Changes in cerebral autoregulation, stroke‐related blood biomarkers, and autonomic regulation after patent foramen ovale closure in severe migraine patients

Abstract Aims This study aimed to investigate changes in dynamic cerebral autoregulation (dCA), 20 stroke‐related blood biomarkers, and autonomic regulation after patent foramen ovale (PFO) closure in severe migraine patients. Methods Patent foramen ovale severe migraine patients, matched non‐PFO severe migraine patients, and healthy controls were included. dCA and autonomic regulation were evaluated in each participant at baseline, and within 48‐h and 30 days after closure in PFO migraineurs. A panel of stroke‐related blood biomarkers was detected pre‐surgically in arterial‐and venous blood, and post‐surgically in the arterial blood in PFO migraineurs. Results Forty‐five PFO severe migraine patients, 50 non‐PFO severe migraine patients, and 50 controls were enrolled. The baseline dCA function of PFO migraineurs was significantly lower than that of non‐PFO migraineurs and controls but was rapidly improved with PFO closure, remaining stable at 1‐month follow‐up. Arterial blood platelet‐derived growth factor‐BB (PDGF‐BB) levels were higher in PFO migraineurs than in controls, which was immediately and significantly reduced after closure. No differences in autonomic regulation were observed among the three groups. Conclusion Patent foramen ovale closure can improve dCA and alter elevated arterial PDGF‐BB levels in migraine patients with PFO, both of which may be related to the preventive effect of PFO closure on stroke occurrence/recurrence.


| INTRODUC TI ON
Previous studies have reported that ischemic stroke is more common in patients with migraine, with 5.4% and 0.7% in migraineurs and controls, respectively, 1,2 the mechanisms are largely unknown.
Several recent studies have found a strong association among migraine, stroke, and patent foramen ovale (PFO), 3,4 and thus PFO may be the link between migraine and stroke. It is important to explore whether the presence of PFO in the migraine population causes specific pathophysiological changes that predispose migraine patients to stroke. The brain is highly dependent on constant and stable cerebral blood flow (CBF) despite changes in cerebral perfusion pressure and arterial blood pressure (ABP). Dynamic cerebral autoregulation (dCA) is critical in regulating cerebral hemodynamics and its dysfunction plays a role in the occurrence and progression of stroke, 5 as an impaired dCA may increase an individual's susceptibility to stroke and impair the cerebrovascular function of clearing away the microemboli. 6 Interestingly, a previous study found that dCA is impaired in migraine patients with PFO, especially in those with a large rightto-left shunt. 7 Thus, dCA impairment may be an intermediate link between PFO and stroke in a migraine patient. Accordingly, whether the PFO closure can alleviate impaired dCA is a key concern and is still unclear.
In patients with a PFO, venous blood can enter the arterial blood circulation through the oval foramen, which may produce alterations in several stroke-related blood biomarkers in the arterial blood, such as in homocysteine, 8 related to the increased risk of stroke occurrence. Identifying the changes in stroke-related blood biomarkers in migraine patients with PFO before and after PFO closure may help us understand the mechanisms of PFO-associated stroke and pinpoint the targets for its management.
Traditional risk factors for ischemic stroke include body stressors that may lead to autonomic dysfunction, which can be reflected by beat-to-beat heart rate variability (HRV). 9 Previous studies have reported that PFO is closely related to autonomic dysfunctional disorders such as irritable bowel syndrome. 10 Thus, it is worth exploring the characteristics of autonomic function in migraine patients with PFO and the effect of PFO closure on autonomic functioning.
In the present study, we investigated whether PFO closure affects dCA function, blood biomarkers, and autonomic function in severe migraine patients.

| ME THODS
The raw data of this study are available from the corresponding author upon reasonable request. This prospective study was approved by the ethics committee of the First Hospital of Jilin University (2012-111). Written informed consent was obtained from all participants. The participants had the right to withdraw at any time point during the course of the study.

