Electronic cigarette refill liquids are commercially provided with a wide variety of flavoring agents. A recent study suggested that several common flavors may scavenge nitric oxide (NO) and cause endothelial dysfunction. It was the aim of the present study to investigate the effects of these flavors on NO/cyclic GMP-mediated signaling and vascular relaxation. We tested the flavoring agents for effects on Ca2+-induced cGMP accumulation and NO synthase activation in cultured endothelial cells. NO scavenging was studied with NO-activated soluble guanylate cyclase and as NO release from a NO donor, measured with a NO electrode. Blood vessel function was studied with precontracted rat aortic rings in the absence and presence of acetylcholine or a NO donor. Cinnamaldehyde inhibited Ca2+-stimulated endothelial cGMP accumulation and NO synthase activation at ≥0.3 mM. Cinnamaldehyde and diacetyl inhibited NO-activated soluble guanylate cyclase with IC50 values of 0.56 (0.54–0.58) and 0.29 (0.24–0.36) mM, respectively, and caused moderate NO scavenging at 1 mM that was not mediated by superoxide anions. The other compounds did not scavenge NO at 1 mM. None of the flavorings interfered with acetylcholine-induced vascular relaxation, but they caused relaxation of pre-contracted aortas. The most potent compounds were eugenol and cinnamaldehyde with EC50 values of ~0.5 mM. Since the flavors did not affect endothelium-dependent vascular relaxation, NO scavenging by cinnamaldehyde and diacetyl does not result in impaired blood vessel function. Although not studied in vivo, the low potency of the compounds renders it unlikely that the observed effects are relevant to humans inhaling flavored vapor from electronic cigarettes.
Citation: Wölkart G, Kollau A, Stessel H, Russwurm M, Koesling D, Schrammel A, et al. (2019) Effects of flavoring compounds used in electronic cigarette refill liquids on endothelial and vascular function. PLoS ONE 14(9): e0222152. https://doi.org/10.1371/journal.pone.0222152
Editor: Michael Bader, Max Delbruck Centrum fur Molekulare Medizin Berlin Buch, GERMANY
Received: March 12, 2019; Accepted: August 18, 2019; Published: September 9, 2019
Copyright: © 2019 Wölkart et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The raw data of all experiments shown in the manuscript body and the supporting information are freely available at zenodo: https://zenodo.org/record/3377414#.XWOsOS35xXs.
Funding: This work was supported by the Austrian Science Fund (FWF) grant [P24946] to BM. website: https://www.fwf.ac.at/en/. The FWF had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Electronic cigarettes (e-cigs) vaporize liquids, typically containing propylene glycol, glycerol, nicotine, and flavoring compounds, at a temperature of 200–250°C. Current and former smokers inhale the resulting vapor instead of tobacco smoke . The combustion of tobacco leaves and cigarette paper at >800°C leads to the emission of several thousand compounds, many of which are cancerogenic or otherwise toxic. According to the WHO, about 6 million smokers die each year from tobacco-associated diseases, and cigarettes kill every second user. Since e-cigs do not burn tobacco, vaping is not associated with the inhalation of combustion products. The analytical studies currently available show that at normal usage of the products the harmful constituents of tobacco smoke are either absent or present only in trace amounts in the inhaled aerosol [2, 3] and in the air of closed rooms after vaping [4, 5]. Some researchers described e-cig vapor as being similarly harmful as tobacco smoke because of the emission of particulate matter without considering the difference between the harmful solid particles in smoke and benign liquid droplets in e-cig aerosols . The switch from smoking to vaping was shown to result in the pronounced reduction of biomarkers for toxicity and improvement of several clinical parameters [7, 8], demonstrating the reduction of the health risks of consumers. Clinical studies showing significant improvement of lung function of smokers with asthma  as well as reduced symptom load of COPD patients , who used e-cigs instead of combustible tobacco cigarettes, further support this view.
Despite the apparent potential of e-cigs in tobacco harm reduction, some health professionals are still reluctant to recommend these products to smokers because of several caveats. According to a report by Stanton Glantz and co-workers, e-cigs do not aid in smoking cessation but, in contrast, decrease the rate of successful quit attempts . This claim contradicts a recent randomized clinical trial  and observational studies [13, 14], indicating that e-cigs are superior over approved nicotine replacement therapy. Another objection is the continued consumption of nicotine, which could cause adverse cardiovascular effects due to sympathetic stimulation and support the dependence of users. However, nicotine is not responsible for the high morbidity and mortality of smoking and does not increase the risk for severe cardiovascular events . Observation of never smokers vaping for 3.5 years provided no evidence for impaired lung or cardiovascular function . Concerning dependence, studies with rodents show that non-nicotine ingredients of tobacco smoke, e.g. inhibitors of monoamine oxidase and acetaldehyde, significantly enhance the positive reinforcing effects of nicotine . In humans, the addictive potential of e-cigs appears to be similar to that of nicotine-containing chewing gums approved for therapeutic smoking cessation . In addition to the substance-induced release of dopamine, airway sensation and conditioned smoking behavior significantly contribute to the dependence of smokers . Vaping allows consumers to maintain their conditioned behavior and provides the throat hit necessary for satisfaction. This may explain the greater appeal and success rates of e-cigs as compared to nicotine replacement therapy .
Published surveys show that availability of a wide array of flavors is essential for the satisfaction of consumers and that the majority of vapers prefer fruit and other non-tobacco flavors [21, 22]. It is beyond the scope of this article to discuss the current controversy about the alleged nicotine epidemic caused by flavored e-cigs among youth, and we wish to focus on the potential health risks of inhaled flavors. Virtually all of the compounds used in refill liquids are approved as food flavorings and widely used without evidence for adverse effects. However, it cannot be excluded that some of these compounds are harmful if inhaled. For most flavors, this assumption is not supported by evidence, but there are two notable exceptions. Diacetyl, used in certain sweet liquids, was proposed to cause Bronchiolitis obliterans ("popcorn lung"), due to damage to lung epithelial cells . Although the diacetyl concentrations in e-cig vapor are probably below occupational limits, the addition of diacetyl to e-liquids was banned in some countries, including Germany and Austria. For safety reasons, consumers are advised to refrain from using diacetyl containing liquids. The second suspicious compound is cinnamaldehyde, which was shown to be toxic to some types of cultured cells, including lung epithelial cells . Whether the results of in vitro experiments with cultured cells exposed to e-cig vapor can be translated to the in vivo situation of human consumers remains to be established.
