Hyper-Activation of Endoplasmic Reticulum Stress PERK/Calcineurin/RyR2 Signaling Pathway is Involved in AGEs-Exacerbated Post- Myocardial Infarction Ventricular Arrhythmias

Background: The current study was aimed to investigate the mechanisms of advanced glycation end products (AGEs) in exacerbating post- myocardial infarction (MI) ventricular arrhythmias (VAs). Methods: Correlation between premature ventricular contractions (PVCs) and serum AGEs concentrations was analyzed in a cohort consisted of 101 STEMI patients with culprit vessel of left anterior descending artery (LAD). Established MI rat model were treated with AGEs and/or anti- receptor for AGE (RAGE) antibody. Electrocardiography was used to record VAs. Myocytes were isolated from adjacent area around infracted region. Immunouorescent stains were used to evaluate association between FKBP12.6 (FK506-bindingprotein 12.6) and ryanodine receptor 2 (RyR2). Calcium sparks were evaluated by confocal microscope. Protein expression and phosphorylation were assessed by Western blotting. A colorimetric method was used to determine the enzymatic activity of calcineurin (CaN). [ 3 H]-ryanodine binding assay was carried out to detect the RyR2 channel activity. Results: In the cohort study, signicantly increased amount of (PVCs) were found in STEMI patients with diabetes (P<0.05). Serum AGEs concentration was signicantly positively correlated with PVCs amount in STEMI patients (r=0.416, P<0.001). Multivariate analysis showed serum AGEs concentration was independently and positively related to frequent PVCs (adjusted hazard ratio, 1.86; 95% CI, 1.09-3.18, P=0.022). In the animal study, increased glucose regulated protein 78 (GRP78) expression, protein kinase RNA-like ER kinase (PERK) phosphorylation, CaN enzymatic activity, FKBP12.6-RyR2 disassociation, RyR2 channel opening and endoplasmic reticulum (ER) calcium releasing were found in MI animals exposed to AGEs, which were attenuated by anti-RAGE antibody treatment. This RAGE blocking also signicantly lowered the VAs amount in MI animals exposed to AGEs. Conclusions: Hyper-activation Continuous variables are presented as mean ± SD; categorical variables are presented as numbers or percentages.


Background
Higher incidence of lethal arrhythmias such as R-on-T premature ventricular contractions (PVCs), ventricular tachycardia and ventricular brillation were found in diabetes, which has been identi ed as a risk factor elevating the incidence of ventricular arrhythmia-related adverse events [1]. In our previous investigation, elevated amount of VAs was found in diabetic animals [2]. Ventricular arrhythmia (VA) is a common complication after myocardial infarction (MI). Increased risks of occurrence of life-threatening VAs were identi ed in acute MI (AMI) patients, which elevated in-hospital mortality. However, the underlying mechanisms of the exacerbating role of diabetes in post-MI VAs are still unclear.
Advanced glycation end products (AGEs) are characterized metabolites of DM which participated in many DM-associated pathological processes. AGEs are fostered by non-enzymatic reactions among nucleic acids, amino groups and lipids in DM and closely correlated with major adverse cardiovascular events (MACE) including lethal arrhythmias [3]. A recent study pointed out that inhibiting AGEs formation attenuated the vulnerability to tachyarrhythmias in diabetic animals [4]. In this regard, we hypothesized that AGEs might be the cause facilitating post-MI VAs.
Calcium homeostasis is important for maintaining the normal electrophysiological integrity of myocytes.
Previous investigation suggested intracellular calcium metabolism dys-regulation contributed to arrhythmias in diabetic hearts [5]. Delayed afterdepolarizations (DADs) have been recognized as the mechanism triggering VAs. Appearing posterior to the completion of repolarization, DADs are induced by oscillation of intracellular calcium released from endoplasmic reticulum (ER), which further trigger a greater amount calcium release from ER via channels such as ryanodine receptor 2 (RyR2), which was termed as calcium-induced calcium release (CICR) [6].
ER is the major cellular organelle maintaining intracellular calcium homeostasis [7]. ER stress is triggered when challenged by strong and sustained harmful stimuli [8]. Protein kinase RNA-like ER kinase (PERK) is a sensor transducting ER stress signals. By direct contacting, PERK activates calcineurin (CaN) which further facilitates ER calcium releasing via promoting the disassociation of FK506-binding protein 12.6 (FKBP12.6) from RyR2 [9]. Our previous investigation indicated that ER stress PERK signaling was activated in diabetic hearts, inducing ventricular arrhythmias through the PERK/CaN signaling [10].
Notably, results from an investigation suggested that AGEs induced RyR2 channel hyperactivationmediated ER calcium releasing after interacting with the receptor for AGE (RAGE) in cultured myocytes [11]. Moreover, interestingly, several previous studies pointed out ER stress PERK signaling activations were RAGE-dependent [12,13]. Thus, it is reasonable for us to speculate that AGEs cause aggravation of post-MI VAs in DM. The RAGE-dependent ER stress mediated PERK/CaN/RyR2 signaling pathway could be possibly involved. We believe that results from this study would enrich our current understanding of the mechanisms of exacerbated post-MI VAs in DM.