| Participants
Severe migraine was defined as Headache Impact Test-6 score > 55. 11 Severe migraine patients (aged 18-65 years) with a large right-to-left shunt (contrast-enhanced transcranial Doppler >10 microbubbles) 12 were screened from the Department of Neurology, and who were scheduled to undergo PFO closure at the Department of Cardiac Surgery from January 2013 to June 2018 were included as PFO migraineurs group in this study. The effect of PFO closure surgery was determined by consultation with the cardiac surgeons, neurologists, cardiologists, radiologists, and sonographers taking into consideration the patient's wishes.
The exclusion criteria include: (1) insufficient bilateral temporal bone window for insonation of the middle cerebral artery (MCA); (2) extracranial/intracranial artery stenosis (including mild, moderate, and severe stenosis) or occlusion 13 [measured with an EMS-9 PB transcranial Doppler detector (Delica) and an iU22 ultrasound (Phillips)]; (3) patients who had brain tumor, stroke, white matter lesions, and other structural abnormities of brain screened with brain CT scan, myocardial infarction, and unstable angina within the past 6 months, or atrial fibrillation; (4) who were administered medications known to affect vasomotion within 1 week before the study; (5) who had large residual shunting measured with contrast-enhanced transcranial Doppler (>10 microbubbles) at the 1-month follow up; (6) who did not agree to participate in the study.
Age-and sex-matched severe migraineurs without a PFO (determined by contrast-enhanced transcranial Doppler) and healthy migraine-free volunteers were also enrolled as non-PFO migraineurs and healthy controls. Informed consent was obtained. The exclusion criteria applied were (1)-(4) and (6), like patients in the PFO migraineurs.
Simultaneously, volunteers with a right-to-left shunt (screened from healthy volunteers) were included in PFO non-migraineurs group.
Conclusion: Patent foramen ovale closure can improve dCA and alter elevated arterial PDGF-BB levels in migraine patients with PFO, both of which may be related to the preventive effect of PFO closure on stroke occurrence/recurrence.

K E Y W O R D S
autonomic regulation, dynamic cerebral autoregulation, migraine, patent foramen ovale, stroke Basic demographic and clinical characteristics were collected from all participants, and dCA assessment, biomarker measurement, and autonomic function evaluation were performed on all participants.

| Contrast-enhancedtranscranial Doppler protocol
We performed contrast-enhanced transcranial Doppler before PFO closure, as previously described. 7 The insonation of one MCA was accomplished via transcranial Doppler (MultiDop X4, DWL) with the patient in a supine position. An 18-gauge needle was inserted into the patient's cubital vein, and a mixed contrast agent (9 mL normal saline, 1 mL air, and a drop of the patient's blood) was injected as a rapid bolus. Testing was performed once at rest and twice during the Valsalva maneuver, and the maximum right-to-left shunt was recorded. The presence of >10 microbubbles was indicative of a large right-to-left shunt or large residual shunting. 12

| TranscatheterPFOclosure
Patients in the PFO migraineurs group underwent PFO closure surgery. Procedures were performed by a chief surgeon and involved a femoral approach with the patient under local anesthesia. A long sheath was advanced into the left atrium and an Amplatzer PFO Occluder (AGA Medical Corporation), which was a self-expanding double-disk device, was used to occlude the PFO under fluoroscopic guidance, and closure was confirmed via intra-cardiac echocardiography. Patients were administered heparin (80-100 U/kg) during the procedure and they subsequently underwent routine electrocardiography, contrast transthoracic echocardiography, and chest radiography 24 h after the PFO closure procedure. Aspirin (100 mg/day) alone or in combination with clopidogrel (75 mg/day) was administered as antiplatelet therapy for 6 months following the procedure.

