Contributions of charcoal, tobacco, and syrup to the toxicity and particle distribution of waterpipe tobacco smoke.

Abstract

The use of waterpipes in the United States is increasing in a largely unregulated market. The shisha smoked in a waterpipe is a complex matrix of tobacco, flavorings, and humectant with smoke generated by an external heat source. This study explored the relationship between shisha components and the particulate matter size distributions and toxicity of smoke generated with heating. Standard waterpipe puff topography of charcoal- or electronic- heated whole shisha and shisha components generated smoke particulate matter that was characterized using a TSI Engine Exhaust Particle Sizer. Relative toxicity of the whole smoke was determined via measurement of lysosomal integrity and measures of membrane integrity following acute exposure of type II alveolar cells at the air-liquid interface. All waterpipe aerosols exhibited a unimodal particle size distribution, the peak and concentration of which varied depending upon the shisha components present. Acute exposure to charcoal-heated whole shisha, flavoring syrup, or humectant smoke, or electronic-heated whole shisha smoke caused significant alveolar cell damage and death, indicating neither tobacco nor charcoal are needed for these cytotoxic effects to occur.

1. INTRODUCTION

During the first two decades of the 21^st century, waterpipe usage has increased sharply in the United States and around the world, even though cigarette usage has decreased.(Maziak et al., 2015) Producers and advertisers of waterpipes have taken advantage of predominantly lenient, and often non-existent advertising laws to market waterpipes as healthy and natural alternatives to cigarettes.(Khalil et al., 2009) Waterpipe’s rise in US popularity followed the mass production of shisha, a mixture of shredded tobacco, humectant and flavoring. (Maziak et al., 2015) The commercial availability of hundreds of flavors such as mint, fruit, and other desserts has been recognized as a strong enticement to new users.(Jawad et al., 2015; Shihadeh et al., 2015) However, relatively little is known about how these non-tobacco components of shisha contribute to the composition of waterpipe tobacco smoke (WTS) and its toxicity.

Shisha is not the only factor that contributes to WTS composition. In fact, bulk characterization of WTS indicates that charcoal contributes significant toxicants to mainstream smoke during a typical waterpipe smoking session.(Shihadeh et al., 2015) Today, the most commonly used heat source is charcoal derived from coconut hulls which has been demonstrated to contribute heavy metals, polycyclic aromatic hydrocarbons (PAHs), CO, and ash to WTS.(Elsayed et al., 2016) Other novel heat sources, including heated bowls and electronic charcoal (electronic) are harder to obtain but have the potential to eliminate charcoal’s contribution to the composition of WTS and warrant further research.

Targeted investigations for individual chemicals within WTS provide one avenue of characterizing the smoke’s potential to cause harm, i.e. chemical toxicity. An additional avenue is through the characterization of WTS particulate matter (PM). Regardless of chemical composition, the diameter of PM provides information about how far aerosolized particles can penetrate into the human respiratory system and how long they may persist in that compartment, quantified as lung deposited surface area (LDSA). Since the narrowest ducts in the lungs have diameters of approximately 12 μm, particulate matter ≤ 2.5 μm (PM_2.5) and ≤ 100 nm (ultrafine particles; UFP) can efficiently penetrate into alveoli and may cross tight junctions to enter the circulatory system.(Whimster, 1970) Of the over 300 compounds identified in WTS, at least seven are carcinogenic and many are depressants or irritants; therefore, characterizing the particulate matter profile of the aerosols carrying these compounds is critical to understanding what may be introduced into the user’s body.(Elsayed et al., 2016) In addition, the physical properties of WTS affect the extent of damage caused by particles impacting lung cells and blocking gas exchange, i.e. physical toxicity. Previous efforts to characterize the physical properties of hookah smoke have produced inconsistent results, with mean particle diameters as low as 0.04 μm and as high as 0.24 μm.(Becquemin et al., 2008; Markowicz et al., 2014; Monn et al., 2007)

This study characterizes how the heating mechanism and shisha constituents impact the particle size distribution, concentration and toxicity of WTS. To study how different ingredients, contribute to the smoke, a non-reactive matrix (NRM) and whole shisha were smoked with both synthetic and electronic charcoal and NRM coated with glycerol or extracted shisha syrup were smoked with charcoal. Physical properties of the particulate in the smoke, including number concentration, size distribution, mass and LDSA, were determined, and the effect of acute exposure of rat alveolar Type II cells to the smoke was measured via membrane integrity and lysosomal function metrics for each condition. Ultimately, the information gathered serves to inform regulation of shisha constituents and charcoal as a heating method.

