Effects of Electronic Cigarette E-liquid and Device Wattage on Vascular Function

Abstract

Electronic cigarettes (e-cigs) are customizable tobacco products that allow users to select e-liquid composition, flavors, and (in some devices) adjust wattage or heat used to generate e-cig aerosol. This study compared vascular outcomes in a conducting vessel (thoracic aorta) and a resistance artery (middle cerebral artery, MCA) in C57Bl/6 mice exposed to e-cig aerosol generated from either pure vegetable glycerin (VG) or pure propylene glycol (PG) over 60-min (Study 1), and separately the effect of using 5- vs. 30-watt settings with an exposure of 100-min (Study 2). In Study 1, aortic endothelial-dependent-dilation (EDD) was only impaired with PG- exposure (p<0.05) compared with air. In the MCA, EDD response was impaired by ~50% in both VG and PG groups compared with air (p<0.05). In Study 2, the aortic EDD responses were not different for either 5- or 30-watt exposed groups compared with air controls; however, in the MCA, both 5- and 30-watt groups were impaired by 32% and 55%, respectively, compared with air controls (p<0.05). These pre-clinical data provide evidence that chronic exposure to aerosol produced by either VG or PG, and regardless of the wattage used, leads to vascular dysfunction at multiple levels within the arterial system. For all exposures, we observed greater impairment of arterial reactivity in a resistance artery (i.e. MCA) compared with the aorta. These data could suggest the smaller arteries may be more sensitive or first to be affected, or that different mechanism(s) for impairment may be involved depending on arterial hierarchy.

INTRODUCTION

The popularity of electronic cigarettes (e-cigs) has increased steadily since being introduced to the global market ~15 years ago. In 2020, the global e-cig market was calculated to be worth $15 billion U.S. dollars and is expected to grow to over 5 times that value by 2028 (Berg et al., 2022). E-cigs, unlike traditional cigarettes, are customizable products that allows users to select e-liquid components (i.e. base solution composition, flavors, and nicotine concentration) and in some cases (i.e. 3^rd generation ‘mod’ devices) the wattage or temperature used to generate the e-cig aerosol. The customizable nature of e-cigs adds to their appeal, but at the same time the consumer has the ability to alter and vary the toxicity potential. A recent example (from 2019) reflects the hospitalizations of >2000 individuals with E-cigarette/Vaping Associated Lung Injury (EVALI) that were linked to e-liquid mixtures (either self-made or purchased from unlicensed dealers) that contained vitamin E acetate as the base solution. These cases resulted in over 60 deaths (Blount et al., 2019; Lozier et al., 2019) and highlight the fact that users’ ability to modify the e-cig e-liquid can be a threat to public health.

Proponents of e-cigs suggest that they are safer, or healthier than cigarettes, but this claim is not often evaluated from a broad health perspective. Indeed, tobacco-industry studies suggest the risk for lung disease (such as emphysema) may be lower with e-cigs than with cigarettes (Kogel et al., 2014; Phillips et al., 2015). However, many human and animal studies that have evaluated the cardio- or cerebro-vascular effects between e-cigs and smoking have shown similar adverse health outcomes and concerns (Conklin et al., 2018; Nabavizadeh et al., 2018; Olfert et al., 2018; Mills et al., 2022; Rao et al., 2022; Wold et al., 2022). Given cardiovascular disease (CVD) is the global leading cause of morbidity and mortality in humans, it seems premature to tout e-cigs as safer than cigarettes, or even as a harm-reduction tobacco product, when the overall impact on longevity and health (especially towards blood vessels) remains in question.

Given that users can customize e-liquid and/or device settings, and given the growing evidence that e-cigs pose adverse risks to vascular health, we investigated the consequences of chronic exposure to the different e-liquid base solutions, as well as varying wattage used to create e-cig aerosol on vascular function. We report the effects on a conduit artery (e.g. aorta) and resistance artery (e.g. middle cerebral artery, MCA); and tested the hypothesis that: 1) the two most common base solutions used in e-liquid (VG or PG) would both negatively affect vascular function; and 2) aerosol produced from higher wattage would result in greater vascular dysfunction compared with that produced at a lower wattage.

MATERIALS AND METHODS

Ethical Approval

All procedures were conducted with approval from West Virginia University Animal Care and Use Committee (IACUC) approval (#1605003053), and to established principles and regulations for animal experimentation (Grundy, 2015). All surgical procedures were terminal and performed under deep anesthesia (inhaled isoflurane) to minimize pain or discomfort.

Study design and exposure system

Male and female C57BL/6J mice (5-month-old) were obtained from Jackson Laboratories (Bar Harbor, ME) and housed in an AAALAC-compliant vivarium facility on the Health Science Campus at West Virginia University. Mice were allowed acclimation to the facility for at least 1-wk prior to exposure and provided standard chow and reverse-osmosis filtered water and kept on 12:12 day:night cycle throughout the study.

