Febuxostat Increases Ventricular Arrhythmogenesis Through Calcium Handling Dysregulation in Human-Induced Pluripotent Stem Cell-Derived Cardiomyocytes

Abstract

Febuxostat is a xanthine oxidase inhibitor used to reduce the formation of uric acid and prevent gout attacks. Previous studies have suggested that febuxostat was associated with a higher risk of cardiovascular events, including atrial fibrillation, compared with allopurinol, another anti-hyperuricemia drug. Whereas in our clinical practice, we identified 2 cases of febuxostat-associated ventricular tachycardia (VT) events. The proarrhythmogenic effects of febuxostat on human cardiomyocytes and underlined mechanisms remain poorly understood. In this study, we employed real-time cell analysis and calcium transient to investigate the effects of febuxostat on the cytotoxicity and electrophysiology properties of human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs). Up to 10 μM febuxostat treatment did not show toxicity to cell viability. However, 48-h febuxostat exposure generated dose-dependent increased irregular calcium transients and decreased calcium transient amplitude. Furthermore, RNA-seq analysis indicated that the MAPK signaling pathway was enriched in the febuxostat-treated group, especially the protein kinases c-Jun N-terminal kinase (JNK). Western blotting of 3 main protein kinases demonstrated that JNK activation is related to febuxostat-induced arrhythmia rather than extracellular signal regulated kinases (ERK) or p38. The dysfunctional calcium dynamics of febuxostat-treated hiPSC-CMs could be ameliorated by SP600125, the inhibitor of JNK. In conclusion, our study demonstrated that febuxostat increases the predisposition to ventricular arrhythmia by dysregulating calcium dynamics.

Febuxostat, a urate-lowering drug inhibiting xanthine oxidase (XO), is widely administered for the treatment of hyperuricemia in patients with gout (Schumacher et al., 2008). Previous animal studies have shown that febuxostat exerts cardiovascular protective function by inhibiting atrial electrical and structural remodeling (Fan et al., 2019), attenuating endothelial dysfunction (Li et al., 2017), protecting against myocardial ischemia/reperfusion injury (Wang et al., 2015), and ameliorating renal tubulointerstitial fibrosis (Lu et al., 2019). However, the cardiovascular safety of febuxostat is controversial from the real-world evidence (Mackenzie et al., 2020; Zhang et al., 2018). A multicenter randomized trial (CARES trial) focused on the cardiovascular safety of febuxostat or allopurinol, another XO inhibitor, concluded that all-cause mortality and cardiovascular mortality were higher in the febuxostat group than in the allopurinol group among patients with coexisting hyperuricemia and cardiovascular disease (White et al., 2018). In addition, a propensity-matched cohort study of gout patients found that the administration of febuxostat was associated with a higher risk of atrial fibrillation than that of allopurinol (Singh and Cleveland, 2019). In our clinical practice, we identified 2 patients prescribed with febuxostat (40 mg qd) had febuxostat-associated VT. The mechanisms responsible for ventricular arrhythmogenicity related to febuxostat have yet to be elucidated.

Because human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) are able to recapitulate normal and pathologic cardiac electrophysiological properties, they are now extensively adopted for in vitro drug screening and cardiotoxicity study (Savoji et al., 2019). Optical mapping of calcium transients in hiPSC-CMs was previously reported to provide an optimal noninvasive high-throughput platform for drug testing (Jiang et al., 2019; Shafaattalab et al., 2019; Suzuki et al., 2018). Intracellular calcium movement captured by fluorescent calcium-sensitive indicator reflects the rise and decay of cytoplasm Ca²⁺ concentrations during action potentials. Arrhythmic events could be reflected by Ca²⁺ traces, such as early after-depolarization (EAD)-like events (Cai et al., 2019; additional spikes or bursts during a single Ca²⁺ transient cycle) or fibrillation-like events (Kopljar et al., 2018; small, rapid-rate Ca²⁺ transients), which represent the risk of VT in humans. Hence, recent studies have included optical mapping techniques as a mainstream research method of drug screening (Bedut et al., 2016; Lu et al., 2015; Pioner et al., 2019).

In the present study, we evaluated the cytotoxicity and electrophysiological profile of hiPSC-CMs after 0–10 μM febuxostat exposure. Additionally, RNA sequencing was performed to seek a potential explanation for the arrhythmogenic effects of febuxostat.

