Comparative Genotoxicity and Mutagenicity of Cigarette, Cigarillo, and Shisha Tobacco Products in Epithelial and Cardiac Cells

Abstract

Epidemiology studies link cigarillos and shisha tobacco (delivered through a hookah waterpipe) to increased risk for cardiopulmonary diseases. Here we performed a comparative chemical constituent analysis between 3 cigarettes, 3 cigarillos, and 8 shisha tobacco products. The potency for genotoxicity and oxidative stress of each product’s generated total particulate matter (TPM) was also assessed using immortalized oral, lung, and cardiac cell lines to represent target tissues. Levels of the carcinogenic carbonyl formaldehyde were 32- to 95-fold greater, while acrolein was similar across the shisha aerosols generated by charcoal heating compared to cigarettes and cigarillos. Electric-mediated aerosol generation dramatically increased acrolein to levels exceeding those in cigarettes and cigarillos by up to 43-fold. Equivalent cytotoxic-mediated cell death and dose response for genotoxicity through induction of mutagenicity and DNA strand breaks was seen between cigarettes and cigarillos, while minimal to no effect was observed with shisha tobacco products. In contrast, increased potency of TPM from cigarillos compared to cigarettes for inducing oxidative stress via reactive oxygen radicals and lipid peroxidation across cell lines was evident, while positivity was seen for shisha tobacco products albeit at much lower levels. Together, these studies provide new insight into the potential harmful effects of cigarillos for causing tobacco-associated diseases. The high level of carbonyls in shisha products, that in turn is impacted by the heating mechanism, reside largely in the gas phase which will distribute throughout the respiratory tract and systemic circulation to likely increase genotoxic stress.

Cigarillos and shisha (delivered through a hookah waterpipe) are 2 tobacco products whose popularity has been growing steadily among teenagers and young adults due to low cost and the social networking that the hookah bar environment promotes. Cigarillos are viewed by users as less of a health risk because of misperceptions that they are more natural and that levels of nicotine are lower than in cigarettes, making them less addictive (Jolly, 2008; Malone et al., 2001). World-wide epidemiological studies link cigarillos and shisha to lung cancer, coronary heart disease, aortic aneurysm, and chronic obstructive pulmonary disease (COPD; Aljarrah et al., 2009; Chang et al., 2015; Shihadeh et al., 2015). Cigarillo use of at least once monthly is reported for 2.2%, 7.7%, and 4.4% of middle school, high school, and 18- to 24-year-old adults (Centers for Disease Control and Prevention, 2017; Phillips et al., 2017). A recent study of adult cigarillo users 23–68 years old reported weekly average use of 20 cigarillos (range, 1.5–84) with most inhaling the product (≥80%; Claus et al., 2018). We conducted a comparative study between common brands of cigarettes and cigarillos using total particulate matter (TPM) generated from these products to evaluate potency for cytotoxicity and genotoxicity (Crosby et al., 2021). Equivalent or greater cytotoxicity dependent on the puffing protocol (ISO versus Canadian Intensive) was observed for cigarillos compared to cigarettes (Crosby et al., 2021). Micronuclei formation was also significantly greater for cigarillos compared to cigarettes at the highest dose of TPM. Thymidine kinase mutation frequency was also higher for the cigarillo products, while in the Ames assay, both tobacco products exhibited significant dose-dependent increases in mutations in the presence of metabolic activation (rat liver S9 fraction). This study provided new insights supporting equivalent or greater health risks for people using cigarillo products (Crosby et al., 2021).

Shisha tobacco use of at least monthly is reported in the United States for 2.0%, 4.8%, and 3.4% of middle school, high school, and 18- to 24-year-old adults (Centers for Disease Control and Prevention, 2017; Phillips et al., 2017). The shisha tobacco is incapable of burning and is smoked using charcoal as a heat source that is placed on top of the tobacco with a thin, perforate sheet of aluminum separating the 2 products, or by placing the shisha in an electrically heated bowl. The hookah waterpipe then draws air and combustion products from the burning charcoal through the tobacco to create aerosols consisting of volatilized and pyrolyzed tobacco components. Alternatively, with the electric source, the aerosols are generated solely from the burning of the tobacco. Shisha smoking in hookah bars is associated with significant nicotine intake and carcinogen exposure resulting from inhaling the tobacco smoke (Cobb et al., 2015). Smoking behavior and physiological effects (nicotine intake, carbon monoxide, development of dependence) change with increased frequency of use (Cobb et al., 2015; Primack et al., 2016; Shishani et al., 2014). There are few studies characterizing toxicant chemical properties within these tobacco aerosols and limited information regarding effects on target cells (oral, lung, and cardiac) with respect to cytotoxicity, genotoxicity, and oxidative stress responses (Chang et al., 2015; Shihadeh et al., 2015). Relative to smoking a cigarette over an average 6-min period, a waterpipe use episode with a mean smoking time of 43 min resulted in similar peak nicotine exposure, 3.75-fold greater carboxyhemoglobin, and 56-fold greater inhaled smoke volume (Cobb et al., 2011). A major concern is the high exposure of the shisha smoker to volatile aldehydes and the 27 known carcinogens identified in the hookah smoke (Helen et al., 2014; Shihadeh et al., 2015). In addition, genotoxicity studies report detection of DNA damage in the form of sister chromatid exchange, chromosomal aberrations, an increase in immune regulators, and the detection of acrolein, a toxic respiratory irritant in urine from shisha smoking (Alsatari et al., 2012; Khabour et al., 2011; Kassem et al., 2018; Rammah et al., 2012). Importantly, no comparative dose-response studies of cigarillo and hookah products to cigarettes have been conducted, a research gap that is imperative to address for providing pharmacodynamic outcomes for these addictive nicotine-containing tobacco products.

The goal of this study was to extend our original work on cigarillos to include shisha tobacco products and to also incorporate sensitive assays of oxidative stress (reactive oxygen radicals, lipid peroxidation), and DNA damage for comparative risk assessment (Crosby et al., 2021). Furthermore, given the risk for health effects that involve the oral cavity, lungs, and cardiovascular system, immortalized cell lines representing these target organs were selected for these studies. Aerosols were generated from 3 each cigarettes and cigarillos and 8 shisha tobacco products using charcoal and electric to characterize levels of nicotine, cotinine, nitrosamines, polyaromatic hydrocarbons (PAHs), carbonyls, and particle size distributions. TPM was also generated from these products for dose-response studies to evaluate cell toxicity, mutagenicity, oxidative stress, and DNA damage using the immortalized cell lines.

MATERIAL AND METHODS

Cell Lines

Three immortalized cell lines were selected for study rather than primary cells that senesce after 7–10 population doublings given the magnitude of the studies and the need for rigor with respect to comparing experimental results across products and endpoints. The MOE1B oral epithelial cell line was obtained from a 32-year-old male never smoker and immortalized by introducing hTERT, CDK4R2C, cyclin D1, and P53C234 (Kibe et al., 2011). The HBEC4 bronchial epithelial cell line was obtained from a 71-year-old female smoker and immortalized by introducing hTERT and CDK4 (Damiani et al., 2008). The H9c2 cardiac cell line derived from embryonic rat cardiomyocytes was spontaneously immortalized and obtained from ATCC (Hescheler et al., 1991). Cell lines were grown as described previously (Damiani et al., 2008; Hescheler et al., 1991; Kibe et al., 2011).

Tobacco Products

The cigarettes selected for study were Kentucky reference 3R4F, Marlboro Red, and Camel 99. Cigarillos selected were Swisher Sweets Classic Original, Black & Mild Middleton’s, and Dutch Masters Palma Natural. The shisha tobaccos and specific flavors were Al Fakher Bubble Gum, Starbuzz Candy, Fantasia Cotton Candy, Nakhla Fruits, Tangiers Green Apple Candy, Nirvana Candy Baby, Romman Sweet Start, and Fumari Tropical Punch. Coco Nara coals or an electric bowl were used for generating the aerosol.