We conducted two different evaluations, where (Study 1) the effect of base solution (i.e. VG vs PG, N=19) and (Study 2) the effect of wattage (i.e. 5- vs 30-watt, N=19) was assessed with vascular function and compared with air controls.

Study 1:

Mice were randomly assigned to ambient air control (n=7; 3M, 4F), 100% VG (n=6; 3M, 3F) or 100% PG (n=6; 3M, 3F) groups. E-cig exposure consisted of 20 puffs/d (1 puff every 3 min in 1-h), 5 d/wk, for 4–5 wks using a whole-body exposure system. Each aerosol was generated using Joyetech atomizer (0.5 Ω) with a custom e-cig controller system that allowed precise and reliable activation of the e-cig device. Puff duration was set to 5s with an inhalation draw rate of ~1 LPM. Wattage was set at 17.5w, and a 20 L/min bias flow of air was used to continuously flush the chamber throughout the exposure which provided an episodic exposure pattern.

E-liquid composition was either VG (glycerin #2143–01, Avantor, J.T. Baker) or PG (propylene glycol, Fisher #P355–1) with no nicotine and no flavoring. Temperature and humidity were measured with probes inside the chamber. Air groups received transport and handling at the same time and manner as e-cig exposed groups to replicate any associated transport or handling stress. The air group remained in the same room as the e-cig groups but were only exposed to ambient air. Real-time aerosol concentration was measured using condensation particle counters (CPC Model 3775, TSI Inc.).

At least 2 d after the final exposure, mice were euthanized (via exsanguination under deep anesthesia - described below) followed by organ removal to collect thoracic aorta and MCA vessel for ex-vivo studies. Aortas were assessed with wire myography and MCAs were assessed using pressure myography.

Study 2:

Mice were randomly assigned to ambient air control (n=7; 3M, 4F), e-cig at 5-watt (n=6; 3M, 3F) or e-cig at 30-watt (n=6; 3M, 3F) groups. The exposure consisted of 100 puffs/d (1 puff every 1 min over 100 min), 5 d/wk for 8 wks using a whole-body exposure system. E-cig aerosol was generated with the same e-cig device and custom e-cig controller system in Study 1, except the Joyetech atomizer was different for 5- vs 30-watt, where 1.2 Ω atomizer was used for 5w, and 0.5 Ω for 30w, and E-liquid composition was 50:50 VG:PG with no nicotine and no flavoring. Puff duration was also set at 5 s with an inhalation draw rate of ~1 LPM with and a 20 L/min bias flow of air was used to continuously flush the chamber throughout the exposure which provided an episodic exposure pattern.

The procedures for air group handling, chamber measurements, and post-exposure processing (i.e. euthanasia, surgical procedure, tissue collection, etc.) were otherwise identical to Study 1, except in this study we incorporated a real-time assessment of volatile chemicals using a photoionization detector (PID: MiniRAE3000+, RAE Systems by Honeywell, San Jose, CA) equipped with a 11.7 eV lamp. The specification was considered to assess a broad range of volatile organic chemicals (VOCs) along with, but not limited to, reactive aldehydes present in the e-cig emission. The PID monitor was calibrated prior to each use with 100 ppm isobutylene, and a correction factor of 1.6 was applied for all measurements.

Measurements of Vascular Reactivity

Mice were anesthetized with 2–5% isoflurane inhalation and euthanized by exsanguination (via cardiac puncture) by flushing the vascular system with phosphate buffer saline (PBS) solution. Immediately thereafter, the head was removed to allow careful extraction of the brain and isolation of the MCA for pressure myography. At same the time the thoracic aorta was surgically excised and processed for wire myography.

Wire myography (Aorta):

Thoracic aorta was isolated, cleaned of perivascular adipose tissue and cross-sectioned into 2–3 mm rings and mounted on the pins of a myobath chamber for wire myography (DMT system, AD Instruments, CO, USA). Myobath contained 5mL of Krebs-Henseleit buffer solution (pH = 7.38–7.42) which was warmed to 37°C and aerated with 95% O₂ and 5% CO₂. Vessels were equilibrated for 10 min at zero tension and then stretched incrementally by 0.25g to a resting tension of ~1.0g. One pin was connected to an isometric strain gauge and the other was connected to a micrometer. Tension signals were recorded by PC with LabChart software.

After a 30- min acclimation period at resting tension, KCl (60 mM) is added to the myobath to examine the viability and contractility of the aortic ring. Thereafter, three consecutive washes with 5 mL of Krebs Henseleit occurred between assessments of reactivity to vasoactive drugs. To assess endothelial-dependent dilation (EDD), vessels were pre-contracted with thromboxane A2 agonist U46619 (7.5 × 10⁻⁹ M) to develop vascular tone, and then change in tension (g) was recorded to increasing concentrations of methacholine (MCh, 10⁻⁹ to 10⁻⁵ M). To assess endothelial-independent dilation (EID), changes in tension were assessed with increasing concentrations of SNP (10⁻⁹ to 10⁻⁵). Vasorelaxation responses for both EDD and EID are reported as a percentage change from baseline:

$$\%\text{relaxation}=\left(z-x/z-y\right)\text{x}100,$$ Eq1.

where z = tension after U46619 (7.5 × 10⁻⁹ M), x = tension after a given concentration of MCh or SNP, and y = baseline tension.

