Long-term cerebrovascular dysfunction in the offspring from maternal electronic cigarette use during pregnancy

Abstract

Electronic cigarettes (E-cigs) have been promoted as harm-free or less risky than smoking, even for women during pregnancy. These claims are made largely on E-cig aerosol having fewer number of toxic chemicals compared with cigarette smoke. Given that even low levels of smoking are found to produce adverse birth outcomes, we sought to test the hypothesis that vaping during pregnancy (with or without nicotine) would not be harm-free and would result in vascular dysfunction that would be evident in offspring during adolescent and/or adult life. Pregnant female Sprague Dawley rats were exposed to E-cig aerosol (1 h/day, 5 days/wk, starting on gestational day 2 until pups were weaned) using e-liquid with 0 mg/mL (E-cig0) or 18 mg/mL nicotine (E-cig18) and compared with ambient air-exposed controls. Body mass at birth and at weaning were not different between groups. Assessment of middle cerebral artery (MCA) reactivity revealed a 51%–56% reduction in endothelial-dependent dilation response to acetylcholine (ACh) for both E-cig0 and E-cig18 in 1-mo, 3-mo (adolescent), and 7-mo-old (adult) offspring (P < 0.05 compared with air, all time points). MCA responses to sodium nitroprusside (SNP) and myogenic tone were not different across groups, suggesting that endothelial-independent responses were not altered. The MCA vasoconstrictor response (5-hydroxytryptamine, 5-HT) was also not different across treatment and age groups. These data demonstrate that maternal vaping during pregnancy is not harm-free and confers significant cerebrovascular health risk/dysfunction to offspring that persists into adult life.

NEW & NOTEWORTHY These data established that vaping electronic cigarettes during pregnancy, with or without nicotine, is not safe and confers significant risk potential to the cerebrovascular health of offspring in early and adult life. A key finding is that vaping without nicotine does not protect offspring from cerebrovascular dysfunction and results in the same level of cerebrovascular dysfunction (compared with maternal vaping with nicotine), indicating that the physical and/or chemical properties from the base solution (other than nicotine) are responsible for the cerebrovascular dysfunction that we observed.

Listen to this article's corresponding podcast at https://ajpheart.podbean.com/e/maternal-vaping-impairs-vascular-function-in-theoffspring/.

INTRODUCTION

Electronic cigarettes (E-cigs) are a new and increasingly popular nicotine delivery system. Proponents for E-cigs suggest they are a healthier alternative to smoking and therefore should be considered a harm reduction tool and an aid to smoking cessation (1, 2), including during pregnancy (3). However, the limited knowledge stemming from chronic exposure, particularly in vulnerable populations like youth and during pregnancy, is a cause for great concern. Moreover, the benefits of vaping for smoking cessation are being questioned as more robust clinical studies are conducted (4–6). The case for harm reduction may also be viewed as somewhat dubious since 1) the majority of current users are young (<25 yr old) who vape for pleasure rather than smoking cessation (7) and 2) the consequence of vaping is increasingly found to affect multiple organ systems (8). For example, immune (9) and platelet function (10, 11) are found to be altered, where vaping can enhance platelet activity resulting in enhanced aggregation and cell-signaling thus increasing the risk for thrombogenic events (10). There is evidence of oxidative stress and epigenetic modifications that occur in response to vaping (12, 13). Furthermore, in both humans and animals, vaping has been shown to increase blood vessel stiffness and induce vascular dysfunction (14–18), with some studies showing similar damage/dysfunction between tobacco and E-cigs (16, 17). Impaired vascular function is notable, as it is widely recognized with advanced aging (19) and is also associated with neurocognitive decline (19–22), development of cerebral microbleeds (23), lower cerebral blood flow (24, 25), dysfunction of resistance arteries (19, 26), Alzheimer’s disease (20, 22, 27), hypertension (28), and greater overall risk for cardiovascular and cerebrovascular disease (29, 30).

It is estimated that almost half of all women who smoke before becoming pregnant will continue smoking during and after pregnancy (31, 32). A growing number of pregnant women who smoke are being encouraged to switch to use E-cigs (3, 33, 34) based on the perception that vaping is “safer” than smoking (1). This is concerning since little is known about the overall health consequences of long-term E-cig usage, and even less in the context of pregnancy. Given that smoking (35–37), and even ambient air pollution (38), are known to trigger adverse birth and adolescent outcomes (e.g., low birth weight, impaired lung and brain development, impaired adolescent learning/neurocognitive performance, and vascular dysfunction), it is critical to understand the potential threat of vaping on vascular development/function during this vulnerable period. Given that vaping is known to induce vascular dysfunction (14–18), we tested the hypothesis that maternal vaping during pregnancy, with or without nicotine, would result in impaired cerebrovascular reactivity in progeny and that the impairment would be evident during early adolescent and adult life.

MATERIALS AND METHODS

Study Design and Exposure System

All procedures were approved by the West Virginia University Animal Care and Use Committee. Male (250–275 g) and female (200–250 g) Sprague–Dawley rats were purchased for breeding (Charles River, Wilmington, MA) and housed in pathogen-free vivarium facility at West Virginia University. They were allowed to acclimate to the new facility for at least 7 days before breeding, provided standard rat chow and tap water, and kept on 12-h:12-h day/night cycle, throughout the study. Estrous was confirmed from vaginal smear, after which male and female rat were housed together (in the morning) for up to 24 h. Evidence of pregnancy and determination of gestational day (GD)-0 was made with observation of sperm and/or vaginal plug on the female. Once pregnant, rat dams were randomly assigned to receive exposure to each of the following groups: 1) E-cig aerosol with no nicotine (E-cig0, n = 5); 2) E-cig aerosol with 18 mg/mL e-liquid nicotine (E-cig18, n = 5); or 3) ambient air (control, n = 5; Table 1). Pregnant dams with the same exposure conditions were housed together, up to 2 per cage, until just before giving birth (i.e., GD20), at which time dams were individually housed. E-liquid used in this study was obtained from a local E-cig distributor (i.e., VapeHut), and we used 75/25 vegetable glycerine (VG)/propylene glycol (PG) composition with French Vanilla flavor with and without nicotine (as noted above).

