ArticlePDF Available

Toxicity of the main electronic cigarette components, propylene glycol, glycerin, and nicotine, in Sprague-Dawley rats in a 90-day OECD inhalation study complemented by molecular endpoints

Authors:

Abstract and Figures

While the toxicity of the main constituents of electronic cigarette (ECIG) liquids, nicotine, propylene glycol (PG), and vegetable glycerin (VG), has been assessed individually in separate studies, limited data on the inhalation toxicity of them is available when in mixtures. In this 90-day subchronic inhalation study, Sprague-Dawley rats were nose-only exposed to filtered air, nebulized vehicle (saline), or three concentrations of PG/VG mixtures, with and without nicotine. Standard toxicological endpoints were complemented by molecular analyses using transcriptomics, proteomics, and lipidomics. Compared with vehicle exposure, the PG/VG aerosols showed only very limited biological effects with no signs of toxicity. Addition of nicotine to the PG/VG aerosols resulted in effects in line with nicotine effects observed in previous studies, including up-regulation of xenobiotic enzymes (Cyp1a1/Fmo3) in the lung and metabolic effects, such as reduced serum lipid concentrations and expression changes of hepatic metabolic enzymes. No toxicologically relevant effects of PG/VG aerosols (up to 1.520 mg PG/L + 1.890 mg VG/L) were observed, and no adverse effects for PG/VG/nicotine were observed up to 438/544/6.7 mg/kg/day. This study demonstrates how complementary systems toxicology analyses can reveal, even in the absence of observable adverse effects, subtoxic and adaptive responses to pharmacologically active compounds such as nicotine.
Content may be subject to copyright.
Toxicity of the main electronic cigarette components, propylene
glycol, glycerin, and nicotine, in Sprague-Dawley rats in a 90-day
OECD inhalation study complemented by molecular endpoints
Blaine Phillips
a
,
1
, Bjoern Titz
b
,
1
, Ulrike Kogel
b
,
1
, Danilal Sharma
a
,
1
, Patrice Leroy
b
,
Yang Xiang
b
,Gr
egory Vuillaume
b
, Stefan Lebrun
b
, Davide Sciuscio
b
, Jenny Ho
a
,
Catherine Nury
b
, Emmanuel Guedj
b
, Ashraf Elamin
b
, Marco Esposito
b
,
Subash Krishnan
b
, Walter K. Schlage
c
, Emilija Veljkovic
a
, Nikolai V. Ivanov
b
,
Florian Martin
b
, Manuel C. Peitsch
b
, Julia Hoeng
b
, Patrick Vanscheeuwijck
b
,
*
a
Philip Morris International Research Laboratories Pte. Ltd. (part of Philip Morris International Group of Companies), 50 Science Park Road, Singapore
117406, Singapore
b
Philip Morris International Research and Development (part of Philip Morris International Group of Companies), Philip Morris Products S.A., Quai
Jeanrenaud 5, 2000 Neuchatel, Switzerland
c
Biology Consultant, Max-Baermann-Str. 21, 51429 Bergisch Gladbach, Germany
article info
Article history:
Received 19 April 2017
Received in revised form
23 August 2017
Accepted 1 September 2017
Available online 5 September 2017
Keywords:
Organization for Economic Cooperation and
Development (OECD) 413 guideline
Electronic cigarette
Systems toxicology
Propylene glycol
Glycerin
Nicotine
Inhalation toxicity
abstract
While the toxicity of the main constituents of electronic cigarette (ECIG) liquids, nicotine, propylene
glycol (PG), and vegetable glycerin (VG), has been assessed individually in separate studies, limited data
on the inhalation toxicity of them is available when in mixtures. In this 90-day subchronic inhalation
study, Sprague-Dawley rats were nose-only exposed to ltered air, nebulized vehicle (saline), or three
concentrations of PG/VG mixtures, with and without nicotine. Standard toxicological endpoints were
complemented by molecular analyses using transcriptomics, proteomics, and lipidomics. Compared with
vehicle exposure, the PG/VG aerosols showed only very limited biological effects with no signs of toxicity.