Pressure myography (MCA):

The MCA was separated from Circle of Willis, cleaned, and placed into a microvessel chamber (Living Systems, Scintica Inc.) filled with cold physiological salt solution (PSS; 4°C). Both ends of the isolated MCA were cannulated within a heated chamber (37°C) that allowed the lumen and exterior of the vessel to be perfused and superfused, respectively, with heated PSS from separate reservoirs. When excising the MCA special attention was made in selecting and assessing MCA segments more proximal to the Circle of Willis and areas without 2^nd order branching nearby. PSS was equilibrated with a 21% O₂, 5% CO₂, and 74% N₂ gas mixture and had the following composition (mM): 119 NaCl, 4.7 KCl, 1.17 MgSO₄, 1.6 CaCl₂, 1.18 NaH₂PO₄, 24 NaHCO₃, 0.026 EDTA, and 5.5 glucose.

Following cannulation, MCAs were extended to their in-situ length and equilibrated to ~70 mmHg (Lombard et al., 1999). Then the vessel was challenged using 10⁻⁴M phenylephrine to check viability; thereafter, vessel reactivity was assessed in response to increasing concentrations of an endothelial-dependent dilator (acetylcholine, ACh; 10⁻⁹ to 10⁻⁴M), an endothelial-independent dilator (sodium nitroprusside, SNP; 10⁻⁹ to 10⁻⁴M), and the vasoconstrictor phenylephrine (PE, 10⁻⁹ to 10⁻⁴M). MCA responses to ACh were also measured following acute incubation (30 min) of vessels with nitro-L-arginine methylester (L-NAME, 10⁻⁴M; an inhibitor of nitric oxide (NO) synthase; Sigma Aldrich) to assess contributions of NO (Butcher et al., 2012).

A video dimension analyzer connected to the inverted microscope was used to measure lumen diameter (Table 2). Resistance vessels (such as the MCA) have endogenous smooth muscle tone, therefore pre-constriction prior to assessing dilator responses is not required. Thus, dilation and constriction are assessed by subtracting the starting lumen diameter before each treatment, and chambers were washed between each drug.

Data and Statistical Analyses

All data are presented as mean ± SE, except when noted. Vessel reactivity to changes in drug concentration and exposure were first analyzed by two-way analysis of variance (ANOVA)(Prism 9.3.1, GraphPad, Boston, MA). If main or interaction effects were observed, post-hoc (Bonferroni) test was used to determine differences between conditions (i.e. time and exposure group). Maximal responses were also independently analyzed using one-way analysis of variance (ANOVA) to evaluate exposure group differences. Tukey post-hoc test was used to determine differences between groups, when appropriate. In all cases, p ≤ 0.05 was taken to reflect statistical significance.

RESULTS

Study 1: E-liquid composition.

Chamber conditions are reported in Table 1. Although statistical differences in temperature were seen between the groups, the differences in absolute environmental temperature (up to 1.5 °C) are not expected to be clinically or biologically relevant. Humidity was lower in e-cig exposed groups since air supplied to the exposure system is filtered and dehumidified (Table 1). As expected, the number of particles produced by e-cigs was significantly greater in the VG and PG groups compared with air (P<0.001). When comparing VG and PG with each other, VG produced more particles (i.e. greater aerosol density) than PG (p<0.05). At sacrifice, we recorded the age and body (and organ, i.e. heart and lung) mass of the mice. Normalized lung (to body) mass was not significantly different between groups, but normalized heart mass was greater in VG, and lower in PG, compared with air (p<0.05, Table 1).

Table 1.

Exposure and Animal Characteristics for E-liquid exposure (Study 1)

	Air	100% VG	100% PG	ANOVA p-value
Chamber Conditions
Temperature (C)	22.5 ± 1.2	21.4 ± 0.7*	22.9 ± 0.8* ┼	<0.001
Relative Humidity (%)	72 ± 6	23 ± 1*	22 ± 1*	<0.001
CPC Particle Total (#/cc)	2.6 × 104	20.6 × 109 *	7.5 × 109* ┼	<0.001
CPC Particle per puff (#/cc)	1.2 × 103	1.0 × 109 *	0.38 × 109 * ┼	<0.001
VOC (ppm) per puff	na	na	na	-
Mice
n=	7	6	6
Age (months)	4.9 ± 0.1	4.8 ± 0.1	4.9 ± 0.1	0.15
Body Mass (g)	27.2 ± 6.4	24.0 ± 4.1	25.5 ± 6.7	0.63
Change in Body Mass (g)	2.3 ± 1.5	1.6 ± 0.71	1.5 ± 1.2	0.38
Heart (g) / Body mass (g) %	0.56 ± 0.03	0.63 ± 0.12*	0.46 ± 0.02* ┼	<0.01
Lung (g) / Body mass (g) %	0.92 ± 0.22	0.85 ± 0.19	0.71 ± 0.08	0.32
Aorta
EC50 - MCh (M)	1.7 ± 0.7 (×10−7)	1.6 ± 0.7 (×10−7)	2.1 ± 1.6 (×10−7)	0.53
EC50 - SNP (M)	2.3 ± 1.1 (×10−8)	2.1 ± 1.2 (×10−8)	2.8 ± 0.9 (×10−8)	0.22
MCA
EC50 - ACh (M)	4.2 ± 2.0 (×10−7)	3.5 ± 2.5 (×10−7)	2.4 ± 1.7 (×10−7)	0.44
EC50 - SNP (M)	3.5 ± 1.7 (×10−7)	3.6 ± 1.2 (×10−7)	4.4 ± 3.2 (×10−7)	0.76
MCA resting diameter (μm)	62 ± 8	56 ± 11	61 ± 3	0.14
MCA % tone	39 ± 4	31 ± 7	31 ± 5	0.58