Table 1.

Dams and exposure conditions

	Air	E-cig0	E-cig18	ANOVA P value	E-cig0 vs. E-cig18 P value
Dams, n	5	5	5
Dam age @ birth, mo	5.8 ± 1.1	5.1 ± 2.1	5.4 ± 1.9	0.87	0.85
Litter size, no. of pups	14 ± 2	12 ± 2	12 ± 4	0.25	0.59
Litter size, range no. of pups	12–16	10–16	9–16
Total particle count,particles/cm3	2.8 ± 0.4 × 104	1.4 ± 0.5 × 1010	1.5 ± 0.3 × 1010	<0.001	0.23
Ave particle concentration, particles/cm3	1.56 ± 0.2	1.36 ± 0.36 × 106	1.47 ± 0.26 × 106	<0.001	0.24
TPM, mg/m3	nd	117 ± 49	134 ± 51	<0.001	0.34
Ambient temperature, °C	20.2 ± 0.2	22.8 ± 0.7	21.2 ± 0.8	<0.001	<0.001
Relative humidity, %	44 ± 14	62 ± 6	57 ± 7	<0.001	0.12

Values are means ± SD. Ave, average; E-cig18, E-cig with 18 mg/mL of nicotine; E-cig0, E-cig without nicotine; nd, not detected; TPM, total particulate matter measured gravimetrically.

Maternal exposure began on gestational day 2 (GD2) using a whole body exposure system (Scireq inExpose, Montreal, QC) and continued (for dams only) until pups were weaned on postnatal day 21 (PD21). We choose to perform exposure only during pregnancy and lactation to solely evaluate the influence of vaping while pregnant and eliminate the potential for any influence that vaping before pregnancy might have. E-cig0 and E-cig18 exposures were performed concurrently using two separate, but identical E-cig devices and exposure chambers that were independently operated and monitored. The rat pups themselves were never placed in the exposure chamber and were never exposed to E-cig aerosol. Offspring were euthanized 1-, 3-, and 7 mo of age and ex vivo vessel function studied using pressure myography. Selection of offspring for each time point was made randomly within each litter, with the exception that we sought (as much as possible based on sex availability) to attain balanced representation of male and female rats at each time point for each group. One to three pups were used from each litter at each time point from each group. See Table 2 for sex distribution and total number of offspring studied in each group.

Table 2.

Offspring anthropometric and organ assessments

	1 Mo				3 Mo				7 Mo				ANOVA	ANCOVA
	Air	E-cig0	E-cig18	ANOVA P=	Air	E-cig0	E-cig18	ANOVA P=	Air	E-cig0	E-cig18	ANOVA P=	Exposure/Age Interaction P=	Exposure/Sex Interaction P=
Data at birth/weaning
Pups in study (n=)	5	11	13		9	9	7		7	7	7
Sex (male, female)	3, 2	6, 5	6, 7		4, 5	5, 4	3, 4		3, 4	2, 5	4, 3
Pup body mass at P21, g	44.9 ± 0.7	43.5 ± 2.2	44.4 ± 2.1	0.91	42.5 ± 2.8	41.2 ± 3.5	40.0 ± 3.4	0.89	37.9 ± 2.0	40.9 ± 3.8	39.5 ± 3.9	0.94	0.97	0.53
Data at sacrifice
Body mass, g	184 ± 23	193 ± 20	280 ± 50	0.59	388 ± 47	395 ± 47	360 ± 33	0.86	519 ± 85	429 ± 70	517 ± 62	0.61	0.74	0.47
Heart mass, g	0.69 ± 0.04	0.73 ± 0.04	0.75 ± 0.10	0.89	1.13 ± 0.10	1.13 ± 0.10	1.20 ± 0.06	0.84	1.07 ± 0.24	1.19 ± 0.10	1.43 ± 0.18	0.37	0.71	0.79
Heart/body mass ratio, mg/mg	4.4 ± 0.2	4.2 ± 0.2	3.7 ± 0.1*	0.03	3.0 ± 0.1	3.1 ± 0.2	3.4 ± 0.1	0.23	2.6 ± 0.3	3.2 ± 0.1*	2.7 ± 0.1	0.09	0.01	0.47
Lung, g	1.04 ± 0.11	1.05 ± 0.10	1.15 ± 0.14	0.79	1.41 ± 0.10	1.49 ± 0.13	1.76 ± 0.15	0.15	1.70 ± 0.05	1.42 ± 0.08	1.72 ± 0.10	0.06	0.52	0.55
Lung/body mass ratio, mg/mg	6.9 ± 1.2	6.2 ± 0.8	5.9 ± 0.3	0.65	3.8 ± 0.3	4.2 ± 0.4	4.9 ± 0.3	0.09	3.9 ± 0.9	4.2 ± 0.2	3.8 ± 0.3	0.87	0.43	0.28
Liver mass, g	7.0 ± 0.6	8.2 ± 0.7	9.0 ± 1.6	0.62	13.2 ± 1.5	12.9 ± 2.0	11.5 ± 1.5	0.78	12.9 ± 3.0	14.1 ± 2.3	16.2 ± 4.2	0.79	0.83	0.89

Values are means ± SE, unless indicated otherwise. ANCOVA, analysis of covariance; E-cig0, E-cig device without nicotine; E-cig18, E-cig device with 18 mg/mL nicotine; P21, postnatal day 21 (i.e., weaning).