2. MATERIALS AND METHODS

2.1. Equipment.

A TSI Engine Exhaust Particle Sizer (EEPS) (Model 3090, TSI, Shoreview, MN, USA) was used to measure the size distribution of particulate matter with diameters between 5.6 and 560 nm in 32 bins at a sampling rate of 10 Hz and a flow rate of 10 liters per minute (LPM). A TSI Environmental Particle Counter (EPC) (Model 3783, TSI, Shoreview, MN, USA) was used to measure the total aerosol particle concentration of particles between 7 nm and 10 μm at a flow rate of 3.0 L/min and a sampling rate of 1 Hz. Both instruments reported concentrations in particles/cm³. 35 mm Starbuzz Coconut Shell Instant Light charcoal and a multivolt resistor element from a Master Heat Gun, controlled through a powerstat to provide a temperature profile that matched the charcoal, were used as the heat source for charcoal and electronic smoking trials, respectively. The hookah pipe used for all trials was an Inhale Junior 1 CP hookah pipe (36 cm pipe height) filled with distilled water, 530 ml, maintaining 100 ml head space, equipped with a 183 cm hose (I.D. ½” Tygon 2375).

2.2. Preparation of Head Contents.

10 g Starbuzz^® Exotic Apple Americano shisha (Hookah-Shisha.com) was used for shisha trials. Constituents without tobacco required a non-reactive substrate approximating tobacco in volume and absorptive capacity. For these trials, approximately 0.5 g of shredded glass fiber filters (Filter #61631, Pall Corporation, Port Washington, NY, USA) and 1 g of shredded Teflon (Virgin PTFE 0.15”: Fluoro-Plastics Inc.), were used as the non-reactive matrix (NRM). Trials were performed with charcoal ± NRM to confirm that the NRM does not contribute to the physical or chemical properties of the particulate generated during the smoking session. In the case of glycerol+NRM trials, the head was prepared with 1.5 g NRM and 8.0 g of 92% glycerol in water (Fisher Scientific, USA).

Syrup trials consisted of 8.0 g of shisha syrup and 1.5 g NRM. Shisha syrup was prepared following an adapted protocol from Clutterbuck, et al.(Clutterbuck et al., 2015) Briefly, 10 g Starbuzz^® Exotic Apple Americano shisha and 20 mL, 60°C dH₂O were mixed, followed by three cycles of shaking vigorously by hand for 1 minute, vortexing for 2 minutes and heating to 60°C for 2 minutes. The contents were vacuum filtered with 10 mL of warm dH₂O to remove large tobacco particulate, and the filtrate passed through Whatman #1 filter paper to remove remaining tobacco particles. This final filtrate was allowed to evaporate, rehydrated with approximately 0.9 mL dH₂O to the viscosity of the original syrup and mixed with the NRM to approximate shisha.

2.3. Smoking Session.

Before each session the bowl was filled with 530 mL dH₂O, a clamp was used to secure the bowl to the neck, and Parafilm (PM-992, Bemis, Neenah, WI, USA) was used to secure the Tygon hose to the pipe. After filling the head, a sheet of aluminum foil perforated with 18 holes was placed on top.(Shihadeh, 2003) The preheated charcoal or electronic heat source was placed on top of the foil layer. The apparatus was dismantled and cleaned between trials.