Potential toxic effects of flavors on alveolar epithelial cells may occur in the absence of significant levels in the systemic circulation, but this does not apply to effects on vascular and endothelial function. In a recent study , it has been claimed that certain flavors may cause endothelial dysfunction due to scavenging of nitric oxide (NO). Therefore, we tested the effects of the flavoring agents used by Fetterman et al.  on endothelial NO signaling, NO scavenging and NO-induced relaxation of rat aortas.
Materials and methods
L-[2,3,4,5-3H]arginine (57 Ci/mmol) and [α-32P]GTP (800 Ci/mmol) were from Perkin Elmer Life and Analytical Sciences (Vienna, Austria), 2,2-Diethyl-1-nitroso-oxyhydrazine (DEA/NO) was from Enzo Life Sciences (Lausen, Switzerland) obtained via Eubio (Vienna, Austria). All other chemicals, including the flavorings listed in Table 1, were from Sigma-Aldrich (Vienna, Austria). Stock solutions of flavorings (1 M) were prepared in DMSO and diluted as required with deionized water. The following cinnamaldehyde containing liquids were obtained from happy liquid (Munich, Germany): Auszogne mit Zimt, Glühwein, and Apfelstrudel.
Cell culture and drug treatment
Porcine aortic endothelial cells were isolated as described  and cultured at 37°C, 5% CO2 for up to 3 passages in Dulbecco´s modified Eagle´s medium, containing 10% heat-inactivated fetal calf serum (FCS), 100 U/ml penicillin, 0.1 mg/ml streptomycin, and 1.25 μg/ml amphotericin B. Prior to experiments, cells were subcultured into 24-well plates (determination of cGMP formation) or 6-well plates (determination of L-citrulline formation).
Determination of endothelial cGMP formation
Accumulation of intracellular 3',5'-cyclic GMP (cGMP) was determined as previously described . Briefly, endothelial cells were washed and preincubated with the compounds to be tested for 15 minutes at 37°C in 50 mM Tris buffer, pH 7.4, containing 100 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2.5 mM CaCl2, 1 mM 3-isobutyl-1-methylxanthine, and 1 μM indomethacin. Reactions were started by addition of the Ca2+ ionophore A 23187 (final concentration 1 μM) or the NO donor DEA/NO (final concentration 1 μM) and terminated after 4 minutes by removal of the incubation medium and addition of 0.01 M HCl. Within 1 h, intracellular cGMP was completely released into the supernatant and measured by radioimmunoassay.
Determination of endothelial L-citrulline formation
NO synthase (NOS) activity in intact cells was determined by monitoring the conversion of L-[3H]arginine into L-[3H]citrulline as previously described . Briefly, endothelial cells were washed and preincubated with the compounds to be tested for 10 minutes at 37°C in 50 mM Tris buffer, pH 7.4, containing 100 mM NaCl, 5 mM KCl, 1 mM MgCl2 and 2.5 mM CaCl2. Reactions were started by addition of L-[2,3,4,5-3H]arginine (~ 106 dpm) and A 23187 (final concentration 0.3 μM) and terminated after 4 min by washing the cells with chilled incubation buffer. Subsequent to lysis of the cells with 0.01 M HCl, an aliquot was removed for determination of incorporated radioactivity. To the remaining sample, 200 mM sodium acetate buffer (pH 13.0) containing 10 mM L-citrulline was added (final pH ~5.0), and L-[3H]citrulline was separated from L-[3H]arginine by cation exchange chromatography.
Determination of soluble guanylate cyclase (sGC) activity
The activity of sGC purified from bovine lung  was determined as formation of [32P]cGMP from [α-32P]GTP. The purified enzyme (50 ng) was incubated at 37°C for 10 min in a final volume of 0.1 ml in the absence or presence of flavorings with the NO donor DEA/NO at concentrations indicated in the figure legends, followed by isolation of [32P]cGMP and quantification of radioactivity (Cerenkov radiation) in a β counter.
To test for reversibility of inhibition, sGC (50 ng) was preincubated in the absence or presence of the respective compounds for 5 minutes at 37°C, followed by 50-fold dilution in the assay mixtures and incubation with DEA/NO at 37°C for 10 minutes.
Determination of cytotoxicity
The viability of endothelial cells was judged by incubation with 1 mM of the flavoring compounds, followed by determination of 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide (MTT) to formazan conversion using the Cayman MTT Cell Proliferation Assay Kit purchased through VWR International (Vienna, Austria).
To test for membrane integrity, endothelial cells were assayed for release of lactate dehydrogenase (LDH) into the culture medium. Subsequent to incubation with flavoring compounds (1 mM each) for 1 hour, 20-μl aliquots of the culture media were mixed with 200 μl of reagent buffer (1 mM pyruvate and 0.1 mM NAD+ in PBS). The decrease in light absorbance at 340 nm was monitored using a diode array spectrophotometer (Hewlett Packard 8452A). Maximum LDH release was determined by treating the cells with 1% Triton X-100.
Electrochemical determination of NO
Scavenging of NO by flavoring compounds was studied with a nitric oxide sensor (ISO-NOP, World Precision Instruments, Berlin, Germany), calibrated with acidified nitrite as described previously . Free NO radical was generated by incubation of DEA/NO (1 μM) in 1 ml of 100 mM phosphate buffer, pH 7.4, in the absence or presence of superoxide dismutase (2500 U/ml) at ambient temperature. At the peak concentration of NO, flavoring compounds (10 μl of DSMO stock solutions) were added, DMSO served as vehicle control.
Determination of cinnamaldehyde in commercial liquids
The concentration of cinnamaldehyde in three commercial liquids was quantified by high-performance liquid chromatography (Agilent 1260 Infinity with diode array detection) on a reversed-phase poroshell 120 EC-C18 column (Agilent Technologies) using 40% methanol and 60% water as mobile phase. Liquids were diluted 1:100 in mobile phase, aliquots of 6 μl were applied onto the column and eluted isocratically with a flow rate of 1.0 ml/min at a column temperature of 25°C. The method was calibrated with authentic cinnamaldehyde (10–300 μM), based on the area of the peak detected at 290 nm (retention time 4.5 min).