MATERIALS AND METHODS

Maintenance of hiPSC-CMs

Human-induced pluripotent stem cell-derived ventricular cardiomyocyte cell lines purchased from HELP Innovation were used in this study. Cells were maintained in RPMI 1640 basic medium (Thermo Fisher Scientific, 22400089) supplemented with B27 supplement (Thermo Fisher Scientific, 17504044) in a 5% CO₂ incubator at 37°C. Human-induced pluripotent stem cell-derived cardiomyocytes were cultured for 8 weeks for further maturation and prepared for the following experiments.

Compounds and treatment

Febuxostat (Selleckchem, S1547) was dissolved in dimethyl sulfoxide (DMSO) to make a stock solution of 10 mM. The working solutions were freshly prepared in the respective external medium. Cells were treated with 5 different concentrations (blank, 0.1, 0.4, 1, and 10 μM) of febuxostat around the IC50 (114–210 nM). SP600125 (Selleckchem, S1460), a broad-spectrum JNK inhibitor, was prepared as a 1-mM stock solution in DMSO. A working concentration of 1 μM SP600125 was chosen to completely block the JNK activation according to previous report (Canedo-Antelo et al., 2018). Xanthine oxidase activity kit (Abbkine, China, KTB1070) was used to examine the effectiveness of febuxostat. DMSO (0.1%) here served as vehicle control.

Cytotoxicity assay by real-time cell analyzer

To test the cytotoxicity of febuxostat, lactate dehydrogenase release from hiPSC-CMs under different dosages of febuxostat was tested using assay kit (Beyotime, China, C0016) and the cell cytotoxicity was evaluated according to manufacturer’s protocol. The real-time cell analyzer (RTCA) xCELLigence System (Roche Applied Science, Mannheim, Germany) was also applied to dynamically monitor the cellular response to febuxostat. This system allows the recording of electrical impedance across microelectrodes integrated on the bottom of the E-plates at set intervals and durations to provide a real-time readout of cell number, viability, morphology, and attachment (Schott et al., 2012). The electronic readout of detected impedance is displayed as cell index (CI) values, which reflected the whole-cell status in culture. Briefly, hiPSC-CMs were seeded at a density of 5000 cells/well in 96-well E-Plates after background normalization. After seeding for 2 h, cells were incubated with corresponding concentrations of febuxostat. Dynamic CI was then monitored every 10 min for 48 h after treatment to generate time-dependent cell response dynamic curves. The assays were conducted in triplicate and 3 times for each dosage. Data were analyzed using the RTCA built-in software for calculating the temporal dynamics of CI changes.

Measurement and assessment of calcium transients

Intracellular Ca²⁺ recording was measured as described previously (Zhu et al., 2021). For this assay, hiPSC-CMs were treated with corresponding concentrations of febuxostat (0, 0.1, 0.4, 1, and 10 μM) for 48 h and then loaded with 5 µM calcium-sensitive dye fluo-4 AM (Life Technologies, F14201) for 60 min at 37°C. After washouts, the hiPSC-CMs were subsequently maintained in Tyrode’s solution. The spontaneous and stimulated calcium signals (1 Hz) from individual cells were recorded through the scientifc complimentary metal-oxide-semiconductor camera (sCMOS camera, Tucsen, Dhyana95) mounted on an inverted fluorescence microscope (Olympus IX51, Japan) at a sampling interval of 33 ms for 30 s (excitation wavelengths: 488 nm, emission wavelengths: 505 nm). The Ca²⁺ transients include amplitude and proportion of irregular events (EAD-like or fibrillation-like) were further analyzed. Early after-depolarization-like Ca²⁺ events were defined as additional calcium bursts or spikes observed during calcium cycles, whereas fibrillation-like events referred to fast and low amplitude traces compared with others. Data were analyzed using ImageJ as previously described (Zhu et al., 2021). Briefly, the fluorescence intensity in the region of interest was determined based on the formula: △F/F0 = (F−F0)/(F0−Fb) to calibrate the cell-to-cell variability and background disturbance. Herein, F0 represents the baseline intensity, F represents the intensity at any point in time, whereas Fb stands for background intensity.