Collection of TPM and Aerosol Generation

Cigarettes and cigarillos were burned using a smoking machine using the Health Canadian Intensive Regimen that involves 50 ml puff volumes at 30-s intervals, with 2 s/puff in addition to occluding the entire filter with cellophane tape. The cigarillos did not have filters thus, 23 mm was used as the butt length. A programmable relay was used to draw air through a commercial hookah waterpipe (total height 24 inches) filled with 1 l of water. Twenty grams of each shisha tobacco was placed in the ceramic clay bowl separated by aluminum foil and 4 charcoals previously heated for 30 min to cover the surface area of the bowl (charcoal heater from Gold Star). Alternatively, the shisha was placed in an electric bowl from Ren Headstream that was heated to 120°C prior to initiating the Beirut puffing protocol while maintaining a temperature of 180°C during aerosol generation. The Beirut puffing protocol mimics the topography of humans using the hookah waterpipe and consists of 530 ml puffs of 2.6-s duration at a frequency of 2.8 puffs/min (Brinkman et al., 2016; Shihadeh et al., 2004). The hookah waterpipe was obtained from a local smoke shop and the generated aerosol was bubbled through 1 l of water in the waterpipe. Cigarettes or cigarillos were puffed 8 or 9 times and shisha was puffed 10 times to collect approximately 100 mg TPM on Cambridge filters. The TPM was extracted from filters with DMSO to achieve a final concentration of 40 mg/ml. Following extraction, aliquots of TPM were stored in glass vials at −20°C for in vitro assays. Storage time of cigarette, cigarillo, and shisha product TPMs prior to use did not differ. In addition, samples were randomized and the technical staff was blinded to product type.

For chemical characterization, 3 puffs (3, 4, and 5) were collected from each cigarette or cigarillo. Two puffs were collected from the Hookah pipe with charcoal after 15 puffs, and for Hookah electric after 6 puffs. This allowed for stabilization of the temperature for generating the aerosol. Independent collections were performed on filters for analysis of nicotine/cotinine, tobacco-specific nitrosamines (TSNAs), and PAHs. Carbonyls which reside largely in the gas phase were collected through passing the smoke aerosol through a filter to trap the particles and then through a silica gel cartridge coated with 2-,4-dinitrophenylhydrazine (DNPH) that traps the aldehydes. The carbonyls were eluted from the DNPH cartridge with acetonitrile. These puffing protocols were repeated 2 more times for an n = 3 for nicotine/cotinine, TSNAs, and PAHs and 3 more times for an n = 4 for carbonyls.

Chemical Characterization of the Shisha Tobacco Products

Nicotine and Cotinine

A gas chromatography–mass spectrometry (GC-MS) method was used to determine the amount of nicotine and cotinine in the aerosol samples after collection on 25 mm glass fiber filter pads (Massadeh et al., 2009). Post-exposure, 5 µg nicotine-D4, and 5 µg cotinine-D3 were spiked onto the filters. After drying, filters were extracted with 1 ml of dichloromethane. Limits of detection for nicotine and cotinine were 10 and 0.25 µg/ml, respectively.

Tobacco Specific Nitrosamines

An LC-MS/MS method was used to determine the amount of NNN and NNK in the tobacco aerosols captured on glass fiber filters (Wu et al., 2008). The TSNAs are extracted from the filters by adding ammonium acetate in water to each filter in a glass vial, following the addition of the internal standard (NNN-d5). The vials are mixed on a rotator after which they are vortexed with a benchtop vortexer. The sample extracts are passed through a 0.2 µm filter and diluted 1:50 in 90/10 water/methanol (v/v) prior to analysis. Limit of detection for all analytes was 1 ng/ml.

Polyaromatic Hydrocarbons

A gas chromatography triple-quad mass spectrometry (GC-MS) method was used to determine the amounts of PAH from aerosols captured on glass fiber filters and prepared similarly to that for TSNAs (Ding et al., 2007). The 2 PAHs selected for analysis (BaA and BaP) were based on carcinogenic potency among PAHs and continuity with prior studies with cigarettes and cigarillos (Vu et al., 2015). Limit of detection for each analyte was 2.5 ng/ml.

Carbonyls

An LC-MS/MS method was used to determine the amount of formaldehyde, acetaldehyde, acrolein, acetone, diacetyl, and crotonaldehyde in the aerosol generated from each product (Zhou et al., 2021). A DNPH solution (50 µl) was added to the vials followed by the addition of 280 µL of acetonitrile (Reaction Blank), unknown samples, quality controls, or the internal standard crotonaldehyde DNPH-d₃. The linear range for the standard curves was 50–5000 ng/ml (formaldehyde), 10–1000 ng/ml (acetaldehyde), and 2–1000 ng/ml (acrolein, crotonaldehyde, diacetyl). The lower limits of detection expressed as ng/ml for each analyte were as follows: formaldehyde, 260; acetaldehyde, 52; acrolein, 10.4; diacetyl, 10.4; and crotonaldehyde, 10.4.

Particle Size Determination

Particle size distribution was measured using a TSI Aerodynamic Particle Sizer (APS) with an Aerosol Diluter and a TSI Fast Mobility Particle Sizer (FMPS). The APS provides high-resolution, real-time aerodynamic measurements of particles from 0.5 to 20 μm, while the FMPS spectrometer measures submicrometer aerosol particles in the range from 5.6 to 560 nm in diameter (Chen et al., 1990). The mass median aerodynamic diameter (MMAD) and count median aerodynamic diameter (CMAD) with standard deviations for APS and FMPS were determined. In addition, an In-Tox 7-stage cascade impactor was used for comparison to provide an integrated assessment of the aerodynamic particle size distribution with effective cutoff aerodynamic diameter from 0.22 to 5.03 microns (Mercer et al., 1970). The particle size distribution was determined in terms of MMAD and geometric standard deviation (GSD) with SigmaPlot software.

Standardization of Cell-Based Exposures to Tobacco Product TPM

Our studies were designed to evaluate under similar exposure conditions to each cell line, the dose response for cytotoxicity, oxidative stress, and genotoxicity from the cigarette, cigarillo, and shisha tobacco product TPMs compared to air control. This entailed defining the number of cells plated for each cell line (based on growth rate) to achieve 70% ± 10% confluence 96 h following plating, the time point at which cells were exposed to the TPM. Each assay was performed in a different size culture dish with the volume of media stoichiometric to the plate size and number of cells plated adjusted to achieve the 70% confluence. The NRU assay was conducted in a 96-well plate with 100 µl of media and the number of cells seeded was 3500, 2500, and 2000 for the HBEC4, MOE1B, and H9C2 lines. The ROS-Glo assay was conducted in a 48-well plate with 200 µl of media and the number of cells seeded was 10 000, 7500, and 6000 for the HBEC4, MOE1B, and H9C2 lines. The comet assay was conducted in a 6-well plate with 1000 µl of media and the number of cells seeded was 78 000, 55 000, and 45 000 for the HBEC4, MOE1B, and H9C2 lines. The TBARs assay was conducted in a 60 mm dish with 2000 µl of media and the number of cells seeded was 105 000, 75 000, and 60 000 for the HBEC4, MOE1B, and H9C2 lines.

Neutral Red Uptake Assay

The neutral red uptake (NRU) assay provides a quantitative estimation of the number of viable cells in a culture based on the ability of viable cells to incorporate and bind the supravital dye neutral red in the lysosomes (Repetto et al., 2008). Three replicates of each cell line at approximately 80% confluence were exposed to vehicle (DMSO) or increasing dose (4, 8, 16, 32, 64, and 128 µg/ml) of TPM from each tobacco product. Following 24 h post-exposure, the cells were incubated with growth media containing neutral red dyes for 2 h, washed with phosphate-buffered saline and the dye was solubilized by addition of neutral red solubilization solution (1% acetic acid in 50% ethanol) and the absorbance was read at 540 nm. Exposures were repeated 3 times.

Oxidative Stress Assays

Oxygen species are highly reactive (e.g., superoxide, singlet oxygen) and have a short half-life in solution with most being spontaneously or enzymatically converted to hydrogen peroxide (H₂O₂) by mitochondria with temporal kinetics influenced by the extent of oxidative stress (Lushchak, 2014). The ROS-Glo^TM assay (Promega) uses an H₂O₂ substrate that directly reacts with H₂O₂ to produce a luciferin precursor that is then reacted with a detection solution to convert the precursor to luciferin. This provides luciferase to generate a light signal proportional to the level of H₂O₂ present. Assays were performed with 4 replicates over the dose range (32, 64, 128, and 256 µg/ml) of TPM from each tobacco product. Cell lines were exposed to DMSO or TPM for 30 min followed by the addition of the ROS-Glo substrate for 60 min. The detection solution was then added and H₂O₂ was quantified as relative light units (RLUs). The number of viable cells quantified by Cell-Titer assay (Promega) was used to normalize the RLUs detected.