Addition of nicotine to the PG/VG aerosols resulted in effects in line with nicotine effects observed in
previous studies, including up-regulation of xenobiotic enzymes (Cyp1a1/Fmo3) in the lung and meta-
bolic effects, such as reduced serum lipid concentrations and expression changes of hepatic metabolic
enzymes. No toxicologically relevant effects of PG/VG aerosols (up to 1.520 mg PG/L þ1.890 mg VG/L)
were observed, and no adverse effects for PG/VG/nicotine were observed up to 438/544/6.6 mg/kg/day.
This study demonstrates how complementary systems toxicology analyses can reveal, even in the
absence of observable adverse effects, subtoxic and adaptive responses to pharmacologically active
compounds such as nicotine.
©2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
Nicotine, propylene glycol (PG), and vegetable glycerin (VG),
together with distilled water and avors, are the main constituents
of the liquids used to generate aerosols by electronic cigarettes
(ECIGs) (Brown and Cheng, 2014; Iskandar et al., 2016). While the
in vivo toxicity of nicotine, PG, and glycerin has been individually
assessed in separate studies (Phillips et al., 2015; Renne et al., 1992;
Suber et al., 1989; Werley et al., 2011), only limited published
toxicology data are available on mixtures of the main compounds
administrated via inhalation (e.g., Werley et al., 2016).
ECIGs provide an alternative to cessation for cigarette smokers
who are unable or unwilling to quit (Breland et al., 2017). Several
studies have reported potential positive health effects of switching
from combustible cigarettes to ECIGs. It has been reported, for
Abbreviations: DEG, differentially expressed gene; DEP, differentially expressed
protein; ECIGs, electronic cigarettes; FDR, false discovery rate; Nic, Nicotine; PG,
Propylene glycol; VG, vegetable glycerin.
*Corresponding author.
E-mail address: Patrick.vanscheeuwijck@pmi.com (P. Vanscheeuwijck).
1
B.P., B.T., U.K., and D.S. contributed equally to this manuscript.
Contents lists available at ScienceDirect
Food and Chemical Toxicology
journal homepage: www.elsevier.com/locate/foodchemtox
http://dx.doi.org/10.1016/j.fct.2017.09.001
0278-6915/©2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Food and Chemical Toxicology 109 (2017) 315e332
example, that toxicant levels were much lower in ECIG aerosol than
in cigarette smoke (Flora et al., 2016; Goniewicz et al., 2014; Marco
and Grimalt, 2015; Margham et al., 2016), and toxicant and
carcinogen metabolites have been shown to be signicantly
reduced in the urine of ECIG users versus that of smokers (Hecht
et al., 2015). Public Health England estimated that ECIGs are 95%
less harmful than combustible cigarettes (McNeill et al., 2015).
However, other studies have highlighted potential health hazards
associated with the use of ECIGs (e.g., Hiemstra and Bals, 2016).
Formaldehyde can be formed from PG when the liquid is over-
heated (Jensen et al., 2015), ECIG aerosols have been reported to be
contaminated with oxidants and copper ions (Lerner et al., 2015a),
and to induce oxidative stress and inammatory responses in hu-
man lung epithelial cells (Lerner et al., 2015b). To provide a
comprehensive hazard evaluation of these ECIG liquids, a stepwise
approach has been proposed (Iskandar et al., 2016), in which the
potential toxicity of the aerosolized basic components of the ECIG
liquids is assessed rst. Subsequently, the toxicity of added avor
components should be evaluated, and nally the aerosols formed
are assessed, including potential effects of aerosolization (e.g., by
heating) in a particular ECIG design.