2.4. Characterizing the physical properties of WTS particles.

Smoke was drawn into a 5L plexiglass box (the volume of an adult male lung), before sampling by the instruments (Fig 1). A 3-way polyvinyl chloride (PVC) valve was used to switch between respiration of filtered air and respiration of waterpipe smoke, following a puff pattern variant of the commonly used Beirut method:(Shihadeh, 2003) 64 cycles of 3 sec of 12 L/min draw on the pipe (‘puff’), 10 sec of 12 L/min particle sampling by the EEPS and EPC while drawing filtered air, and 15 sec of 12 L/min draw of filtered air by house vacuum to purge the ‘lung’ during the interpuff interval, representing a 30-minute smoking session. An adjustable, calibrated dilution bridge containing three filters (HEPA Capsule Filter #1602051, TSI, Stillwater, MN, USA) in series and filtered nitrogen was placed in line between the 5L ‘lung’ and the instruments to maintain concentrations below saturation. Flow rates were measured using TSI 41403 flow meters. Trials ± humidification did not yield differences in particle distribution during the sampling period. Humidification was, therefore, not incorporated for the particulate analysis trials, which focused on puff characteristics and not particle aging.

Figure 1.

Experimental Apparatus. Smoke was generated for particle physical characterizations and ALI cell exposures separately using apparatuses with common components (left) and specific components (right).

2.5. Particle data analysis.

A dilution factor was calculated for each session using the flow rates through each filter on the calibrated dilution bridge and the nitrogen flow needed to maintain concentrations below saturation. EPC data were adjusted for dilution by multiplying by the appropriate dilution factor. For the EEPS data, three-second concentration maxima corresponding to the 3-second hookah smoke puffs were identified throughout each trial, added to produce a total puff concentration for each bin size during the session and multiplied by the appropriate dilution factor.

2.6. Cell maintenance and preparation.

Rat alveolar type II (L2) cells (ATCC#: CCL-149) were grown in 75 cm² flasks at 37° C, 5% CO₂ with 4.5 g/L Dulbecco’s Modified Eagle Medium (DMEM) +10% fetal bovine serum (FBS) +1x Antibiotic-Antimycotic (Gibco™ #15240062). This cell line was chosen because sex and cell lineage may affect response to lung toxicants and, unlike other available alveolar lines, CCL-149 cells are from a female animal and are not derived from a carcinoma. (Chalfant and Bernd, 2014) Cultures were passaged upon reaching 80-100% confluence and discarded after 10 passages. Cells were prepared for air-liquid interface (ALI) exposure by seeding 12-well Corning Transwell™ plates with 7.5×10⁵ cells in Low Glucose media (LGM: (1 g/L) DMEM, 10% FBS, 1x Antibiotic-Antimycotic (Gibco™#15240062)) and incubating 48 hours at 37 °C, 5% CO₂ before exposures.

2.7. Alveolar cell WTS exposure.

Each treatment included simultaneous mock-exposed and exposed conditions. Before a treatment, cells were washed with 300 μL Hanks buffered saline solution (HBSS) (BioWhittaker, #10-527F). Mock-exposed cells remained in 37 °C, 5% CO₂ with 1 mL basolateral LGM and served to set maximum response levels for that specific exposure. Exposed cells were placed in 37° C British American Tobacco (BAT)-style exposure chambers (alveoli samplers) containing 20 mL LGM on the basolateral side only (Fig. 1). Exposed cells were subjected to 0.045L/min humidified ambient air (non-smoking control) or smoke drawn from the 5L aluminum ‘lung’ chamber following a modified Beirut puff topography of 171 cycles of 3 sec of 12.3 L/min drawn through the pipe (‘puff’) plus 17 sec interpuff interval (IPI). To model comingling of smoke and ambient air while breathing, 5 L/min of humidified air was drawn into the ‘lung’ during in the IPI. Immediately following each puff, 3 sec of 0.045 L/min commingled smoke was drawn from the 5 L ‘lung’ into the alveoli samplers. The alveoli sampler received 0.045 L/min humidified ambient air during the intervening 17 sec intervals. After 30 minutes, a new piece of preheated charcoal replaced the old charcoal. When the exposure was complete, all cell inserts were returned to 37° C, 5% CO₂ for a 24-hr recovery period with apical air and 1 mL basolateral LGM. Each exposure included six technical replicates. Data from three biological replicates per exposure condition are presented. Ambient air drawn through a waterpipe containing bowl water but an empty head was defined as the ‘non-smoking’ referent condition.

