In Vitro Consequences of Electronic-Cigarette Flavoring Exposure on the Immature Lung

Background: The developing lung is uniquely susceptible and may be at increased risk of injury with exposure to e-cigarette constituents. We hypothesize that cellular toxicity and airway and vascular responses with exposure to flavored refill solutions may be altered in the immature lung. Methods: Fetal, neonatal, and adult ovine pulmonary artery smooth muscle cells (PASMC) were exposed to popular flavored nicotine-free e-cigarette refill solutions (menthol, strawberry, tobacco, and vanilla) and unflavored solvents: propylene glycol (PG) or vegetable glycerin (VG). Viability was assessed by lactate dehydrogenase assay. Brochodilation and vasoreactivity were determined on isolated ovine bronchial rings (BR) and pulmonary arteries (PA). Results: Neither PG or VG impacted viability of immature or adult cells; however, exposure to menthol and strawberry flavored solutions increased cell death. Neonatal cells were uniquely susceptible to menthol flavoring-induced toxicity, and all four flavorings demonstrated lower lethal doses (LD50) in immature PASMC. Exposure to flavored solutions induced bronchodilation of neonatal BR, while only menthol induced airway relaxation in adults. In contrast, PG/VG and flavored solutions did not impact vasoreactivity with the exception of menthol-induced relaxation of adult PAs. Conclusion: The immature lung is uniquely susceptible to cellular toxicity and altered airway responses with exposure to common flavored e-cigarette solutions.


Introduction
Electronic nicotine delivery devices (ENDS), or electronic cigarettes, have been developed as a safer and less toxic substitute to smoking. Use of these products is growing rapidly, and predictions based on current sales suggest that consumption of e-cigarettes could surpass that of conventional cigarettes within the next decade with estimates of a $50 billion global market by 2025 [1,2]. While established smokers can avoid known tobacco toxicities with e-cigarette use, vulnerable populations may exist who are more susceptible to exposure. Pregnant women as well as parents and caregivers of infants may rationalize a switch to these products to reduce risks for children. Indeed, studies have shown that e-cigarette use is increasing sharply among women of reproductive age and that they are perceived by many as safer alternatives during pregnancy [3][4][5][6]. An online survey of 455 pregnant women recruited from a national website survey service (Amazon Mechanical Turk) identified higher rates of vaping as compared to smoking with 6.5% of respondents reporting sole use of e-cigarettes, 5.6% reporting sole leading to concerns for secondhand exposure of newborn and infants in proximity of caregivers who vape [35][36][37].
Flavoring solutions commonly contain a mixture of numerous chemicals; as an example, analysis of 28 representative e-liquids identified over 140 different flavoring chemicals present in one published panel [38]. While many of these chemicals are commonly ingested with food products and have been certified as "Generally Recognized as Safe" (GRAS) by the Flavoring and Extracts Manufacturers Association (FEMA), this designation is in reference to safety with ingestion or topical exposure only and would not apply to inhalation. Indeed, the bioavailability and the consequence of exposure with vaping of flavoring chemicals deemed as safe for ingestion remain unknown.
With growing appreciation for toxicity of flavoring chemicals [30,32,[39][40][41][42] and a specific concern for the consequences of exposure on the vulnerable developing lung, we sought to determine the relative toxicity and the impact on physiologic responses in immature lungs cells and tissue exposed to popular nicotine-free e-cigarette flavored solutions [34,43]. Unique access to fetal, newborn, and adult ovine pulmonary cells lines as well isolated neonatal and adult ovine lung tissue allowed evaluation of developmental susceptibility. We hypothesized that the fetal and the neonatal lung would be more vulnerable to flavoring solution-induced toxicities and that airway and pulmonary vascular response with exposure may also be developmentally regulated.