The ECIG liquid component PG is an aliphatic alcohol used as a
solvent in many pharmaceuticals and avors, and in vaporizers for
delivery of pharmaceuticals by inhalation. It is generally recognized
as a safe food additive (C.F.R., 2014). The reported oral lethal dose
for 50% of a tested population (LD
50
) for PG in rats is 20 g/kg body
weight (RTECS, 1985). A sub-chronic 90-day nose-only inhalation
study in rats exposed to PG at 0.16, 1.0, or 2.2 mg/L of aerosol
showed no treatment-related histological changes in the trachea,
larynx, or lung. However, a decrease in body weight in female rats
exposed to 2.2 mg/L PG, an increase in mucin secretion, and mild
hemorrhages from nasal passages were detected (Suber et al.,
1989). Werley et al. assessed PG in a 28-day rat inhalation study;
the only biologically relevant ndings included clinical signs of
ocular and nasal irritation, indicated by minor bleeding around the
eyes and nose, and minimal laryngeal squamous metaplasia
(Werley et al., 2011).
The ECIG liquid component VG is also a food ingredient recog-
nized as safe by the FDA (C.F.R., 2014), and the present human
threshold limit value for inhaled glycerin is 10 mg/m
3
(ACGIH,
1989). A sub-chronic 90-day nose-only inhalation study in
Sprague-Dawley (SD) rats exposed to 0.03, 0.16 and 0.66 mg/L
glycerin revealed no treatment-related toxicity other than minimal
metaplasia of the epithelium lining at the base of the epiglottis in
rats exposed to 0.66 mg/L glycerin (Renne et al., 1992).
Compared with cigarette smoke, the inhalation toxicity of
nicotine-containing aerosols and vapors has been investigated only
sparsely (Phillips et al., 2015; Salturk et al., 2015; Waldum et al.,
1996; Werley et al., 2014, 2016). In one study, Chowdhury et al.
exposed male SD rats to aerosols of nebulized saline or nicotine
dissolved in saline twice daily for 15, 30, 45, and 60 min for 21 days.
The authors investigated plasma levels of nicotine and histopa-
thology of the pancreas, whereas effects on the respiratory tract
(proximal target tissue and site of nicotine absorption) and liver
(site of major metabolism of nicotine) were not investigated
(Chowdhury et al., 1992). In another study, the acute toxicity (LC
50
test) and pharmacokinetics of inhaled nicotine aerosol after a single
bolus-like exposure of 20-min duration were determined (LC
50
[20 min] ¼2.3 mg nicotine/L aerosol) (Shao et al., 2013). Recently,
we evaluated the toxicity of three equimolar concentrations of
nicotine and pyruvic acid (nicotine concentrations of 0.018, 0.025,
and 0.05 mg/L) in a 28-day rat inhalation study (Phillips et al.,
2015). In this study, rats exposed to nicotine-containing aerosols
displayed decreased body weight gains (males only) and
concentration-dependent increases in liver weight. Blood
neutrophil counts increased in rats exposed to nicotine compared
with sham; alkaline phosphatase and alanine aminotransferase
activities were higher, and cholesterol and glucose concentrations
were lower. The only histopathologic nding in non-respiratory
tract organs was increased liver vacuolation and glycogen con-
tent. Respiratory tract ndings following nicotine exposure (but
also some phosphate-buffered saline [PBS] aerosol effects) were
observed only in the larynx, and were limited to adaptive changes.
To further study the effects of aerosols from combined compo-
nents of e-liquids, we characterized the toxicity following a sub-
chronic (90-day) inhalation of an aerosol generated from a liquid
mixture containing PG and VG, with and without nicotine, on SD
rats. PG and VG are used in variable amounts in e-liquids, and the
mixtures were selected to be generally representative of marketed
ECIG products (e.g., El-Hellani et al., 2016). The nicotine concen-
tration (23
m
g/L) was selected as an intermediate concentration that
induces effects but would not be too high to mask potential effects
of PG/VG (Table 1). The inhalation protocol was performed ac-
cording to the Organization of Economic Cooperation and Devel-
opment (OECD) Test Guideline 413 (OECD, 2009b). To complement
the endpoints suggested by the OECD, a systems toxicology analysis
was conducted with groups of additional female rats that included
transcriptomics, proteomics, and lipidomics analyses.