2.8. Measuring cellular health and viability metrics.

Changes in membrane integrity were determined by following cleavage of GF-AFC or fluorescein diacetate (FDA) by cytoplasmic proteases or esterases, respectively. The GF-AFC assay (CellTiter Fluor; Promega) was performed per manufacturer’s instructions and fluorescence determined at ex. 380 nm; em. 505 nm (Synergy H1 Hybrid Reader, BioTek). To measure fluorescein fluorescence filters were washed with HBSS and cells were incubated with 10 μM FDA in HBSS at 37 °C, 5% CO₂. The cells were washed with HBSS and fluorescence determined (ex 485 nm; em 528 nm: Synergy H1 Hybrid Reader, BioTek).

Changes in lysosomal activity were determined via neutral red dye uptake (NRU) as per Thorne et al. adapted to accommodate cells grown on permeable filters with larger surface area.(Thorne et al., 2015) Dye incorporation in functional lysosomes was determined by measuring Abs_540nm (Bio-Rad 680 Microplate Reader).

Values were background adjusted and normalized to mock-exposed samples within an exposure. If greater than three biological replicates were collected for a treatment condition, a random number generator (Excel) was used to determine which biological replicates would be included in further analysis. One-way ANOVA and Tukey HSD post hoc testing were performed to reveal differences between treatments. T-tests were utilized to determine whether a treatment condition’s damage satisfied the International Organization for Standardization (ISO)/National Institutes of Health (NIH) standard for cytotoxicity (≤70% of viability seen in non-smoking exposure). Statistical analyses were performed in RStudio statistical software. p≤0.05 was considered significant.

3. RESULTS

3.1. Waterpipe Tobacco Smoke

A 30-minute session with shisha and 35 mm charcoal generates 6×10¹³ particles with diameters between 5.3 nm and 530 nm and a mean diameter of 130 nm (Fig 2; corresponding values for mass and LDSA provided in Table 1). The calculated particulate mass is comparable to that observed by Shihadeh 2003 for total particulate matter (TPM) via filter measurements, given the variations in smoking protocol. Mean particle diameters observed in this study are significantly higher than those measured by Monn, 2007 and Marcowicz, 2014, possibly due to evaporation effects in those studies. The particle distribution is indicative of primary nucleation followed by agglomeration and hydroscopic growth. By extrapolation, the estimated particle dose per session in the alveolae sampler would be 1.3×l0¹⁴. However, due to differences between puffing protocols used in our physical properties and cellular exposure studies the true cellular particle dose is higher. In our physical property studies, to understand differences in particles generated with each puff, ‘lung’ contents were exchanged between puffs and sessions were 30 min. In the alveolar exposure studies, as in a living lung, puffs contents and ambient air mixed and particles accumulated during a 57-min smoking session. Therefore, the actual particle dose delivered to cells during a session is greater than the estimate provided above.

Figure 2.

Particle Size Distributions for Shisha smoked with Charcoal and Electronic heating (primary axis) and Charcoal with the non-reactive matrix (secondary axis).

Table 1.

Particle physical characteristics for each condition, per session.

	Particle Conc.	Mean particle diam. (nm)	Mass (mg/m3)	LDSA (mm2/cm3)
Charcoal+Shisha	6.0 × 1013	130	6,000	2 × 107
Electronic+Shisha	6.7 × 1012	160	1,000	3 × 106
Charcoal+NRM	2.4 × 1012	40	7.8	2 × 105
Charcoal+Syrup	1.3 × 1014	130	12,000	4 × 107
Charcoal+Glycerol	4.3 × 1013	120	3,000	1.2 × 107

3.2. Contributions of heat source to WTS

3.2.1. Smoke physical characteristics

To investigate the contribution of charcoal to WTS, the waterpipe was smoked with charcoal as the heating source and non-reactive matrix (NRM) in the head (Fig 2). Charcoal smoke contains approximately 2.4×10¹² total particles with a mean diameter of 40 nm (Table 1). The contribution of charcoal particulate to WTS is further verified by smoking shisha with an electronic heat source that models the temperature profile of charcoal. The electronic-generated WTS has a higher mean particle diameter than charcoal- generated WTS and a lower concentration.