Cohort study
In the period from May, 2017 to May, 2019, 115 patients diagnosed as STEMI with the culprit vessel of LAD at Shaanxi Provincial People's Hospital were initially included. The nal cohort consisted of 101 STEMI patients. 3 patients were excluded due to identi cation of malignant cancer; 1 patient was excluded due to diagnosis of systemic lupus erythematosus; 10 patients were excluded because of rapid development of cardiac systolic dysfunction which were eventually supported by extra-corporeal membrane oxygenation (ECMO) or left ventricular assist devices. STEMI was diagnosed by the current guideline [14]. Culprit vessel was determined by ECG manifestation and coronary angiography. Diabetes was diagnosed according to current guideline [15]. Post-MI VAs were recorded by Holter ECG. Frequent premature ventricular contractions (PVCs) was de ned as ventricular premature beats take ≥ 10% of total cardiac beats during 24 hour-Holter ECG recording [16]. The exclusion criteria were: age below 18 years or above 80 years; pregnant women; previous myocardial infarction; previous PCI/CABG history; New York Heart Association (NYHA) functional class III or IV, renal failure, hepatic dysfunction, known history of cancer, immune-mediated disorders, mental disorders, recent/current use of anti-arrhythmic drugs. The peripheral venous blood samples were collected in a fasting state. Serum AGEs concentrations were determined by using a Human AGEs ELISA kit (CUSABIO, China) per the protocols provided by the manufacturer. All of the patients gave informed consent to participate in this study, which was approved by the ethics committee of Shaanxi Provincial Peoples' Hospital. Speci cally, patients signed the informed consent and agreed the collection and research purposes of medical records for our investigation.

AGEs-BSA preparation
The protocol was in accordance with our previous investigations [17]. Brie y, 0.1 mmol/L glyceraldehydes (Sigma) and bovine serum albumin (BSA) were incubated in 0.2 mmol/L NaPO 4 buffer solution (pH = 7.4) at 37℃ at sterile condition for 7 days. BSA prepared without glyceraldehydes by the same protocol was used as control.
MI model establishment and animal treatments SPF class Sprague-Dawley (SD) rats (9-week old, male/femal = 1, weighted 255 ± 6 g) were provided by Animal Experimental Center of Xi'an Jiaotong University. Animals were maintained in independent polypropylene cages under controlled conditions (12 h/12 h arti cial light/dark cycle, humidity at 56% ±4%, temperature at 25℃±1℃). Animals were accessible to standard chow and sterilized water freely. Animals were accommodated 1 week prior to experiments.
Rats were anaesthetized by inhalation of iso urane (2% for introducing and 4% for continuous anesthesia) in oxygen at 0.6L/min. MI model was induced by ligation of LAD following the protocol adopted from our previous investigation [18]. Before establishment of MI model, AGEs-BSA and/or anti-RAGE IgG 3 (Abcam) were administrated to animals via remaining needles implemented in tail veins. The dosages anti-RAGE antibody were selected according to previous reports [19,20]. Treatments of each group were listed in Table 1.  Continuous variables are presented as mean ± SD; categorical variables are presented as numbers or percentages.