Lipid peroxidation is a well-defined mechanism of cellular damage in animals and plants. Lipid peroxides are unstable indicators of oxidative stress in cells that decompose to form more complex and reactive compounds such as malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE), natural bi-products of lipid peroxidation (Ghani et al., 2017). Thiobarbituric Acid Reactive Substances (TBARs) are a well-established assay for screening and monitoring lipid peroxidation. MDA forms an adduct with thiobarbituric acid that can be measured colorimetrically. TBARs levels were determined from an MDA equivalence standard curve that was linear over a range of 0.98–125 µM. Assay results from 4 replicates (32, 64, 128, and 256 µg/ml TPM) were normalized by measuring total protein by bicinchoninic acid (BCA) in the cell lysate.

Genotoxicity Assays

The Modified Ames ISO kit was used due to the increased sensitivity over the Ames pour-plate assay by allowing testing of higher concentrations of test articles. The TA100 (detects base-pair substitutions primarily at GC pairs) and TA98 (detects frame-shift mutation associated with exposure to 2-nitrofluorene and aromatic nitrosoderivatives of amine carcinogens) were used as these strains have shown mutagenic activity in response to many different tobacco smoke constituents (Aufderheide and Gressmann, 2008; Kilford et al., 2014; Thorne et al., 2015, 2016). Briefly, reagents were mixed with or without the rat liver S9 fraction for metabolic activation, bacteria, and TPM at 10, 40, 130, or 400 µg/ml or DMSO was added to a 24 well plate and incubated for 100 min. The reversion media was added and the suspension was then dispersed into 48-wells on a 96-well plate and incubated for 3 days. Plates were removed and scored by visualizing a change in color from purple to yellow (indicative of a mutation) in each of the 48 wells.

The alkaline comet chip assay was used for the detection of DNA strand breaks in cells or nuclei following exposure to potentially genotoxic materials (Ge et al., 2014). Under alkaline conditions (>pH 13), the comet assay can detect single and double-stranded breaks, resulting, for example, from direct interactions with DNA, alkali labile sites, or as a consequence of transient DNA strand breaks resulting from DNA excision repair (Ge et al., 2014). Cell lines were exposed in triplicate to each TPM dose or DMSO and the Trevigen Alkaline Comet Assay kit was used to conduct the comet assay. Comet images were captured at ×4 magnification on a microscope equipped with epifluorescence and a CCD camera. All slides for analysis were coded and scored “blinded” so the scorer was unaware of the treatment condition. At least 150 scoreable comets that have no overlapping tails and were not located at the edge of slides per treatment per cell line were analyzed using Trevigen Comet Analysis Software. The percent tail DNA (also known as percent tail intensity) was used for the evaluation and interpretation of results, and was determined by the DNA fragment intensity in the tail expressed as a percentage of the cell's total intensity (Bowden et al., 2003).

Data Analysis

Parametric one-way analysis of variance (ANOVA) was used to detect products’ difference from the Kentucky reference cigarette for levels of nicotine, cotinine, TSNAs, PAHs, carbonyls, and particle size. In addition, ANOVA was used for comparing these endpoints between Shisha tobacco products generated with charcoal versus electric. Generalized estimating equations with nested correlation matrix were used to correct the correlation of the repeated measurements. For carbonyl data, measurements below quantitative limits were imputed at the lower limit divided by the square root of 2 to enable comparisons between products (Hornung and Reed, 1990). Imputation was performed on only 6 and 10 measurements for crotonaldehyde and acrolein for different shisha tobacco products generated with charcoal. All reported p-values are based on 2-sided tests at significance <0.05.

The extent of cytotoxicity determined by the NRU assay was assessed by dividing absorbance readings of the different doses and replicates by the absorbance seen with the DMSO control. The IC50 and/or maximum cytotoxicity for each e-liquid per cell line was determined with probit analysis, a type of regression used to analyze binomial response variables in which the cytotoxicity’s sigmoid dose-response curves are transformed to a straight line to allow for assessment by least squares or maximum likelihood.

ROS, TBARs, and comet values were tested separately for each cell line with ANOVA to evaluate the hypothesis that each exposure level would show an increased effect versus the DMSO control, with Dunnett’s adjustment for multiple comparisons of exposures for each product. The increasing dose response for increase in fold change for each product across exposure levels was tested with ANOVA. The dose response of different products was also compared to the dose response of the Kentucky reference cigarette. In the comparison, doses and products were included in the model and adjusted as covariates, and the product/dose interaction term was assessed to determine the differential product dose-response effect. TBARs and Comet data were and log-transformed for analysis to reduce skewness. For all data, each cell line was analyzed separately, and fold change for each product was calculated by normalizing to the average of the DMSO control for that cell line. All statistical calculations were performed using SAS V.9.4.

RESULTS

Chemical Characterization of Cigarette, Cigarillo, and Shisha Tobacco Products

Nicotine levels ranged from 83 to 136 µg/puff across cigarettes and cigarillos with significant difference from the Kentucky reference seen for Camel, Marlboro, and Black & Mild (Table 1). Cotinine levels in cigarillos were ∼2.5-fold higher than in cigarettes. In contrast, with exception of Tangiers (charcoal) and Nirvana shisha, nicotine levels were 50%–90% lower in shisha tobacco products compared to Kentucky reference cigarette and cotinine levels were below limit of quantitation (Table 1). Nicotine levels were greater in 5 of 8 shisha tobacco aerosols generated by charcoal compared to electric (Table 1).

Table 1.

Quantification of Nicotine, Cotinine, and Tobacco Carcinogens in Cigarettes, Cigarillos, and Shisha Tobacco Brands

		Mean µg/puff ±SD		Mean ng/puff ±SD
Product	Brand	Nicotine	Cotinine	NNK	NNN	BaP	BaA
Cigarette	Kentucky 3R4F	107 ± 7	0.8 ± 0.1	20 ± 4	20 ± 5	3.0 ± 0.6	2.1 ± 0.1
	Camel	136 ± 28a	1.2 ± 0.3b	13 ± 9	14 ± 5	4.7 ± 1.5a	3.0 ± 0.7b
	Marlboro	115 ± 6a	1.1 ± 0.1c	16 ± 4	19 ± 4	3.5 ± 0.7	2.9 ± 0.5
Cigarillo	Black & Mild	83 ± 4c	2.5 ± 0.1c	20 ± 2	18 ± 2	2.0 ± 0.7a	6.2 ± 0.2c
	Dutch Masters	121 ± 13	2.5 ± 0.1c	14 ± 4a	40 ± 8c	3.3 ± 1.3	6.3 ± 1.0c
	Swisher Sweets	97 ± 20	3.3 ± 0.1c	76 ± 21c	44 ± 4c	4.1 ± 0.5b	7.2 ± 0.3c
Shisha	Al Fakher	54 ± 13c	ND	ND	ND	ND	1.6 ± 0.9
(Charcoal)	Starbuzz	37 ± 3c	ND	ND	ND	ND	1.8 ± 0.8
	Fantasia	29 ± 2c	ND	ND	ND	ND	1.5 ± 0.9
	Nakhla	48 ± 3c	ND	ND	ND	ND	2.0 ± 0.4
	Tangiers	118 ± 6a	ND	3 ± 1c	1 ± 1c	ND	2.2 ± 0.1a
	Nirvana	134 ± 12c	ND	8 ± 1c	9 ± 1c	ND	1.6 ± 0.3b
	Romman	23 ± 1c	ND	ND	ND	ND	1.6 ± 1.2
	Fumari	48 ± 16c	ND	ND	ND	ND	1.2 ± 0.1c
Shisha	Al Fakher	29 ± 2c–	ND	ND	ND	ND	2.2 ± 0.4
(Electric)	Starbuzz	13 ± 1c–	ND	ND	ND	ND	1.9 ± 0.3
	Fantasia	44 ± 10c+	ND	ND	ND	ND	2.3 ± 0.1c
	Nakhla	93 ± 1c+	ND	ND	ND	ND	1.7 ± 0.5
	Tangiers	53 ± 4c–	ND	3 ± 1c	1 ± 1c	ND	1.9 ± 0.1c
	Nirvana	153 ± 24c	ND	3 ± 1c–	3 ± 1c–	ND	2.1 ± 0.4+
	Romman	12 ± 1c–	ND	ND	ND	ND	2.1 ± 0.3
	Fumari	29 ± 4c–	ND	ND	ND	ND	1.9 ± 0.1a+

ND, not detected.

^a p <.05,

^b p < .01,

^c p < .001 compared to the Kentucky reference cigarette.