2. Material and methods
2.1. Experimental design
A 90-day repeated-dose OECD TG413 (OECD, 2009b) inhalation
study was conducted using male and female rats to determine
potential toxicological effects after exposure to three concentra-
tions of PG and VG mixtures, with or without nicotine. The study
comprised two control groups: one sham group exposed to ltered
conditioned fresh air, and one vehicle group exposed to aerosolized
PBS. For the measurement of endpoints listed in the OECD guide-
line, 10 male and 10 female rats were allocated to each group, the
OECDgroups. To characterize exposure-related effects at the
molecular level (transcriptomics, proteomics, and lipidomics), six
female rats per group were allocated to eight additional experi-
mental groups, the Systems Toxicology arm (see Table 1), exposed
concomitantly with the OECDgroups. Male rats were not
included in these endpoints because of space limitations in the
exposure chambers. Therefore, females were selected, as other
studies have suggested that females show a stronger response to
test aerosols, possibly due to larger aerosol uptake relative to their
body weight (e.g., Kogel et al., 2016; Oviedo et al., 2016; Wong et al.,
2016).
The study was conducted in compliance with the OECD Princi-
ples on Good Laboratory Practice (GLP) (OECD, 2009a), with the
exception of bronchoalveolar lavage uid (BALF) analysis using
RodentMAP and the transcriptomics, proteomics, and lipidomics
investigations.
2.2. Animals
Six-week-old outbred male and nulliparous, non-pregnant fe-
male SD rats [Crl:CD(SD)], bred under specic pathogen-free con-
ditions, were obtained from Charles River Laboratories (breeding
area Raleigh R04, NC, USA). The study was performed in American
Association for the Accreditation of Laboratory Animal Care-
approved and Agri-Food &Veterinary Authority of Singapore-
licensed facilities at Philip Morris International Research Labora-
tories (PMIRL, Singapore). Care and use of animals were in accor-
dance with guidelines set by the National Advisory Committee for
Laboratory Animal Research in 2004 (NACLAR, 2004). All protocols
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332316
were approved by the Animal Care and Use Committee of PMIRL.
2.3. Aerosol generation and animal exposure
Inhalation was selected as the route of administration in this
study. Aerosols were generated as previously described (Phillips
et al., 2015) by using commercially available 6-jet Collison nebu-
lizers (BGI, Butler, NJ, USA) that function by applying pressure to
force liquid solutions through small apertures, resulting in the
production of a ne aerosol. Stock solutions were stored in a cold
room (2e10
C) and used within 2 weeks of preparation. The
relative proportions, concentrations, and molarities of the stock
solution constituents for the target aerosol specications are given
in Table 2. Stock solutions were equilibrated for up to 30 min in a
water bath at 25
C before nebulization; for the high concentra-
tions, the nebulizer was warmed to 30
C to generate aerosols of
appropriate concentrations and particle sizes (Fig. 1A). The gener-
ated aerosols were further diluted to the target concentrations
using ltered conditioned air.
The nose-only exposure method using a ow-past chamber
(FPC1-132) was used to expose rats to the generated aerosols
(Cannon et al., 1983). This ensured daily reproducibility of aerosol
uptake, minimizing aerosol deposition on the animal fur (often
observed when whole body exposure is used), which could be
subsequently taken up by grooming and interfere with the uptake
by inhalation. The duration of the exposure was 13 weeks, for 5
days per week, 6 h per day. A time-adaptation phase was included
in week 1 when animals were exposed to increasing exposure
durations over 7 days (1.5 h for the rst and second days, 3 h for the
third and fourth days, and 4.5 h for days 5e7). Weekend exposures
(7 days per week exposure) were performed prior to scheduled
dissection as necessary in order to ensure that all animals were
subjected to a minimum of two consecutive exposure days before
dissection. The setup used for aerosol generation and exposure is
shown in Fig. 1A.