*P < 0.05 compared with control (i.e., air) in same age group.

Maternal exposure consisted of 20 puffs, using one puff every 3 min for 60 min, from identical third-generation, tank-style, E-cig devices purchased online (Joyetech eGrip OLED). Atomizers were changed once a week. Our E-cig device was controlled using a custom-made cradle and computer-controlled solenoid (i.e., artificial hand and thumb) to allow precise and reliable activation of the E-cig device (without modification to the E-cig device itself). E-cig puff duration was set to 5 s and with watts set at 17.5 W. An inhalation draw of ∼1 L/min was generated by the computer-controlled exposure system. A continuously bias flow of 5 L/min of air was used throughout the exposure in the chambers.

Aerosol Analysis

The concentration and size of the aerosol particles were analyzed separately using condensation particle counters (CPC Model 3775, TSI Inc.) and an electrical low-pressure impactor (ELPI+, Dekati Ltd), respectively. Assessment of the vape cloud showed a complex but similar distribution pattern between the respective devices/chambers, resulting in a median particle diameter of 0.395 μm and 0.336 μm for E-cig0 and E-cig18, respectively (Fig. 1). Control conditions showed very wide and negligible detection of particles in ambient air (Fig. 1). Total particulate matter (TPM) concentrations were determined by conducting gravimetric filter readings from the exposure chambers during exposures (0.45-µm pore size, 37-mm-diameter polytetrafluoroethylene filters with 1.5-L/min sample flow).

Figure 1.

Representative distribution of particle size obtained using a commercially available third-generation, tank-style, Joyetech eGrip OLED E-cig device without nicotine (E-cig0, B) and E-cig with 18 mg/mL of nicotine (E-cig18, C). Compared with ambient air conditions (A), both E-cig0 and E-cig18 produced >3 orders of magnitude in aerosol concentration; however, similar droplet size distributions were noted for both of the E-cig exposure groups (count median diameter E-cig-0 = 0.395 vs. E-cig18 = 0.336 μm) with devices settings at 5-s puff duration at 17.5 W.

Collection and Baseline Assessments of Middle Cerebral Arteries

At the appropriate study group age, offspring were deeply anesthetized by isoflurane and then euthanized by exsanguination using PBS solution to flush the vascular system of blood via intracardiac (left ventricle) puncture. The brain was removed from the skull and placed in cold physiological salt solution (PSS; 4°C). Both middle cerebral arteries (MCA), which supply ∼50% of the cerebral blood flow (39), were dissected from their origin at the Circle of Willis and placed into an isolated microvessel chamber filled with PSS. Each MCA was subsequently doubly cannulated within a heated chamber (37°C) that allowed the lumen and exterior of the vessel to be perfused and superfused, respectively, with PSS from separate reservoirs. The 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. Any arterial branches were ligated using a single strand teased from 6-0 suture. A video dimension analyzer connected to the arteriograph system was used to measure wall thickness (WT) and lumen diameter (LD) at pressures ranging from 5 to 140 mmHg, in 20-mmHg increments. The first measurement was taken at 5 mmHg because negative pressure is generated at 0 mmHg, causing the vessel to collapse.

Measurements of Vascular Reactivity in Isolated MCA

Following cannulation, MCAs were extended to their in situ length and were equilibrated to ∼70 mmHg to approximate in vivo mean arterial perfusion pressure (40). Following equilibration, the MCA’s dilator reactivity was assessed in response to increasing concentrations of an endothelial-dependent dilator (acetylcholine, ACh; 10⁻⁹ M–10⁻⁴ M), endothelial-independent dilator (sodium nitroprusside, SNP; 10⁻⁹ M–10⁻⁴ M), and a potent cerebrovascular constrictor [5-hydroxytryptamine, 5-HT (serotonin); 10⁻⁹ M–10⁻⁴ M]. MCA responses to ACh were also measured following acute incubation (30 min) with N^G-nitro-l-arginine methyl ester (l-NAME, 10⁻⁴ M; an inhibitor of NO synthase, Sigma Aldrich) and Tempol (10⁻⁴ M), to assess the contributions of nitric oxide (^•NO) and oxidative stress, respectively, in modulating MCA reactivity (41).

Following completion of all procedures, the myogenic and passive responses of the MCA were examined. Vessels were exposed to each pressure point for 5 min before readings were recorded. Pressure inner and outer diameter curves were obtained first in the presence of Ca²⁺ to observe the vessels’ contractile properties and then in Ca²⁺-free PSS to evaluate the vessels’ passive properties.

All calculations of passive arteriolar wall mechanics (used as indicators of structural alterations to the individual microvessel) are based on those used previously (42), with minor modification. Media wall thickness, lumen, and outer diameters (used as indicators of structural alterations to the individual microvessel) were determined as follows: media thickness (WT, μm) = outer diameter − lumen diameter (i.e., OD − LD); media-to-lumen (M:L) ratio = MT/LD; and percentage myogenic tone (percentage tone) 1 − (active OD/passive OD) × 100.