^–/+ p < .05 for decrease/increase when electric heating compared to charcoal by individual brand.

The formation of the TSNA NNK was similar across cigarettes and the Black & Mild cigarillo. NNK levels were ∼4-fold greater in Swisher Sweets and 1.4-fold less in Dutch Masters compared to the Kentucky reference (Table 1). NNN formation was also similar across cigarettes and increased 2-fold in Dutch Master and Swisher Sweets cigarillos (Table 1). NNK and NNN were only detected albeit at significantly reduced amounts in the Tangiers and Nirvana shisha independent of method for TPM generation (Table 1). Levels of BaP were generally comparable between cigarettes and cigarillos, varying by approximately 35% around the median value. BaA levels in cigarillos exceeded that seen for cigarettes by 2- to 3-fold (Table 1).

Formaldehyde levels of 19–25 ng/puff for cigarettes and cigarillos increased dramatically to 802–2399 ng/puff for shisha tobaccos using charcoal (Table 2). Shisha tobacco aerosols produced with electric heating contained similar amounts of formaldehyde, the exception being Starbuzz, Tangiers, Nirvana, and Fumari (1.5- to 7-fold greater, Table 2). Acetaldehyde levels were also similar across cigarettes and cigarillos (∼20 000 ng/puff) and reduced by 35%–90% in Shisha products irrespective of the mode for aerosol generation (Table 2). Acrolein levels were also similar between the Kentucky reference cigarette and cigarillos, but significantly lower (66%) in Camel and Marlboro cigarettes (Table 2). Levels of this carbonyl varied dramatically across the different shisha tobacco aerosols generated by the charcoal heating method with most products lower or equivalent to the Kentucky reference cigarette, while a 4-fold greater amount was seen in Nirvana (Table 2). Interestingly, electric-mediated aerosol generation dramatically increased acrolein to levels exceeding those in cigarettes and cigarillos by up to 43-fold (Starbuzz; Table 2). Levels of crotonaldehyde were similar (∼400–500 ng/puff) across cigarettes and cigarillos and reduced by ≥85% in shisha products generated by charcoal. In general, levels remained low in electric-generated aerosols compared to cigarettes and cigarillos, but statistical increases of 2- to 10-fold were seen for most products compared to charcoal heating (Table 2). Finally, the highest levels of diacetyl were found in the Kentucky reference cigarette and the Black & Mild and Dutch Masters cigarillos (Table 2). Levels in other cigarettes and Swisher Sweets were lower by 30%–50%. Diacetyl levels were significantly lower (62%–99%) across shisha product aerosols generated by charcoal and markedly increased with electric heating for all products (Table 2).

Table 2.

Quantification of Carbonyls in Cigarettes, Cigarillos, and Shisha Tobacco Brands

		Mean ng/puff ±SD
Product	Brand	Formaldehyde	Acetaldehyde	Acrolein	Cr-aldehyde	Diacetyl
Cigarette	Kentucky 3R4F	19 ± 1	23 173 ± 4312	93 ± 25	519 ± 42	20 407 ± 2553
	Camel	25 ± 6a	19 237 ± 5263	36 ± 24c	478 ± 165	14 475 ± 237c
	Marlboro	20 ± 4	21 689 ± 5662	36 ± 19c	398 ± 129a	13 542 ± 2817c
Cigarillo	Black & Mild	24 ± 6	22 126 ± 3668	109 ± 48	509 ± 64	20 143 ± 3345
	Dutch Masters	20 ± 2	22 457 ± 1547	99 ± 39	552 ± 76	17 463 ± 2901
	Swisher Sweets	23 ± 5a	18 715 ± 5064	117 ± 5	480 ± 39	9983 ± 5360c
Shisha	Al Fakher	1679 ± 603c	4795 ± 1915c	28 ± 13c	10 ± 4c	98 ± 55c
(Charcoal)	Starbuzz	802 ± 75c	2672 ± 819c	10 ± 3c	13 ± 15c	98 ± 10c
	Fantasia	1705 ± 201c	4610 ± 528c	22 ± 12c	20 ± 6c	78 ± 22c
	Nakhla	851 ± 141c	3124 ± 814c	14 ± 4c	15 ± 3c	215 ± 48c
	Tangiers	2399 ± 298c	10 076 ± 1509c	136 ± 28b	34 ± 8c	864 ± 398c
	Nirvana	2143 ± 337c	7312 ± 844c	457 ± 104c	88 ± 18c	8748 ± 1009c
	Romman	1031 ± 95c	2090 ± 446c	15 ± 10c	6 ± 4c	91 ± 40c
	Fumari	2374 ± 122c	10 531 ± 1113c	133 ± 34a	15 ± 3c	797 ± 201c
Shisha	Al Fakher	2470 ± 727c	3457 ± 866c	1047 ± 393c+	45 ± 14c+	4404 ± 1186c+
(Electric)	Starbuzz	5716 ± 1065c+	13 087 ± 1706c+	4090 ± 558c+	172 ± 31c+	6928 ± 568c+
	Fantasia	413 ± 169c–	4974 ± 1294c	360 ± 276a+	56 ± 8c+	3816 ± 987c+
	Nakhla	2167 ± 754c+	2991 ± 1321c	535 ± 389b+	48 ± 32c+	1788 ± 1010c+
	Tangiers	3793 ± 1299c+	9460 ± 3610c	1491 ± 380c+	82 ± 25c+	6128 ± 1592c+
	Nirvana	5629 ± 492c+	13 321 ± 1725c+	1748 ± 357c+	151 ± 38c+	9030 ± 1646c+
	Romman	938 ± 208c	2403 ± 713c+	472 ± 177c+	53 ± 12c	2899 ± 223c+
	Fumari	3667 ± 422c+	9855 ± 1948c	3652 ± 306c+	127 ± 15c+	8222 ± 107c+

Cr-aldehyde, crotonaldehyde.

^a p <.05,

^b p < .01,

^c p < .001 compared to the Kentucky reference cigarette.

^–/+ p < .05 for decrease/increase when electric heating compared to charcoal by individual brand.

Temperature for Aerosol Generation Affects Particle Size Distribution across Tobacco Products

The temperature generated by burning a cigarette/cigarillo can reach 900°C at the tip, while the temperature of the charcoal and electric bowl is approximately 450°C and 180°C. These temperature differences could affect the distribution of particle size within the vapor between products. Therefore, APS was used to obtain the average of the size distribution ranging from 0.5 to 20 microns (MMAD±SD), while FMPS assessed submicrometer aerosol particles in the range of 5.6–560 nm (count median diameter ±SD). In addition, the In-Tox cascade impactor was used to provide an overall integrated picture of the particle size distribution. Large particle size assessed by APS was similar (1 µm) across cigarettes and cigarillos, increased significantly in shisha product aerosols generated by charcoal (1.2- to 1.8-fold) and electric bowl (1.9- to 3.5-fold) compared to cigarettes, with aerosol size generated by the electric bowl significantly greater than charcoal (Table 3). In contrast, small particle size distribution while again similar (204–215 nm) across cigarettes and cigarillos exceeded that seen for the shisha product aerosols generated by charcoal by 1.3- to 3.6-fold (Table 3). Electric generated aerosols from 4 shisha products (Al Fakher, Starbuzz, Fantasia, Nakhla) had small particle size similar to cigarettes, while particle size for the other products was reduced similarly to their charcoal generated aerosol. The contrasting effects on large and small particle distribution resulted in a similar integrated size distribution when assessed by the cascade impactor across the 3 product types.

Table 3.

