore, we observed in our RAS biomarkers analysis a statistically significant increase in Renin and significant decrease in Ang-II in the AF cohort relative to the healthy controls. This is an interesting finding as one would typically expect that as Renin levels increase there would also be an elevation in Ang-II. A potential explanation for this result can be outlined using the RAS pathway with focus on the ACE/Ang-II/AT1R axis and ACE2/A1-7/MAS1 axis that Mascolo et al neatly summarized in their review. ³⁶ Renin converts angiotensinogen into angiotensin-I (Ang-I) which is a substrate for two downstream pathways. One pathway is the classical conversion of Ang-I into Ang-II by angiotensin converting enzyme (ACE). The other pathway is the conversion of Ang-I by the monocarboxypeptidase angiotensin converting enzyme 2 (ACE2) into the peptides angiotensin 1-9 (A1-9) which are later cleaved by ACE into A1-7. ACE2 is also capable of generating A1-7 directly from Ang-II. A1-7 through their interaction with the MAS1 receptor have been summarized to play a significant role in the reduction of cardiac fibrosis and their ability to reverse remodeling.^36–38 Additionally, ACE2 gene expression has been described to be mechanoresponsive in vascular disease. A study by Song et al illustrated that vascular shear stress and stretch on vascular smooth muscle cells increased ACE2 promotor activity leading to increased ACE2 expression, increased production of A1-7, and decreased levels of Ang-II. ³⁹ Furthermore, it has been described that in bicuspid aortic valve patients with ascending aortic dilatation there is increased expression of ACE2 which the authors considered to be a compensatory response to minimize vascular injury. ⁴⁰ In our studies, we hypothesize that in established AF there may be compensatory upregulation of ACE2, increasing the production of A1-7 in attempt to limit or reverse cardiac fibrosis. This could explain the significant decrease of Ang-II in our AF population secondary to heightened ACE2 conversion of Ang-II to A1-7.

Across our studies, we discovered statistically significant correlations between these biomarkers in AF patients, which together best explain an immunothrombotic state. The positive correlation between the inflammatory biomarkers TNFα vs IL-6 is likely secondary to proinflammatory processes incited by turbulent blood flow stress on the vascular endothelium during active cardiac arrythmia. ⁴¹ In our studies, we also found a positive correlation between ICAM-1 vs TNFα. TNFα is a core proinflammatory cytokine that upregulates the expression of endothelial ICAM-1, locally recruiting leukocyte mediators.^42,43 At the site of the injured endothelium, platelets adhered to the exposed matrix can activate recruited neutrophils leading to the development of neutrophil extracellular traps (NETs). These NETs can activate the coagulation cascade via degradation of TFPI and further promote a thrombogenic state.^44,45

We also saw a positive correlation between Annexin V vs VEGF. It has been described that in AF, several mechanisms, such as high ventricular heart rate, low or oscillatory shear stress, stretch, hypoxia, inflammation and oxidative stress, are potent inducers of apoptotic cell death, which leads to the shedding of procoagulant endothelial derived microparticles (EMPs) within the vasculature. ⁴⁶ The elevated apoptotic ECs and EMPs in AF are both marked by upregulated levels of phosphatidylserine in their outer plasma membrane,^47,48 likely represented by elevated levels of Annexin V in our AF cohort. This cellular dysfunction influences endothelial and other local cells to secrete VEGF, stimulating various immunothrombotic processes through increased vascular permeability, recruitment and migration of inflammatory cells, forward modulation of procoagulant activity, and wound healing mechanisms.^49–52

Our studies also depicted a positive correlation between CRP vs VEGF. It has been described by Schneeweis and colleagues that chronic CRP exposure inhibits VEGF-induced endothelial cell migration in analyses of cultured human umbilical vein endothelial cells. ⁵³ Their thoughtfully crafted studies elucidated two distinct signaling pathways of VEGF-mediated endothelial cell migration, which are the phosphatidylinositol 3-kinase (PI3K) → Protein kinase B (Akt) → Endothelial NO synthase (eNOS) and Extracellular Signal-Regulated kinases 1 and 2 (ERK1/2) pathways respectively. They observed that CRP exclusively inhibits the Akt → eNOS segment of the signaling cascade with no effect on the phosphorylation and activation of the ERK1/2 pathway, which may explain why VEGF-induced endothelial cell migration was only partially inhibited by CRP. Thus in our investigations, the vascular endothelium of patients with AF potentially upregulate VEGF secretion in attempt to enhance induction of endothelial cell migration via the ERK1/2 pathway, as the PI3K → Akt → eNOS pathway could likely be inhibited by increased levels of CRP. Overall, patients with AF may be predisposed to inadequate wound healing due to chronic elevation of CRP, reducing the ability to repair the injured endothelium via EC migration. This may be a potential mechanism that impedes resolution of endothelial damage and dysfunction, and further drives thromboinflammation in AF.

There was also a negative correlation between biomarkers Ang-II vs tPA. This relationship may be explained by the aberrant prothrombotic state in AF that leads to increased endothelial activation of the fibrinolytic pathway via secretion of tPA. Thus, hemostatic mechanisms may be skewed towards downstream fibrinolytic processes represented by elevated levels of tPA and decreased levels of Ang-II. Endothelial cell injury likely promotes the secretion of another endothelial derived fibrinolytic factor, PAI-1. Although ECs may secrete tPA to circumvent the thrombogenic state of AF, there may be inappropriate secretion of PAI-1 that leads to impaired fibrinolytic function and subsequent maintenance of thrombotic disease. ⁵⁴ Our AF cohort illustrated a positive correlation between tPA vs PAI-1, potentially explained by increased tPA/PAI-1 complex formation as previously described in studies investigating AF, disseminated intravascular coagulation, and thrombotic thrombocytopenic purpura.^54,55

This study has a few limitations. The role of anticoagulation in the study in the treatment of AF has not been included. Several studies have discussed the role of anticoagulation, and it will be valuable to design additional studies.^56–58 Our analyses performed were limited by small cohort size secondary to total patient enrollment and the nature of ELISA methodology. We were also unable to obtain demographic information of our normal human plasma population to arrange age and gender-matched controls. Additionally, there was a significantly higher number of males than females in the AF population. This study was based on a single sample analysis, and due to logistical reasons, follow-up samples were not collected or analyzed. Future studies should consider collecting sequential samples for follow-up analysis to assess changes in biomarkers after the treatment of AF. We also could not demonstrate the relationship between the biomarkers and AF severity or its complications. Furthermore, the patients’ comorbidities such as hypertension, diabetes mellitus, obesity, and others are risk factors for atrial fibrillation and may each have a unique effect on biomarker results which is difficult to outline.

In conclusion, our study demonstrated significant elevation of endothelial, inflammatory, and thrombotic biomarkers in AF patients compared to healthy controls, with significant correlations between these biomarkers in the AF population. Our investigation illustrates the relationship between endothelial dysfunction and immunothrombosis in the pathogenesis of AF. Future studies are required to obtain a better mechanistic understanding between endothelial dysfunction and thromboinflammation in AF, which may provide novel therapeutic targets that in addition to current anticoagulation therapies, can best manage thrombogenicity in atrial fibrillation.