3.2.2. Smoke toxicity.

Because most waterpipe smokers use charcoal-heating and charcoal smoke contains known toxicants, it was important to determine whether heating shisha with electronic instead of charcoal heat translated into decreased cellular harm. Charcoal smoke (Charcoal+NRM) and smoke generated by applying either heat source to whole shisha produced smoke that significantly decreased two of the three cell health metrics measured (Fig 3). Compared to ‘non-smoking’ control treatments (exposure to 0.045 L/min ambient air), charcoal-generated WTS decreases lysosomal function to 35% (Fig 3A) and the cytoplasmic esterase activity to 21% (Fig 3B). Protease activity showed no significant differences with the control (Fig 3C). Thus, we report only the esterase activity measure of membrane integrity for all other treatment conditions. Electronic-generated WTS reduced lysosomal function to 54% (Fig 3A) and membrane integrity measures to 31% of those seen in non-smoking control cells (Fig 3B). Therefore, according to the ISO/NIH definition of cytotoxic substances, both charcoal-generated WTS and electronic-generated WTS are cytotoxic.

Figure 3.

Effect of heat source on WTS toxicity as measured by A) lysosomal function (NRU), B) cytoplasmic esterase activity, and C) cytoplasmic protease activity. Alveolar Type II cells were exposed to smoke generated from the heat source and head content indicated following a modified Beirut protocol. NRM= non-reactive matrix. Dashed line=NIH cytotoxic threshold. ¥ p≤0.04; * p≤0.02 vs non-smoking; # p≤0.05 vs. NIH cytotoxic threshold.

3.3. Contributions of shisha components to WTS

3.3.1. Syrup and humectant physical characteristics

Shisha contains tobacco and a syrup composed of flavorings and humectant. To investigate the contribution of the syrup to the physical properties of the smoke, syrup from a 10 g sample of shisha was extracted, mixed with the NRM, and smoked using charcoal. As shown in Figure 4, the particle concentration is higher when smoking the syrup without the tobacco and the mean particle size does not change (Table 1).

Figure 4.

Particle Size Distributions for Shisha, Syrup+NRM and Glycerol+NRM smoked with Charcoal.

Previous studies have found shisha to contain the humectants propylene glycol and glycerol at between 5 and 60% by mass and that these compounds lead to the presence of toxic carbonyls such as formaldehyde, acetaldehyde and acrolein in the smoke.(Schubert 2011, Schubert 2012) To investigate the role of the humectant in generating particulate matter, we measured particle distributions for glycerol at a viscosity similar to syrup (92% with water). Compared to syrup, glycerol generated a similar size distribution but fewer total particles (Fig 4, Table 1).

3.3.2. Syrup and Humectant toxicity.

Since, by mass, shisha smoke particulates are primarily derived from the syrup, the toxicity of smoke produced by charcoal heating NRM coated with extracted syrup or humectant alone (92% glycerol) was investigated. Neither treatment caused changes in lysosomal function (Fig 5A). However, both syrup smoke and humectant smoke exposure significantly decreased alveolar membrane integrity resulting in 35% and 30% survival compared to non-smoking treatments (Fig 5B).

Figure 5.

Effect of shisha constituents on WTS cytotoxicity as measured by A) lysosomal function (NRU) and B) membrane integrity (esterase activity). Dashed line=NIH cytotoxic threshold. In A and B: *p≤0.02 vs non-smoking; # p≤0.05 vs. NIH cytotoxic threshold.