Electrocardiography (ECG) evaluation
Anasthetized rats were xed in a supine position, the standard limb leads were attached to the upper and lower limbs of the rats and connected to Powerlab 4/25 Biological Analysis System (AD Instruments) which recorded the ECG. Occurring amount of VAs, including ventricular premature beats and tachycardia were recorded.

Primary myocytes isolation
The primary myocytes were isolated from the hearts harvested from the rats in accordance with the protocol described in our previous study [2]. Speci cally, the myocytes were carefully extracted from the adjacent area around infracted region in the anterior left ventricular wall under a dissecting microscope. Myocytes were maintained in Dulbecco Modi ed Eagle's Medium (DMEM) supplemented fetal bovine serum (15%, FBS, Hyclone) and antibiotic mix (Sigma) at 37℃ in an atmosphere composed of 95% fresh air and 5% CO 2 . When cell populations reached con uence at 50%-60%, the cells were used for subsequent experiments.

Calcium spark detection
The calcium spark was detected in isolated myocytes by confocal optical calcium imaging according to the methods described previsouly [21]. Isolated myocytes were loaded with 10 µmol/l Fluo3/AM (Sigma-Aldrich) for 30 minutes at 37℃ in a humidi ed dark chamber. After washed by PBS, calcium sparks were observed with a SP8 STED confocal microscope (Leica) equipped with an argon laser at wave length at 488 nm. Liner scan was used to acquire the line scan images (512 pixels/line) at sampling rate of 2 ms/line. The scanning frequency was 600 Hz. Calcium sparks were analyzed by Image J with SparkMaster Plugin according to previous described method [22].

Cytosol and ER membrane fraction preparation and Western blotting
Subfraction isolation protocols were carried out according to ours and others' previous descriptions [10,23]. With Protein Extraction kits (Beyotime), proteins were extracted from cytosol preparation and ER membrane preparation according to the protocol provided by the manufacturer respectively. A BCA kit (Beyotime) was used to determine the protein concentrations of the samples which were then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The separated proteins were then transferred to polyvinylidene uoride (PVDF) membranes. After treated with blocking buffer (Abcam), primary antibodies against GRP78 (Abcam, 1:4000), phosphorylated PERK (p-PERK, Cell Signaling Tech, 1:2000), PERK (Cell Signaling Tech, 1:2000), FKBP12.6 (Invitrogen, 1:2000), Sigma receptor 1 (SigmaR1, Invitrogen, 1:1000) and GAPDH (Abcam, 1:1000) were used to incubate the membranes at 4℃ for 12 hours. After TBST washing, the membranes were incubated by corresponding HRP conjugated secondary antibodies (Abcam) at room temperature for 2 h. After developed with Super Signal West Pico chemiluminescence reagent (Thermo Scienti c), the immunobands were visualized on X-ray lms and then analyzed with ImageJ2x software (Rawak Software).
[H 3 ]-ryanodine bind assay [H 3 ]-ryanodine bind assay was used to assess the channel activity of RyR2 according to the protocol described previously [10].

Calcineurin (CaN) activity detection
The CaN enzymatic activity was determined in total protein samples with a colorimetric method according to the protocol described previously [10]. A Calcineurin Activity Assay kit (Merck) was used per manufacturer's instructions.

Statistics
Data acquired in this study were presented in (mean ± standard deviations) or percentage manner. Number of independent experiments carried out was indicated as n. NSK tests were performed as posthoc tests. Participants were divided into 2 groups (diabetes and non-diabetes). The baseline characteristics among the 2 groups were analyzed by t test for parametric variables, the Mann-Whitney U test for nonparametric variables, and the chi-square test for categorical variables. The Spearman rank correlation coe cient was computed to assess correlation between continuous variables. The association between serum AGEs and frequent PVCs were estimated with univariate and multivariate logistic regression models. Model 1 was unadjusted. Model 2 was adjusted for age, sex, body mass index, cigarette smoking. Model 3 was adjusted for age, sex, body mass index, cigarette smoking, creatinine, UA, low density lipoprotein, K + , Na + , troponin I, BNP, QTC interval. All statistical testing was 2-sided. Results were considered statistically signi cant at a level of P < 0.05. All analyses were performed with PASW Statistics 20.0 software.