Mean ± SD; VG, vegetable glycerin e-liquid; PG, propylene glycol e-liquid;

CPC, Condensation Particle Counter; VOC, Volatile Organic Compound;

MCA, middle cerebral artery; na, not available;

^*different compared with Air, p<0.05;

^┼different than VG.

Aortic EDD response to MCh was impaired in PG-, but not VG-exposed mice (Figure 1A). In PG-exposed mice, maximal response (e.g. MCh 10⁻⁵ M) was reduced by ~22% compared with controls (p<0.05, Figure 1B). Aortic EID responses to SNP were only different at submaximal concentrations for PG (10⁻⁹ to 10⁻⁷), but not VG, compared with air (Figure 1C). However, the EC₅₀ was not significantly different for either MCh or SNP under any exposure conditions (Table 1). Maximal responses for EID in the aorta were not different between e-cig and air groups (Figure 1D).

Figure 1.

Aorta vasodilation responses to increasing concentration of methacholine (MCh, panels A & B) and sodium nitroprusside (SNP, panels C & D) in mice exposed to air (control), 100% vegetable glycerin (VG) or 100% propylene glycol (PG) e-liquid base solution (no nicotine, no flavor). Maximal responses ([10⁻⁵]M) for MCh and SNP are shown in panels B and D, respectively. Mean±SE. ^ signifies PG different than Air.

MCA EDD responses to ACh were significantly blunted in both PG- and VG-exposed groups compared with controls (Figure 2A). For both PG- and VG- exposure groups, the greatest decline in EDD was observed in the maximal concentration of ACh (10⁻⁴ M) leading to a ~57% deficit compared with air controls (Figure 2B). A significant main effect for exposure was seen with EID responses to SNP in the MCA, but post-hoc testing did not reveal differences with either PG- or VG-exposed groups compared with air (Figure 2C&D). MCA vasoconstrictor responses to increasing concentration of PE were not different (data not shown). Shown in Table 1, we report EC₅₀ for ACh and SNP, and also vessel starting characteristics (i.e. resting diameters and starting vessel percent tone relative maximal response) showing no underlying functional differences in the starting point of the vessels.

Figure 2.

Middle cerebral artery (MCA) vasodilation responses to increasing concentration of acetylcholine (ACh, panels A & B) and sodium nitroprusside (SNP, panels C & D) in mice exposed to air (control), 100% vegetable glycerin (VG) or 100% propylene glycol (PG) e-liquid base solution (no nicotine, no flavor). Maximal responses ([10⁻⁴]M) for MCh and SNP are shown in panels B and D, respectively. Mean±SE. * signifies VG different than Air, p<0.05; ^ signifies PG different than Air, p<0.05.

Study 2: Wattage

Chamber conditions are reported in Table 2. No significant differences in temperature were observed between exposure groups. Humidity was lower in e-cig exposed groups since air supplied to the exposure system is filtered and dehumidified (Table 2). As expected, the CPC was significantly increased in e-cig aerosol compared with air (P<0.001). The presence of volatile organic compounds (VOCs) was also measured and found to be greater at 30w compared with controls (p<0.05). For 5w, VOCs were present and were lower than 30w exposure (p<0.05) but were not statistically different compared with controls (Table 2). Normalized heart (to body) mass was not significantly different between groups, but normalized lung mass was lower in both 5w and 30w groups compared with air (p<0.05, Table 2). An explanation for why this lung mass might be lower in the context of this e-cig exposure is not presently clear.

Table 2.