Flavored E-Cigarette Refill Solutions
Convenience samples of flavored solutions used for these studies were labeled as made by Ecto World (https://www.ectoworld.com) and purchased locally from a single vender (Cloud 9 Smokes and Vapors, 476 Elmwood Ave, Buffalo, NY, USA). Products were labeled as Menthol, Vanilla Cupcake, American Tobacco, and Strawberry Blast with 0 mg nicotine. These four were selected as a pilot to both assess actual products in use and to include a range of popular flavorings based on published literature [34,38,43]. Non-targeted analysis of e-liquids was performed as previously published using gas chromatography-mass spectrometry (GCMS) to determine constituents and confirm that samples were nicotine free [30]. Briefly, 10 µL of each sample was diluted with 1 mL methanol (Fisher Scientific; Waltham, MA) and analyzed by GCMS using an Agilent 7890B GC and an Agilent 5977A MS (Santa Clara, CA). An HP−5 30 m × 0.320 mm × 0.25 m capillary column (Agilent) with a flow rate of helium of 1.7 mL/min was utilized. Temperature of the injector and the detector was 250 • C, and the column temperature increased from 110 to 250 • C (10 • C/min) with a hold for 1 min. The injection volume was 1 µL with a split ratio of 40:1. Qualitative analysis of the flavored liquids was carried out using the NIST 14 Mass Spectral Library (U.S. Department of Commerce) as well as the Flavors and Fragrances of Natural and Synthetic Compounds 3 flavoring library.

Cell Culture
Primary pulmonary artery smooth muscle cell (PASMC) cultures were prepared from intrapulmonary arteries collected from healthy 125 (gestational) day fetal, 2 day newborn lambs, and adult ewes as previously published [44]. Numerous publications support our use of this large animal model, as ovine pulmonary development and physiology parallel human, while cellular metabolism and signaling in isolated tissue and cell lines correlate with findings in human disease [44][45][46][47][48][49][50][51][52][53]. For toxicity studies, cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS), 2 mM l-glutamine and 100 units/mL penicillin, 100 mg/mL streptomycin (P/S) under standard cell culture conditions (37 • C with 5% CO 2 ). All experiments were carried out between passages 3 and 8. Cells were seeded onto 12 well plates at approximately 8000 cells per cm 2 and grown to 60-70% confluency. Fetal, newborn, and adult PASMC were subsequently synchronized by transfer to serum-free DMEM + P/S for one hour prior to treatment in parallel with select nicotine-free flavored solutions, pure propylene glycol (PG), or pure vegetable glycerin (VG) (Fisher Scientific, NH) at 1:100, 1:1000, or 1:10,000 dilutions. Cell death assays were performed following 24 h incubation.

Cell Death
Cell death was determined using a lactate dehydrogenase (LDH) assay (Biovision, San Francisco, CA). Using an enzymatic coupling reaction, LDH was quantified in the supernatant and the tritonized cell lysate. LDH activity was determined by spectrophotometric reading at OD 450 . Percent cell death was calculated as concentration in supernatant/concentration in lysate × 100. Dilution at which 50% cell death would occur (lethal dose or LD50) was calculated assuming linear dose dependency (from studies in which testing of two doses concurrently was possible) with extrapolation of values out to a 1:10 dilution.

Isolated Airway and Vessel Studies
Airways and vessels from pregnant ewe and newborn lambs were collected following sacrifice of animals utilized for a separate experimental protocol evaluating transitions at birth. For those studies, newborn lambs were delivered by c-section as part of an Institutional Animal Care and Use Committee at the State University of New York at Buffalo approved protocol (PED10085N). C-section was performed under 4% isoflurane anesthesia. Newborn lambs were ventilated for 4-6 h prior to sacrifice, during which time they received fentanyl titrated from 0-5 mcg/kg/h. Ewes and lambs were euthanized following delivery and completion of the study with pentobarbital (100 mg/kg). Animals were exposed to purified air or oxygen only prior to sacrifice.
The intrapulmonary arteries (PA) and the fifth generation bronchial rings (BR) were dissected, washed in Krebs solution (NaCl 119 mM, KCl 5.4 mM, CaCl 2 2.5 mM, KH 2 PO 4 0.6 mM, MgSO 4 1.2 mM, NaHCO 3 25 mM, and glucose 11.7 mM), and hung in vessel baths aerated with 21% O 2 and 6% CO 2 to achieve a pH of 7.4, as described previously [50]. The samples were mounted and stretched to a passive tension of 800 mg (PA) and 1000 mg (BR) to optimize reactivity to experimental agents. A continuous recording of isometric force generation was obtained by tying each vessel ring to a force displacement transducer (model UC2, Statham Instruments, NY) that was connected to a recorder (Gould Instrument, OH). PA were pretreated with propranolol (10 −6 M) to block the beta adrenergic effects and constricted with increasing titrations (10 −8 to 10 −5 M) of norepinephrine (propranolol and NE, Sigma Aldrich, MO). BR were constricted with increasing titrations (10 −8 to 10 −5 M) of 5-hydroxytryptamine (5-HT, Sigma Aldrich, MO).
To study responses to flavored solutions, PA and BR were preconstricted with half maximal effective concentration (EC50) of NE or 5-HT followed by treatment with a flavoring or 50/50 PG/VG at a 1:1000 dilution with companion preparation (5-HT or NE) alone used as time controls. Preconstriction to EC50 allowed tested e-liquids to further constrict or relax as is physiologically relevant. Following experimental treatment, PA and BR were washed with Krebs solution. Maximal contractile response to 118 mM/L of potassium chloride (KCl) was obtained. Wet tissue weights were obtained at the end of each experiment, and constriction responses were normalized to tissue weight and compared as percentages of relaxation of the EC50 constriction. Data were analyzed as force of constriction per unit of tissue weight (grams of force per gram of tissue weight) of the vasoconstrictor agent. Response to flavored solutions was expressed as a percentage of maximal constriction of 5-HT or NE.