Serum AGEs concentration elevation correlated with frequent PVCs in STEMI patients
Data were successfully obtained from 101 STEMI patients. The baseline data was listed in Table 2. STEMI patients were divided into diabetes group and non-diabetes group. The amount of 24 h-PVCs was signi cantly higher in diabetes group than non-diabetes group (Fig. 1a). Spearman rank correlation was used to analyze the association between 24-h PVCs amount and serum AGEs concentrations. Results showed that 24-h PVCs amount and serum AGEs concentrations was signi cantly and positively related (r = 0.461, P < 0.001) (Fig. 1b). The univariate and multivariate logistic regression models were used to further evaluate the association between frequent PVCs and serum AGEs concentration in STEMI patients. As demonstrated in

RAGE antibody administration attenuated AGEsexacerbated ventricular arrhythmias in MI rats
As demonstrated in Fig. 2a, VAs were presented as PVCs and ventricular tachycardias. In some particular cases, bigeminal rhythm, polymorphic PVCs and paroxysmal ventricular tachycardia could be observed.
Amount of VAs increased signi cantly in MI rats compared with control, which was further dramatically increased by AGEs exposure. However, the RAGE antibody treatment signi cantly reduced VAs amount in AGEs-exposed MI rats (Fig. 2b).

RAGE blocking suppressed activation of ER stressmediated PERK/CaN signaling in AGEs-exposed MI rats
As demonstrated in Fig. 3, we found that the expression level of GRP78, phosphorylation level of PERK and enzymatic activity of CaN increased signi cantly in MI compared with control. AGEs exposure further caused signi cant elevation of GRP78 expression level, PERK phosphorylation level as well as enzymatic activity of CaN in myocytes isolated from MI hearts, which were dramatically decreased by RAGE antibody administration.

RAGE antibody treatment impaired ER membrane-cytosol translocation of FKBP12.6 and its disassociation with RyR2
The co-localization of FKBP12.6 and RyR2 was evaluated by calculating the Pearson's correlation coe cients based on analyzing the captured uorescent images. The coe cients could indicate the degree of contact between two detected molecules. As demonstrated in Fig. 4a, lowered co-localization level between FKBP12.5 and RyR2 was found in MI compared with control. AGEs exposure further dramatically impaired the co-localization between FKBP12.5 and RyR2 in MI hearts, which was improved by RAGE antibody administration. CaN facilitates the disassociation between RyR2 and FKBP12.6 which would translocate from ER membrane to cytosol. As demonstrated in Fig. 4b, compared with control, signi cantly promoted ER membrane-cytosol translocation of FKBP12.6 was found in MI. AGEs facilitated more ER membrane-to-cytosol translocation of FKBP12.6. The RAGE antibody treatment, however, reduced the ER membrane-to-cytosol translocation in MI hearts exposed to AGEs.
Administration of RAGE antibody alleviated calcium releasing by inhibiting RyR2 channel activity in AGEsexposed MI hearts As demonstrated in Fig. 6a, results from [ 3 H]-ryanodine binding assay indicated that the RyR2 channel activity increased signi cantly in myocytes isolated from MI hearts. The AGEs exposure was proved to further increase the RyR2 channel activity which was reduced by administration of RAGE antibody. The opening of RyR2 channel would result in calcium releasing from ER which was assessed by calcium sparks evaluations (Fig. 6b). Compared with control, promoted calcium sparks were found in MI. The AGEs exposure further promoted calcium sparks in myocytes isolated from MI hearts. The administration of RAGE antibody, however, dramatically attenuated calcium sparks in myocytes isolated from MI hearts exposed to AGEs.