4. DISCUSSION

Consistent with previous research we found that charcoal smoke alone is cytotoxic, causing a >80% decrease in lysosomal function and >50% decrease in membrane integrity as measured via esterase activity (Fig 3A and B; (Elsayed et al., 2016; Monzer et al., 2008; Sepetdjian et al., 2010). Smoke can cause cellular damage due to particles hitting the cell surface (physical toxicity) or reacting with cell components (chemical toxicity). Comparing the physical properties of the particles in WTS to the smoke’s toxicity provides insight into the physical versus chemical toxicity of the smoke and the relative sensitivity of the cell viability assays to these different sources of toxicity. Charcoal smoke had the smallest mean particle diameter, and the lowest particle dose, total mass, and predicted lung deposited surface area (LDSA; Fig 2, Table 1) of any of the conditions tested, but it caused the greatest decrease in alveolar lysosomal function (Fig 3A). This suggests that the toxicity of charcoal smoke measured by the NRU/ lysosomal function assay is due more to the chemical properties of the smoke than the physical properties and the assay is sensitive to cellular changes in response to chemical toxicants.

Smoke derived from the shisha components alters WTS toxicity as cells exposed to charcoal+shisha smoke (i.e. whole WTS) have significantly lower membrane integrity than cells exposed to charcoal smoke alone (Fig 3B). Whole WTS’s particle concentration per session is 25 times greater than charcoal+NRM smoke (Table 1). The more complex matrix, furthermore, results in charcoal+shisha smoke having a mean particle diameter three times larger than that of charcoal+NRM smoke (130 nm vs 40 nm), a total particle mass 750 times larger (6000 mg/m³ vs 8 mg/m³) and a predicted LDSA that is 100 times greater (10⁷ vs 10⁵). Comparing these physical properties with changes in measures of cell health reveals an inverse relationship between particle mass and number and measures of membrane integrity. This suggests that, in addition to its chemical toxicity, charcoal+shisha smoke is acting as a physical toxicant, and that the esterase-based membrane integrity assay is more sensitive to cellular changes due to physical damage than the NRU/lysosomal function assay (Fig 3).

To investigate whether WTS toxicity was inherent to charcoal combustion components an alternative heat source was used. Smoke generated by electronic heating continued to cause alveolar cell damage, fulfilling the definition of a cytotoxic substance (Fig 3B). The particle number concentration per session decreases when shisha is heated using an electronic source in place of charcoal and heating shisha with charcoal increases the number of particles by more than just the number present in charcoal smoke (Fig 2). Microparticles generated during charcoal combustion serve as condensation nuclei for more volatile components in shisha syrup, resulting in a synergistic increase in total number.(Hinds, 1999) Because electronically-produced smoke lacks heterogeneous nucleation and requires supersaturation to facilitate nucleation, electronic heating would produce a smaller number of larger particles with similar chemical composition to those in charcoal heated shisha smoke. The electronic+shisha smoke causes cell damage resulting in lysosomal activity and membrane integrity measures that are not statistically different from those seen after cells are exposed to charcoal+shisha smoke. These results support that, independent of the toxicants generated by burning charcoal, shisha volatilization products are responsible for WTS toxicity.

WTS causes significant cellular harm and since negligible amounts of tobacco are consumed during a smoking session this points to the syrup, not the tobacco, as the source of toxicants. Charcoal+syrup smoke decreased lysosomal activity by 20%. However, alveolar cells exposed to syrup smoke at the air liquid interface show a 65% decrease in membrane integrity, a decrease that is not statistically different from whole shisha smoke (Figure 5B). Compared to whole shisha smoke, syrup smoke contains particles with a similar mean diameter (130nm), but two times greater particle number concentration and predicted LDSA (Table 1). As the chemical toxicity of tobacco-derived smoke is well established, and syrup smoke lacks these components, these data further support the conclusion that the lysosomal function assay is sensitive to chemical toxicants but relatively insensitive to particle dose, while the esterase-based membrane integrity assay is sensitive to the physical properties and dose of the particulate treatment.