2.4. Analytical characterization of test atmospheres and aerosol
uptake
To characterize the test atmosphere and to check the repro-
ducibility of aerosol generation, concentrations of total particulate
matter (TPM), nicotine, PG, and VG, temperature, ow rate through
the exposure chamber, conductivity (ion concentration of aerosol),
relative humidity, and particle size distribution were determined at
the animalsbreathing zone during exposure (Supplementary
Table 1). Nicotine concentrations were determined by capillary
gas chromatography after trapping on sulfuric acid-impregnated
silica gel (Majeed et al., 2014). TPM was determined by gravim-
etry after trapping on Cambridge lters. For the determination of
PG and VG concentrations in the diluted test atmosphere, samples
were collected using a Cambridge lter and analyzed by gas chro-
matography (GC) with a ame ionization detector (see Supple-
mentary Material and Methods for details).
For the measurement of biomarkers of exposure in blood, blood
was collected from all animals, 10 male and 10 female rats per
group, from the facial vein under isouorane anesthesia during
study days 39e51 within 6e7 min after the rats were removed from
the exposure, to minimize clearance from the plasma. Plasma was
isolated, and the nicotine, cotinine, PG, and glycerin levels were
determined by Analytisch-biologisches Forschungslabor GmbH
(Munich, Germany) using LC-MS/MS.
2.5. Biological parameters
Biological parameters were determined as described previously
(Phillips et al., 2015). In brief, respiratory physiology was measured
once in the study using head-out plethysmography (EMKA Tech-
nologies, Paris, France) from each rat, 10 male and 10 female rats
Table 1
Exposure groups and target concentrations in the test atmospheres of 90-day repeated-dose OECD TG413 inhalation toxicity study. For the OECD rats, endpoints
suggested in the OECD TG 413 were determined (e.g., body weight, clinical chemistry, organ weight, measurement of inammatory markers in the bronchoalveolar lavage uid,
histopathological changes). For the rats of the Systems Toxicology arm, transcriptomics, proteomics and lipidomics analyses were performed from selected organs.
Group Carrier [mg/L] Nicotine [mg/L] OECD Arm Systems Toxicology Arm
PG VG Male Female Female
Sham (ltered air) 0 0 0 10 10 6
Vehicle (saline) 0 0 0 10 10 6
Low (PG/VG) 0.174 0.210 0 10 10 6
Med (PG/VG) 0.520 0.630 0 10 10 6
High (PG/VG) 1.520 1.890 0 10 10 6
Nic þLow (PG/VG) 0.174 0.210 0.023 10 10 6
Nic þMed (PG/VG) 0.520 0.630 0.023 10 10 6
Nic þHigh (PG/VG) 1.520 1.890 0.023 10 10 6
Table 2
Relative proportions, concentrations, and molarities of the stock solution constituents for the target aerosol specications.
