Trimethylamine N-Oxide (TMAO) Impairs Purinergic Induced Intracellular Calcium Increase and Nitric Oxide Release in Endothelial Cells

Trimethylamine N-oxide (TMAO) is a diet derived compound directly introduced through foodstuff, or endogenously synthesized from its precursors, primarily choline, L-carnitine, and ergothioneine. New evidence outlines high TMAO plasma concentrations in patients with overt cardiovascular disease, but its direct role in pathological development is still controversial. The purpose of the study was to evaluate the role of TMAO in affecting key intracellular factors involved in endothelial dysfunction development, such as reactive oxygen species, mitochondrial health, calcium balance, and nitric oxide release using bovine aortic endothelial cells (BAE-1). Cell viability and oxidative stress indicators were monitored after acute and prolonged TMAO treatment. The role of TMAO in interfering with the physiological purinergic vasodilatory mechanism after ATP stimulation was defined through measurements of the rise of intracellular calcium, nitric oxide release, and eNOS phosphorylation at Ser1179 (eNOSSer1179). TMAO was not cytotoxic for BAE-1 and it did not induce the rise of reactive oxygen species and impairment of mitochondrial membrane potential, either in the basal condition or in the presence of a stressor. In contrast, TMAO modified the purinergic response affecting intracellular ATP-induced calcium increase, nitric oxide release, and eNOSSer1179. Results obtained suggest a possible implication of TMAO in impairing the endothelial-dependent vasodilatory mechanism.


Introduction
Endothelial dysfunction (ED) is a pathological condition, strictly related to thrombosis and atherosclerosis development, characterized by endothelium impairment due to oxidative stress, inflammation and improper balance between vasodilator and vasoconstrictor mediators [1][2][3]. In particular, ED is characterized by endothelial cells' compromised ability to control the vasodilatory mechanism through nitric oxide (NO) release [3]. In physiological conditions, NO is synthesized by endothelial cells principally through the activation of endothelial nitric oxide synthase (eNOS) by both a calcium-dependent and a calcium-independent mechanism. The first process is activated by the rise of intracellular calcium [Ca 2+ ] i due to its entry from extracellular environment or to the opening of intracellular stores during specific agonist stimulation. Ca 2+ -calmodulin interaction activates eNOS which catalyzes the conversion of L-arginine and O 2 to L-citrulline and NO [4,5]. Endothelial nitric oxide synthase activation can be also mediated by post-translational modifications of the enzyme, such as phosphorylation or dephosphorylation of specific residues. In particular, phosphorylation at Ser1177 (human), or Ser1179 (bovine), causes conformational changes of the enzyme that promote calmodulin interaction with eNOS and the subsequent NO synthesis [4][5][6]. Improper calcium handling in response to extracellular

TMAO Does Not Affect Cells Viability
To assess if TMAO has any cytotoxic effect on endothelial cells, BAE-1 were treated with different concentrations (1 μM, 10 μM, 100 μM, 1 mM, and 10 mM) of the compound for 24 h, 48 h, and 72 h. TMAO did not influence cell viability, monitored through the MTS assay, at any time and concentration tested ( Figure 1). Because TMAO had no effects on BAE-1 viability, subsequent experiments were performed with a concentration of 100 μM, considered the highest plasma level registered in pathological settings [29,30], both in basal condition and in the presence of stressors.

TMAO Does Not Impair Mitochondrial Membrane Potential
Mitochondrial membrane potential (∆Ψm) was assessed through confocal microscopy and the ratiometric probe JC-1. Variations of 568/488 nm fluorescence ratio were used as indicators of ∆Ψm depolarization after 1 h or 24 h of treatment with TMAO 100 μM in basal condition or in the presence of Menadione (MEN) 100 μM added in the last hour of TMAO treatment. As Figure 2 shows, TMAO did not induce any effect on ∆Ψm at both times tested, furthermore it did not modify the effect of MEN that, as expected, exerted a reduction in 568/488 nm ratio.