Particle Size Distribution of Tobacco Product Aerosols

		APS MMAD (µm)	FMPS CMD (nm)	Impactor (µm)
Product	Brand	Mean ± SD	Mean ± SD	MMAD ± GSD
Cigarette	Kentucky 3R4F	1.0 ± 0.1	207 ± 8	1.2 ± 1.7
	Camel	1.0 ± 0.1	204 ± 6	1.3 ± 1.6
	Marlboro	1.0 ± 0.0	207 ± 7	1.3 ± 1.6
Cigarillo	Black & Mild	1.1 ± 0.1	221 ± 11a	1.6 ± 1.4
	Dutch Masters	1.1 ± 0.1	215 ± 11	1.4 ± 1.6
	Swisher Sweets	1.2 ± 0.1c	210 ± 17	1.6 ± 1.6
Shisha	Al Fakher	1.6 ± 0.1c	116 ± 31c	1.8 ± 1.6
(Charcoal)	Starbuzz	1.7 ± 0.2c	84 ± 18c	1.6 ± 1.5
	Fantasia	1.4 ± 0.0c	90 ± 16c	1.7 ± 1.5
	Nakhla	1.3 ± 0.1c	58 ± 11c	1.3 ± 1.6
	Tangiers	1.8 ± 0.0c	95 ± 8c	2.0 ± 1.5
	Nirvana	1.5 ± 0.0c	111 ± 12c	2.1 ± 1.5
	Romman	1.5 ± 0.1c	79 ± 5c	1.4 ± 1.5
	Fumari	1.2 ± 0.1c	100 ± 9c	2.0 ± 1.4
Shisha	Al Fakher	2.3 ± 0.0c+	188 ± 51+	2.3 ± 1.3
(Electric)	Starbuzz	2.9 ± 0.4c+	144 ± 53a+	4.8 ± 1.4
	Fantasia	2.5 ± 0.1c+	209 ± 5+	1.6 ± 2.7
	Nakhla	1.9 ± 0.1c+	152 ± 58a+	1.9 ± 1.5
	Tangiers	3.2 ± 0.5c+	77 ± 32c	1.7 ± 1.5
	Nirvana	3.5 ± 0.2c+	61 ± 4c–	1.3 ± 2.5
	Romman	2.4 ± 0.3c+	80 ± 19c	3.5 ± 2.0
	Fumari	2.8 ± 0.4c+	105 ± 20c	1.4 ± 1.8

^a p < .05,

^c p < .001 compared to Kentucky reference cigarette.

^–/+ p < .05 for decrease/increase when electric heating compared to charcoal by individual brand.

Cytotoxic Effects of Tobacco Products

Potent cytotoxicity detected by the NRU assay was induced by cigarettes and cigarillos with average IC₅₀s ranging from 17 to 58 µg/ml TPM across products and cell lines (Table 4). The HBEC4 line was most sensitive to the cigarette and cigarillos with individual product IC₅₀s of 13–23 µg/ml TPM (Supplementary Table 1). The H9C2 cell line was more resistant to the toxic effects of these products compared to HBEC4 (Supplementary Table 1). Maximal cytotoxicity exceeded 80% with 128 µg/ml TPM (highest dose studied) for all cigarettes and cigarillos for the 3 cell lines (Table 4, Supplementary Table 1). In contrast, none of the shisha tobaccos evaluated across the same dose range of TPM induced cell toxicity that exceeded 50% with charcoal or the electric bowl heating to generate the TPM (Table 4, Supplementary Table 1).

Table 4.

Differential Induction of Cytotoxicity by Tobacco Products

		Cell Line (Mean ± SD)
Endpoint	Product	MOE1B	HBEC4	H9C2
	Cigarette	31.8 ± 13.3a	19.6 ± 3.4	57.7 ± 8.3b
IC50	Cigarillo	18.4 ± 3.5	17.5 ± 4.9	43.3 ± 16.5b
(µg TPM/ml)	Shisha (Charcoal)	NR	NR	NR
	Shisha (Electric)	NR	NR	NR
	Cigarette	81.1 ± 10.8	82.5 ± 4.9	89.6 ± 6.5
Maximum	Cigarillo	88.7 ± 10.2	88.0 ± 3.3	95.8 ± 5.2
Cytotoxicity1	Shisha (Charcoal)	25.8 ± 9.0c	33.1 ± 13.2c	28.8 ± 4.7c
(%)	Shisha (Electric)	30.0 ± 10.1c	33.1 ± 13.3c	43.5 ± 4.5c

¹Maximum dose studied was 128 µg/ml.

NR, not reached.

^a p < .05,

^b p < .001 compared to HBEC4.

^c p < .001 compared to cigarette.

Induction of Oxidative Stress by Tobacco Products

Cigarettes and cigarillos induced significant dose-dependent increases in oxidative stress assessed by the RosGlo assay in all 3 cell lines with linearity seen beginning at the second dose (64 µg/ml TPM; Table 5). Marlboro and Dutch Masters were the most potent products with an 18–23-fold and 24–36-fold induction seen at the highest level of TPM across cell lines (Figure 1). All shisha products generated with charcoal showed dose-dependent increases in oxidative stress with significance seen at 3 or more doses for most products (Table 6). However, the fold-increases were modest compared to cigarettes and cigarillos with maximal fold changes at the highest TPM dose ranging from 1.5 to 4.2 (Tables 5 and 6). Levels of oxidative stress were generally higher across TPM doses in H9C2 compared to MOE1B and HBEC4 cells for Starbuzz, Fantasia, and Tangiers (Table 6). Oxidative stress was also increased in shisha aerosols generated by electric heating; however, dose-response changes were significantly lower in H9C2 cells compared to charcoal generation for most products (Table 6, Figure 2).

Table 5.

Induction of Oxidative Stress by Cigarette and Cigarillo Products

		Exposure	Cell Line (Fold Change ± SD)
Product	Brand	µg TPM/ml	MOE1B	HBEC4	H9C2
Cigarette	Kentucky 3R4F	32	2.0 ± 0.1c	2.8 ± 0.1c	1.3 ± 0.1
		64	3.8 ± 0.1c	5.0 ± 0.2c	2.0 ± 0.1c
		128	8.0 ± 0.2c	9.4 ± 0.8c	4.2 ± 0.3c
		256	15.2 ± 0.5c↑	19.7 ± 0.7c↑	9.0 ± 0.5c↑
	Camel	32	2.4 ± 0.1	2.2 ± 0.2c	2.1 ± 0.1c
		64	3.4 ± 0.5b	3.4 ± 0.2c	2.6 ± 0.2c
		128	7.6 ± 0.7c	6.8 ± 0.4c	5.4 ± 0.4c
		256	15.3 ± 2.0c↑	12.7 ± 0.3c↑–	9.2 ± 0.2c↑
	Marlboro	32	2.5 ± 0.1a	2.6 ± 0.2c	2.0 ± 0.1b
		64	3.7 ± 0.2c	3.8 ± 0.2c	2.7 ± 0.2c
		128	9.5 ± 0.3c	10.3 ± 0.2c	7.9 ± 0.3c
		256	18.6 ± 1.5c↑+	23.6 ± 0.7c↑+	18.7 ± 0.7c↑+
Cigarillo	Black & Mild	32	1.9 ± 0.1	1.9 ± 0.2c	2.2 ± 0.1a
		64	2.9 ± 0.2b	2.8 ± 0.2c	3.0 ± 0.2c
		128	6.8 ± 0.5c	6.1 ± 0.2c	6.1 ± 0.3c
		256	19.1 ± 1.4c↑+	15.8 ± 0.3c↑–	13.6 ± 1.4c↑+
	Dutch Masters	32	2.7 ± 0.2	2.9 ± 0.2a	2.3 ± 0.2
		64	3.8 ± 0.2	4.1 ± 0.4b	3.6 ± 0.3
		128	7.3 ± 0.8b	7.7 ± 0.3c	7.3 ± 0.7b
		256	36.6 ± 5.8c↑+	24.8 ± 2.3c↑+	29.7 ± 5.8c↑+
	Swisher Sweets	32	2.6 ± 0.3c	2.9 ± 0.2b	2.1 ± 0.1c
		64	3.8 ± 0.1c	3.9 ± 0.2c	2.8 ± 0.1c
		128	6.9 ± 0.1c	6.7 ± 1.0c	4.8 ± 0.4c
		256	16.7 ± 0.6c↑	15.9 ± 0.9c↑–	15.7 ± 0.5c↑+

Fold change calculated by dividing the ROS value at each exposure by average of the DMSO control.

^a p < .05,

^b p < .01,

^c p < .001 compared to DMSO control.

^↑ p < .05 for dose response.

^–/+ p < .05 for dose response decrease/increase compared to Kentucky reference cigarette.

Figure 1.

Dose-dependent increase in ROS activity in MOE1B, HBEC4, and H9C2 cells exposed to increasing amounts of TPM from Marlboro or Dutch Masters. ^ap < .05, ^bp < .01, ^cp < .001 compared to vehicle; ^↑p < .05 for dose response; and ⁺p < .05 for dose response trend increase compared to Kentucky reference cigarette (Table 5).

Table 6.