RNA-seq

Three different batches of hiPSC-CMs were used in this experiment. Total RNA was extracted from control and 1 μM febuxostat-treated hiPSC-CMs using TRIzol reagent (Invitrogen, 15596026) following the manufacturer’s instructions. The RNA sample was quantified by NanoDrop (Thermo Fisher Scientific, A30221) and the purity was checked by 1% agarose gel electrophoresis. After confirmation of RNA quality, the cDNA library was constructed using the TruSeq RNA LT Sample Prep Kit v2 (Illumina) and amplified on cBOT (Illumina). Sequencing was performed on an Illumina Hiseq2000 platform (Illumina). The DESeq2 package was used to screen differentially expressed genes (DEGs) between the experimental group and the control group. Candidate genes that satisfied |log2FC| ≥ 1 and adjusted p value (False Discovery Rates, FDR) ≤.05 were included for the following Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses.

qRT-PCR

The quantitative reverse transcription PCR (qRT-PCR) reactions were repeated in 3 biological and 3 technical replications. Total RNA was isolated from treated cells as mentioned above and converted into cDNA through reverse transcription using the Hifair III First-Strand cDNA Synthesis SuperMix kit (Yeasen, 11141ES60). Reverse transcription was performed at 37°C for 15 min, followed by 85°C for 5 s for inactivation. For quantitative real-time PCR, cDNA samples were mixed with oligos synthesized by TSINGKE biological technology and Hieff qPCR SYBR Green Master Mix (Yeasen, 11202) according to the manufacturer’s instructions. The amplification procedure was performed using a QuantStudio 5 (ABI) and divided into the following 3 steps: predenaturation at 95°C for 5 min; 40 cycles of denaturation at 95°C for 10 s, annealing at 55°C for 20 s, and extension at 72°C for 20 s; and a melting curve to confirm product specificity. The 2^−ΔΔCT method was applied to calculate the relative value of target gene expression compared to the control group. Additionally, 2 reference genes (GAPDH and β-actin) were used. The primer sequence list is provided in the Supplementary files (Supplementary Table 1 and Resource 1).

Western blot

Total protein was extracted from hiPSC-CMs lysed in radioimmunoprecipitation assay buffer mixed with Phenylmethanesulfonylfluoride, protease inhibitors and phosphatase inhibitors (from the membrane protein, nuclear protein, and cytoplasmic protein extraction kit, KegGEN Bio, China, Cat: KGBSP002). The bicinchoninic acid method was used to calculate the protein concentration. SDS-PAGE and polyvinylene difluoride (Millipore, IPVH00010) membranes were used to separate and block the proteins, respectively. The primary antibody used here was p-JNK (1:1000, CST, 4668S), p-ERK (1:2000, CST, 4370S), p-p38 (1:1000, CST, 4511S), p-CAMKII (1:1000, CST, 12716S), JNK (1:1000, CST, 9252S), ERK (1:1000, CST, 4695S), p38 (1:1000, CST, 8690S), CAMKII (1:1000, CST, 4436S), and GAPDH (1:1000, CST, 5174S), which was detected by antirabbit IgG, HRP-linked antibody (1:3000, CST, 7074S). Bands were visualized with exposure machine (Invitrogen, iBright CL1000). The protein expression levels were normalized to those of GAPDH.

siRNA-mediated knockdown

Xanthine oxidase was converted from xanthine dehydrogenase, encoded by the gene XO (also known as: XDH Gene ID: 7498). Human-induced pluripotent stem cell-derived cardiomyocytes were transfected with XO siRNA (constructed by Hippo Biotechnology, China) at the final concentration of 50 nM using Lipofectamine RNAiMAX (Life Technologies). qRT-PCR was used to detect the knockdown efficiency after 48-h transfection and total protein was extracted from hiPSC-CMs cotreated with febuxostat and siRNA for 48 h. Cells transfected with siRNA negative control were used for comparison.

Statistical analysis

Data were expressed as percentage or mean ± SD. Statistical comparisons were performed using One-way ANOVA or χ² test as appropriate. Post-hoc multiple comparisons were corrected using Tukey’s multiple comparisons test. Two-sided p values <.05 were considered statistically significant. All analyses were conducted in SPSS (version 28.0.1).