Extracellular Vesicles

Blood was obtained by cardiac puncture at euthanasia, immediately spun at 3,000 rpm, 4°C, for 10 min. After centrifugation, plasma removed in 200-μL aliquots, flash frozen in liquid N₂ and stored at −80°C until processed. To obtain extracellular vesicles (EVs), frozen samples were thawed on ice and spun at 1,500 rpm, 4°C, for 10 min to separate plasma and cell debris. The supernatant was removed and placed in a new tube. Plasma EVs were purified by centrifugation at 16,500 rpm, 4°C for 1 h. EVs pellet was washed with sterile and filtered 1× PBS and suspended in 200 µL 1× PBS with 1 µL taken and diluted in 1 mL of 1× PBS. The diluted sample was immediately visualized with a Malvern Panalytical Nanosight NS300 for particle size and quantity. The remaining plasma was spun at max speed (13,000 rpm) for 2 h at 4°C. The supernatant was removed and 500 µL of 1× PBS was added. The sample was spun for another hour at max speed at 4°C. These samples were later used in the Exo-check antibody array (No. EXORAY200A-4, Thermo Fisher Scientific) to confirm the presence of EVs. Markers for exosomes represented in the kit included CD63, CD81, ALIX, FLOT1, ICAM1, EpCam, ANXA5, and TSG101. GM130, a cis-Golgi marker, was also present to visualize any cellular contaminations within the samples.

Zeta potential (ZP) was used to measure the surface electrostatic potential of EVs as an indicator of surface charge and colloidal stability influenced by surface chemistry. It was conducted by diluting the purified EVs (1:1,000–1:10,000 with pure 1× PBS) to reach a number per frame of 50–500, ideal for Nano Tracking Analysis (NTA) and measured by Zetasizer Nano Z (Zetasizer Nano Z) at WVU core laboratory. Temperature was set at 25°C, and five cycles at high sensitivity settings were performed per measurement for a total of three to five measurements per sample.

Data and Statistical Analyses

All data are presented as means ± SE, except when noted. Normality was evaluated by the Kolmogorov–Smirnov test. Anthropometric assessments were analyzed by analysis of variance (ANOVA) for exposure condition, as well as by analysis of covariance (ANCOVA) for exposure by sex. The vascular reactivity in the MCA was analyzed by repeated-measures two-way ANOVA, with a Tukey’s post hoc test to determine differences between doses of ACh, 5-HT, or SNP. Differences in passive and active mechanical characteristics and descriptive characteristics between groups were assessed using a multifactorial ANOVA with an interaction term (time-by-group), and a Tukey’s post hoc test to determine differences between groups, as appropriate. In all cases, P ≤ 0.05 was taken to reflect statistical significance.

RESULTS

Maternal Exposures

Table 1 provides comparison dams and exposure conditions. Fecundity was not altered, as litter size was not different between exposed (E-cig0 or E-cig18) and control (air) rat dams (Table 1). There were no significant differences in particle counts and aerosol concentration between E-cig0 and E-cig18 exposure chambers, and as expected, a minimal number of particles were detected in ambient air (control) group (Table 1). Average TPM concentration was calculated at 117 ± 49 mg/m³ (±SD) and 134 ± 51 mg/m³ for E-cig0 and E-cig18 chambers, respectively (P = 0.34). Small differences in ambient and chamber temperature (±2.6°C) and relative humidity (±14%) were observed between the exposure groups during the 1-h exposure time period (Table 1). However, neither the time spent nor the magnitude of difference under these conditions would be expected to be biologically relevant or alter outcomes that we have reported.

Table 2 shows anthropometric and organ assessments for rat pups. There were no significant differences in anthropometric measures in rat pups between the exposure groups at birth or at weaning (postnatal day 21, P21; Table 2). Body and organ (heart, lung, liver) were not different within any age or by exposure group (Table 2).

MCA Endothelial Function and Myogenic Tone

Endothelial-dependent dilator (EDD) responses of the MCA to increasing concentrations of ACh are shown in Fig. 2. At 1-, 3-, and 7 mo of age, a significant time-by-group interaction was evident (Fig. 2A), whereby the E-cig0 and E-cig18 rat pups had >50% reduction in maximal MCA EDD compared with control offspring at 1-, 3-, and 7 mo of age (Fig. 2B, deficits range from −51% to −56%, P < 0.01). In contrast, endothelial-independent dilation (EID) response of the MCA to increasing concentrations of SNP did not differ between exposure groups at any age (Fig. 3A). Likewise, the vasoconstrictor response of the MCA to serotonin (5-HT) did not differ between exposure groups at any age (Fig. 3B), suggesting the impaired ACh dilation responses induced by vaping were not altered by EID modulation.

Figure 2.

A: ex vivo pressure myography data examining the endothelial-dependent dilatory (EDD) response of the middle cerebral artery (MCA) to acetylcholine (ACh). The EDD response of the MCA was ∼50% impaired compared with controls (in offspring at 1-, 3-, and 7 mo of age) from in utero exposure to E-cig aerosol with nicotine (18 mg/mL, E-cig18) or without nicotine (0 mg/mL, E-cig0) due to maternal vaping during pregnancy. Controls (Con) are offspring with maternal exposure to ambient air. Mean ± SD. ANOVA for age × exposure interaction for maximal ACh response (10⁻⁴ M) is P < 0.0001. Within each age group, ANOVA main effect for exposure group × drug [conc] where ++P < 0.05 vs. Ecig 18, and ##P < 0.01 vs. Ecig0. B: age-related summary of offspring’s maximal MCA response (ACh 10⁻⁴ dose) shown as % change relative to baseline tone. All data shown are means ± SD; n = 5–13 rat pups/group. Fischer’s post hoc testing for group differences, *P < 0.05 vs. controls (within age group).