Exposure and Animal Characteristics for e-cig wattage exposure (Study 2)

	Air	5w	30w	ANOVA p-value
Chamber Conditions
Temperature (C)	20.6 ± 1.2	20.5 ± 0.9	20.5 ± 1.0	0.97
Relative Humidity (%)	60 ± 8	26 ± 5*	26 ± 6*	<0.001
CPC Particle Total (#/cc)	2.5 × 104	4.3 × 107 *	7.1 × 107*┼	<0.001
CPC Particle per puff (#/cc)	2.5 × 102	4.3 × 105 *	7.1 × 105 *┼	<0.001
VOC (ppm) per puff	0 ± 0	4.7 ± 3.2	11.5 ± 9.7 *	<0.001
VOC range (lowest - highest) ppm	0 – 0	0.6 – 9.8	0.5 – 40.4	n/a
Mice
n=	7	6	6
Age (months)	5.5 ± 0.3	5.7 ± 0.1	5.7 ± 0.1	0.07
Body Mass (g)	26.5 ± 5.6	26.1 ± 5.9	24.7 ± 3.8	0.83
Change in Body Mass (g)	2.5 ± 1.6	2.0 ± 0.9	0.44 ± 0.47	0.11
Heart (g) / Body mass (g) %	0.56 ± 0.03	0.59 ± 0.09	0.50 ± 0.07	0.08
Lung (g) / Body mass (g) %	0.85 ± 0.23	0.61 ± 0.10 *	0.69 ± 0.16 *	<0.01
Aorta
EC50 - MCh (M)	1.6 ± 0.4 (×10−7)	1.6 ± 0.5 (×10−7)	1.6 ± 0.3 (×10−7)	0.98
EC50 - SNP (M)	2.3 ± 0.9 (×10−8)	4.7 ± 5.8 (×10−8)	2.9 ± 0.65 (×10−8)	0.22
MCA
EC50 - ACh (M)	3.8 ± 2.1 (×10−7)	2.7 ± 4.2 (×10−7)	11.3 ± 10.7 (×10−7)	0.07
EC50 - SNP (M)	5.2 ± 6.3 (×10−7)	3.3 ± 2.3 (×10−7)	2.3 ± 1.6 (×10−7)	0.46
MCA resting diameter (μm)	62 ± 8	64 ± 8	66 ± 8	0.58
MCA % tone	39 ± 4	38 ± 6	39 ± 5	0.90

Mean ± SD; CPC, Condensation Particle Counter; VOC, Volatile Organic Compound;

MCA, middle cerebral artery; na, not available;

^*different compared with Air, p<0.05;

^┼different than 5 (w)att.

Aortic EDD responses were not different with either 5w or 30w e-cig groups compared with air controls (Figure 3A&B). A significant main effect for exposure was seen with EID responses to SNP in the aorta, but post-hoc testing revealed only some submaximal concentration conditions in which either 5w or 30w were different from air (Figure 3C). However, EC₅₀ was not different (for either SNP or MCh) between any of the exposure conditions compared with controls (Table 2). Maximal responses for EID in the aorta were not different between e-cig and air groups (Figure 3D).

Figure 3.

Aorta vasodilation responses to increasing concentration of methacholine (MCh, panels A & B) and sodium nitroprusside (SNP, panels C & D) in mice exposed to air (control), e-cig device at 5- or 30-(W)atts. Maximal responses ([10⁻⁵]M) for MCh and SNP are shown in panels B and D, respectively. E-liquid was 50:50 VG:PG with no nicotine and no flavor. Mean±SE. * signifies 5w different than Air, p<0.05; ^ signifies 30w different than Air, p<0.05.

MCA EDD responses to ACh were significantly blunted in 5w and 30w exposed groups (p<0.05) compared with air controls (Figure 4A). For both 5w and 30w exposure groups, the greatest decline in EDD was observed at maximal concentration (ACh 10⁻⁴ M) leading to a 32% and 55% decline, respectively (Figure 4B). A significant main effect for exposure was seen with EID responses to SNP in the MCA, but post-hoc testing only revealed the two highest concentrations (SNP 10⁻⁵ and 10⁻⁴) for the e-cig 5w-exposed group to be different compared with air (Figure 4C). Maximal EID responses in MCA were not statistically different between either e-cig exposure group (5w and 30w) compared with air controls (Figure 4D). EC₅₀ for SNP was not different between exposure conditions or compared with controls (Table 2). MCA vasoconstrictor response to increasing concentrations of PE was not different (data not shown). Shown in Table 2, we report vessel starting characteristics (i.e. resting diameters and starting vessel percent tone relative maximal response) showing no underlying functional differences in the starting point of the vessels.

Figure 4.

Middle cerebral artery (MCA) vasodilation responses to increasing concentration of acetylcholine (ACh, panels A & B) and sodium nitroprusside (SNP, panels C & D) in mice exposed to air (control), or e-cig device operating at either 5- or 30-(W)atts. Maximal responses ([10⁻⁴]M) for MCh and SNP are shown in panels B and D, respectively. E-liquid was 50:50 VG:PG with no nicotine and no flavor. Mean±SE. * signifies 5w different than Air, p<0.05.

Incubation with nitric oxide synthase (NOS) inhibitor L-NAME.

For both studies, we also investigated the bioavailability of NO toward maximal EDD by incubating the Aorta and MCA with L-NAME. Under all exposure conditions (i.e. Studies 1 & 2), the addition of L-NAME did not completely eliminate vessel reactivity, but with ~70% decline in MCA response and ~82% decline in aorta response in the control group. These findings support that impaired NO bioavailability is a major mechanism for EDD dysfunction (Figure 5).