Stock solutions Relative proportions of stock
solution constituents (% V/V)
Concentrations of the stock
solution constituents (g/L)
Molarities of stock
solution constituents (M)
Target concentrations of
aerosol constituents (mg/L)
Estimated doses of
constituents (mg/kg)
a
Nic/PG/VG Nic/PG/VG Nic/PG/VG Nic/PG/VG Nic/PG/VG
Low (PG/VG) 0.00/3.78/3.56 0.00/39.14/44.41 0.00/0.50/0.50 0.000/0.174/0.210 0/50/60
Med (PG/VG) 0.00/11.33/10.67 0.00/117.41/133.23 0.00/1.50/1.50 0.000/0.520/0.630 0/150/181
High (PG/VG) 0.00/34.00/32.00 0.00/352.24/399.68 0.00/4.60/4.30 0.000/1.520/1.890 0/438/544
Nic þLow (PG/VG) 0.50/3.78/3.56 5.05/39.14/44.41 0.03/0.50/0.50 0.023/0.174/0.210 6.6/50/60
Nic þMed (PG/VG) 0.50/11.33/10.67 5.05/117.41/133.23 0.03/1.50/1.50 0.023/0.520/0.630 6.6/150/181
Nic þHigh (PG/VG) 0.50/34.00/32.00 5.05/352.24/399.68 0.03/4.60/4.30 0.023/1.520/1.890 6.6/438/544
a
Doses of aerosol constituents were calculated according to the adapted formula (Alexander et al., 2008): D ¼(C RMV d)/BW), where D is the dose (mg/kg), C is the
concentration of constituent in aerosol (mg/L), RMV is the respiratory minute volume (for rats, 0.2 L/min was used), d is the duration of exposure (min), and BW is body weight
(kg) (for rats, 0.25 kg was used). PG, propylene glycol; VG, vegetable glycerin; Nic, nicotine.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 317
per group, for evaluation of breathing frequency, minute volume,
tidal volume, and peak inspiratory ow.
All rats in the OECD study were euthanized after 13 weeks of
exposure (approximately 16e24 h after the nal exposure without
diet deprivation) according to OECD TG 413. For hematology anal-
ysis, blood samples were collected from the retro-orbital venous
plexus at dissection time under pentobarbital anesthesia and
analyzed using a Sysmex blood analyzer (Sysmex, Kobe, Japan). For
clinical blood chemistry analysis, serum was collected from the
abdominal aorta during necropsy, and evaluation was performed
using a UniCel
®
DxC 600i clinical analyzer system (Beckman
Coulter, Brea, CA, USA). The weights of the OECD TG 413-specied
organs were measured during dissection as absolute values, and
the relative weights were then calculated according to body
weights after exsanguination.
BALF was collected as described previously (Kogel et al., 2014)
and the cellular content was analyzed by ow cytometry. For this
assay, the right lung was cannulated and instilled with lavage
medium for collection of free lung cells from the rst-to-fth cycle
of bronchoalveolar lavagefrom the right lung of all animals (cycle 1,
PBS; cycles 2e5, PBS þ0.325% bovine serum albumin). Cells were
collected and analyzed by ow cytometry for viability and differ-
ential leukocyte count, as described previously (Kogel et al., 2014;
Phillips et al., 2015). Cell-free BALF from cycle 1 (remaining after
centrifugation of cells for BALF) was frozen, and submitted to
Myriad Rules-Based Medicine (Austin, TX, USA) for exploratory
(non-GLP) analyses of a panel of 60 selected proteins (RodentMAP
v. 3.0; Life Technologies, Carlsbad, CA, USA) (Phillips et al., 2015).
2.6. Histopathology evaluation
Respiratory tract organs (nose, larynx, trachea, left lung) were
xed in ethanol glycerol acetic acid formaldehyde solution (EGAFS).
Histological sections of respiratory tract organs were prepared at
Fig. 1. Exposure setup and characterization. (A) Aerosol generation and exposure setup. For the high concentrations of (PG/VG) ±nicotine, aerosol generation required the
nebulizer to be warmed with warm water (at 30 C). Black circles indicate the positions of the sampling ports.
D
P, change in pressure; TPM, total particulate matter; FPC1-132, ow-
past chamber (1 stack of 32 ports). (B) Mean concentrations of nicotine, PG, and VG in the aerosols over the 90-days exposure period (mean ±SD). Target concentrations are
indicated by the midlines in the gray areas, which represent ±10% of the target concentrations. (CeF) Uptake of aerosol constituents in plasma. Mean concentrations of nicotine (C),
cotinine (D),PG(E), and glycerin (F) from male (left panels) and female (right panels) rats ±SEM (N ¼10). Statistically signicant differences from the vehicle group are represented
by asterisks (*); statistically signicant differences between groups with and without nicotine (at the same PG/VG concentration) are represented by carets (^)(p-value <0.05). (G)
Total nicotine metabolites in 24-h urine. Mean concentrations ±SEM of absolute (total) nicotine metabolites in urine samples from male (left panel) and female (right panel) rats.