Induction of Oxidative Stress by Shisha Products

	Exposure	Cell Line (Fold Change ± SD)
Brand$	µg TPM/ml	MOE1B		HBEC4		H9C2
		Charcoal	Electric	Charcoal	Electric	Charcoal	Electric
Al Fakher	32	1.2 ± 0.1	1.0 ± 0.1	1.1 ± 0.1	1.1 ± 0.1	1.4 ± 0.1c	1.2 ± 0.2
	64	1.4 ± 0.1b	1.1 ± 0.1	1.5 ± 0.1b	1.2 ± 0.1	1.8 ± 0.1c	1.1 ± 0.2
	128	1.8 ± 0.2c	1.2 ± 0.1	2.0 ± 0.3c	1.3 ± 0.1a	2.3 ± 0.1c	1.2 ± 0.2
	256	3.3 ± 0.2c↑	1.3 ± 0.3–	3.4 ± 0.2c↑	1.3 ± 0.1b↑–	3.0 ± 0.1c↑	1.4 ± 0.3–
Starbuzz	32	1.0 ± 0.1	1.2 ± 0.2a	1.1 ± 0.1a	1.0 ± 0.2	1.2 ± 0.1a	1.1 ± 0.1
	64	1.0 ± 0.1	1.2 ± 0.1a	1.3 ± 0.1c	1.1 ± 0.1	1.5 ± 0.1c	1.3 ± 0.2
	128	1.2 ± 0.1b	1.3 ± 0.1c	1.6 ± 0.1c	1.3 ± 0.2	2.5 ± 0.1c	1.5 ± 0.2b
	256	1.8 ± 0.2c↑	1.8 ± 0.1c↑	2.2 ± 0.1c↑	1.7 ± 0.2c↑	4.2 ± 0.1c↑	1.7 ± 0.2c↑–
Fantasia	32	1.0 ± 0.1	0.9 ± 0.2	1.1 ± 0.1	1.1 ± 0.1a	1.6 ± 0.1c	1.0 ± 0.2
	64	1.0 ± 0.1	1.0 ± 0.1	1.2 ± 0.1a	1.2 ± 0.1b	2.2 ± 0.1c	1.0 ± 0.1
	128	1.2 ± 0.1a	1.1 ± 0.1	1.4 ± 0.1c	1.4 ± 0.1c	3.0 ± 0.1c	1.3 ± 0.1
	256	2.9 ± 0.2c↑	1.5 ± 0.1c↑	3.6 ± 0.2c↑	1.8 ± 0.1c↑	4.2 ± 0.2c↑	1.7 ± 0.3c↑–
Nakhla	32	1.2 ± 0.1a	0.9 ± 0.1	1.2 ± 0.1	1.3 ± 0.3	1.2 ± 0.1a	1.2 ± 0.1
	64	1.4 ± 0.1c	0.9 ± 0.1	1.3 ± 0.1a	1.5 ± 0.4b	1.5 ± 0.1c	1.4 ± 0.2a
	128	1.9 ± 0.1c	1.2 ± 0.1	1.4 ± 0.1b	1.8 ± 0.4a	1.9 ± 0.2c	1.8 ± 0.2c
	256	2.6 ± 0.2c↑	1.3 ± 0.1a↑-	2.0 ± 0.1c↑	2.1 ± 0.6b↑	2.9 ± 0.1c↑	2.4 ± 0.4c↑
Tangiers	32	1.1 ± 0.1	1.1 ± 0.1b	1.2 ± 0.1a	1.2 ± 0.1c	1.3 ± 0.1b	1.2 ± 0.1a
	64	1.2 ± 0.1a	1.2 ± 0.1c	1.3 ± 0.1c	1.3 ± 0.1c	1.4 ± 0.1c	1.5 ± 0.1c
	128	1.3 ± 0.1c	1.2 ± 0.1c	1.3 ± 0.1c	1.4 ± 0.1c	2.0 ± 0.1c	1.3 ± 0.1c
	256	1.5 ± 0.2c↑	1.4 ± 0.1c↑	1.9 ± 0.1c↑	1.6 ± 0.1c↑	2.8 ± 0.1c↑	1.7 ± 0.1c↑–
Nirvana	32	1.2 ± 0.2	1.3 ± 0.1b	1.1 ± 0.1	1.0 ± 0.1	1.2 ± 0.1b	1.2 ± 0.2
	64	1.3 ± 0.2a	1.3 ± 0.1c	1.1 ± 0.1	1.0 ± 0.1	1.6 ± 0.1c	1.1 ± 0.1
	128	1.3 ± 0.1a	1.3 ± 0.1c	1.3 ± 0.2b	1.2 ± 0.1a	2.1 ± 0.1c	1.4 ± 0.1b
	256	2.1 ± 0.2c↑	1.6 ± 0.1c↑	1.8 ± 0.2c↑	1.6 ± 0.1c↑	3.3 ± 0.2c↑	1.6 ± 0.1c↑–
Romman	32	1.0 ± 0.2	1.3 ± 0.1b	1.2 ± 0.1a	1.3 ± 0.2b	1.3 ± 0.1c	1.1 ± 0.1
	64	1.2 ± 0.1a	1.4 ± 0.1b	1.5 ± 0.1c	1.3 ± 0.1a	1.8 ± 0.1c	1.2 ± 0.1b
	128	1.4 ± 0.1c	1.4 ± 0.1c	1.8 ± 0.1b	1.5 ± 0.1c	2.2 ± 0.1c	1.2 ± 0.1c
	256	2.0 ± 0.1c↑	2.5 ± 0.2c↑	2.9 ± 0.2c↑	2.1 ± 0.1c↑	2.8 ± 0.1c↑	1.5 ± 0.1c↑–
Fumari	32	0.9 ± 0.1	1.1 ± 0.1	1.0 ± 0.1	1.1 ± 0.1	1.1 ± 0.1	1.2 ± 0.1a
	64	1.0 ± 0.1	1.2 ± 0.1a	1.1 ± 0.1	1.2 ± 0.1a	1.2 ± 0.1b	1.3 ± 0.1c
	128	1.1 ± 0.1	1.2 ± 0.1b	1.1 ± 0.1a	1.2 ± 0.1a	1.2 ± 0.1c	1.3 ± 0.1c
	256	1.3 ± 0.1c↑	1.3 ± 0.1c↑	1.4 ± 0.1c↑	1.4 ± 0.2c↑	1.4 ± 0.1c↑	1.8 ± 0.1c↑

Fold change calculated by dividing the ROS value at each exposure by average of the DMSO control.

^a p < .05,

^b p < .01,

^c p < .001 compared to DMSO control.

^↑ p < .05 for dose response.

^– p < .05 for dose response decrease comparing electric to charcoal by individual brand.

^$All Shisha test articles had dose response significantly decreased (p < .05) compared to Kentucky reference cigarette.

Figure 2.

Comparison of ROS activity across MOE1B, HBEC4, and H9C2 cells exposed to increasing amounts of TPM generated by charcoal versus electric heating of Fantasia shisha tobacco. ^ap < .05, ^bp < .01, ^cp < .001 compared to vehicle; ^↑p < .05 for dose response; and ^–p < .05 for dose response decrease comparing electric to charcoal.

There was considerable heterogeneity for dose-response and level of induction of lipid peroxidation across cell lines and by the cigarette and cigarillo products (Table 7). Camel and Black & Mild were the most potent products and showed the greatest effect in the HBEC4 cells with 27- and 19-fold increases in lipid peroxidation seen at the highest dose (Table 7). In general, linearity for increased lipid peroxidation was seen beginning at the second or third dose of the TPM for these products. In contrast, H9C2 cells appeared most resistant, only Black & Mild induced lipid peroxidation that exceeded 3-fold. No significant dose-dependent increases in lipid peroxidation were seen in shisha TPMs generated by charcoal, although Al Fakher induced significant lipid peroxidation at 2 or more doses across the 3 cell lines (Supplementary Table 2). Similar findings were also observed with electric heating with exception to Nirvana that caused significant dose-dependent increases in MOE1B and HBEC4 cells (Supplementary Table 2).

Table 7.