Statistical Analysis
Continuous variables were analyzed using a Student's t-test or ANOVA as appropriate with Tukey's post hoc analysis. A p value of ≤0.05 was considered statistically significant. For cellular experiments, 2-3 wells on a plate were treated for each condition, and the average result considered one n with >6 n from independent cell platings for all studies. For vessel and airway reactivity, replicate tissues from 4 adult and 6 neonatal lambs were studied.

Chemical Analysis of Flavored Liquid Solutions
Non-targeted GCMS analysis confirmed that the Ecto flavored solutions did not include nicotine in two batches of the four flavors used in the study. The two batches were purchased nine months apart with the second batch being nine months old at chemical analysis. Consistent with prior publications [33,38], non-targeted GCMS analysis identified a large number of chemicals present in each of the four samples with variability between batches (Table 1).
Flavorings common to all four included benzophenone (CAS 119-61-9), hydroxyacetone (CAS 116-09-6), and ethyl maltol (CAS 11-8-4940). Unique flavoring chemicals were identified with each flavored solution; however, there was significant overlap in chemicals present in the four studied. Menthol had the most unique chemical flavorings present while tobacco had the least (Figure 1 and Supplementary Table S1). GCMS did not allow for quantification of the specific chemicals present. Table 1. Chemical analysis by gas chromatography-mass spectrometry (GCMS). GCMS was performed on two batches of Ecto flavoring solutions, confirming absence of nicotine but chemical variability with differences in the number of chemicals present in batches of the same flavored solution.

Menthol
Strawberry Tobacco Vanilla   Table A1 for specific chemicals.

Cell Toxicity Studies
As pulmonary blood vessels actively promote alveolar development and contribute to the maintenance of alveolar structures throughout postnatal life [54], we chose to evaluate the toxicity of flavored solutions on pulmonary artery smooth muscle cells (PASMC). To determine relative toxicity in immature versus mature cells, fetal, neonatal, and adult PASMCs were exposed to a select panel of flavored solutions, pure PG, or VG for 24 h at variable dilutions (1:100, 1:1000, 1:10,000). Cell death was flavoring dependent, with the greatest to the least toxicity observed with menthol, strawberry, tobacco, and vanilla flavorings. Exposure to 1:100 dilution of menthol and strawberry solutions induced significant cell death in fetal, neonatal, and adult cells (control: 15.2 ± 3.5, 13.6 ± 2.7, 15.6 ± 2.9%; menthol: 57.9 ± 6.2, 74.5 ± 4.9, 46.3 ± 7.7%; strawberry: 47.6 ± 6.2, 47.2 ± 5.5, 44.8 ± 7.9% for fetal, neonatal and adult PASMC, respectively). Immature cells were more susceptible to menthol toxicity with the greatest cell death documented in neonatal cells (Figure 2A). In contrast, exposure to tobacco and vanilla solutions did not significantly impact viability as compared to untreated controls. In addition, treatment with unflavored, nicotine-free VG and PG resulted in no significant increase in cell death ( Figure 2B), implicating additional flavoring or non-flavoring chemicals present in the refill solutions for the death observed in Figure 2A. Calculated LD50s suggested increased susceptibility to flavored solutions in immature PASMC as compared to adults, with lower calculated concentration noted for fetal and neonatal PASMCs in all four tested flavorings as compared to adults ( Table 2).  Table S1 for specific chemicals.