PG, propylene glycol; VG, vegetable glycerin; Nic, nicotine. Note that nicotine uptake and its metabolism was only assessed for the sham group (negative control) and the three
nicotine-containing aerosol groups, as only for these nicotine uptake could be expected based on the aerosol measurements (Fig. 1B).
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332318
dened levels and stained with hematoxylin and eosin (H&E).
Histopathological evaluations of respiratory tract organs were
performed on digitalized histological slides. Four nasal cavity levels,
including the nasopharyngeal duct, four larynx levels, trachea
(transversal and longitudinal section at the bifurcation, including
the carina), and three lung sections were evaluated by Dr. Klaus
Weber (AnaPath GmbH, Liestal, Switzerland).
Non-respiratory tract organs were collected according to OECD
413 guidelines and xed in 4% formaldehyde solution except for the
sternum and testes, which were xed in Schaffer's solution and
Bouin solution, respectively. The slides of non-respiratory tract
organs were prepared by LPT (Hamburg, Germany) and evaluated
by Dr. Ansgar Buettner (Histovia GmbH, Overath, Germany). For
these non-respiratory endpoints, the histopathological evaluation
was performed initially only in the high test and reference item
groups, and further analysis of the low and medium were per-
formed only if ndings were present.
2.7. Tissue preparation for omics endpoints
Dissection of the animals of the Systems Toxicology arm was
also performed 16e24 h after the last exposure. Prior to organ
removal, a whole-body perfusion with cold saline was performed.
Respiratory nasal epithelium (RNE) for transcriptomics/proteomics
analysis was isolated from the left/right side of the nose, snap
frozen in liquid nitrogen, and stored at 80
C. For transcriptomics
and proteomics, the left lung lobe was frozen on dry ice and stored
at 80
C. The lung lobe was cryosectioned into 40-
m
m slices, and
the slices were collected alternately for transcriptomics and pro-
teomics analysis. Liver dissections were collected for tran-
scriptomics, proteomics, and lipidomics. The liver sections for
transcriptomics were placed in MagNa Lyser tubes (Roche, Basel,
Switzerland), and all sections were snap-frozen in N
2
(l) and stored
at 80
C. For lipidomics, blood was collected into an EDTA tube,
and the plasma was then isolated by centrifugation.
2.8. Transcriptomics analysis
See Supplementary Material and Methods for details. Briey,
RNA from respiratory nasal epithelium (RNE), left lung and liver
samples was isolated, processed in randomized order, and hy-
bridized on GeneChip
®
Rat Genome 230 2.0 arrays (Affymetrix,
Santa Clara, CA, USA). Raw data les were processed in the custom
Chip Description File environment (Dai et al., 2005) and normalized
using frozen robust microarray analysis (fRMA) (McCall et al., 2010).
Quality controls were performed with the affyPLM package (Bio-
conductor suite) (Bolstad et al., 2013). Raw p-values were generated
for contrasts between vehicle and exposed groups with the limma
package (Smyth, 2004), and adjusted using the Benjamini-
Hochberg false discovery rate (FDR) multiple test correction
(Gentleman et al., 2004). In addition, the overall nicotine effect was
statistically assessed as the contrast between the average
Nicotine þ(PG/VG) vs. vehicle and the average (PG/VG) vs. vehicle
effect.