Induction of Lipid Peroxidation by Cigarette and Cigarillo Products

		Exposure	Cell Line (Fold Change ± SD)
Product	Brand	µg TPM/ml	MOE1B	HBEC4	H9C2
Cigarette	Kentucky 3R4F	32	1.0 ± 0.1	1.1 ± 0.3	1.8 ± 0.1a
		64	1.6 ± 0.2a	1.6 ± 0.1	1.3 ± 0.3
		128	1.5 ± 0.2a	1.2 ± 0.1	1.3 ± 0.1
		256	3.1 ± 0.3c↑	3.2 ± 0.3b↑	1.5 ± 0.1
	Camel	32	1.1 ± 0.1	3.0 ± 0.1c	0.9 ± 0.1
		64	1.1 ± 0.1	4.6 ± 0.2c	1.0 ± 0.1
		128	1.2 ± 0.2	10.9 ± 2.8c	1.1 ± 0.1
		256	3.1 ± 0.6c↑	27.0 ± 1.6c↑+	1.2 ± 0.1
	Marlboro	32	0.9 ± 0.2	1.2 ± 0.3	1.0 ± 0.1
		64	1.1 ± 0.4	1.2 ± 0.1	1.0 ± 0.1
		128	1.4 ± 0.3	2.1 ± 0.3b	1.1 ± 0.1
		256	1.5 ± 0.1–	3.3 ± 0.7b↑	1.6 ± 0.2b↑
Cigarillo	Black & Mild	32	3.1 ± 1.1b	1.1 ± 0.1	0.7 ± 0.2
		64	3.7 ± 0.3b	5.7 ± 2.3c	0.9 ± 0.1
		128	9.6 ± 2.5c	8.1 ± 0.7c	1.1 ± 0.1
		256	15.4 ± 2.0c↑+	19.3 ± 2.3c↑+	1.5 ± 0.6
	Dutch Masters	32	1.2 ± 0.1a	1.1 ± 0.1	1.2 ± 0.2
		64	1.3 ± 0.1a	1.2 ± 0.3	1.2 ± 0.5
		128	2.2 ± 0.2c	1.5 ± 0.1	1.2 ± 0.4
		256	3.0 ± 0.3c↑	2.9 ± 0.6b↑	2.6 ± 1.5
	Swisher Sweets	32	4.4 ± 1.7b	0.3 ± 0.1	1.3 ± 0.1
		64	2.1 ± 0.6a	0.7 ± 0.1	1.9 ± 0.2b
		128	2.9 ± 0.4a	1.4 ± 0.5	2.2 ± 0.2b
		256	9.8 ± 1.3c↑	3.2 ± 0.3a↑	4.1 ± 0.8c↑+

Fold change calculated by dividing the TBARs value at each exposure by average of the DMSO control.

^a p < .05,

^b p < .01,

^c p < .001 compared to DMSO control.

^↑ p < .05 for dose response.

⁺ p < .05 for dose response decrease/increase compared to Kentucky reference cigarette.

Genotoxic Effects of Tobacco Products

Strong similar dose response increases in mutagenicity detected by the Ames assay were seen with the TA98 strain in the presence of S9 across all cigarette and cigarillo products compared to the DMSO control (Table 8). For some products, toxicity was apparent at the highest TPM dose (bacterial growth was reduced) leading to a reduction in number of revertants compared to the next lower dose. There was no significant mutagenicity detected in this strain in the absence of S9. Mutagenicity was also largely not seen with the TA100 strain irrespective of S9 status for these tobacco products with the exception of Black & Mild. With respect to shisha tobacco products, only sporadic mutagenicity at 1 dose was observed irrespective of S9 status for TPM generated by charcoal, while electric generated TPM products were all negative (Supplementary Table 3).

Table 8.

Induction of Genotoxicity by Cigarette and Cigarillo Products

		Exposure	TA98		TA100
Product	Brand	µg TPM/ml	+S9	-S9	+S9	- S9
Cigarette	Kentucky 3R4F	0	1	1	5	7
		10	3	0	6	8
		40	9b	0	9	2
		130	21c	1	8	4
		400	9b	2	9	3
	Camel	0	0	3	3	7
		10	8b	8	5	12
		40	27c	8	3	7
		130	31c	6	13b	11
		400	34c↑	7	8	6
	Marlboro	0	1	1	5	7
		10	4	0	3	6
		40	17c	3	6	5
		130	24c	1	14a	3
		400	10b	0	11	2
Cigarillo	Black & Mild	0	0	3	3	7
		10	2	6	4	8
		40	14c	6	7	10
		130	30c	5	9a	10
		400	32c↑+	7	11a↑	7
	Dutch Masters	0	0	1	7	12
		10	8b	4	4	9
		40	17c	4	6	11
		130	26c	5a	8	6
		400	13c	4	7	10
	Swisher Sweets	0	1	1	5	7
		10	9c	0	1	4
		40	17c	0	7	7
		130	24c	4	8	2
		400	22c↑+	1	6	6

^a p < .05,

^b p < .01,

^c p < .001 compared to DMSO control.

^↑ p < .05 for dose response.

⁺ p < .05 for dose response decrease/increase compared to Kentucky reference cigarette.

All cigarettes induced significant dose-dependent increases in DNA damage detected by the Comet assay in the 3 cell lines. The fold changes were comparable across products and cell lines with 4- to 8.5-fold increases seen at the highest TPM dose compared to the DMSO control (Table 9). All cigarillo products also induced significant levels of DNA damage that increased with dose. Black & Mild and Dutch Masters were more genotoxic than Swisher Sweets and all cigarette products for inducing DNA damage (Table 9, Figure 3). In marked contrast, TPM generated by charcoal heating for only one shisha product, Fantasia, induced DNA damage in HBEC4 (Supplemenatry Table 4). Similarly, only TPM generated by electric heating for Tangiers showed positivity in the MOE1B cell line.

Table 9.

Induction of DNA Damage by Cigarette and Cigarillo Products

		Exposure	Cell Line (Fold Change ± SD)
Product	Brand	µg TPM/ml	MOE1B	HBEC4	H9C2
Cigarette	Kentucky 3R4F	32	2.5 ± 0.2c	2.9 ± 0.3c	1.4 ± 0.1
		64	4.5 ± 1.7c	4.9 ± 0.1c	1.5 ± 0.4a
		128	3.4 ± 0.4c	5.0 ± 0.4c	2.1 ± 0.3b
		256	7.5 ± 0.6c↑	8.5 ± 0.4c↑	6.6 ± 0.1c↑
	Camel	32	2.3 ± 0.3c	2.5 ± 0.1c	2.0 ± 0.2
		64	3.1 ± 0.3c	3.5 ± 0.1c	1.4 ± 1.4
		128	3.6 ± 0.6c	5.4 ± 0.7c	1.9 ± 0.4
		256	5.1 ± 1.1c↑	6.7 ± 0.8c↑	3.9 ± 0.8b↑
	Marlboro	32	2.9 ± 0.1c	2.8 ± 0.3c	1.3 ± 0.1a
		64	4.1 ± 0.5c	4.1 ± 0.8c	1.7 ± 0.2b
		128	5.6 ± 0.8c	5.7 ± 0.9c	4.6 ± 0.3c
		256	7.4 ± 0.4c↑	6.9 ± 0.1c↑	7.4 ± 0.1c↑
Cigarillo	Black & Mild	32	4.3 ± 0.7c	4.1 ± 0.2c	2.0 ± 0.4c
		64	6.1 ± 1.0c	6.1 ± 0.7c	2.5 ± 0.1c
		128	6.8 ± 0.7c	7.7 ± 0.6c	3.2 ± 0.8c
		256	12.1 ± 0.7c↑	11.0 ± 0.2c↑	9.2 ± 0.3c↑
	Dutch Masters	32	6.3 ± 0.3c	5.6 ± 0.2c	3.3 ± 0.5b
		64	9.6 ± 1.4c	8.3 ± 1.1c	4.7 ± 1.3c
		128	10.2 ± 0.7c	9.0 ± 0.3c	34.9 ± 0.6c
		256	12.3 ± 0.4c↑	13.0 ± 0.3c↑	35.6 ± 3.3c↑+
	Swisher Sweets	32	2.3 ± 0.2c	1.7 ± 0.1c	1.0 ± 0.2
		64	3.3 ± 0.5c	2.5 ± 0.4c	1.2 ± 0.1
		128	3.0 ± 0.3c	2.9 ± 0.1c	1.6 ± 0.1b
		256	4.9 ± 0.6c↑	5.1 ± 0.8c↑	4.8 ± 0.2c↑

Fold change calculated by dividing the Comet value at each exposure by average of the DMSO control.

^a p < .05,

^b p < .01,

^c p < .001 compared to DMSO control.

^↑ p < .05 for dose response.

^–/+ p < .05 for dose response decrease/increase compared to Kentucky reference cigarette.

Figure 3.

Dose-dependent increase in DNA damage detected by the Comet assay. MOE1B, HBEC4, and H9C2 cells were exposed to increasing amounts of TPM from Black & Mild or Dutch Masters tobacco products. ^bp < .01, ^cp < .001 compared to vehicle, and ^↑p < .05 for dose response.