Figure 3.

Ex vivo pressure myography data examining the endothelial-independent dilatory (EID) response of the middle cerebral artery (MCA) to sodium nitroprusside (SNP, A–C) and the vasoconstrictor responses to serotonin (5-HT, D–F). No differences were noted in EID or the vasoconstrictor of the MCA in offspring at 1-, 3-, and 7 mo of age with in utero exposure to E-cig aerosol with (18 mg/mL) or without (0 mg) nicotine (E-cig18 and E-cig0, respectively) due to maternal vaping during pregnancy. Controls are offspring with maternal exposure to ambient air. All data shown are means ± SD; n = 5–13 rat pups/group. ANOVA for age × exposure interaction for maximal SNP and 5-HT response (10⁻⁴ M) are P = 0.99 and P = 0.63, respectively. Con, controls.

Acute coincubation of the MCA with a NO inhibitor (l-NAME) revealed that 56%–80% of the ACh EDD response was accounted for by NO (as seen by a 56%, 75%, and 80% reduction in MCA dilation to ACh in control vessels in 1-, 3-, and 7-mo-old offspring, respectively, P < 0.05; Fig. 4). Given that 20%–44% of MCA reactivity was present with acute l-NAME incubation suggests that other non-NO-dependent pathways remain involved. To further explore the role of oxidative stress on MCA dilation, we also acutely coincubated the MCA with Tempol (a SOD mimic; Fig. 4). Tempol was effective in restoring maximal EDD of the MCA to ACh in both E-cig0- and E-cig18-exposed rats (Fig. 4).

Figure 4.

Data showing maximal acetylcholine dose (ACh 10⁻⁴ M) responses using pressure myography on ex vivo middle cerebral arteries (MCA) treated with nitric oxide inhibitor (l-NAME) and superoxide dismutase mimetic (Tempol) in offspring at 1- (A), 3- (B), and 7 mo of age (C) with in utero exposure to E-cig aerosol with nicotine (18 mg/mL, E-cig18) or without nicotine (0 mg/mL, E-cig0) due to maternal vaping during pregnancy. Controls are offspring with maternal exposure to ambient air. ACh data are same as in Fig. 2B, shown here for comparison of individual inhibitor effects. All data shown are means ± SD; n = 5–13 rat pups/group. ANOVA for age × exposure interaction for maximal ACh + l-NAME and ACh + Tempol response (10⁻⁴ M) is P < 0.0001 and P < 0.001, respectively. *P = 0.05, **P < 0.01. l-NAME, N^G-nitro-l-arginine methyl ester.

The active pressure diameter at 60 mmHg (Table 3) indicated that at 1 mo of age, the MCAs from E-cig18 group were larger than the air-controlled rat pups. However, at 3 and 7 mo of age, no differences were evident in active pressure diameter at 60 mmHg between groups. Furthermore, the myogenic tone was similar between groups at all ages (Table 3). The MCA passive ID, OD, WT, and WLR obtained under calcium-free conditions were similar between groups at all ages (Table 3).

Table 3.

MCA active and passive vessel tone (60 mmHg)

	1 Mo			3 Mo			7 Mo
	Air	E-cig0	E-cig18	Air	E-cig0	E-cig18	Air	E-cig0	E-cig18
Active OD, µm	82 ± 4	95 ± 4	109 ± 4*	129 ± 3	132 ± 7	147 ± 7	145 ± 5	157 ± 7	165 ± 8
Myogenic tone, %	48 ± 6	45 ± 4	41 ± 4	34 ± 1	36 ± 5	42 ± 2	32 ± 2	34 ± 3	30 ± 3
Passive ID, µm	116 ± 13	133 ± 9	133 ± 8	131 ± 3	140 ± 10	140 ± 12	137 ± 5	144 ± 8	149 ± 7
Passive OD, µm	140 ± 18	166 ± 9	171 ± 10	174 ± 4	184 ± 16	196 ± 13	194 ± 7	207 ± 11	210 ± 9
Passive WT, µm	12 ± 3	17 ± 2	19 ± 2	22 ± 1	22 ± 3	28 ± 2	28 ± 2	31 ± 2	31 ± 2
WLR	0.19 ± 0.03	0.26 ± 0.03	0.30 ± 0.04	0.33 ± 0.02	0.31 ± 0.03	0.42 ± 0.03	0.42 ± 0.04	0.44 ± 0.03	0.42 ± 0.02

Values are means ± SD. E-cig0, E-cig device without nicotine; E-cig18, E-cig device with 18 mg/mL nicotine; ID, inner diameter; OD, outer diameter; WLR, wall-to-lumen ratio; WT, wall thickness.

*P < 0.05 vs. control (air) within age group.

Extracellular Vesicles

Plasma from 1-mo-old offspring show elevated number and diverging EV population (based on size) in E-cig0 and E-cig18 offspring compared with air controls (Fig. 5). In general, however, E-cig0 maintained consistency in the peak distribution across the age groups but had showed fewer number of EVs at 7 mo compared with the younger 1- and 3-mo-old counterparts (Fig. 5). In contrast, the age groups for E-cig18 animals were less consistent, where offspring showed lower number of EVs at 3 mo, and then greater number of EVs at 7 mo (compared with control).

Figure 5.

Data showing extracellular vesicles (EVs) number and size distribution in plasma obtained from offspring with in utero exposure (due to maternal vaping) with E-cig aerosol with nicotine (18 mg/mL, E-cig18) or without nicotine (0 mg/mL, E-cig0). Controls are offspring with maternal exposure to ambient air.