Cell Toxicity Studies
As pulmonary blood vessels actively promote alveolar development and contribute to the maintenance of alveolar structures throughout postnatal life [54], we chose to evaluate the toxicity of flavored solutions on pulmonary artery smooth muscle cells (PASMC). To determine relative toxicity in immature versus mature cells, fetal, neonatal, and adult PASMCs were exposed to a select panel of flavored solutions, pure PG, or VG for 24 h at variable dilutions (1:100, 1:1000, 1:10,000). Cell death was flavoring dependent, with the greatest to the least toxicity observed with menthol, strawberry, tobacco, and vanilla flavorings. Exposure to 1:100 dilution of menthol and strawberry solutions induced significant cell death in fetal, neonatal, and adult cells (control: 15.2 ± 3.5, 13.6 ± 2.7, 15.6 ± 2.9%; menthol: 57.9 ± 6.2, 74.5 ± 4.9, 46.3 ± 7.7%; strawberry: 47.6 ± 6.2, 47.2 ± 5.5, 44.8 ± 7.9% for fetal, neonatal and adult PASMC, respectively). Immature cells were more susceptible to menthol toxicity with the greatest cell death documented in neonatal cells (Figure 2A). In contrast, exposure to tobacco and vanilla solutions did not significantly impact viability as compared to untreated controls. In addition, treatment with unflavored, nicotine-free VG and PG resulted in no significant increase in cell death ( Figure 2B), implicating additional flavoring or non-flavoring chemicals present in the refill solutions for the death observed in Figure 2A. Calculated LD50s suggested increased susceptibility to flavored solutions in immature PASMC as compared to adults, with lower calculated concentration noted for fetal and neonatal PASMCs in all four tested flavorings as compared to adults ( Table 2).

Airway and Vessel Reactivity Studies
Bronchial Rings: To determine whether airway response to flavored solutions was also developmentally regulated, both neonatal and adult intrapulmonary bronchial rings (BR) were preconstricted with EC50 doses of 5-HT. Once plateau contraction was achieved, BR were exposed to nicotine-free, unflavored 50%/50% PG/VG as well as selected flavored solutions at 1:1000 dilution. Treatment with PG/VG did not impact EC50 constriction of neonatal or adult BR (92.6 ± 3.5% and