TMAO Does Not Impair Mitochondrial Membrane Potential
Mitochondrial membrane potential (∆Ψm) was assessed through confocal microscopy and the ratiometric probe JC-1. Variations of 568/488 nm fluorescence ratio were used as indicators of ∆Ψm depolarization after 1 h or 24 h of treatment with TMAO 100 µM in basal condition or in the presence of Menadione (MEN) 100 µM added in the last hour of TMAO treatment. As Figure 2 shows, TMAO did not induce any effect on ∆Ψm at both times tested, furthermore it did not modify the effect of MEN that, as expected, exerted a reduction in 568/488 nm ratio.

TMAO Does Not Induce the Rise of Reactive Oxygen Species
Reactive oxygen species (ROS) increase was monitored through the microplate reader and the CellROX ® green probe. Variation in mean cell fluorescence at 485 nm was considered as a predictor of intracellular ROS levels. Data obtained showed similar ROS levels between control condition and TMAO 100 μM treatment both at 1 h and 24 h. Treatment with MEN 100 μM induced, as expected, the rise of ROS, and this effect was maintained in the presence of TMAO (Figure 3).

TMAO Does Not Induce the Rise of Reactive Oxygen Species
Reactive oxygen species (ROS) increase was monitored through the microplate reader and the CellROX ® green probe. Variation in mean cell fluorescence at 485 nm was considered as a predictor of intracellular ROS levels. Data obtained showed similar ROS levels between control condition and TMAO 100 µM treatment both at 1 h and 24 h. Treatment with MEN 100 µM induced, as expected, the rise of ROS, and this effect was maintained in the presence of TMAO (Figure 3).

TMAO Interferes with ATP-Induced Intracellular Calcium Increase
Intracellular calcium variations after purinergic stimulation with ATP were monitored in confocal microscopy in time course acquisitions using the Fluo-3 AM probe. As Figure 4a shows, only pretreatment with TMAO for 24 h induced a significant decrease in the calcium signal lengthening with respect to basal ATP stimulation. Intracellular calcium variations were expressed as the difference of the maximal mean fluorescence and the mean fluorescence at t/2 of the total acquisition time (indicated by the vertical dotted line in Figure 4b), (Fmax-Ft/2), for each experimental condition. Figure 4b presents mean fluorescence curves aligned with respect to the peak value (Fmax) of one representative experiment, in which more pronounced intracellular calcium signal decay is visible in 24 h TMAO-treated cells stimulated with ATP.

TMAO Interferes with ATP-Induced Intracellular Calcium Increase
Intracellular calcium variations after purinergic stimulation with ATP were monitored in confocal microscopy in time course acquisitions using the Fluo-3 AM probe. As Figure 4a shows, only pretreatment with TMAO for 24 h induced a significant decrease in the calcium signal lengthening with respect to basal ATP stimulation. Intracellular calcium variations were expressed as the difference of the maximal mean fluorescence and the mean fluorescence at t/2 of the total acquisition time (indicated by the vertical dotted line in Figure 4b), (F max -F t/2 ), for each experimental condition. Figure 4b presents mean fluorescence curves aligned with respect to the peak value (F max ) of one representative experiment, in which more pronounced intracellular calcium signal decay is visible in 24 h TMAO-treated cells stimulated with ATP.

TMAO Reduces Nitric Oxide Release in Purinergic Response to ATP
As in endothelial cells, ATP-induced calcium signal triggers eNOS activation and consequently nitric oxide release, thus leading to vasodilation. The next goal was to interlink the inhibitory effect of TMAO on the calcium signal with a TMAO-dependent impairment of ATP-stimulated nitric oxide production. NO release after ATP stimulation was monitored in time course imaging using confocal microscopy and the DAR-4M AM probe. Figure 5a presents fluorescence variations after ATP stimulation and, as the graph shows, only TMAO pretreatment for 24 h induced an alteration in NO release after purinergic stimulation. Figure 5b shows single cell fluorescence variations in a representative experiment, pointing out that ATP stimulation in basal condition and after TMAO pretreatment for 1 h induced fluorescence slope changes, that are directly associated with NO increase. On the contrary, TMAO pretreatment for 24 h inhibited NO release after ATP stimulation, indeed, no fluorescence slope changes were detected. the calcium signal lengthening with respect to basal ATP stimulation. Intracellular cal cium variations were expressed as the difference of the maximal mean fluorescence and the mean fluorescence at t/2 of the total acquisition time (indicated by the vertical dotted line in Figure 4b), (Fmax-Ft/2), for each experimental condition. Figure 4b presents mean fluorescence curves aligned with respect to the peak value (Fmax) of one representative experiment, in which more pronounced intracellular calcium signal decay is visible in 24 h TMAO-treated cells stimulated with ATP.