Animals and tissue preparation
Animal experiments were performed in compliance with the Austrian law on experimentation with laboratory animals (last amendment 2013) based on the European Union guidelines for the Care and the Use of Laboratory Animals (European Union Directive 2010/63/EU). Unsexed Sprague-Dawley rats (obtained from Charles River, Sulzfeld, Germany) were housed at the local animal facility in approved cages. They were fed standard chow (Altromin 3023; obtained from Königshofer Futtermittel, Ebergassing, Austria) and received water ad libitum. Animals were euthanized in a box that was gradually filled with CO2 until vital signs had ceased (cessation of respiration and circulation). Subsequently, the thorax of the animals was opened, the thoracic aorta was removed, cleaned from connective tissue and immediately used for assessment of vessel function.
The killing of animals solely for the use of their organs or tissues is explicitly excluded in Article 3 (page L276/39) of EU Directive 2010/63/EU on protection of animals used for scientific purposes.
Isometric tension vasomotor studies
For isometric tension measurements, vessel rings with intact endothelium were suspended in 5-ml organ baths containing oxygenated Krebs-Henseleit buffer (concentrations in mM: NaCl 118.4, NaHCO3 25, KCl 4.7, KH2PO4 1.2, CaCl2 2.5, MgCl2 1.2, D-Glucose 11; pH 7.4), as previously described in detail . After equilibration for 60 min at the optimal resting tension of 20 mN, maximal contractile activity was determined with a depolarizing solution containing 60 mM K+. Rings that did not elicit adequate and stable contraction to high K+ were considered damaged and omitted from the study. After washout, tissues were precontracted with the thromboxane mimetic 9,11-dideoxy-9α,11α-epoxy-methanoprostaglandin F2α (U-46619) to an equivalent level (i.e. ~90% of maximal contraction obtained with high K+). When contraction had reached a stable plateau (~after 20 min) cumulative concentration-response curves were established with flavorings. Where indicated, the effects of the flavorings on vascular tone were tested in the presence of the NOS inhibitor NG-nitro-L-arginine methyl ester (L-NAME, 0.2 mM).
To examine a possible effect of the tested flavoring compounds on endothelial function and NO bioavailability, cumulative concentration-response curves to acetylcholine (1 nM—10 μM) or DEA/NO (1 nM—10 μM) were performed on U-46619 precontracted rings that had been preincubated with the flavoring compounds for 30 min. The concentrations of the tested flavoring compounds in these experiments represent the highest concentration that did not induce relaxation per se (i.e. 100 μM for eugenol and cinnamaldehyde, and 300 μM for all other compounds).
Effects are expressed as percent of relaxation of U-46619-induced contraction. Concentration-response curves of different ring segments from a single animal were averaged and counted as an individual experiment.
Data are presented as mean values±SEM of n experiments. Where applicable, concentration-response curves were fitted to a Hill-type model giving estimates of agonist potency (EC50) and efficacy (Emax) using Kaleidagraph (Synergy Software, version 4.5). EC50 values are reported as geometric means with 95% confidence limits (CI). Analysis of variance (ANOVA) with post hoc Dunnett's test was used for comparison between groups using the Kaleidagraph software. Significance was assumed at p<0.05.
Fig 1 shows the effects of 9 flavoring compounds (1 mM each) on endothelial cGMP accumulation in response to the Ca2+ ionophore A23187 (A) and the NO donor DEA/NO (B). The effect of A23187 was completely inhibited by the NOS inhibitor NG-nitro-L-arginine (NNA), confirming activation of the endogenous NO/cGMP pathway. Cinnamaldehyde reduced DEA/NO- and A23187-induced cGMP formation to about 35% of controls, while the other compounds had no significant effects.
Endothelial cells preincubated for 15 min with NNA (0.3 mM), DMSO (0.1%, as solvent control) or the compounds listed (1 mM, each) were stimulated for 4 min with 1 μM A23187 (A) or 1 μM DEA/NO (B), and cGMP formation was determined as described under Materials and Methods. Data are expressed as mean values±SEM (n = 3). *p<0.05 (ANOVA with Dunnett's post hoc test).
Based on these results we studied the concentration-dependent effects of cinnamaldehyde on A23187-stimulated formation of the endothelial NOS products NO (measured as accumulation of cGMP) and L-citrulline. As shown in Fig 2A, endothelial NOS activity, measured as formation of L-citrulline, was significantly inhibited by 0.3 and 1.0 mM cinnamaldehyde, and the effect of the highest concentration was enhanced when the cells were pretreated with the aldehyde for 70 instead of 10 minutes (p<0.05). Direct activation of endothelial sGC with DEA/NO (Fig 2B) was more sensitive to cinnamaldehyde with significant inhibition to 25.3±3.9 and 10.4±2.4% of controls upon incubation with 1 mM of the aldehyde for 15 and 75 minutes, respectively. Lack of effect of superoxide dismutase argues against the involvement of superoxide-mediated oxidation of NO.
A, Endothelial cells were preincubated with increasing concentrations of cinnamaldehyde for 10 min in incubation buffer (open circles) or for 60 min in culture medium and 10 min in incubation buffer (filled circles). Then, cells were stimulated with 1 μM A 23187 (1 μM), and L-citrulline was determined as described under Materials and Methods. Data were normalized to the conversion of L-[3H]arginine into L-[3H]citrulline measured in non-pretreated cells (= 100%) and are expressed as mean values±SEM (n = 4). B, Endothelial cells were preincubated with increasing concentrations of cinnamaldehyde for 15 min in incubation buffer in the absence (open circles) and presence of 1000 U/ml SOD (filled circles), or for 60 min in culture medium and 15 min in incubation buffer. Then, cells were stimulated with DEA/NO (1 μM) and cGMP formation was determined. Data were normalized to cGMP production measured in non-pretreated cells (= 100%) and are expressed as mean values±SEM (n = 4). *p<0.05 vs. controls in the absence of cinnamaldehyde (ANOVA with Dunnett's post hoc test).
As shown in Table 2, neither of the flavoring compounds, including cinnamaldehyde, exhibited cytotoxicity at up to 1 mM, measured as release of LDH and reduction of MTT after incubation of the cells for 60 minutes.
As shown in Fig 3A, at a concentration of 1 mM, 7 out of the 9 tested flavorings had no effect on the activity of purified sGC stimulated with increasing concentrations of DEA/NO (0.01–10 μM). Cinnamaldehyde and diacetyl (1 mM each) caused significant rightward shifts of the DEA/NO concentration-response curve from 37 (29–47) to 71 (43–118) and 155 (106–129) nM, respectively (p<0.05). Cinnamaldehyde slightly reduced maximal sGC activity from 28±0.9 to 24±0.8 μmol cGMP x min-1 x mg-1 (Fig 3B), the other flavorings had no significant effects. The Emax and EC50 values obtained with the various compounds are listed in S1 Table.