Airway and Vessel Reactivity Studies
Bronchial Rings: To determine whether airway response to flavored solutions was also developmentally regulated, both neonatal and adult intrapulmonary bronchial rings (BR) were preconstricted with EC50 doses of 5-HT. Once plateau contraction was achieved, BR were exposed to nicotine-free, unflavored 50%/50% PG/VG as well as selected flavored solutions at 1:1000 dilution. Treatment with PG/VG did not impact EC50 constriction of neonatal or adult BR (92.6 ± 3.5% and 98.4 ± 5.1% constriction of time control, respectively). Treatments with 1:1000 dilution of menthol, strawberry, vanilla, and tobacco e-liquids all resulted in significant airway relaxation in neonatal BR as compared to untreated time controls (menthol: 8.0 ± 9.1; strawberry: 88.2 ± 2.8; vanilla: 54.6 ± 13.2; tobacco: 69.7 ± 8.2% constriction of neonatal time control) ( Figure 3A). In contrast, exposure to the same concentrations of 1:1000 dilution of strawberry, vanilla, and tobacco did not impact airway relaxation in adult BRs, although some relaxation with menthol was observed (menthol: 85.5 ± 2.2% constriction of adult time control). Neonatal airway bronchodilation with exposure to menthol and tobacco flavored solutions was significantly greater than that observed in adult BR, with marked bronchodilation noted with menthol ( Figure 3B). 98.4 ± 5.1% constriction of time control, respectively). Treatments with 1:1000 dilution of menthol, strawberry, vanilla, and tobacco e-liquids all resulted in significant airway relaxation in neonatal BR as compared to untreated time controls (menthol: 8.0 ± 9.1; strawberry: 88.2 ± 2.8; vanilla: 54.6 ± 13.2; tobacco: 69.7 ± 8.2% constriction of neonatal time control) ( Figure 3A). In contrast, exposure to the same concentrations of 1:1000 dilution of strawberry, vanilla, and tobacco did not impact airway relaxation in adult BRs, although some relaxation with menthol was observed (menthol: 85.5 ± 2.2% constriction of adult time control). Neonatal airway bronchodilation with exposure to menthol and tobacco flavored solutions was significantly greater than that observed in adult BR, with marked bronchodilation noted with menthol ( Figure 3B). Contraction to EC50 5-HT elicited a force of 103 ± 14 g/g with a maximum constriction with KCl of 171 ± 23 g/g in adult BRs. Similarly, neonatal constriction to 5-HT was 54 ± 12 g/g with a maximum constriction with KCl of 98 ± 19 g/g. These data support comparable contractions, as 5-HT elicited 60% and 55% of max contraction for neonatal and adult BR, respectively.
Pulmonary Arteries: To evaluate pulmonary vascular response to flavored refill solutions, neonatal and adult pulmonary arteries (PA) were preconstricted to EC50 with NE. Once stable constriction was obtained, PA were similarly treated with 1:1000 nicotine-free, unflavored 50%/50% PG/VG and selected flavored solutions. The 1:1000 dilution of PG/VG did not impact vasoconstriction in either the neonate or the adult (101.4 ± 6.6 and 98.9 ± 2.4% constriction for neonate and adult, respectively). While comparable dosing of flavored solutions resulted in relaxation of bronchial rings, no significant change was noted with exposure of neonatal or adult PA to 1:1000 dilution of strawberry, vanilla, or tobacco. However, treatment with 1:1000 dilution of menthol induced relaxation of adult but not neonatal PA (85.5 ± 2.2 and 102.6 ± 3.2% constriction of time control for adult and neonate) (Figure 4).
Contraction to NE after propranolol elicited a force of 512 ± 6 g/g with a maximum constriction with KCl of 538 ± 53 g/g in adult PAs. Similarly, neonatal constriction to NE was 126 ± 37 with a maximum constriction with KCl of 148 ± 49 g/g. These data support comparable contractions, as NE elicited 95% and 85% of max contraction for neonatal and adult PA, respectively. Contraction to EC 50 5-HT elicited a force of 103 ± 14 g/g with a maximum constriction with KCl of 171 ± 23 g/g in adult BRs. Similarly, neonatal constriction to 5-HT was 54 ± 12 g/g with a maximum constriction with KCl of 98 ± 19 g/g. These data support comparable contractions, as 5-HT elicited 60% and 55% of max contraction for neonatal and adult BR, respectively.
Pulmonary Arteries: To evaluate pulmonary vascular response to flavored refill solutions, neonatal and adult pulmonary arteries (PA) were preconstricted to EC50 with NE. Once stable constriction was obtained, PA were similarly treated with 1:1000 nicotine-free, unflavored 50%/50% PG/VG and selected flavored solutions. The 1:1000 dilution of PG/VG did not impact vasoconstriction in either the neonate or the adult (101.4 ± 6.6 and 98.9 ± 2.4% constriction for neonate and adult, respectively). While comparable dosing of flavored solutions resulted in relaxation of bronchial rings, no significant change was noted with exposure of neonatal or adult PA to 1:1000 dilution of strawberry, vanilla, or tobacco. However, treatment with 1:1000 dilution of menthol induced relaxation of adult but not neonatal PA (85.5 ± 2.2 and 102.6 ± 3.2% constriction of time control for adult and neonate) (Figure 4).
Contraction to NE after propranolol elicited a force of 512 ± 6 g/g with a maximum constriction with KCl of 538 ± 53 g/g in adult PAs. Similarly, neonatal constriction to NE was 126 ± 37 with a maximum constriction with KCl of 148 ± 49 g/g. These data support comparable contractions, as NE elicited 95% and 85% of max contraction for neonatal and adult PA, respectively.