Figure 5.

Maximal dilation response of middle cerebral artery (MCA) to acetylcholine (ACh[10⁻⁴]) (A & C) and aorta to methacholine (MCh [10⁻⁵](B & D) showing the effect of NOS inhibition using L-NAME from Study 1 (E-liquid composition, top panels A & B) and Study 2 (Watts, bottom panels C & D). Mean±SE. *signifies p<0.05 for ACh+LNAME compared to ACh within the same condition. § signifies p<0.05 for the difference in ACh minus ACh+LNAME compared Air.

In Study 1 (e-liquid composition), the decline with L-NAME for VG and PG were similar (41% and 34%, respectively), and the absolute effect of L-NAME on MCA diameter was the same between all groups (i.e. Air+LNAME, VG+LNAME, and PG+LNAME), but the NO-mediated contribution to vasodilation significantly impaired in the e-cig exposure groups compared with air (Figure 5A). The decline in aortic response with L-NAME was similar across all exposure conditions (Air=82%, VG=84% and PG=94%, Figure 5B).

For Study 2 (wattage effect), the absolute effect of L-NAME on MCA diameter was also similar between 5w and 30w e-cig groups compared with air (Figure 5C), but the magnitude of decline was greater for 5w (40% decrease) compared with 30w (22% decrease). The NO-mediated contribution to vasodilation was also significantly impaired in the 5w and 30w groups compared with air (Figure 5C). The decline in aortic response with L-NAME was similar across all exposure conditions (Air=81%, 5w=85%, and 30w=92%, Figure 5D).

DISCUSSION

To our knowledge, this is the first study, in the context of e-cig aerosol exposure, to report arterial assessments at multiple points in the vascular system from the same animals in response to either; varying e-liquid (i.e. VG vs PG) composition or varying watts (5w vs 30w). An important finding from this work is that, regardless of the e-liquid solution or wattage used, using e-cigs leads to evidence of arterial dysfunction in both conducting vessels (e.g. aorta) and smaller resistance arteries (e.g. MCA), but the magnitude of dysfunction is greater in the smaller arteries compared with the aorta. In general, chronic exposure to e-cig aerosol from either pure VG, or PG, induced similar vascular EDD dysfunction in the MCA. However, in the aorta, only PG exposure showed evidence of vascular dysfunction compared with controls. It is also notable that using e-cigs at either wattage level we used (i.e. 5- vs 30-watts) produced significant vascular dysfunction. While there was a trend for higher wattage to have greater EDD dysfunction in MCA vessels, it was not statistically significant. The take-home message is that: 1) the base solution used in e-cigs (e-liquid without any additives, such as flavors or nicotine) is responsible for the etiology of vascular dysfunction; and 2) the vascular impairment induced from even a low wattage (i.e. heating temperature), similar to the pod-style devices (such as JUUL, Elf bar, Puff Bar, etc.) was similar to a higher wattage, meaning that both low and higher wattage settings on the e-cig device results in similar vascular dysfunction.

Endothelial dysfunction is widely recognized in the early stages of CVD and cerebrovascular disease and is well-established with smoking-related pathology (Celermajer et al., 1993; Celermajer et al., 1996; Hankey, 1999; Durazzo et al., 2015; Golbidi et al., 2018). To our knowledge, this is the first report of differential responses in aorta vs MCA with exposure to e-cigs. While the reason for greater MCA vs aortic dysfunction is not known yet, it is tempting to speculate that resistance arteries (since they are smaller, have thinner walls, and are meant to be very reactive) could be more sensitive. Therefore, perhaps the MCA shows evidence of dysfunction before larger vessels (which have thicker walls, are less distensible, and designed to withstand high pulsatile pressures). The fact that we see limited aortic dysfunction from the exposure paradigm and timeline used should not be viewed as the aorta is protected or not impacted by e-cigs. On the contrary, numerous in vivo studies using flow-mediated dilation (FMD) show that e-cig exposure induces rapid endothelial dysfunction in conduit vessels (e.g., brachial artery in humans; femoral artery in rats)(Carnevale et al., 2016; Rao et al., 2020; Nabavizadeh et al., 2022; Rao et al., 2022). Moreover, acute (Jin et al., 2021) and chronic (Olfert et al., 2018) e-cig exposures in mice also show evidence of impaired aortic endothelial function, as well as increased arterial stiffness. One explanation for the difference in this work from other studies may be the timing of when our assessments were made. We have recently reported that it takes 48 to 72 hours for endothelial function to return to normal following a single 1-h e-cig exposure session (Mills et al., 2022). Thus, any vascular reactivity measures made within 48–72 hours of the last exposure may not reflect baseline changes in vascular function, but rather responses from the last exposure. In this study, we waited at least 48 hours from the last exposure before making our assessments to avoid this concern, but this could explain why we see minimal (or no) aortic impairment from this work compared to our prior work and others in the field.