DISCUSSION

These studies provide new insight regarding the presence of high levels of the carcinogenic carbonyls, formaldehyde, and acrolein in shisha tobacco product aerosols that greatly exceed that seen in cigarettes and cigarillos. Increased potency of TPM from cigarillos compared to cigarettes for inducing oxidative stress via reactive oxygen radicals and lipid peroxidation across oral, lung, and cardiac cell lines was also evident, while positivity was seen for shisha products albeit at much lower levels. Equivalent cytotoxicity and dose response for genotoxicity through induction of mutagenicity and DNA strand breaks with positivity at the lowest dose of TPM evaluated was seen for cigarette and cigarillo products. Together, these findings reinforce the potential for cigarillos to cause cardiopulmonary health effects, while the likely deleterious genotoxic effect of carbonyls found largely in the gas phase of aerosols supports future studies to fully capture the harmful effects of the shisha tobacco products.

The shisha tobacco is approximately one-third tobacco along with humectants and flavorings that account for the lower level of nicotine seen for many of the products characterized and the low to undetectable amounts of the TSNAs (Shihadeh, 2015). PAHs, of which 90% originate in the charcoal, were also significantly lower or not detected in shisha tobacco aerosols most likely due to collecting only 2 puffs for analysis and the large dilution stemming from interaction of the coal combustion products with the heated shisha to generate vapors that cool and recondense to ultimately form the inhalable vapor (Monzer et al., 2008). However, PAHs at levels that often exceed that within aerosol from a single smoked cigarette were detected over a 1-h puffing regimen (Beirut) with charcoal as the heating source to mimic exposures in a Hookah/Shisha bar (reviewed in Shihadeh et al., 2015). In contrast, with the same design using an electrical heating element there was an 80% reduction in PAHs compared to charcoal mediated generation (Hourani, 2019).

Electrical heating elements are being marketed for waterpipe use as a reduced harm charcoal substitute. While this may be the situation with generation of PAHs, this was clearly not the case for some carbonyls such as acrolein whose levels were increased for all shisha tobaccos and up to 400- and 40-fold (Starbuzz) compared to charcoal generation and cigarettes or cigarillos, respectively. Levels of formaldehyde were also increased 40- to 120-fold across shisha tobacco product aerosols generated with charcoal compared to cigarettes and cigarillo, while electric generation augmented this carbonyl further in some products. Significant increases in crotonaldehyde and diacetyl were also seen across shisha products with electrical heating. The increased levels of these carbonyls with electric generation can be explained by Reilly et al. (2018) that showed the addition of an in-line charcoal filter placed downstream of a cigarette could reduce delivery of these toxicants into the smoke aerosol. Collectively, these results underscore the likely increased harm that electrical heating may facilitate. Our studies of oxidative stress, lipid peroxidation, and genotoxicity were conducted with TPM that contains <5% of most carbonyls, thus precluding inclusion of their contribution to these outcomes (Pang and Lewis, 2011). Acrolein can induce DNA strand breaks and crosslinks to form 2 pairs of 1, N²-propano-dexyguanosine adducts; and increase levels of reactive oxygen species leading to formation of the 8-oxo-,8-2′-deoxyguanosine adduct (Li et al., 2008; Zhang et al., 2007). The α-OH-Acr-deoxyguanosine adduct is more mutagenic and induces mainly G–>T transversions that were preferentially formed in mutational hotspots for the p53 tumor suppressor gene in human lung cancers and lung tissue from smokers (Feng et al., 2006; Zhang et al., 2007). Co-exposure to formaldehyde and acrolein mixtures at individual doses below no effect levels have been shown to induce significant cytotoxicity and genotoxicity in lung A549 cells (Zhang et al., 2018). Finally, 114 volatiles were identified in 13 flavored shisha tobacco products comprising major classes of potentially harmful constituents such as esters, alcohols, and aldehydes that likely contribute to the increased levels of carbonyls compared to cigarettes and cigarillos (Farag et al., 2018). Our new findings regarding high levels of hazardous carbonyls in shisha tobacco aerosols should help dispel the misconception that this product is a safer alternative to cigarettes (Palamar et al., 2014).

In vitro toxicology tests have been routinely used to assess smoke-induced toxicological effects and the battery recommended includes NRU, Ames, micronucleus, and TK-lymphoma assays (CORESTA In vitro Toxicology Task Force, 2004). Our recent study comparing 10 conventional cigarettes and 10 cigarillos using TPM from these products found greater cytotoxicity and genotoxicity by cigarillos (Crosby, 2021). The current study extended this work to include assays of oxidative stress, and DNA damage in an oral, bronchial airway, and a cardiac cell line to provide more insight into potency for cigarillos for causing cancer, COPD, and cardiovascular effects. These endpoints along with the NRU and Ames assays were extended to shisha tobacco products where little in vitro toxicology data exists. Potent mutagenicity similar to that observed in our prior study was seen for cigarettes and cigarillos (Crosby et al., 2021). The lack of mutations detected by the Ames assay for shisha tobacco product TPMs is most likely due to their low levels of PAHs and aromatic amines (Shihadeh et al., 2015).

Oxidative stress and associated DNA damage are an important component for initiating cardiopulmonary diseases associated with chronic tobacco use (Ayaori et al., 2000; Faux et al., 2009; Lourenco et al., 2018). Cigarillos produce even more chemicals than cigarettes and this may account for their potency for inducing oxidative stress and DNA damage detected by the TBARs, ROS-Glo, and Comet assays that were equivalent or exceeded that seen for the 3 cigarettes across all cell lines (Ghosh et al., 2017). In contrast, the low level of oxidative stress and lack of DNA damage seen by the shisha tobacco products likely stems from the lack of mutagenic chemicals present in the TPM. However, waterpipe smokers show increased levels of DNA damage and stress through detection of the 8-hydroxy-2′-deoxyguanosine adduct in plasma and reduced expression of DNA repair genes (Alsaad et al., 2019). In addition, chronic exposure of mice for 6 weeks to shisha tobacco aerosol increased recruitment of inflammatory cells to the airways and altered levels of cytokines in lung tissue (Khabour et al., 2018). These effects are most likely mediated by the high levels of carbonyls such as acrolein present in the gas phase of the aerosol that can inflict oxidative stress and DNA damage to lung tissue. Mice exposed nose-only 30 min a day for 5 days to shisha aerosol showed increases in cardiac levels of lipid peroxidation, interleukin 6, and tumor necrosis factor-α (Nemmar et al., 2015). These effects likely manifest from carbonyls in the gas phase that can also pass through the airway epithelial barrier to enter the systemic circulation via the pulmonary circulation to stress endothelial and cardiac cells (Conklin et al., 2017; Horinouchi et al., 2016). Thus, it is likely that the high levels of the toxic and carcinogenic carbonyls present in gas phase of the shisha tobacco aerosol will cause genotoxic and oxidative stress.

Our in vitro exposures to characterize the chemical profiles, oxidative stress, and DNA damage potential for cigarillos and shisha tobacco products have provided valuable new insight into the health risks associated with these perceived reduced-harm products. There are clear differences in smoking behavior that will likely impact relative risk for diseases associated with these tobacco products. Waterpipe smoking occurs often in a social environment within a hookah bar, thus frequency of use is lowest for this product. Cigarillos are often consumed daily but at lower frequency (3–4/day) compared to cigarettes (10–20/day). All these products are inhaled and dose with respect to smoke volume is greatest for shisha, while genotoxicity is greatest for cigarillos. The extension of our approaches to in vitro and chronic in vivo dose-response studies evaluating the complete tobacco aerosol (particulate fraction and gases) in pulmonary, endothelial, and cardiac target cells for disease manifestation and rodent animal models should provide the needed comprehensive data to support implementation of new evidence-based regulatory policies for these products.

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.

AUTHORS’ CONTRIBUTIONS

S.A.B. designed the studies with input from C.S.T., S.K., and T.K. C.S.T., D.E.J., L.M.P., K.D., C.L.T., R.W., H.I., and Y.Z. conducted the studies. G.W. and W.W.D. conducted the analyses. S.A.B. wrote the manuscript and all authors reviewed the manuscript and provided feedback.

FUNDING

This work was supported largely by National Institute of Health grant (R01 ES029448 to S.A.B.) and in part by (P30CA11800 to C. Willman) from the National Institutes of Health.

DECLARATION OF COMPETING INTERESTS

The authors declared no conflicts of interest with respect to the research, authorship, and/or publication of this article.