Purified bovine lung sGC (50 ng) was stimulated with increasing concentrations of DEA/NO (0.01–10 μM) in the presence of DMSO (0.1% as solvent control) or flavorings (1 mM each). Formation of cGMP was determined as described in Materials and methods. Data are expressed as mean values±SEM (n = 3). For clarity, the effects of cinnamaldehyde and diacetyl compared to control are shown separately in panel B.
The effects of cinnamaldehyde and diacetyl on DEA/NO-stimulated sGC were concentration-dependent with IC50 values of 0.56 (0.54–0.58) and 0.29 (0.24-.36) mM, respectively (Fig 4A) and largely (cinnamaldehyde) or completely (diacetyl, p<0.05) reversed by 50-fold dilution of the enzyme that had been pre-incubated for 5 minutes in the presence of 1 mM of these compounds (Fig 4B).
Effects of cinnamaldehyde and diacetyl concentration on sGC activity (A) and reversibility of inhibition (B). A, Purified bovine lung sGC (50 ng) was stimulated with DEA/NO (0.05 μM) in the presence of DMSO (0.1% as solvent control) or increasing concentrations of cinnamaldehyde and diacetyl (0.01–3 mM each). B, The enzyme (50 ng) was pre-incubated with DMSO (0.1% as solvent control) or flavorings (1 mM each) for 5 minutes and subsequently diluted 50-fold in reaction mixture, followed by determination for enzyme activity by incubation for 10 minutes in the presence of 0.05 μM DEA/NO in the absence or presence of the flavorings (1 mM each). Data are expressed as mean values±SEM (n = 3). Note that the error bars in panel A are smaller than the symbols. *p<0.05 vs. controls in the absence of cinnamaldehyde (ANOVA with Dunnett's post hoc test).
The data with purified sGC indicated that high concentrations of cinnamaldehyde and diacetyl may scavenge NO. This was confirmed by measuring the concentration-time profile of NO released from DEA/NO in the absence and presence of the flavoring compounds with a NO-sensitive electrode. As shown in Fig 5, at 10 mM both compounds led to a pronounced decrease of the recorded NO signals, while only moderate effects were observed at 1 mM. Thus, we cannot exclude that direct effects on sGC contribute to reduced cGMP formation in the presence of the flavorings. Scavenging of NO was not prevented by superoxide dismutase, indicating that it was due to a direct reaction with NO rather than generation of superoxide anions and peroxynitrite formation.
Inactivation of NO by cinnamaldehyde (A) and diacetyl (B). NO formation was induced by adding DEA/NO (final concentration 1 μM) to 100 mM phosphate buffer, pH 7.4. At time points indicated by arrows, cinnamaldehyde or diacetyl was added. Experiments performed in the presence of SOD (2500 U/ml) are marked with open circles. Original traces each representative for 3 experiments are shown.
The effects of the flavoring compounds on blood vessel function were tested in organ bath experiments with isolated rat aortic rings. As shown in Fig 6, the compounds (0.1–0.3 mM) had no significant effects (p>0.05, ANOVA) on vascular relaxation to acetylcholine (A) or DEA/NO (B) that occurred with EC50 values of 55 (32–94) nM and 43 (31–61) nM, respectively. The Emax and EC50 values obtained with the various compounds are given in S2 Table.
Effects of flavorings on relaxation of aortic rings to acetylcholine (A) or DEA/NO (B). The concentration-response to acetylcholine (1 nM—10 μM) or DEA/NO (1 nM—10 μM) was measured with U-46619 pre-contracted rings that had been preincubated with the flavoring compounds for 30 min. The compounds were applied at the highest concentrations that did not induce relaxation on their own (0.1 mM for eugenol and cinnamaldehyde, 0.3 mM for all other compounds; see Fig 7). Concentration-response curves obtained with different ring segments from a single animal were averaged and counted as an individual experiment. Data are expressed as mean values±SEM obtained in 6 (vehicle), 5 (cinnamaldehyde, diacetyl) or 3 experiments (all other flavorings).
Fig 7 shows that most of the flavorings caused relaxation of pre-contracted vessels, with EC50 values ranging from about 0.5 mM [eugenol 0.4 (0.3–0.6) mM and cinnamaldehyde 0.5 (0.4–0.7) mM] to 1.7–4.2 mM (vanillin, menthol, acetylpyridine, and 2,5-dimethylpyrazine). Diacetyl, isoamyl acetate and eucalyptol had no effects below 10 mM. Neither of these effects were blocked by the NOS inhibitor L-NAME (0.2 mM; see S3 Fig, indicating that the effects of the flavorings on vascular smooth muscle tone are not mediated by the endothelial NO synthase/cGMP pathway. The EC50 values obtained with the various compounds in the absence and presence of L-NAME are given in S3 Table.
The concentration-response to the various compounds (1 μM—10 mM) or vehicle (DMSO) was measured with U-46619 precontracted aortic rings. Concentration-response curves obtained with different ring segments from a single animal were averaged and counted as an individual experiment. Data are expressed as mean values±SEM (n = 6).
To elucidate the potential mechanisms of relaxation induced by cinnamaldehyde and eugenol we studied their effects on accumulation of cAMP and cGMP in endothelium-denuded aortic rings. As shown in S4 Table, neither of the two compounds had significant effects on cyclic nucleotide levels at up to 10 mM. However, similarly to verapamil, both flavorings significantly inhibited CaCl2-induced contraction of rat aortas (S2 Fig), indicating interference with Ca2+ entry into vascular smooth muscle cells.
Since Cinnamaldehyde was shown to slowly decompose to yield benzaldehyde upon heating , we tested the potency of benzaldehyde to relax aortic rings. As shown in S4 Fig, benzaldehyde exhibited similar potency as the parent compound, suggesting a common mechanism of action and excluding pronounced vascular effects of thermal degradation of cinnamaldehyde.
Three commercially available liquids were purchased and analyzed for cinnamaldehyde content by HPLC. As shown in Table 3, the three liquids contained 1.5 to 6.9 mM cinnamaldehyde, corresponding to 0.2–0.91 mg/ml. The measured values were strikingly different from the information provided by the manufacturer on the respective data sheets (see Table 3). After independent triple checking of our results, the discrepancy was communicated to the manufacturer but has remained unresolved. Most likely the values provided on the data sheets refer to the concentrated flavors added to nicotine base and not to the cinnamaldehyde concentrations of the finished liquids.