It should also be noted that we used different agonists when assessing aortic vs MCA reactivity (i.e. MCh vs ACh, respectively). While both agonists are widely used in arterial myography experiments and act via ACh receptors (AChRs), MCh is more selective to muscarinic AChRs while ACh binds both muscarinic and nicotinic AChRs. But both act to stimulate vessel relaxation via the NO pathway. This is evident in Fig. 5 where L-NAME significantly impaired relaxation in the responses from both MCh (in the aorta) and ACh (in the MCA). Thus, the general lack of aortic impairment (Figs. 1 & 3) compared to MCA (Figs. 2 & 4) suggest that NO bioavailability was not likely impaired at the aortic level (or at least not during the time frame post-exposure that we studied these vessels). Based on our prior work (Mills et al., 2022; Pitzer et al., 2023), and others (Nabavizadeh et al., 2018; Rao et al., 2022), it is clear that blood vessels react within minutes of inhalation exposure to either cigarette smoke or e-cig aerosol, and ultimately (at least in an acute context) can resolve or recover from the insult. Since we waited at least 2 days after the last exposure, we believe the effects we report most likely represent a change in baseline function and not a residual effect from the last exposure, which suggests the normal function of resistance arteries may be first to be impacted by chronic smoking or e-cig use. What remains unknown, is when (or even if) the transient endothelial dysfunction observed from repeated exposures becomes permanent. Notwithstanding the vital influence of NO throughout the vascular system, other mechanisms can also influence vessel tone and reactivity. Indeed, emerging evidence suggests that e-cigs are likely to trigger multiple mechanisms (Conklin, 2022; Mohammadi et al., 2022). For example, a recent study showed that bilateral vagotomy prevents impairment of FMD in the femoral artery in rats’ exposure to either cigarette smoke, aldehyde gases, or inert carbon particles. These findings indicate the nervous system (likely through airway irritant receptors) plays a vital role in the FMD response (Nabavizadeh et al., 2022). Whether the same protection from bilateral vagotomy occurs in smaller resistance arteries is not currently known, but such studies could help identify mechanistic differences within the arterial hierarchy.

It was interesting to see that exposure to e-liquid consisting of pure PG had a greater impact (compared with VG) on aortic endothelial dysfunction. The reason why PG, and not VG, affected aortic function is not fully understood. But there are some differences in the aerosol produced by PG vs VG. For example, PG (or e-liquid with a higher portion of PG) is reported to produce aerosol with smaller-sized aerosol particles (i.e. greater number of nanosized particles that <100 μm) in contrast to VG (Baassiri et al., 2017; Lechasseur et al., 2019; Stefaniak et al., 2022). Inhalation of nano-sized (i.e. ultrafine) particles is known to have a greater influence on cardiovascular outcomes, which would be consistent with our findings. Another explanation could stem from differences in the concentration of chemical constituents produced from VG and PG thermal degradation. For example, the formation of reactive aldehydes is different as a function of PG and VG, where compounds such as formaldehyde, acetaldehyde, acetone, and propionaldehyde are found to be higher with PG (and less with VG)(Li et al., 2021). Inhaled formaldehyde (5 ppm, 6h/d, for 4 d) has also recently been implicated in the pathophysiology of aortic endothelial dysfunction (Jin et al., 2021). Therefore, it is possible that e-liquid with a higher portion of PG could create or initiate pathologic responses differently than e-liquid with a higher composition of VG. But it should also be noted, other compounds (e.g. acrolein) are found to be higher in VG than PG (Li et al., 2021), and many of the effects from these chemicals toward vascular function remain poorly understood. For example, although we did find greater VOCs with 30w exposures compared with 5w and air controls (Table 2), there was no difference in aortic or MCA EDD responses between 5w and 30w exposures. But, it is also important to recognize, the PID monitor with 11.7 eV lamp provides a broad range of VOCs (including reactive aldehydes) assessment but cannot selectively identify the concentration of specific compounds (e.g., formaldehyde, acetaldehyde, etc.), but rather provides an overall assessment of total concentration from all compounds. So, any interpretation with respect to specific compounds must be viewed with caution. Regardless of the compound or agent responsible, most people use some combination of both PG and VG; and perhaps more importantly, both PG and VG alone created the same level of endothelial dysfunction in the MCA (Fig. 2). Notwithstanding the smaller effect we saw with PG in the aorta (Fig. 1), the overall message from these data suggests e-cigs devices and the constituents used as the base solution plays a significant role in triggering vascular impairment. It is also important to emphasize, even using e-cigs at the lowest possible wattage (i.e. 5-watts) induces significant vascular dysfunction in vital peripheral arteries (e.g. MCA). Although we do not report a significant difference between the 5w and 30w exposed groups, it is evident that there was a trend for greater endothelial dysfunction in the MCA with 30w compared with 5w. It is possible that even higher wattage (e.g.>30w), or even longer exposure to 30w, could result in more severe vascular impairment, but this will require further investigation and additional studies to determine. Nonetheless, these data add to a growing body of literature that suggests that e-cig use induces adverse vascular health risks that are identical to those known to occur with traditional cigarette smoking (Olfert et al., 2018; Wolkart et al., 2019; Jin et al., 2021; Nabavizadeh et al., 2022; Rao et al., 2022).