DISCUSSION

To our knowledge, this is the first report to describe the effects of vaping during pregnancy and lactation on the health and cerebrovascular function in adolescent and adult offspring. The significance of our findings is that 1) vaping during pregnancy confers risk and harm potential to the cerebrovascular health of offspring in adolescent and adult life stages and 2) vaping without nicotine does not protect the offspring from cerebrovascular dysfunction, suggesting that chemicals in the base solution (other than nicotine) induce cerebrovascular dysfunction.

It is important to emphasize that the offspring in this study were never directly exposed to E-cig aerosol. They were only indirectly exposed via maternal vaping. Given that E-cigs have only been widely available since 2007, the long-term consequences of E-cig usage in humans are still unclear, with some suggesting an overall public health benefit with vaping compared with smoking (43, 44). We studied the effects of maternal vaping with and without nicotine, and although it seems improbable that nicotine-dependent humans would vape without nicotine, it is interesting to note (in the context of vascular function) there was no difference in the cerebrovascular dysfunction observed with or without nicotine (Fig. 2). The importance of this finding is that the components of base solution (i.e., VG or PG, and not nicotine) likely account for the vascular dysfunction that we observed.

It is worth noting that maternal vaping without nicotine would have the benefit of eliminating any nicotine-induced harm to the developing fetus. For example, evidence from rats shows that exposing pregnant dams to E-cig aerosol with nicotine (as with smoking cigarettes) leads to a reduction in body mass and size of offspring and decreases in maternal uterine and fetal umbilical blood flow, whereas these outcomes did not occur in offspring born to dams exposed to nicotine-free E-cig aerosol (45). Indeed, the effects of nicotine during pregnancy have been well studied and are known to adversely affect offspring growth and development (35). We, however, did not observe any differences in body mass between our control and E-cigs groups. The most likely explanation for this discrepancy is the relatively low exposure (only 20 puffs within 1-h exposure window each day) and nicotine level in the e-liquid (18 mg/mL) that we used. Whereas Orzabal et al. (45) using a more intense vaping paradigm (1-s puffs every 20 s for a 3 h each day from GD 5 to GD 19) and up to 100 mg/mL nicotine have observed reduced lower birth weights associated with vaping. Additional factors relating to puff topology and choices of flavors may also have an added influence. To better understand the etiology of the harm that is developing and partition effects associated with e-liquid components, it will be important for future studies to have chemical/compound analysis of the aerosol exposure(s) to allow for more direct comparison (both within a study as well as across studies) to any exposure paradigms.

Despite the relatively low exposure to our dams, we observed significant (>50%) vaping-induced cerebrovascular impairment that is typically associated with overt vascular disease (46–48). Whether vaping has the same effect on human progeny is not clear yet. It is worth noting that in some acute exposures, only smoking and vaping with nicotine showed increased arterial stiffness (15, 49). Nonetheless, several studies report that vaping (with or without nicotine) produce similar unfavorable effects on vascular function as seen with smoking (15, 16, 49, 50). Collectively, these data suggest that E-cigs are not likely to be useful approach to reduce the harmful actions of smoking in the context of cardiovascular and cerebrovascular disease.

From a toxicological point of view, there are two principal sources of potential harm from E-cigs. These are: 1) toxic chemical or compounds founded in, or produced from, the vehicle/base solution (such as carbonyls, volatile organic compounds, etc.) and/or 2) the generation of fine particulate matter (PM) in the respirable range (i.e., 2.5 µm or less, PM_2.5). As a nicotine-delivery device, E-cigs (including the one we used) produce almost exclusively particles in the PM_2.5 range (Fig. 1) that are optimal for parenchymal lung deposition and subsequent delivery to the vascular system (51, 52). Our data indicated that ∼10%–20% of the particles produced are in the ultrafine range (PM <0.1 μm; Fig. 1). Ultrafine particles (PM_0.1) are of particular interest, as they can have cardiovascular effects that are independent of their effects in the lung (see review Ref. 53). Although both PM_2.5 and PM_0.1 are recognized in the etiology of vascular dysfunction (51, 54, 55), the mechanisms relating to the developmental origins of vascular dysfunction/disease from maternal exposure are still poorly understood (53). Moreover, the real effects of PM_0.1 are also not clear, as E-cig aerosol will contain mostly droplets of PG and VG, and whose breakdown products are readily soluble compared with solid particles that have been studied from other environmental sources (52, 56).

There is a large body of literature establishing the presence of carcinogenic and toxic compounds in E-cig aerosol (57–62). Although it is true that the number of toxic compounds are fewer in E-cig aerosol than cigarette smoke, it must be emphasized the dose/concentration (for any given compound) is more relevant to harm potential than simply the presence of the compound. Although we did not measure the chemical signatures of our exposure and therefore cannot know which compounds may be responsible for the vascular dysfunction we observed, numerous studies have shown that carbonyls compounds are produced in E-cig aerosol (57, 63–65). Jin et al. (18) have recently reported that formaldehyde (produced by heating VG or PG) can induce vascular endothelial cell dysfunction that is independent of nicotine or flavorants. In addition, chronic E-cig use reported in humans and animals are associated with greater oxidative stress (66–68) and a change in autonomic balance to greater sympathetic predominance (66), which is similar to that observed with smoking (69). Both oxidative stress and greater sympathetic activation are associated with increased risk for cardiovascular events. Indeed, we have previously reported that chronic exposure of mice to either cigarette smoke or E-cig aerosol increase aortic stiffness and impaired endothelial-dependent aortic reactivity (17). Although in vitro studies using e-liquid aerosol extracts have produced varying results, with some suggesting reduced harm in cultured cardiomyocytes (70), other studies show high cytotoxicity and oxidative stress to endothelial cells, epithelial cells, and fibroblasts (71–74) and often show similar responses as smoking. The observation that ACh stimulation with l-NAME (Fig. 4) reduced MCA reactivity in control animals, but did not alter the impaired E-cig0 and E-cig18 responses, indicates that reduced bioavailability of NO has a major influence in the response we report. This finding is consistent with evidence that both smoking (75) and vaping (50) lead to reduced NO bioavailability and likely explains most (but not all) of the cerebrovascular dysfunction that we see in offspring with maternal E-cig exposure (Fig. 4).