TMAO Reduces Nitric Oxide Release in Purinergic Response to ATP
As in endothelial cells, ATP-induced calcium signal triggers eNOS activation and consequently nitric oxide release, thus leading to vasodilation. The next goal was to interlink the inhibitory effect of TMAO on the calcium signal with a TMAO-dependent impairment of ATP-stimulated nitric oxide production. NO release after ATP stimulation was monitored in time course imaging using confocal microscopy and the DAR-4M AM probe. Figure 5a presents fluorescence variations after ATP stimulation and, as the graph shows, only TMAO pretreatment for 24 h induced an alteration in NO release after purinergic stimulation. Figure 5b shows single cell fluorescence variations in a representative experiment, pointing out that ATP stimulation in basal condition and after TMAO pretreatment for 1h induced fluorescence slope changes, that are directly associated with NO increase. On the contrary, TMAO pretreatment for 24 h inhibited NO release after ATP stimulation, indeed, no fluorescence slope changes were detected.

TMAO Impacts on Endothelial Nitric Oxide Synthase Phosphorylation (eNOS Ser1179 )
Western blot analyses were performed to evaluate if TMAO could influence eNOS phosphorylation in purinergic response to ATP. As Figure 6a shows, pretreatment of BAE-1 with TMAO for 1 h did not influence the eNOS Ser1179 /eNOS ratio after purinergic

TMAO Impacts on Endothelial Nitric Oxide Synthase Phosphorylation (eNOS Ser1179 )
Western blot analyses were performed to evaluate if TMAO could influence eNOS phosphorylation in purinergic response to ATP. As Figure 6a shows, pretreatment of BAE-1 with TMAO for 1 h did not influence the eNOS Ser1179 /eNOS ratio after purinergic stimulation, while pretreatment with TMAO 100 µM for 24 h before ATP stimulation significantly reduced eNOS phosphorylation, thus suggesting a possible implication of the molecule in the impairment of physiological vascular tone.

Discussion
ED is a pathological condition characterized by improper balance between vasodilatory and vasoconstrictory mechanisms [1][2][3]7]. Prevention of its development is fundamental to maintain a proper physiological functionality of the cardiovascular system. Several factors have been outlined to favor the development of ED and, among others, genetic, environmental, and life-style dependent are the most studied [2]. In particular, wrong dietary habits can exacerbate a pre-existing unstable condition, thus aggravating oxidative stress and inflammatory response that contribute to the onset of ED [10]. Several studies point out a possible involvement of the diet-derived compound TMAO in cardiovascular diseases development. The aim of the study was to evaluate the role of TMAO in the onset of endothelial dysfunction in an in vitro model using bovine aortic endothelial cells. First results show no cytotoxic effects of TMAO in cells treated for 24 h, 48 h, and 72 h with different concentrations (1 μM, 10 μM, 100 μM, 1 mM, and 10 mM) of the compound (Figure 1). This result is supported by data obtained by Ma et al., that show no reduction in human umbilical vein endothelial cell (HUVEC) viability after 24 h of treatment with TMAO (0, 10, 50, 100 μmol/L) [25]. Further experiments were addressed to the analysis of the TMAO contribution in two conditions that steer towards endothelial damage: impairment of the mitochondrial membrane potential and the rise of reactive oxygen species [1]. The role of MEN in inducing these deleterious effects on endothelial cells has already been studied [31,32], so, in the second part of this work, we focused on the comparison between the impact of TMAO 100 μM, considered the highest plasma concentration registered in patients with chronic kidney disease [29,30], and MEN on ∆Ψm and ROS production. Results presented in Figures 2 and 3 show no effects of TMAO alone in inducing mitochondrial membrane depolarization and reactive oxygen species increase, both after 1 h or 24 h of treatment, and these results are comparable to our previous data obtained in adult rat cardiomyocytes [33]. Furthermore, and according to our report on cardiomyocytes, the effects of MEN were unchanged in cells pretreated with TMAO for 1 h or 24 h [33]. Contrasting results on TMAO and ROS production, obtained by Sun and