The present results indicate that NO scavenging does not necessarily reflect endothelial dysfunction and impaired vascular tone, as claimed in a previous study . Among the flavoring compounds tested, only cinnamaldehyde and diacetyl reduced NO bioavailability, evident as reduced activity of NO-stimulated purified sGC and scavenging of the NO signal recorded with a NO-sensitive electrode. Lack of effect of superoxide dismutase argues against superoxide-mediated NO inactivation and suggests a direct reaction of the compounds with free NO radical. Our observation that the effects of both compounds were fully reversed upon dilution of the reaction mixtures after pre-incubation of the enzyme with cinnamaldehyde or diacetyl (see Fig 4B) indicate that sGC is not modified by these compounds in an irreversible manner.
Fetterman et al.  reported significant NO scavenging by vanillin, menthol, cinnamaldehyde, and acetylpyridine. Except for cinnamaldehyde neither of these compounds had any effects on the concentration-dependent stimulation of sGC, the most sensitive physiological target of NO. The discrepancy may be related to the use of diaminofluorescein (DAF) fluorescence as an indicator for NO formation in the previous study, a method which can result in artifacts [32–34] due to reactions of the dye with the test compounds.
Endothelium-derived NO is an important modulator of blood vessel function. Damage of the vascular endothelium, resulting in the impaired release of NO and other vasoactive agents, may contribute to hypertension, thrombotic events, and inflammation [35, 36]. However, in some cases direct effects on vascular smooth muscle prevail over reduced NO bioavailability as illustrated by the pronounced relaxation of rat aortic rings to cinnamaldehyde, which showed NO scavenging at similar concentrations. It has been shown that cinnamaldehyde exhibits a wide variety of beneficial effects, including improved glucose handling in type-II diabetes , protection against M. Alzheimer , antibacterial activity  and general anti-oxidant effects that confer protection against natural and artificial toxicants . In addition, it is well established that cinnamaldehyde causes endothelium-independent vasodilation [41–43]. However, the role of the endothelium is controversial, as it was reported that activation of endothelial transient receptor potential ankyrin-1 (TRPA1) contributes to relaxation . Since cinnamaldehyde did not increase endothelial L-citrulline formation in the absence of A23187 (see S1 Fig) but inhibited Ca2+ activation of NOS at higher concentrations, TRPA1-mediated activation of endothelial NO synthesis appears to play no role or may be masked by other processes in porcine aortic endothelial cells. Given the low potency of the aldehyde, we did not further investigate this effect.
Cinnamaldehyde has a high margin of safety and is generally recognized as safe (GRAS) by the Food and Drug Administration. We observed no toxicity by up to 1 mM cinnamaldehyde on endothelial cells, judged by LDH release and MTT reduction. Similarly, Farsalinos et al. found only moderate toxicity of undiluted cinnamon liquid on a cardiomyoblast cell line . However, it has been shown that cinnamaldehyde containing vapor is toxic to lung epithelial cells , indicating that inhalation of aerosols containing high concentrations of cinnamaldehyde, as reported recently , could be harmful to the lungs.
With an EC50 of about 0.5 mM, eugenol was the most potent vasorelaxant among the tested compounds. Eugenol is a natural ingredient of a wide variety of plants, including beans and coffee, and is used in the food industry as preservative and flavoring agent. Eugenol exhibits pronounced anti-inflammatory and anti-oxidant activities (for review see Barboza et al. ). Several previous studies have demonstrated vascular relaxation and reduction of blood pressure by eugenol and eugenol containing essential oils [48–52]. There is general agreement that eugenol acts mainly via inhibition of Ca2+ influx, but some studies reported on partial inhibition of vasodilation by NOS inhibitors, indicating that endothelium-derived NO might contribute to relaxation [48, 51]. However, in agreement with other studies [50, 52], we observed no effect of the NOS inhibitor L-NAME on eugenol-induced relaxation (see S3 Fig), and at 1 mM neither eugenol nor cinnamaldehyde increased NOS-catalyzed L-citrulline formation in cultured endothelial cells (see S1 Fig) but blocked Ca2+-induced contraction of endothelium-denuded blood vessels (see S2 Fig). The reason for the conflicting results is unknown.
Vanillin, menthol, acetylpyridine, and 2,5-dimethylpyrazine caused vasorelaxation with EC50 values in the range of 2 to 5 mM. Inhibition of Ca2+ influx appears to mediate vasodilation to vanillin  and menthol [54–56], the vascular effects of acetylpyridine and 2,5-dimethylpyrazine have not been reported so far. Relaxation observed with 10 mM isoamyl acetate and eucalyptol is probably unspecific and considered as irrelevant.
Previous in vitro studies indicate that some refill liquids of e-cigs may be cytotoxic to cultured endothelial cells, but the role of flavors has not been addressed [57–59]. Of note, since harmful effects on blood vessels would occur independently of the route of administration, toxicity to endothelial cells would question the safety of compounds used for decades as food flavors. The present results do not support the claim of Fetterman et al.  that flavoring compounds used in refill liquids of e-cigs impair blood vessel function. If having any in vivo effects at all, some of these compounds may be vasoprotective by reducing vascular tone and, according to previous studies, inhibition of inflammatory processes. To judge the biological relevance of the observed effects we estimated the plasma concentration that might occur upon inhalation of aerosols containing cinnamaldehyde, a compound with a known pharmacokinetic profile in rats. The highest concentration of this compound measured in three selected, commercially available cinnamon-related liquids was 6.9 mM. Assuming a plasma volume of 3 liters, 100% bioavailability and lack of distribution, metabolism or clearance of cinnamaldehyde, consumption of 10 ml of liquid (69 μmol) would result in a plasma concentration of 23 μM, a value that is below the range of active concentrations in our experiments. Human pharmacokinetic data are not available, but according to a study with rats , these assumptions are highly unrealistic and the expected plasma concentrations will be much lower. Upon intravenous application, cinnamaldehyde was rapidly and efficiently metabolized to cinnamic acid, which was unstable itself and exhibited a biological half-life of about 7 minutes. Intravenous bolus application of 25 mg/kg (corresponding to 1.5 g or 11.4 mmol for a human with 60 kg body weight) resulted in a maximal plasma concentration of 20 μM . In a more recent study, oral administration of 500 mg/kg to rats (corresponding to 30 g or 228 mmol for a human with 60 kg body weight), resulted in a maximal plasma concentration of about 2 μM cinnamaldehyde . Since consumption of flavoring agents by vaping is distributed over several hours, the effective plasma concentration is expected to remain in the nanomolar range even upon heavy vaping. Similar considerations may apply to eugenol and the other, even less potent compounds tested.