RESULTS

Febuxostat-Induced VT Cases

Two cases of febuxostat-related VT were identified in individuals without history of cardiovascular disease or syncope. Both of the patients were prescribed febuxostat for the management of hyperuricemia and no other drug was used. Case 1 is a 59-year-old man presented to the emergency department with palpitations for 1.5 h. His electrocardiogram (ECG; Figure 1A) showed fast VT, with a ventricular rate of 233 beats/min, which was converted to sinus rhythm by defibrillator (Figure 1B). Febuxostat (40 mg qd) was administered 2 months before admission for inadequate serum uric control by allopurinol. Echocardiography and cardiac magnetic resonance (CMR) exhibited normal cardiac structure (Figure 1C, Supplementary Figure 1 and Resource 1). He underwent further electrophysiological study (EPS). The EPS showed an AH interval (conduction time of the atrioventricular node) of 125 ms and an HV interval (conduction time from the onset His potential to the earliest ventricular activation) of 48 ms. Ventricular and atrial procedural stimulation excluded the possibility of bypass and atrioventricular nodal dual path. After the application of isoproterenol, repeated stimulation did not induce tachycardia. Febuxostat was discontinued thereafter. Case 2 is a 54-year-old man who also presented VT during febuxostat treatment. His echocardiography and coronary computed tomography angiogram were normal, but he refused to undergo CMR and the EPS. Both of the patients had normal serum uric acid levels during the onset of VT. Episodes of VT were not observed after discontinuation of febuxostat during the 2-year follow-up.

Cytotoxicity and XO Activity Test on hiPSC-CMs

Lactate dehydrogenase release assay was conducted to evaluate the cytotoxicity of febuxostat. Sequential increases in febuxostat up to 10 μM did not change membrane integrity of hiPSC-CMs (Figure 2A). Immunofluorescence staining of α-actinin was conducted to evaluate the toxicity of febuxostat. No structural change was found in hiPSC-CMs with sequential increases in febuxostat up to 10 μM (Figure 2A). Additionally, CI was further evaluated using the RTCA xCELLigence System after exposure of febuxostat (Wachter et al., 2012). Figure 2B showed that 10 μM febuxostat becomes toxic after 12 h of treatment as indicated by the decline in CI values, but cellular viability was restored after 48 h. None of the other concentrations of febuxostat showed any deleterious effects on cellular viability during the 48-h treatment. Febuxostat (1 µM) decreased the XO activity to about a half in hiPSC-CMs (Figure 2C), which confirmed the effectiveness of drug.

Effects of Febuxostat on Calcium Transients

Intracellular Ca²⁺ transient in cardiomyocytes is a bridge of excitation-contraction coupling, reflecting the electrophysiological profiles of the cells. Subsequently, we measured the intracellular Ca²⁺ recordings in the control and febuxostat-treated hiPSC-CMs. As shown in Figure 3A, increased ratio of cells with irregular calcium transients (EAD-like and fibrillation-like) were observed in febuxostat-exposed cardiomyocytes in a dose-dependent manner. The occurrence of irregular events was 4.5%, 5.6%, 17.0%, 26.1%, and 25.2% in control, 0.1, 0.4, 1, and 10 μM-treated cells, respectively (Figure 3C). To avoid variability of spontaneous beating cells, we paced the hiPSC-CMs at the frequency of 1 Hz (Figure 3B). Consistently, high concentration of febuxostat treatment (1 or 10 μM) also reduced the Ca²⁺ transient amplitude (Figure 3D). These results strongly suggest that febuxostat compromised calcium homeostasis at the cellular level, predisposing these cells to arrhythmia.

RNA-Seq Reveals MAPK Signaling Pathway Involvement in Febuxostat-Induced Arrhythmia

To explore the cause of febuxostat-induced calcium transient disorder, we performed RNA sequencing on the control group and 1 μM febuxostat-treated group. The principal component analysis plot showed a clear distinction of expression profiles between the 2 groups (Figure 4A). The volcano map and heatmap visually displayed the number of DEGs (Figs. 4B and 4C). Among the 729 DEGs, 474 genes were upregulated and 255 genes were downregulated after febuxostat exposure. Based on the GO and KEGG analyses (Figs. 4D and 4E), the MAPK signaling pathway was enriched among the differential expressed genes (FDR = 5.70E−08). Quantitative real-time PCR of the 8 genes involved in MAPK signaling pathway showed significantly increased in febuxostat-treated group, which were in agreement with the results of enrichment analysis (Figs. 4F and 4G and Supplementary Figure 3).