Discussion
Our data suggest that the fetal and the neonatal lung may be at increased susceptibility to toxicity with exposure to flavored e-cigarette solutions and that physiologic responses to these common additives may also be altered in the immature lung. While flavored solutions have previously been shown to be harmful to lung epithelial cells in some but not all in vitro models [30][31][32]55], these studies are the first to identify a developmental susceptibility with exposure of differentiated lung cells. Further, our results are novel in demonstrating that physiologic response to e-cigarette flavoring solutions may also be developmentally regulated.
GCMS analysis identified a large number of non-flavoring and flavoring chemicals present within each of the solutions with notable differences in products purchased nine months apart. While these differences may in part be attributed to aging and chemical decomposition, our data more likely represent the variability within products produced by a specific vendor [22]. Batch variability has been well described and may account for some of the range in responses that we saw in our experiments as well.
We further noted a wide range of toxicity with exposure to a limited panel of flavored solutions. Both menthol and strawberry were noted to be highly toxic, consistent with prior publications [30,56]. While Bahl et al. suggested that the number of peaks on high-performance liquid chromatography (HPLC) correlate with cytotoxicity [22], we noted a comparable number of total and flavoring chemicals (by GCMS) in both toxic and non-toxic products. Importantly, as no nicotine was detected in all tested products, the observed findings can be attributed to the added chemicals in flavoring solutions.
Of interest, while only two of four flavorings demonstrated significant toxicity at a 1:100 dilution, all four flavorings induced bronchodilation in neonatal bronchial rings at a 1:1000 dilution with limited or no impact on adult airway constriction at the same dose. As menthol has previously been shown to inhibit airway smooth muscle contraction in tracheal tissue isolated from guinea pigs and rats [57,58], we suspected all four flavoring samples might include some quantity of menthol or one of its chemical derivatives. However, GCMS did not confirm this suspicion with identification of (±)-menthol, (±)-menthone, (-)-menthone only in the product referred to as menthol, indicating that additional flavoring chemicals are capable of inducing relaxation of neonatal airways.
Bronchodilation in the neonatal airways suggests that newborns and infants in proximity to secondhand flavored e-cigarette vapors may be at risk of exaggerated exposure to a higher concentration