| EvaluationofdCA
dCA data were collected from all participants, and PFO migraineurs received two additional dCA measurements, one within 48 h postsurgery and the other at 1-month post-surgery.
dCA was evaluated in a dedicated soundproof room, as previously reported. 14 With the participant in the supine position, continuous cerebral blood flow velocity (CBFV) and ABP were recorded simultaneously for approximately 10 min in the bilateral MCA using transcranial Doppler (MultiDop X4, DWL), ABP was measured at the fingertip of the middle finger using a servo-controlled plethysmograph (Finometer Model 1, FMS), respectively. The end-tidal CO 2 was also measured using a capnograph (MultiDop X4, DWL) with a face mask attached to the nasal cannula to confirm the stability of respiration during the experiment, in accordance with the recommendations of the dCA monitoring white paper. 15 The dynamic relationship between the continuous CBFV and ABP was assessed via transfer function analysis (TFA) using MATLAB (Math Works). Evaluation parameters included phase difference (the main parameter, determined as the phase shift angle ranging from 0° to 90°), gain (difference in the amplitude between CBFV and ABP), and the coherence function (indicates signal-tonoise ratio). Each parameter was averaged in 0.06-0.12 Hz, 14 and a cut-off value for coherence was set at 0.34 to establish the validity of TFA estimates as previously recommended (number of windows: 5; critical values of coherence estimate: α = 5% significant level). 15 Generally, a lower phase difference indicates an impaired dCA and a phase difference at 0° indicates a CBFV that passively follows ABP changes with no dCA at all. High gain at the same frequency band is also suggestive of compromised dCA for passively transferring the amplitude of ABP to CBFV, though less sensitive than phase difference. 16

| Beat-to-beatHRVmeasurementandanalysis
The autonomic function was assessed through beat-to-beat HRV analysis. The baseline beat-to-beat HRV of each participant was determined while dCA values were measured. Then, PFO migraineurs underwent two additional HRV measurements: one within the 48-h period following surgery and the other 1 month after surgery. Continuous recordings of beat-to-beat information obtained from the servo-controlled plethysmograph were processed using MATLAB scripts developed by our research team, as previously reported. 17 Ectopic beats and artifacts were automatically detected, visually confirmed, and removed via linear interpolation. The software processed the beat-to-beat recordings and generated heart period values. HRV was calculated using Heart Rate Variability

| Bloodbiomarkermeasurement
Human specimens were obtained from the Department of Biobank, Division of Clinical Research, the first hospital of Jilin University.
Arterial blood (obtained from radial artery) and venous blood (obtained from the cubital vein/basilic vein) were collected in all participants, and arterial blood was collected again immediately post-surgery from the PFO migraineurs. Twenty stroke-related blood biomarkers were analyzed with the Quantibody Human custom array from RayBiotech using a quantitative cytokine chip as previously reported. [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36] A list of the 20 biomarkers as well as their relationships with stroke is shown in Table 1. The measurement was performed according to the manufacturer's instructions, and four technical replicates were applied for each sample. A GenePix 4000B laser scanner (Bio-Rad Laboratories) was used to capture signals (green fluorescence, Cy3 channel, 555 nm excitation, and 565 nm emission). We analyzed the images using GenePix Pro 6.0 microarray analysis software and quantified the levels of biomarkers according to the standard curve calibrated from the same array.