Limitations of the study
The present results obtained in vitro with purified sGC, cultured endothelial cells and isolated blood vessels do not necessarily reflect how vascular NO signaling is affected by the flavoring agents in vivo. It is conceivable that interaction of the compounds with NO/cGMP signaling in vascular smooth muscle cells differs from the effects observed in endothelial cells.
In addition, vaporization of liquids in e-cigs may cause thermal degradation of some compounds, resulting in formation of toxic decomposition products. Although the data shown in S4 Fig argue against a considerable effect of cinnamaldehyde decomposition to benzaldehyde, we cannot exclude harmful effects of unknown degradation products that were not studied.
Finally, it should be mentioned that we tested for single, acute effects and did not consider potential consequences of chronic inhalation. Besides the tested flavorings, relatively large amounts of propylene glycol and glycerol are inhaled by users. Although inhalation of these agents is generally considered as safe, longterm adverse effects cannot be excluded.
Our data indicate that flavorings typically present in e-cig refill liquids do not cause endothelial dysfunction that would result in impaired vasodilation upon acute exposure. In contrast, most of the tested compounds caused endothelium-independent vasorelaxation, albeit at fairly high concentrations that appear to exceed by far the plasma concentrations expected to occur upon vaping flavored liquids.
S4 Table. Effects of cinnamaldehyde and eugenol on cAMP and cGMP levels in endothelium-denuded rat aortic rings.
S1 Fig. Lack of direct effects of eugenol and cinnamaldehyde on endothelial L-citrulline formation.
S2 Fig. Effects of cinnamaldehyde and eugenol on CaCl2-induced contractile response in endothelium-denuded rat aortic rings.
S3 Fig. Lack of effect of the NOS inhibitor NG-nitro-L-arginine methyl ester (L NAME) on vasodilation induced by flavorings.
We thank Evamaria Imp for excellent technical assistance. This work was supported by the Austrian Science Fund [P24946].
- 1. Farsalinos KE, Polosa R. Safety evaluation and risk assessment of electronic cigarettes as tobacco cigarette substitutes: A systematic review. Ther Adv Drug Saf. 2014;5(2):67–86. pmid:25083263
- 2. Farsalinos KE, Voudris V, Poulas K. E-cigarettes generate high levels of aldehydes only in 'dry puff' conditions. Addiction. 2015;110:1352–6. pmid:25996087
- 3. LeBouf RF, Burns DA, Ranpara A, Attfield K, Zwack L, Stefaniak AB. Headspace analysis for screening of volatile organic compound profiles of electronic juice bulk material. Anal Bioanal Chem. 2018;410(23):5951–60. pmid:29974153
- 4. O’Connell G, Colard S, Cahours X, Pritchard JD. An assessment of indoor air quality before, during and after unrestricted use of E-cigarettes in a small room. Int J Environ Res Public Health. 2015;12(5):4889–907. pmid:25955526
- 5. van Drooge BL, Marco E, Perez N, Grimalt JO. Influence of electronic cigarette vaping on the composition of indoor organic pollutants, particles, and exhaled breath of bystanders. Environ Sci Pollut Res. 2018:[Epub ahead of print]. pmid:30560536
- 6. Martuzevicius D, Prasauskas T, Setyan A, O’Connell G, Cahours X, Julien R, et al. Characterisation of the spatial and temporal dispersion differences between exhaled e-cigarette mist and cigarette smoke. Nicotine Tob Res. 2018:[Epub ahead of print].
- 7. Wei B, Goniewicz ML, O’Connor RJ, Travers MJ, Hyland AJ. Urinary metabolite levels of flame retardants in electronic cigarette users: A study using the data from NHANES 2013–2014. Int J Environ Res Public Health. 2018;15(2):E201. pmid:29370113
- 8. Walele T, Bush J, Koch A, Savioz R, Martin C, O'Connell G. Evaluation of the safety profile of an electronic vapour product used for two years by smokers in a real-life setting. Regul Toxicol Pharmacol. 2018;92:226–38. pmid:29248487
- 9. Polosa R, Morjaria JB, Caponnetto P, Caruso M, Campagna D, Amaradio MD, et al. Persisting long term benefits of smoking abstinence and reduction in asthmatic smokers who have switched to electronic cigarettes. Discov Med. 2016;21(114):99–108. pmid:27011045.
- 10. Polosa R, Morjaria JB, Caponnetto P, Prosperini U, Russo C, Pennisi A, et al. Evidence for harm reduction in COPD smokers who switch to electronic cigarettes. Respir Res. 2016;17(1):166. pmid:27986085.
- 11. Kalkhoran S, Glantz SA. E-cigarettes and smoking cessation in real-world and clinical settings: A systematic review and meta-analysis. Lancet Respir Med. 2016;4(2):116–28. pmid:26776875
- 12. Hajek P, Phillips-Waller A, Przulj D, Pesola F, Myers Smith K, Bisal N, et al. A randomized trial of e-cigarettes versus nicotine-replacement therapy. N Engl J Med. 2019:[Epub ahead of print]. Epub 2019/01/31. 30699054. pmid:30699054
- 13. Johnson L, Ma Y, Fisher SL, Ramsey AT, Chen LS, Hartz SM, et al. E-cigarette usage Is associated with increased past-12-month quit attempts and successful smoking cessation in two US population-based surveys. Nicotine Tob Res. 2018:[Epub ahead of print].