Discussion
VAs are critical clinical manifestations of MI and contribute to the major adverse cardiovascular events (MACE). Indeed, MI rat model established in this study was characterized by increased amount of VAs. MACE was reported to be closely associated with AGEs content in high-risk patients with atherosclerotic cardiovascular diseases [24]. In the present study, our cohort analysis suggested that post-MI PVCs were more frequent in STEMI patients complicated with diabetes. More importantly, we found that serum AGEs concentration was signi cantly and independently associated with post-MI PVCs in STEMI patients. Thus, AGEs forged in diabetes could be possibly identi ed as the cause exacerbating post-MI VAs.
AGEs are sets of biochemical byproducts generated by non-enzymatic glycosylations in vivo, contributing to diabetic complications via interacting with its receptor RAGE located on the targeted cells [25]. As a member of immunoglobulin superfamily of cell surface receptors, RAGE is a trans-membrane protein involved in multiple cell functions including proliferation, redox adjustment, in ammation and migration [25]. An investigation suggesting the correlation between RAGE and VAs aroused our interest: RAGE knockdown by speci c small interfering RNA (siRNA) effectively reduced VAs in cardiac ischemia and reperfusion model [26]. Moreover, it has been proved that ER stress activation was RAGE-dependent [27,28]. Therefore, our hypothesis that AGEs could induce VAs through RAGE-dependent ER stress activation was formed and further testi ed in the current study.
Evidenced by GRP78 expression and PERK phosphorylation elevation, ER stress-mediated PERK signaling was activated in myocytes from MI rats which exhibited increased post-MI VAs. The AGEs administration resulted in further elevation of GRP78 expression and PERK phosphorylation, indicating the hyper-activation of ER stress-mediated PERK signaling in AGEs-exposed MI rats which were featured with exacerbated post-MI VAs. Similar to several previous works, we used anti-RAGE IgG 3 antibody to block AGEs-RAGE interactions [11]. Evidenced by decreased GRP78 expression and PERK phosphorylation, RAGE blockage alleviated ER stress-mediated PERK signaling hyper-activation, resulting in attenuation of post-MI VAs in MI rats exposed to AGEs. These results indicated that RAGE-dependent ER stress-mediated PERK signaling played a critical role in exacerbating post-MI VAs after AGEs exposure. As an ER membrane localized protein, phosphorylated PERK could increase enzymatic activity of its down-stream effecter CaN by direct contact [6,10]. CaN would further promote the disassociation between RyR2 and its inhibitor FKBP12.6 on ER membrane to facilitate RyR2 channel opening [23]. Our data showed that RAGE blockage reduced CaN activity, leading to impaired disassociation between FKBP12.6 and RyR2 which was evidenced by reduced ER membrane-to-cytosol translocation behavior of FKBP12.6. As a result, RyR2 channel activity was compromised.
Calcium released through RyRs from ER raised calcium concentration in subspace between plasma membrane and ER membrane, which is visualized as calcium sparks. It has been well established that DADs are mediated by ER calcium release [29]. DADs are voltage oscillations at diastole period causing triggered activity (TA) which is manifested as arrhythmias, typically PVCs. We found that due to elevated RyR2 channel activity, calcium sparks were enhanced signi cantly in a spatio-temporal manner in myocytes isolated from MI hearts, which was further boosted by AGEs exposure. The RAGE blockage effectively suppressed calcium sparks in myocytes isolated from AGEs-exposed MI hearts. These results further supported our theory that AGEs mediated ER RyR2 calcium releasing via RAGE.

Conclusion
Notably, we also found that ER stress was also activated in MI heart not exposed to AGEs. This phenomenon we observed was in consistence with the established opinions [30]. We innovatively reported ER stress was hyper-activated via RAGE/ROS pathway in MI heart exposed to AGEs which made MI individuals more vulnerable to VAs. Our further investigation suggested AGEs-induced hyper-activation of PERK/CaN/RyR2 signaling was the underlying mechanism of exacerbated VAs. Taken together, to some extent, the data we acquired in the present study interpreted the clinical scenario that post-MI VAs are more frequent in patients with diabetes. Ethics approval and consent to participate All of the patients gave informed consent to participate in this study, which was approved by the ethics committee of Shaanxi Provincial Peoples' Hospital. Speci cally, patients signed the informed consent and agreed the collection and research purposes of medical records for our investigation.

Consent for publication
Written informed consent for publication was obtained from all participants.

Availability of data and materials
Original data and materials were available on reasonable requests.