| Statisticalanalysis
Data were analyzed using Statistical Program for Social Sciences version 19.0 (SPSS, IBM Corp.). The distribution of data was assessed using a one-sample Kolmogorov-Smirnov test and numerical variables were analyzed based on normality. Normally distributed data are expressed as mean ± standard deviation (SD) and analyzed with one-way ANOVA or Student's t-test. Nonnormally distributed data are expressed as median (interquartile range) and analyzed with Wilcoxon signed-rank test (for paired continuous data) or Mann-Whitney test (for independent data). A general linear model was used to detect differences among PFO migraineurs of repeated measurements (before, after, and 1-month post-closure), as deemed appropriate based on data distributions. Angiotensinogen, Renin Components of the renin-angiotensin system, implicated in atherosclerosis and associated with increased risk for ischemic stroke 20 CRP An acute phase reactant appears to be predictive of stroke risk 21 lgG Provides important information on the humoral immune status and has been reported to be associated with the occurrence of cardio-cerebrovascular disease events 22 IL-1β, IL-6 Lymphatic factors associated with the risk of ischemic stroke 23,24 Kallikrein 5 A component of the kallikrein-kinin system that protects against vascular injury and stroke 25 LOX-1 A transmembrane endocytosis receptor of oxidized low-density lipoprotein involved in the development of atherosclerotic disease, which is increased in ischemic stroke and transient ischemic attack 26 MIP-1α Plays a potentially important role in the development of inflammatory responses and plays a significant role in the etiopathogenesis of cardiovascular disease 27 MMP-2, MMP-9 The two best-studied matrix metalloproteinases, increase of which results in aberrant proteolysis contributing to blood-brain barrier dysfunction and in part determining the extent of the infarct 28 PAI-I, t-PA Components of the fibrinolysis system; PAI-I inhibits the activity of t-PA, and t-PA activates plasminogen; imbalance of the system leads to stroke 29 PDGF-BB One of the most potent stimulants promoting the development of atherosclerosis 30 S100b A reliable biomarker of neural injury including stroke 31 Tau Involved in the regulation of blood-brain barrier integrity, and opening of the blood-brain barrier is implicated in stroke 32 TNFα A proinflammatory and proatherogenic cytokine that modulates tissue injury in stroke 24,33 VEGF-a Participates in atherosclerosis, neuroprotection, neurogenesis, and angiogenesis; its decrease leads to endothelial dysfunction and increased risk of cardio-cerebrovascular disorders 19,34 VE-Cadherin An endothelial-specific adhesion protein at adherens junctions, related to vascular integrity and associated with progressive ischemic stroke 35,36 Abbreviations: CRP, C reactive protein; IL, interlukin; lgG, immunoglobulin; LOX-1, lectin-like oxidized low-density lipoprotein receptor-1; MIP-1α, macrophage inflammatory protein-1α; MMP-2, matrix metalloproteinase-2 MMP-9; PAI-I, plasminogen activator inhibitor-I; PDGF-BB, plateletderived growth factor-BB; p-TA, tissue-type plasminogen activator; TNFα, tumor necrosis factorα; VE-Cadherin, vascular endothelial-Cadherin; VEGF-a, vascular endothelial growth factor-a.

| Demographicinformation
Of the 61 PFO migraineurs, five had an insufficient bilateral temporal bone window and 11 had a large residual shunting at the 1-  (Tables S1-S3).
The demographic and clinical information of PFO migraineurs, non-PFO migraineurs, and healthy controls are presented in were assessed for 20 stroke-related arterial blood biomarkers.

| Dynamiccerebralautoregulationevaluation
As shown in Table 3, the PFO migraineurs exhibited significantly lower phase differences in the left hemisphere, the right hemisphere, and the whole brain overall (combined left and right hemispheres)  (Table 3).

| Platelet-derived growth factor-BB
The arterial PDGF-BB levels were higher in PFO migraineurs than in healthy controls, which were immediately corrected by PFO closure (  Figure 3A,B], possibly due to the abnormally high arterial PDGF-BB levels. Immediately after PFO closure, the arterial PDGF-BB levels in PFO migraineurs decreased significantly ( Figure 3C and

| Beat-to-beatheartratevariability
No significant differences in HRV values were observed between the healthy controls and PFO migraineurs (