Discussion
Our data suggest that the fetal and the neonatal lung may be at increased susceptibility to toxicity with exposure to flavored e-cigarette solutions and that physiologic responses to these common additives may also be altered in the immature lung. While flavored solutions have previously been shown to be harmful to lung epithelial cells in some but not all in vitro models [30][31][32]55], these studies are the first to identify a developmental susceptibility with exposure of differentiated lung cells. Further, our results are novel in demonstrating that physiologic response to e-cigarette flavoring solutions may also be developmentally regulated.
GCMS analysis identified a large number of non-flavoring and flavoring chemicals present within each of the solutions with notable differences in products purchased nine months apart. While these differences may in part be attributed to aging and chemical decomposition, our data more likely represent the variability within products produced by a specific vendor [22]. Batch variability has been well described and may account for some of the range in responses that we saw in our experiments as well.
We further noted a wide range of toxicity with exposure to a limited panel of flavored solutions. Both menthol and strawberry were noted to be highly toxic, consistent with prior publications [30,56]. While Bahl et al. suggested that the number of peaks on high-performance liquid chromatography (HPLC) correlate with cytotoxicity [22], we noted a comparable number of total and flavoring chemicals (by GCMS) in both toxic and non-toxic products. Importantly, as no nicotine was detected in all tested products, the observed findings can be attributed to the added chemicals in flavoring solutions.
Of interest, while only two of four flavorings demonstrated significant toxicity at a 1:100 dilution, all four flavorings induced bronchodilation in neonatal bronchial rings at a 1:1000 dilution with limited or no impact on adult airway constriction at the same dose. As menthol has previously been shown to inhibit airway smooth muscle contraction in tracheal tissue isolated from guinea pigs and rats [57,58], we suspected all four flavoring samples might include some quantity of menthol or one of its chemical derivatives. However, GCMS did not confirm this suspicion with identification of (±)-menthol, (±)-menthone, (-)-menthone only in the product referred to as menthol, indicating that additional flavoring chemicals are capable of inducing relaxation of neonatal airways.
Bronchodilation in the neonatal airways suggests that newborns and infants in proximity to secondhand flavored e-cigarette vapors may be at risk of exaggerated exposure to a higher concentration of these and additional aerosolized chemicals (including nicotine when present) via increased delivery to the alveoli. Increased minute ventilation in infants would further facilitate the delivery of these aerosolized chemicals, compounding concerns for risk of toxicity in the immature lung.
Vaporized e-cigarette solutions and flavoring chemicals have been shown to contain notable quantities of potentially toxic reactive oxygen species (ROS) [32,59]. Chemicals present in flavored refill solutions may also be capable of inducing intracellular oxidative stress with absorption. The developmental susceptibility to flavoring-induced toxicity and bronchodilation that we observed in these studies may reflect compromised antioxidant capacities present in the immature lung [23], as both cell death and airway responses are mediated by ROS [60][61][62]. Alternatively, altered cellular metabolism, variable cellular absorption, or a difference in the immature inflammatory response [25] could contribute to developmentally regulated responses.
While 1:1000 dilution of each of our studied e-liquids elicited bronchodilation in the newborn airway, this dilution did not impact the pulmonary vasculature with the exception of vasodilation of adult pulmonary arteries with exposure to menthol flavored solutions. Menthol is a common and highly popular flavoring with both vaping and conventional cigarettes, with surveys implying that 27% of smokers report use of mentholated products [63]. Pulmonary vasodilation (as observed in the adult PAs) with menthol inhalation could facilitate absorption of e-cigarette constituents including nicotine, raising concern for potential exacerbation of toxicities and increased gestational exposure. Indeed, some studies suggest that mentholated products may be associated with increased pulmonary pathology [64,65].
While our data are novel in evaluating the consequences of exposure in fetal and neonatal pulmonary cells as well as immature lung tissue, there are several limitations to these studies. Notably, e-cigarette refill solutions were added directly to cell cultures or tissue perfusate, resulting in an exposure that failed to include toxic products generated with heating and aerosolization, avoided the process of thermal degradation, and was potentially at an inaccurate pH [66]. While analytical evaluation demonstrated highly efficient transfer of flavoring chemicals to aerosol (mean transfer 98% at both 3V and 5V) [67], unique chemicals produced with vaping of flavored refill solutions are not represented in these studies. Further, variable temperatures, voltage, devices, and puffing patters have all been shown to modify aerosol toxicity and were not explored [68]. We acknowledge that these limitations impact the interpretation of these studies and fail to meet optimal modeling suggested by Behrsing et al. [69]. We utilized direct exposure with the technical challenges of delivering aerosolized product into the tissue perfusate and to cultured PASMCs.
In addition, our study does not distinguish the effect of specific flavoring chemicals or eliminate the potential contribution of impurities limiting our conclusions. Future studies investigating the specific chemicals of interest individually with known purity and concentrations present in refill solutions would be indicated. We also acknowledge that our small panel fails to characterize the extensive library of now over 7000 unique flavors on the market and may not be generalizable to other e-cigarette products [70]. Finally, the toxicity of the flavoring solutions was not contextualized with comparison to tobacco; ultimately, these products may represent a safer alternative to smoking.
We are further limited by the lack of data regarding typical concentration of flavoring chemicals that lung cells, pulmonary tissue, and a fetus might be exposed to. Online vaping forums suggest that use of 5 mL per day of flavored refill solutions would be common [33]. Our dosing range maxed at 1% to model some degree of dilution and was based on rationale provided in other publications [56]. Specifically, based on particle size, the predicted deposition of e-cigarette aerosol in the lungs is 15-45% [71]. Given that the airway surface liquid volume in the lung is approximately 3 mL [56], vaping of 10 mL would lead to a dilution of 33-60%, suggesting that our experimental dosing of 0.1-1% may be in reasonable range for PASMC and tissue exposure. While the specific concentrations with fetal exposure remain unknown, recent publications have demonstrated that flavoring chemicals can indeed impact fetal development [14].
Finally, while these studies are novel in comparing toxicity and physiologic responses with exposure in both neonatal and adult models, responses were limited to isolated cells and tissue. Our data suggesting a developmental susceptibility to commonly used flavored solutions advocate for further evaluation of these chemicals by way of in vivo exposure to the aerosolized product.

Conclusions
Gestational and postnatal exposure to electronic cigarettes represents rapidly growing threat to the fetal and the newborn population. Use of added flavorings with vaping has far out-paced our understanding of their implications for public health. Our data suggesting a developmental susceptibility with exposure to common flavoring solutions argue the critical need for further thoughtful evaluation of these products.

Conflicts of Interest:
The authors declare no conflict of interest.