We saw no exposure-related differences in cerebrovascular responses to SNP (Fig. 3) demonstrating that the vascular smooth muscle response was not impaired and that the MCA is capable of fully dilating if NO is provided via endothelial-independent sources. However, the fact that ACh with l-NAME did not completely abolish MCA reactivity indicates that the impairment we observed can only be partly explained by reduced NO bioavailability (Fig. 4). When examining the cerebral vessels response to ACh in the presence of l-NAME, we only found a difference in E-cig18 compared with air in the 1-mo-old group. This could suggest an effect of nicotine that was only present in early life, as this difference between the E-cig0 and E-cig18 did not persist in the older (3- and 7-mo-old) offspring. An opposite effect was seen for Tempol (a stable synthetic compound that mimics the superoxide dismutase) in 1-mo offspring, where E-cig0 response was not fully rescued, as was seen for E-cig18 (Fig. 4). The fact that responses to l-NAME and Tempol were different between E-cig0 and E-cig18 (at the 1-mo time point) is, perhaps, not surprising and likely suggests that nicotine is acting on different pathways. These differences, however, did not persist in the older (3- and 7-mo-old) offspring and therefore appear to be lost as the offspring ages. This could suggest an effect of nicotine on offspring may only be present in early life, as there was no difference between E-cig0 and E-cig18 in the older (3- and 7-mo-old) offspring. The finding that Tempol fully restored the impaired cerebrovascular reactivity in both E-cig0 and E-cig18 in the older offspring suggests that the mechanisms underpinning vascular dysfunction involve (at least in part) oxidative stress.

The effect of vaping with nicotine and its potential health consequences is likely to be more complex than most current studies are designed to reveal. On one hand, based on our MCA data (Fig. 2), it may be tempting to conclude that nicotine is not the principal effector in the context of cerebrovascular dysfunction. This is consistent with a broad literature base that says nicotine alone (at least via transdermal application) does not increase the risk of cardiovascular events (76–79). Evidence from in vivo animal embryo models also find minimal effects of nicotine-only exposure, but greater incidence and severity of cardiac defects with exposure to either E-cig aerosol or cigarette smoke extracts (80). The same study also reports that in vitro exposure to E-cig and smoking extracts on human embryonic stem cells (hESCs) lowers expression of contractile and transcription factors (80). On the other hand, we found fewer changes in EVs in E-cig0 compared with E-cig18 offspring (Fig. 5). This is consistent with recent work in humans suggesting the vaping with nicotine likely elicits cellular stress responses (and release of platelet versus endothelial-derived EVs) that are different than vaping without nicotine (45). EVs are released from many different cell types (particularly endothelial cells) and are believed to represent early indicator of cellular changes that lead to chronic disease, including endothelial dysfunction (81). Although we have not identified the source(s) of EVs in our data, platelets-derived EVs have been reported to influence and effect smooth muscle cells (82). Thus, it is tempting to speculate that given that MCA responses were not different when conditioned to SNP (Fig. 3), it is most likely that the changes we see in EVs stem from endothelial cells.

Relevance to Humans and Study Limitations

It is important to recognize that the comparison of animal and human exposure studies can be confounded by difference in biology and exposure methodology. Although no animal model perfectly recapitulates the human conditions, there are decades of research that reliably show that tobacco smoke exposure (and more generally exposure to airborne PM) in rodents induces similar physiological and pathological outcomes seen in humans. Although our vaping exposure was only during pregnancy and lactation and that may not be analogous to human behavior (who likely would have been smoking/vaping before becoming pregnant), we sought to determine whether vaping solely during the most vulnerable period of growth and development would have adverse effects. Although not examined in our study, vaping before becoming pregnant and continued vaping during pregnancy/lactation would reasonably be expected to have the same, or perhaps worse, outcomes for offspring. The fact that we see such robust adverse cerebrovascular effects within this narrow fetal exposure window should raise alarm and concern and serve as harbinger for humans.

An additional concern could arise from the stress that accompanied the daily intermittent separation that pups experienced when dams were removed for vaping exposure each day (∼1 h/day, 5 days/wk) until weaning. It is well recognized that maternal separation from neonatal pups each day during the first weeks in life can influence developmental programming, vascular function, and alter neurobehavioral outcomes (83–86). A recent maternal separation (3 h/day from PND 2 to PND 14) study, using wire myography, has also observed that mesenteric artery function was impaired in rat pups at PND 21 compared with nonseparated controls (87). However, when mesenteric artery reactivity was studied just 2 wk later (at PND 35), or as an adult (5–7 mo), the vessel responses were not different than controls (87). Collectively these studies clearly establish that early life (e.g., postnatal) stress events can influence vascular function; thus, we cannot exclude the possibility that maternal separation in our study may not have had some influence in the MCA responses we report. But it may also be that (at least) some of influences on vessel reactivity may be short lived, as was observed for mesenteric arteries at PND 21 but not beyond PND 35. If the same is true for other vessels (such as the MCA), then we might expect minimal, if any, influence of maternal separation on the 1-, 3-, and 7-mo-old rats we studied.