The Phosphorylation Activation of JNK Might Disrupt Calcium Homeostasis

Previous studies have revealed that the stress-response MAPK signaling pathway was involved in cardiac hypertrophy, fibrosis, and arrhythmia (Auger-Messier et al., 2013; Kumar et al., 2019). c-Jun N-terminal kinase, ERK, and p38 MAPK are 3 predominant members of the MAPK family. Western blot of these 3 protein kinases demonstrated that febuxostat exposure might prompt the phosphorylation of JNK rather than ERK and p38 (Figure 5). To test whether XO was involved in the phosphorylated JNK activation, cells were cotreated with XO siRNA transfection and febuxostat. Western blot indicated that febuxostat increased phosphorylated JNK independent of XO knockdown (Supplementary Figure 2). We also observed increased phosphorylation levels of the CaMKII in the febuxostat group but did not reach statistical significance (Supplementary Figure 3).

JNK Inhibitor Reverses Febuxostat-Induced Abnormal Calcium Activity

Activated JNK has been previously reported to enhance atrial arrhythmogenicity by reducing gap junction channels and impairing action potential conduction velocity in aged animals (Yan et al., 2013). To test whether the activation of JNK participates in febuxostat-induced ventricular arrhythmia, we used the broad-spectrum JNK inhibitor SP600125 on febuxostat-treated hiPSC-CMs. After 48 h of febuxostat and SP600125 treatment, the percentage of irregular Ca²⁺ transient occurrence was reduced to 11.8% compared with that in the febuxostat group (27.1%, p < .05, Figs. 6A and 6B). Meanwhile, the amplitude of Ca²⁺ transients increased in the SP600125 plus febuxostat group compared with the febuxostat treatment group (p < .001, Figs. 6C and 6D), showing that the JNK inhibitor reduced the febuxostat-induced abnormal Ca²⁺ transients.

DISCUSSION

Xanthine oxidase, in addition to catalyzing uric acid synthesis, is also related to reactive oxygen species (ROS) production. Febuxostat, as an XO inhibitor, was reported to maintain atherosclerotic plaque stability by decreasing ROS-induced matrix metalloproteinase expression in macrophages (Wei et al. 2020). Febuxostat alleviated mitochondrial-dependent apoptosis and protected against myocardial ischemia/reperfusion and hypoxia/reperfusion injury-induced ROS generation (Wang et al., 2015). In previous basic studies, febuxostat exhibited a protective effect on cardiovascular disease by decreasing oxidative stress and inflammation. However, clinical studies have reported that febuxostat was associated with an increased risk of heart failure hospitalization, atrial fibrillation hospitalization, and cardiovascular death compared with allopurinol (Mackenzie et al., 2020; Singh and Cleveland, 2019; Su et al., 2019). In our clinical practice, we found 2 hyperuricemia patients characterized by febuxostat-related VT, suggesting an additional target of febuxostat. The side effects of drugs are usually caused by undesirable off-target activity. Thus, we speculate that there are potential off-target effects of febuxostat in addition to its intended effects on XO. Previous studies showed that febuxostat treatment suppressed the expression of USAG-1 (an antagonist of bone morphogenetic protein-7; Lu et al., 2019) and breast cancer resistance protein (Toyoda et al., 2019). In the present study, we reported that febuxostat activates JNK phosphorylation independent of inhibition on XO activity, increasing arrhythmia risk in gout patients. In future studies, using artificial intelligence-guided computer simulation may help to screen binding points of febuxostat and JNK or upstream of JNK.