| DISCUSS ION
In this study, we report that dCA was impaired in migraine patients with PFO, which was promptly and persistently improved upon PFO closure. PFO closure could also immediately and significantly restore the abnormally high levels of PDGF-BB in the arterial blood. Owing to dCA, constant cerebral blood flow could be maintained despite changes in cerebral perfusion pressure and ABP. dCA is an important indicator of a normal cerebrovascular functioning. 14 As Caplan reported, poor cerebrovascular function could diminish the ability of the cerebral circulation to clear thromboemboli and limit the availability of blood flow to the ischemic regions, 6 and thus a dysfunction in dCA may increase an individual's susceptibility to stroke. In the present study, we confirmed the presence of dCA impairment in patients with PFO. Moreover, as shown in Table S2, although the sample size was small, dCA of PFO patients without migraine was significantly lower than those measured in migraine patients without PFO and healthy control. Therefore, we speculated that migraine may be not associated with dCA alteration. And our results were in line with the previous study. 39,40 The possible mechanism of impaired dCA in PFO patients may involve cortical spreading depression induced by cerebral arterial microemboli originating from the venous system, which further results in impaired dCA by affecting vasomotor function. 41 Furthermore, we found that PFO closure could reverse the impaired dCA in patients with PFO, thus supporting the notion that PFO migraineurs could benefit from PFO closure.
In the analysis of blood biomarkers, we found dramatically high arterial PDGF-BB levels, and an altered arteriovenous PDGF-BB ratio in PFO migraineurs. Several in vivo and in vitro studies have shown that PDGF-BB is one of the most potent stimulants for vascular smooth muscle cell proliferation. PDGF-BB is also potent in transforming vascular smooth muscle cells from a contractile phenotype to a synthetic phenotype, a critical step in the development of atherosclerosis. 30 The TOSS-2 trial (Trial of Cilostazol in Symptomatic Intracranial Stenosis-2) reported that the venous TA B L E 4 Biomarkers in the arterial blood in migraineurs with patent foramen ovale (before and after closure) and controls. Abbreviations: CRP, C-reactive protein; IgG, immunoglobulin G; IL-1β, interleukin-1β; IL-6, interleukin-6; LOX-1, lectin-like oxidized low-density lipoprotein receptor-1; MIP-1α, macrophage inflammatory protein-1α; MMP-2, matrix metalloproteinase-2; MMP-9, matrix metalloproteinase-9; PAI-I, plasminogen activator inhibitor-I; PDGF-BB, platelet-derived growth factor-BB; PFO, patent foramen ovale; TNFα, tumor necrosis factorα; t-PA, tissue-type plasminogen activator; VE-cadherin, vascular endothelial-cadherin; VEGF-a, vascular endothelial growth factor-a. a p < 0.05 compared with healthy controls using Mann-Whitney test.
b p < 0.05 compared with before closure using Wilcoxon signed-rank test.
PDGF-BB level is associated with the progression of symptomatic intracranial atherosclerotic stenosis. 42 And a study found that baseline venous PDGF-BB was higher in stroke patients than in controls. 43 Furthermore, PDGF-BB acts as a vasoconstrictor 44   time-domain and absolute frequency-domain HRV after PFO closure decreased significantly, consistent with the previous findings. 46 The immediate post-surgical HRV alteration and autonomic dysfunction are likely caused by cardiac surgery. 47 Notably, the HRV-related parameters recovered to the preoperative levels after 1 month, indicating that the effect of PFO closure on the autonomic nervous system was essentially transient. Interestingly, the nLF, nHF, and LF/ HF values did not change during the process, indicating that PFO closure did not alter the balance between the sympathetic and parasympathetic activities.
This study had some limitations. First, because of patient intolerance, we did not collect venous and arterial blood samples throughout the long-term follow-up from patients who underwent PFO closure. Second, the sample size in our study was relatively small. Thus, our findings warrant further investigation using largescale studies. Third, changes in beat-to-beat variability may not directly and comprehensively reflect cardiac autonomic changes, therefore, the findings of cardiac autonomic regulation in our study should be further verified using other methods such as plasma markers. 48 In conclusion, PFO closure can improve dCA and alter elevated arterial blood PDGF-BB levels in migraine patients with PFO, both of which may be related to the preventive effect of PFO closure on stroke occurrence and recurrence in these patients.

AUTH O RCO NTR I B UTI O N S
YY and Z-NG drafted the initial protocol, which was reviewed with

ACK N O WLE D G E M ENTS
We thank the Department of Biobank, Division of Clinical Research, and the First Hospital of Jilin University for providing human tissues.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflict of interest.

DATAAVA I L A B I L I T YS TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

PATI ENT CO N S ENT S TATEM ENT
Written informed consent was obtained from all participants.