Since syrup significantly contributed to the toxic effect of exposure to WTS, it is important to understand the relative contribution of syrup constituents including humectant and flavorants, similar to electronic cigarettes. Glycerol is the main humectant used in commercial shisha and might be considered only a ‘carrier’ for syrup’s flavorants. However, rat inhalation studies report squamous metaplasia following exposures above 0.167 mg glycerol/L. (Renne et al., 1992) In our system, although glycerol+charcoal smoke mitigated the decrease in lysosomal function caused by charcoal smoke alone (Fig 5A and 3A), perhaps due to its ability to coat charcoal smoke constituents, glycerol was not a balm for all harm as exposure to glycerol smoke still resulted in a 70% decrease in membrane integrity (Fig 5B). While glycerol smoke has one-third the particle number concentration and predicted LDSA of syrup smoke, it has a particle dose per session that is twenty times greater than charcoal smoke (Table 1). Furthermore, our treatments used water to adjust the viscosity of glycerol where commercial shisha syrups are mixtures glycerol, propylene glycol (PG) and flavorants.(Schubert et al., 2012, 2011; Shihadeh et al., 2015) These mixtures include semi-volatile compounds with the capacity to form particles and, based on our findings that the particles act as physical toxicants, increase the toxicity of WTS.

The fact that the two metrics reported here can differentiate between relative levels of cytotoxicity underscores the importance of studies including multiple cellular health metrics. Measuring lysosomal function via incorporation of neutral red dye reveals distinctions between smoke sources. This metric proved to be more sensitive to chemical toxicants than physical toxicity, as similar particle doses cause significantly different levels of cellular harm (charcoal+shisha versus electronic+shisha, charcoal+syrup and charcoal+glycerol) and low particle doses result in higher levels of cellular harm (charcoal+NRM). Measures of relative membrane integrity, using cytoplasmic esterase cleavage of fluorescein diacetate as a proxy, reveal that membrane damage occurs and all WTS is cytotoxic, regardless of heating source and shisha constituent. This assay is sensitive to chemical toxicants but perhaps more sensitive to physical toxicity and, therefore, provides information based on the smoke’s high concentration of UFP that decrease membrane integrity, regardless of composition. Two weaknesses to the approach are, first, the particle concentrations could be beyond the esterase-based assay’s range so that it cannot distinguish between them. Second, the cellular assays used are considered measures of cell death and if the damage caused by one exposure session does not rise to this high threshold the true harm of WTS would be underestimated. Even taking these aspects into consideration, the combination of cellular health metrics supports the hypothesis that heating source and shisha constituents impact the toxicity of the smoke generated by the waterpipe.

5. CONCLUSIONS

This study combines analysis of the physical properties of WTS with measurement of the impact that whole smoke exposure at the air liquid interface has on alveolar cell viability. Our data indicate that the charcoal heat source and each of the constituents of shisha (tobacco, flavorants and humectant) contribute to levels of particulate matter in WTS that greatly exceed all current outdoor PM regulatory levels (WHO 25 μg/m³, EPA 35 μg/m³ 24 hour mean exposure).(EPA, n.d.; World Health Organization, 2006) WTS’s high number density of small particles gives a correspondingly small mass, yet high LDSA, validating the importance of using particle size and concentration information when evaluating the health effects of WTS. The constituents heated individually with charcoal, or as whole shisha smoked using charcoal or an electronic heating source, damaged lung cells that play crucial roles in maintaining alveolar gas exchange. The importance of using multiple metrics in evaluating toxicity is underscored by data showing that smoke generated by charcoal heating of syrup and glycerol decreased only membrane integrity, while smoke generated by charcoal or electronic heating of shisha and charcoal alone decreased both lysosomal function and membrane integrity. Our results support the conclusions that all constituents used in waterpipe smoking--charcoal, tobacco, syrup, and humectants-- contribute to smoke dose and toxicity and warrant regulatory scrutiny.

Highlights.

• Shisha tobacco, flavorants, and humectant each contribute to dose and toxicity.

• WTS from electronic and charcoal heating cause similar alveolar cell damage.

• Particle size and concentration data are critical when evaluating WTS health effects.

• Assessing chemical and physical toxicity of smoke requires multiple metrics.