Calcium release from both L-type channels and the sarcoplasmic reticulum leads to myocyte contraction and action potentials (Bers, 2008). Therefore, calcium influx has been used for contraction assessment (van Marion et al., 2019) and arrhythmia prediction (Lu et al., 2015). Tracing intracellular calcium movement using a calcium-sensitive dye equipped with a CMOS camera is a cost-effective and efficient method (Han et al., 2014). In our study, we observed irregular calcium traces and low CaT amplitudes after febuxostat exposure in a dose-dependent manner. Early after-depolarizations are secondary depolarizations during phase 2 or 3 of AP, causing lethal VT. Here, we evaluated EAD-like events to predict ventricular arrhythmia risk. We also observed fibrillation-like events which represented proarrhythmic abnormalities in this study. These fibrillation-like waves were also described as repolarization failure (Passini et al., 2016) or impulse reentry (Kopljar et al., 2018) in other studies. The low calcium transient amplitude in hiPSC-CMs after febuxostat treatment may represent diastolic calcium leakage risk and weak contraction abilities.

Gene expression sequencing and bioinformatic analysis enabled us to focus on the MAPK signaling pathway to explain the mechanism of febuxostat-induced electrophysiological abnormalities in hiPSC-CMs. The MAPK signaling pathway can be activated by both intrinsic and extrinsic stress, which was reported to be involved in lysosomal-mediated degradation of connexin-43 (Kam et al., 2018) and interleukin 17-mediated left ventricle structural remodeling (Chang et al., 2018). Hall et al. (2021) found that cardiac natriuretic peptides can protect against stress-induced ventricular arrhythmias by regulating cardiomyocyte cAMP response element-binding protein phosphorylation through a p38 MAPK signaling cascade. All these findings indicated that activated MAPKs are critical in cardiac arrhythmogenesis. As a prominent part of MAPKs, stress kinase JNK activation was found to be elevated by western blot in this study. Furthermore, JNK subsequently upregulated CaMKII expression at the transcriptional level and phosphorylated the CaMKII protein at the posttranscriptional level, enhancing diastolic sarcoplasmic reticulum Ca²⁺ leakage (Yan et al., 2018). CaMKII is a classical proarrhythmic molecule. A recent study also suggested that dual modulations of JNK2 in CaMKII-dependent arrhythmic SR Ca²⁺ leakage and CaMKII-independent uptake are responsible for atrial arrhythmogenicity (Yan et al., 2021). Our results were consistent with these findings. Febuxostat induces calcium dyshomeostasis through elevated JNK phosphorylation, although increased CaMKII phosphorylation did not reach statistical significance.

Our study had some limitations that need to be considered. One is that we did not perform whole exon sequencing on these 2 patients, so that the genetic factors could be underestimated. However, we used the hiPSC-CMs obtained from the healthy population instead of patient-specific iPSC-CMs but still identified the proarrhythmic effects of febuxostat. Second limitation is the immature nature of hiPS-CMs (Yang et al., 2014). To compensate this, we extended the culture period to 8 weeks to mature cells as much as possible. Another limitation is the calcium indicator. Constrained by the equipment, we would not able to use fura2 as indicator so that we fail to perform quantitative measurement. Nevertheless, fluo-4 is more sensitive and allows to measure larger calcium elevation with lower saturation. Also, our fluorescent microscope is equipped with an sCMOS camera which allows for capturing images with high signal-to-noise ratio.

CONCLUSIONS

In summary, we examined cardiac toxicity and calcium dynamics of febuxostat using hiPSC-CMs. Our study is the first to demonstrated that febuxostat exposure increased the risk of arrhythmia events by disrupting calcium dynamics. Phosphorylated activation of JNK was involved in this process and the inhibition of phosphorylated JNK attenuated arrhythmogenicity.

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.

The authors certify that all research involving human subjects was done under full compliance with all government policies and the Helsinki Declaration.

ACKNOWLEDGMENTS

We gratitude Chengyi Peng and Chaofeng Chen for their work on cases collection.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

This study involved 2 cases and was previously approved by the Bioethics Committee of the First Affiliated Hospital of Nanjing Medical University (2014-SR-090). Participants provided their verbal informed consent for use and publishment of their clinical data.

FUNDING

This work was supported by the National Natural and Science Foundation of China (grant number 82070343 to M.C., 81900295 to C. Cui, and 82000320 to C. Cai) and Natural Science Foundation of Jiangsu Province of China (grant number BK20191071 to C. Cui).

DECLARATION OF CONFLICTING INTERESTS

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability

The data generated and analyzed during the current study are available from the corresponding author on reasonable request.

REFERENCES

Author notes