- 14. Farsalinos K, Siakas G, Poulas K, Voudris V, Merakou K, Barbouni A. E-cigarette use is strongly associated with recent smoking cessation: an analysis of a representative population sample in Greece. Intern Emerg Med. 2019:[Epub ahead of print]. pmid:30637600
- 15. Mills EJ, Thorlund K, Eapen S, Wu P, Prochaska JJ. Cardiovascular events associated with smoking cessation pharmacotherapies: A network meta-analysis. Circulation. 2014;129(1):28–41. pmid:24323793
- 16. Polosa R, Cibella F, Caponnetto P, Maglia M, Prosperini U, Russo C, et al. Health impact of E-cigarettes: A prospective 3.5-year study of regular daily users who have never smoked. Sci Rep. 2017;7(1):13825. pmid:29150612
- 17. Kapelewski CH, Vandenbergh DJ, Klein LC. Effect of monoamine oxidase inhibition on rewarding effects of nicotine in rodents. Curr Drug Abuse Rev. 2011;4(2):110–21. pmid:21696345
- 18. Etter JF, Eissenberg T. Dependence levels in users of electronic cigarettes, nicotine gums and tobacco cigarettes. Drug Alcohol Depend. 2015;147:68–75. pmid:25561385
- 19. Fagerström K. Determinants of tobacco use and renaming the FTND to the Fagerström test for cigarette dependence. Nicotine Tob Res. 2012;14(1):75–8. pmid:22025545
- 20. Simmons VN, Quinn GP, Harrell PT, Meltzer LR, Correa JB, Unrod M, et al. E-cigarette use in adults: A qualitative study of users’ perceptions and future use intentions. Addict Res Theory. 2016;24(4):313–21. pmid:27725794
- 21. Schneller LM, Bansal-Travers M, Goniewicz ML, McIntosh S, Ossip D, O’Connor RJ. Use of flavored electronic cigarette refill liquids among adults and youth in the US—results from wave 2 of the population assessment of tobacco and health study (2014–2015). PLoS ONE. 2018;13(8):e0202744. pmid:30138412
- 22. Russell C, McKeganey N, Dickson T, Nides M. Changing patterns of first e-cigarette flavor used and current flavors used by 20,836 adult frequent e-cigarette users in the USA. Harm Reduct J. 2018;15(1). pmid:29954412
- 23. Brass DM, Palmer SM. Models of toxicity of diacetyl and alternative diones. Toxicology. 2017;388:15–20. pmid:28232124
- 24. Behar RZ, Luo W, Lin SC, Wang Y, Valle J, Pankow JF, et al. Distribution, quantification and toxicity of cinnamaldehyde in electronic cigarette refill fluids and aerosols. Tob Control. 2016;25(Suppl 2):ii94–ii102. Epub 2016/09/17. pmid:27633763; PubMed Central PMCID: PMC5503843.
- 25. Fetterman JL, Weisbrod RM, Feng B, Bastin R, Tuttle ST, Holbrook M, et al. Flavorings in tobacco products induce endothelial cell dysfunction. Arterioscler Thromb Vasc Biol. 2018;38(7):1607–15. Epub 2018/06/16. pmid:29903732; PubMed Central PMCID: PMC6023725.
- 26. Schmidt K, Mayer B, Kukovetz WR. Effect of calcium on endothelium-derived relaxing factor formation and cGMP levels in endothelial cells. Eur J Pharmacol. 1989;170:157–66. pmid:2559853
- 27. Schmidt K, Werner-Felmayer G, Mayer B, Werner ER. Preferential inhibition of inducible nitric oxide synthase in intact cells by the 4-amino analogue of tetrahydrobiopterin. Eur J Biochem. 1999;259(1–2):25–31. pmid:9914471
- 28. Russwurm M, Koesling D. Purification and characterization of NO-sensitive guanylyl cyclase. Methods Enzymol. 2005;396:492–501. pmid:16291256
- 29. Mayer B, Klatt P, Werner ER, Schmidt K. Kinetics and mechanism of tetrahydrobiopterin-induced oxidation of nitric oxide. J Biol Chem. 1995;270:655–9. pmid:7529763
- 30. Wölkart G, Schrammel A, Koyani CN, Scheruebel S, Zorn-Pauly K, Malle E, et al. Cardioprotective effects of 5-hydroxymethylfurfural mediated by inhibition of L-type Ca2+currents. Br J Pharmacol. 2017;174(20):3640–53. pmid:28768052
- 31. Friedman M, Kozukue N, Harden LA. Cinnamaldehyde content in foods determined by gas chromatography-mass spectrometry. J Agric Food Chem. 2000;48(11):5702–9. Epub 2000/11/23. pmid:11087542.
- 32. Broillet M, Randin O, Chatton J. Photoactivation and calcium sensitivity of the fluorescent NO indicator 4,5-diaminofluorescein (DAF-2): implications for cellular NO imaging. FEBS Lett. 2001;491(3):227–32. Epub 2001/03/10. pmid:11240132.
- 33. Jourd'heuil D. Increased nitric oxide-dependent nitrosylation of 4,5-diaminofluorescein by oxidants: implications for the measurement of intracellular nitric oxide. Free Radic Biol Med. 2002;33(5):676–84. Epub 2002/09/05. pmid:12208354.
- 34. Balcerczyk A, Soszynski M, Bartosz G. On the specificity of 4-amino-5-methylamino-2′,7′- difluorofluorescein as a probe for nitric oxide. Free Rad Biol Med. 2005;39:327–35. pmid:15993331
- 35. Vanhoutte PM, Shimokawa H, Tang EHC, Feletou M. Endothelial dysfunction and vascular disease. Acta Physiologica. 2009;196(2):193–222. pmid:19220204
- 36. Boulanger CM. Highlight on endothelial activation and beyond. Arterioscler Thromb Vasc Biol. 2018;38:e198–e201. pmid:30571176
- 37. Wang F, Pu C, Zhou P, Wang P, Liang D, Wang Q, et al. Cinnamaldehyde prevents endothelial dysfunction induced by high glucose by activating Nrf2. Cell Physiol Biochem. 2015;36(1):315–24. Epub 2015/05/15. pmid:25967970.
- 38. Momtaz S, Hassani S, Khan F, Ziaee M, Abdollahi M. Cinnamon, a promising prospect towards Alzheimer's disease. Pharmacol Res. 2018;130:241–58. Epub 2017/12/21. pmid:29258915.
- 39. Vasconcelos NG, Croda J, Simionatto S. Antibacterial mechanisms of cinnamon and its constituents: A review. Microb Pathog. 2018;120:198–203. Epub 2018/04/28. pmid:29702210.
- 40. Dorri M, Hashemitabar S, Hosseinzadeh H. Cinnamon (Cinnamomum zeylanicum) as an antidote or a protective agent against natural or chemical toxicities: a review. Drug Chem Toxicol. 2018;41(3):338–51. Epub 2018/01/11. PubMed 29319361. pmid:29319361
- 41. Xue YL, Shi HX, Murad F, Bian K. Vasodilatory effects of cinnamaldehyde and its mechanism of action in the rat aorta. Vasc Health Risk Manag. 2011;7:273–80. Epub 2011/05/24. pmid:21603596; PubMed Central PMCID: PMC3096507.