Human Relevance and Study Limitations

We chose not to include nicotine or flavors in our e-liquid. This could be viewed as a limitation of our study; however, we would contend our findings are relevant to all users and devices that use VG and PG. Our data highlight the fact that the delivery vehicle being used, i.e. the base solution, is likely to be the primary contributor to vascular dysfunction. This does not mean that other components like nicotine or flavors are not important, indeed they may also have their own potential for harm (Fetterman et al., 2018). For nicotine, the literature is replete with data evaluating its effect on health. As a sympathomimetic drug, nicotine has a wide range of effects and is known to produce health effects viewed both positively and negatively. In the current context, it is worth noting there are long-standing epidemiological data (based primarily on studies using nicotine replacement therapies, NRTs) that nicotine is not likely a major contributor to CVD (Joseph, 1993; 1994; Tzivoni et al., 1998; Meine et al., 2005; Mills et al., 2014). With traditional cigarettes it was not possible to separate out nicotine’s influence from the delivery vehicle (i.e. smoke), but with e-cigs it is possible. Recent studies from our lab (Burrage et al., 2021; Mills et al., 2022; Aboaziza et al., 2023; Pitzer et al., 2023) and others (Neunteufl et al., 2002; Rao et al., 2022) find that endothelial vascular dysfunction is similar when e-cigs are used with or without nicotine, suggesting that nicotine is not the source of vascular impairment. Given the similarity in vascular dysfunction seen between e-cig use and smoking, it seems likely the same adverse outcomes seen with cigarette smoke will occur with e-cig use.

Individual flavors, when heated, can and will have varying degrees of toxicity (Wolkart et al., 2019; Kerasioti et al., 2020). This means flavors added to e-liquid have the potential to create their own or even an additive influence on vascular dysfunction. A recent study showed some flavors (such as diacetyl, eucalyptol, and menthol) can directly influence aortic reactivity/contractile force, while other flavors may have no effect (Wolkart et al., 2019). Given that >10,000 flavors are reportedly available to e-cig users, it will be difficult to investigate the full impact that flavors may really have. However, the relative importance of this concern may be overshadowed (at least in the context of vascular health) by the fact our study finds that the base solution is a major (and likely primary) source of e-cig induced vascular dysfunction. This is supported by recent evidence from our lab showing that intravital assessments of vascular responses are similar between cigarette and e-cig exposures, and that the inclusion of a Vanilla-based flavor in e-liquid does not change the level of vascular dysfunction (Pitzer et al., 2023).

Since male and female mice were used in our study, we also evaluated the potential of sex difference in our data. Based on the current data we did not find any sex-based differences, but given the small number of male and female mice we had to evaluate it must be recognized our study was not powered for a robust statistical comparison according to sex.

Conclusions

Although proponents of e-cigs argue that they produce less harm compared with cigarette smoking (Benowitz and Burbank, 2016; Benowitz and Fraiman, 2017), there is growing evidence showing a similarity in vascular dysfunction between smoking cigarettes and using e-cigs that cannot be ignored. Our data show significant artery dysfunction is associated with exposure to the primary e-liquid components used in e-cigs (VG or PG). When evaluating the effect of wattage used to generate an e-cig aerosol, we see even when the device operates at the lowest wattage (i.e. 5-watts) it produces similar vascular dysfunction to an exposure at a higher wattage (i.e. 30w). Thus, regardless of the wattage used, a main finding from this work is that the base solution used in e-cigs (without any additives, such as flavors or nicotine) is responsible in the etiology of vascular dysfunction. Our data also suggest that smaller (resistance) arteries may experience greater impairment in vascular reactivity compared with larger vessels (such as the aorta). At the same time aortic dysfunction is still observed, and most likely the time course for impairment and/or mechanisms may be different. When comparing this work to long-established evidence of endothelial dysfunction resulting from smoking cigarettes in humans and pre-clinical animal models, it seems clear that e-cigs are likely to result in a similar risk for the development of cardiovascular and cerebrovascular dysfunction as compared with cigarettes.

Highlights:

• E-cig exposure to either VG or PG alone triggers arterial vascular dysfunction

• E-cig exposure at even very low wattages results in significant vascular impairment compared to air-exposed controls

• E-cig triggers vascular dysfunction at multiple levels within the arterial system

• Resistance arteries have greater dysfunction compared to conducting vessels

Abbreviations:

e-cigs

Electronic cigarettes

MCA

Middle Cerebral Artery

VG

Vegetable Glycerin

PG

Propylene Glycol

ACh

Acetylcholine

SNP

Sodium Nitroprusside

PE

Phenylephrine

MCh

Methacholine

EDD

Endothelial-dependent dilation

EID

Endothelial-independent dilation

CVD

Cardiovascular disease