In our exposure paradigm, we used whole body exposure system that resulted in the 117–134 mg/m³ over 1-h period. This may by criticized for not being strictly identical to the intermittent pattern of human smoking/vaping behavior. However, there are several issues to note. First, our exposure did produce an intermittent exposure pattern. We used a 5-s puff duration to generate the E-cig aerosol once every 3 min. Given that the total volume of the exposure chamber is 5.5 L and that we had bias flow (of 5.0 L/min) of fresh air continually being drawn into the chamber, the resulting steady state time constant is ∼3.6 (meaning that between each E-cig puff ∼95%–98% of the chamber volume was flushed with clean air). The result of this is that our animals received exposure to the aerosol particles in a pulsatile (or intermittent) fashion, similar to humans, with the opportunity to inhale numerous clean air breaths between subsequent E-cig puffs. Second, rodents are obligate nasal breathers, whereas adult humans are not. The consequence of this is not insignificant. In rodents (and humans, if breathing through their nose), we expect ∼80% of the particles inhaled to be filtered via the nasopharynx and upper respiratory airways, with <20% of inhaled particles reaching distal airways (88–90). Given that humans smoke/vape via their mouths (bypassing nasopharynx filtering), this filtering effect on the particles is dramatically less. Humans also take deeper (i.e., larger tidal volume) breaths when smoking/vaping, which increases the bolus number of particles for each breath, which all together result in greater percentage of particle deposition. In essence, if one wants to compare the same aerosol concentration between rodents and humans, the resulting health outcome/consequence in rodents is likely to underestimate the effect on humans. That being said, an average mass concentration with TPM up to 134 mg/m³ is consistent with evidence for a light to moderate smoker (91–93). But it is interesting to note that this mass concentration was achieved with only 20 puffs. When comparing the number of puffs/day to human use, evidence in the literature (94) and from E-cig devices that self-report (95) suggest a median of ∼100 puffs/day (more than five times higher than our exposure to rodents) but with a wide preference range (median absolute deviation, ±72 puffs/day). Given the evidence of vascular dysfunction we report from only 20 puffs, it seems likely our finding may only be the “tip of the iceberg” and likely underestimates the vascular harm potential to the vast majority of humans who vape. The clinical significance of this observation is that there may be no safe level of vape exposure to the user (and potentially even to secondhand exposure) and that even very low levels of vaping during pregnancy (i.e., 20 puffs/day, or less) will have a significant negative impact on the vascular health of offspring exposed in utero.

Although we purposely sought to include both males and females in this work, we did not have equal representation of sexes in all groups and did not have the statistical power to appropriately conduct a separate analysis based on sex, but we can say that qualitatively there is no major differences in responses between male and female offspring in the controls and E-cig18 groups (showing only 1- to 2-µm difference between sexes). However, it should be noted that the 1-mo-old E-cig0 females did have an average of 4 μm greater vessel diameter responses than males, hinting that an underlying sex difference might be present. Thus, we cannot rule out the possibility that sex might not have an influence.

Conclusions

Our data show that subtle, nonlethal, disruptions in maternal/fetal environment can have a significant influence on fetal development and far-reaching consequences later in life, a concept known as “Barker Hypothesis” or Developmental Origins of Health and Disease (96, 97). Vaping, with or without nicotine, during pregnancy produces significant cerebrovascular dysfunction that could (particularly when combined with other potential risk factors, e.g., inborn/genetic vascular anomalies, obesity, diet, sedentary lifestyle, etc.) result in greater risk for cerebrovascular events in offspring with in utero exposure. Our data add to the growing momentum of evidence that counters the notion that E-cigs are “safe,” or even “safer” than cigarettes. These data highlight the need for E-cigs, or for any novel tobacco product, to be evaluated with a more holistic view of health (i.e., beyond just the lungs), and more fully evaluated across multiple organ/biological outcomes before being considered “harm-reduction” and/or recommended by health care systems/providers, particularly during pregnancy.

GRANTS

Funding support was provided by Philip R Dino Innovative Research Grant from the WVU Cancer Institute (I.M.O.), Transition Grant Support from the Office of Research and Graduate Education, WVU Health Sciences Center (I.M.O.), and NIH U54-GM104942-03 (P.D.C.) and American Heart Association 20CSA35320107 (I.M.O., P.D.C., D.D., and E.E.K.). NanoSight NS300 grant: Stroke CoBRE GM109098 and WV-CTSI grant GM103434, ZetaSizer NanoZ grant: Stroke CoBRE GM109098 and WV-CTSI grant GM103434.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.N.B., P.D.C., and I.M.O. conceived and designed research; E.N.B., E.A., L.H., S.R., J.M., E.E.K., A.M., and D.D. performed experiments; E.N.B., P.D.C., I.M.O., E.A., W.T.G., E.E.K., A.M., and D.D. analyzed data; E.N.B., P.D.C., I.M.O., E.E.K., and D.D. interpreted results of experiments; E.N.B., P.D.C., I.M.O., and W.T.G. prepared figures; E.N.B., P.D.C., and I.M.O. drafted manuscript; E.N.B., P.D.C., I.M.O., E.A., W.T.G., A.M., and D.D. edited and revised manuscript; E.N.B., P.D.C., I.M.O., E.A., L.H., J.M., W.T.G., E.E.K., A.M., and D.D. approved final version of manuscript.