- 42. Raffai G, Kim B, Park S, Khang G, Lee D, Vanhoutte PM. Cinnamaldehyde and cinnamaldehyde-containing micelles induce relaxation of isolated porcine coronary arteries: role of nitric oxide and calcium. Int J Nanomedicine. 2014;9:2557–66. Epub 2014/06/07. pmid:24904214; PubMed Central PMCID: PMC4039418.
- 43. Sun L, Liu LN, Li JC, Lv YZ, Zong SB, Zhou J, et al. The essential oil from the twigs of Cinnamomum cassia Presl inhibits oxytocin-induced uterine contraction in vitro and in vivo. J Ethnopharmacol. 2017;206:107–14. Epub 2017/05/24. pmid:28532683.
- 44. Pozsgai G, Bodkin JV, Graepel R, Bevan S, Andersson DA, Brain SD. Evidence for the pathophysiological relevance of TRPA1 receptors in the cardiovascular system in vivo. Cardiovasc Res. 2010;87(4):760–8. Epub 2010/05/06. pmid:20442136.
- 45. Farsalinos KE, Romagna G, Allifranchini E, Ripamonti E, Bocchietto E, Todeschi S, et al. Comparison of the cytotoxic potential of cigarette smoke and electronic cigarette vapour extract on cultured myocardial cells. Int J Environ Res Public Health. 2013;10(10):5146–62. Epub 2013/10/19. pmid:24135821; PubMed Central PMCID: PMC3823305.
- 46. Omaiye EE, McWhirter KJ, Luo W, Tierney PA, Pankow JF, Talbot P. High concentrations of flavor chemicals are present in electronic cigarette refill fluids. Sci Rep. 2019;9(1):2468. Epub 2019/02/23. pmid:30792477.
- 47. Barboza JN, da Silva Maia Bezerra Filho C, Silva RO, Medeiros JVR, de Sousa DP. An overview on the anti-inflammatory potential and antioxidant profile of eugenol. Oxid Med Cell Longev. 2018;2018:3957262. Epub 2018/11/15. pmid:30425782; PubMed Central PMCID: PMC6217746.
- 48. Damiani CE, Rossoni LV, Vassallo DV. Vasorelaxant effects of eugenol on rat thoracic aorta. Vascul Pharmacol. 2003;40(1):59–66. Epub 2003/03/21. pmid:12646411.
- 49. Interaminense LF, Juca DM, Magalhaes PJ, Leal-Cardoso JH, Duarte GP, Lahlou S. Pharmacological evidence of calcium-channel blockade by essential oil of Ocimum gratissimum and its main constituent, eugenol, in isolated aortic rings from DOCA-salt hypertensive rats. Fundam Clin Pharmacol. 2007;21(5):497–506. Epub 2007/09/18. pmid:17868202.
- 50. Lahlou S, Interaminense LF, Magalhaes PJ, Leal-Cardoso JH, Duarte GP. Cardiovascular effects of eugenol, a phenolic compound present in many plant essential oils, in normotensive rats. J Cardiovasc Pharmacol. 2004;43(2):250–7. Epub 2004/01/13. pmid:14716213.
- 51. Peixoto-Neves D, Wang Q, Leal-Cardoso JH, Rossoni LV, Jaggar JH. Eugenol dilates mesenteric arteries and reduces systemic BP by activating endothelial cell TRPV4 channels. Br J Pharmacol. 2015;172(14):3484–94. Epub 2015/04/03. pmid:25832173; PubMed Central PMCID: PMC4507154.
- 52. Soares PM, Lima RF, de Freitas Pires A, Souza EP, Assreuy AM, Criddle DN. Effects of anethole and structural analogues on the contractility of rat isolated aorta: Involvement of voltage-dependent Ca2+-channels. Life Sci. 2007;81(13):1085–93. Epub 2007/09/18. pmid:17869309.
- 53. Raffai G, Khang G, Vanhoutte PM. Vanillin and vanillin analogs relax porcine coronary and basilar arteries by inhibiting L-type Ca2+ channels. J Pharmacol Exp Ther. 2015;352(1):14–22. Epub 2014/10/26. pmid:25344384.
- 54. Cheang WS, Lam MY, Wong WT, Tian XY, Lau CW, Zhu Z, et al. Menthol relaxes rat aortae, mesenteric and coronary arteries by inhibiting calcium influx. Eur J Pharmacol. 2013;702(1–3):79–84. Epub 2013/02/06. pmid:23380688.
- 55. Johnson CD, Melanaphy D, Purse A, Stokesberry SA, Dickson P, Zholos AV. Transient receptor potential melastatin 8 channel involvement in the regulation of vascular tone. Am J Physiol Heart Circ Physiol. 2009;296(6):H1868–77. Epub 2009/04/14. pmid:19363131; PubMed Central PMCID: PMC2716108.
- 56. Silva DF, de Almeida MM, Chaves CG, Braz AL, Gomes MA, Pinho-da-Silva L, et al. TRPM8 channel activation induced by monoterpenoid rotundifolone underlies mesenteric artery relaxation. PLoS One. 2015;10(11):e0143171. Epub 2015/11/26. pmid:26599698; PubMed Central PMCID: PMC4657920.
- 57. Schweitzer KS, Chen SX, Law S, Van Demark M, Poirier C, Justice MJ, et al. Endothelial disruptive proinflammatory effects of nicotine and e-cigarette vapor exposures. Am J Physiol. 2015;309(2):L175–L87. pmid:25979079
- 58. Putzhammer R, Doppler C, Jakschitz T, Heinz K, Förste J, Danzl K, et al. Vapours of US and EU market leader electronic cigarette brands and liquids are cytotoxic for human vascular endothelial cells. PLoS ONE. 2016;11(6). pmid:27351725
- 59. Carnevale R, Sciarretta S, Violi F, Nocella C, Loffredo L, Perri L, et al. Acute Impact of Tobacco vs Electronic Cigarette Smoking on Oxidative Stress and Vascular Function. Chest. 2016;150:606–6012. pmid:27108682
- 60. Yuan JH, Dieter MP, Bucher JR, Jameson CW. Toxicokinetics of cinnamaldehyde in F344 rats. Food Chem Toxicol. 1992;30(12):997–1004. Epub 1992/12/01. pmid:1473801.
- 61. Zhao H, Xie Y, Yang Q, Cao Y, Tu H, Cao W, et al. Pharmacokinetic study of cinnamaldehyde in rats by GC-MS after oral and intravenous administration. J Pharm Biomed Anal. 2014;89:150–7. Epub 2013/12/03. pmid:24291110.