Discussion
ED is a pathological condition characterized by improper balance between vasodilatory and vasoconstrictory mechanisms [1][2][3]7]. Prevention of its development is fundamental to maintain a proper physiological functionality of the cardiovascular system. Several factors have been outlined to favor the development of ED and, among others, genetic, environmental, and life-style dependent are the most studied [2]. In particular, wrong dietary habits can exacerbate a pre-existing unstable condition, thus aggravating oxidative stress and inflammatory response that contribute to the onset of ED [10]. Several studies point out a possible involvement of the diet-derived compound TMAO in cardiovascular diseases development. The aim of the study was to evaluate the role of TMAO in the onset of endothelial dysfunction in an in vitro model using bovine aortic endothelial cells. First results show no cytotoxic effects of TMAO in cells treated for 24 h, 48 h, and 72 h with different concentrations (1 µM, 10 µM, 100 µM, 1 mM, and 10 mM) of the compound (Figure 1). This result is supported by data obtained by Ma et al., that show no reduction in human umbilical vein endothelial cell (HUVEC) viability after 24 h of treatment with TMAO (0, 10, 50, 100 µmol/L) [25]. Further experiments were addressed to the analysis of the TMAO contribution in two conditions that steer towards endothelial damage: impairment of the mitochondrial membrane potential and the rise of reactive oxygen species [1]. The role of MEN in inducing these deleterious effects on endothelial cells has already been studied [31,32], so, in the second part of this work, we focused on the comparison between the impact of TMAO 100 µM, considered the highest plasma concentration registered in patients with chronic kidney disease [29,30], and MEN on ∆Ψm and ROS production. Results presented in Figures 2 and 3 show no effects of TMAO alone in inducing mitochondrial membrane depolarization and reactive oxygen species increase, both after 1 h or 24 h of treatment, and these results are comparable to our previous data obtained in adult rat cardiomyocytes [33]. Furthermore, and according to our report on cardiomyocytes, the effects of MEN were unchanged in cells pretreated with TMAO for 1 h or 24 h [33]. Contrasting results on TMAO and ROS production, obtained by Sun and collaborators, showed increased levels of ROS and lower superoxide dismutase (SOD) activity after HUVEC treatment with TMAO (100, 200, 300 µmol/L) for 1 h, 3 h, and 6 h [26]. These discrepancies could be ascribed to the dissimilarity in the cell model, the setting of time treatments and the TMAO concentration used. Given the lack of changes in redox status and in mitochondrial membrane potential observed in both control and stressed BAE-1 cells treated with TMAO 100 µM, our attention moved towards the potential role of the molecule in affecting the main physiological mechanism of vascular tone regulation, precisely the nitric oxide-mediated pathway. Endothelial dysfunction is indeed strictly related to impaired nitric oxide release. Therefore, the last part of the study was focused on TMAO modulation of the purinergic-dependent intracellular pathway, consisting of the rise of intracellular calcium followed by nitric oxide release, with ATP being used as purinergic agonist. Results obtained show an altered shape of the ATP-induced intracellular calcium signal in cells pretreated with TMAO for 24 h (Figure 4). In particular, we observed significant reduction of the plateau-phase duration. A role of TMAO in the modulation of agonist-induced calcium signals was described by Zhu and coworkers in platelets. Their results showed an increase in IP3-induced calcium release in thrombin-stimulated platelets pretreated with TMAO, and a consequent enhancement of platelet activation and thrombotic event development [27]. Thus, in both studies TMAO seems to affect intracellular calcium homeostasis, even if at different levels, but either way resulting in a detrimental effect on vascular health.
Several cellular proteins have been identified as molecular targets of TMAO with involvement in its pathological effects. Among them, and within the development of vascular dysfunction, proteins of the Nlrp3 inflammasome complex, mitogen-activated protein kinases, protein kinase like ER kinase (PERK), have been recently highlighted. Therefore, a unique and specific molecular mechanism cannot be ascribed to TMAO, and this evidence complies with its general role of protein stabilizer against various stresses. As interestingly proposed by Hong, this property of TMAO could be detrimental in the absence of stressors, as it could affect the conformational flexibility needed for protein functions [34]. In this perspective, the alteration of the ATP-mediated intracellular calcium signal induced by TMAO in our cellular model is time dependent and could be the result of broken conformational changes required both for functionality and mutual interaction of the various calcium channels involved.
As the leading downstream target of purinergic-induced calcium signal in endothelial cells is represented by eNOS, impairment of this key factor could impact NO release. Our results of NO measurement in BAE-1 cells clearly follow this scenario, showing that TMAO treatment for 24 h reduced nitric oxide release ( Figure 5). In agreement with these data, treatment with TMAO for 24 h also induced a significant reduction in eNOS Ser1179 phosphorylation (Figure 6), thus integrating the whole mechanism leading up to eNOS dysfunctional activity.
A TMAO-dependent reduction in NO release was also observed by Sun et al. after HUVEC treatment with TMAO (100, 200, 300 µmol/L) for 1 h, 3 h, and 6 h [26], even if it was associated with a reduction in eNOS expression rather than in its activation.
On the other hand, a reduction in eNOS phosphorylation was detected in aortic rings from mice fed with a TMAO supplemented diet [24].
The present study shows some limitations, such as the experimental model, that is not based on a human-derived cell line, and the short lasting of TMAO treatment, that is no longer than 24 h and could not be comparable to a chronic exposure, as occurs in some pathological conditions. Despite these weaknesses, different points of strength could be outlined: above all, this is the first in vitro study showing a TMAO mediated impairment of the purinergic dependent NO release in endothelial cells, thus suggesting its direct detrimental effect in the physiological control of vascular tone. Moreover, by using TMAO 8 of 12 100 µM, we aimed to keep close to the highest plasma TMAO concentrations detected in patients with atherosclerotic coronary artery disease and chronic kidney disease [29,30,35]. Furthermore, this last hallmark could explain the ineffectiveness of TMAO in our conditions in affecting ROS levels and mitochondrial health, in comparison with other results obtained with higher TMAO concentrations [36].
In conclusion, this study points out the potential role of TMAO in unsettling the purinergic-activated Ca 2+ -eNOS pathway, adding a new element in the complex scenario of the emergent gut-heart axis concept.

Intracellular Calcium in Response to ATP Stimulation
Intracellular calcium variations were monitored through time course acquisitions in confocal microscopy and Fluo-3 AM probe (Thermo Fisher Scientific). BAE-1 were seeded at a density of 3.3 × 10 4 cells/mL on uncoated glass bottom dishes of 35 mm diameter (Ibidi) in culture media and incubated at 37 • C for 24 h. Following incubation, cells were treated for 1 h or 24 h with TMAO 100 µM or were left in culture medium. Fluo-3 AM (2 µM) was added in each dish 30 min before the end of all treatments and cells were then washed twice with Tyrode standard solution. Fluorescence intensity variations in time were measured at 488 nm through an Olympus Fluoview 200 laser scanning confocal system mounted on an inverted IX70 Olympus microscope, and all acquisitions were performed with a 60X Uplan FI (NA 1.25) oil-immersion objective. During the experiments BAE-1 were maintained in Tyrode standard solution containing or not TMAO according to pretreatment protocols; ATP 100 µM and ATP 100 µM + TMAO 100 µM used for stimulation were added through a microperfusion system (pipette diameter 200 µm). Intracellular calcium variations in time were analyzed through the definition of the ROIs, using the software ImageJ. Changes in intracellular calcium concentration were calculated as F/F 0 to normalize the traces; moreover, an additional analysis was applied in order to highlight changes in the ATP-induced calcium signal profile between the control and TMAO-treated cells. First, traces of the same dish were mediated; from each of four independent experiments, mean traces from every condition (ATP, TMAO 1 h + ATP, TMAO 24 h + ATP) were aligned with respect to the peak value (F max ). Then, the difference between F max and mean fluorescence at t/2 (F t/2 ), (F max -F t/2 ), of the aligned curves was compared in order to evaluate if any perturbation in calcium curves occurred.

Nitric Oxide Release after ATP Stimulation
Nitric oxide (NO) release was monitored through time course acquisitions in confocal microscopy and DAR-4M AM probe (Thermo Fisher Scientific). BAE-1 were seeded at a density of 3.3 × 10 4 cells/mL on uncoated glass bottom dishes of 35 mm diameter (Ibidi) in culture media and incubated at 37 • C for 24 h. Following incubation cells were treated for 1 h or 24 h with TMAO 100 µM or were left in culture medium. DAR-4M AM (5 µM) was added in each well 30 min before the end of the treatment and cells were then washed twice with Tyrode standard solution. Fluorescence intensity variations in time were measured at 568 nm through an Olympus Fluoview 200 laser scanning confocal system mounted on an inverted IX70 Olympus microscope, and all acquisitions were performed with a 60X Uplan FI (NA 1.25) oil-immersion objective. During the experiments BAE-1 were maintained in Tyrode standard solution containing or not TMAO according to pretreatment protocols; ATP 100 µM and ATP 100 µM + TMAO 100 µM used for stimulation were added through a microperfusion system (pipette diameter 200 µm). NO release variations in time were analyzed through the definition of the ROIs using the software ImageJ. NO release after different treatments was expressed as percentage change in mean fluorescence detected ((F max − F 0 )/F 0 ) × 100 for each trace. Normalized percentage change values of three independent experiments were then mediated and expressed as folds toward basal ATP treatment to verify if any variation in NO release could be detected in different treatments.

Western Blot
Total eNOS and its phosphorylated form (eNOS Ser1179 ) in purinergic stimulation with ATP 100 µM for 1 min after treatment with TMAO 100 µM for 1 h or 24 h were monitored through Western blot. BAE-1 were seeded at a density of 4 × 10 4 cells/mL on plastic petri dishes with 22.1 cm 2 of growth area in culture media and incubated at 37 • C for 24 h. Following incubation, cells were treated for 1 h or 24 h with TMAO 100 µM or were left in culture medium. After the treatment with ATP 100 µM for 1 min, cells were lysed with RIPA lysis buffer (ThermoFisher Scientific) containing phosphatase inhibitor cocktail (PhosSTOP, Roche, Mannheim, Germany), forced through a 1 mL syringe needle several times, centrifuged at 10,000 rpm for 5 min at 4 • C and stored at −80 • C. Protein lysates (20 µg per lane) were run on 8% SDS-PAGE gel, transferred to a polyvinylidene fluoride membrane (PVDF) and blocked for 1 h in TBST (10 mM Tris-HCl, pH 7.5, 0.1 M NaCl, 0.1% Tween 20) plus 5% non-fat dry milk at 37 • C. PVDF were incubated overnight at 4 • C with primary antibodies (monoclonal anti-eNOS, 1:500 dilution, BD, Biosciences; polyclonal anti-eNOS Ser1179 , 1:250 dilution, Thermo Fisher Scientific; monoclonal anti-βactin, 1:2000 dilution, Sigma-Aldrich). Membranes were then washed three times with TBST and incubated for 1 h at room temperature with horseradish peroxidase-conjugated secondary antibodies (anti-mouse, 1:20,000 dilution, Amersham, for total eNOS and βactin; and anti-rabbit, 1:10,000 dilution, Amersham, for eNOS Ser1179 ) and washed again three times with TBST. Protein bands were localized by chemiluminescence with Western Lightning Plus-ECL (Perkin Elmer, Waltham, MA, USA). Protein levels were determined using the software ImageJ and expressed as mean percentage toward ATP condition of eNOS Ser1179 /eNOS ratio of four independent experiments.

Statistical Analysis
Data are presented as mean ± SEM. All data were analyzed with GraphPad Prism 8.0.1 software using ANOVA followed by Bonferroni's multiple comparison for post hoc tests. Differences with p < 0.05 were considered statistically significant.