2.9. Proteomics analysis
Proteome alterations were assessed by isobaric tag-based
quantication using the iTRAQ
®
approach (Titz et al., 2015a). See
Supplementary Material and Methods for details. Briey, frozen
right RNE samples, lung tissue slices, and liver tissue samples were
homogenized and processed in randomized order for iTRAQ 8-plex
labeling according to the manufacturer's instructions (AB Sciex,
Framingham, MA, USA). Mapping of samples to the replicate iTRAQ
sets and reporter ion channels was randomized. All labeled samples
that belonged to one iTRAQ replicate set were pooled, processed,
and analyzed in randomized order using an Easy nanoLC 1000 in-
strument (Thermo Fisher Scientic, Waltham, MA, USA) connected
online to a Q Exactive mass-analyzer (Thermo Fisher Scientic). The
raw mass spectrometry data were processed using Proteome
Discoverer (Thermo Fisher Scientic) and custom processing scripts
in the R statistical environment (R Development Core Team, 2007).
To detect differentially abundant proteins, a linear model was tted
for each exposure condition and the vehicle group, and p-values
were calculated from moderated t-statistics with the empirical
Bayes approach (Gentleman et al., 2004). In addition, the overall
nicotine effect was assessed statistically as the contrast between
the average Nicotine þ(PG/VG) vs. vehicle and the average (PG/VG)
vs. vehicle effect. The FDR method was then used to correct for
multiple testing effects. Proteins with an adjusted p-value <0.05
were considered differentially abundant.
2.10. Lipidomics analysis
Lipids were extracted from lung tissue, liver tissue, and plasma
and quantied by Zora Biosciences (Espoo, Finland) according to
internal Standard Operating Procedures, Lab Method Sheets, and
Data Processing Method Sheets (Ansari et al., 2016). Lipid levels
were normalized to their respective internal standards and tissue
weight or plasma volume.
Statistical analysis of the lipidomics data was similar to the
analyses of the transcriptomics and proteomics data. Samples with
a total summed lipid concentration below the rst
quartile 1.5 the interquartile range or above the third
quartile þ1.5 the interquartile range were excluded. For the
detection of differentially abundant lipids, only lipid species that
were present/quantied in at least 50% of the samples of each
compared group were included. A linear model was tted for each
exposure condition and the corresponding vehicle group (and for
Nic þHigh (PG/VG) vs. High (PG/VG)), and p-values from a
moderated t-statistic were calculated with the empirical Bayes
approach (Gentleman et al., 2004). The FDR method was used to
correct for multiple testing effects. Lipids with an adjusted p-value
<0.05 were considered differentially abundant.
2.11. Statistical evaluations and additional data analysis methods
To evaluate group differences in biological and histopathological
endpoints, pairwise differences between groups were estimated
separately for each sex. Each exposure group was compared with
the vehicle group of the same sex and, in addition, the PG/VG
groups with and without nicotine were compared for each con-
centration. Continuous variables were compared using a t-test
(accounting for variance heterogeneity) or, depending on the non-
normality of the data (assessed by a Shapiro-Wilk test at 5% applied
on the standardized residuals of both groups being compared), by
the mean of a non-parametric Mann-Whitney-Wilcoxon rank-sum
test. For incidences, this was done using Fisher's exact test. For
ordinal variables, this was done using the Mann-Whitney-
Wilcoxon rank-sum test. These statistical analyses were per-
formed using the SAS software package (SAS Institute, Cary, NC,
USA).
Gene sets were obtained from the Molecular Signatures Data-
base (mSigDB) (Liberzon et al., 2011). Set analysis was supported by
the Piano package in the R statistical environment (Varemo et al.,
2013). Set enrichment was assessed by gene-set analysis with the
(absolute) log2 fold-change as the gene/protein statistic and the
mean as the set statistic. Gene/protein permutation and sample
permutation were used to assess statistical signicance. p-value
adjustment was done with the Benjamini-Hochberg procedure.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 319
Functional associations among proteins were obtained from the
STRING database (Szklarczyk et al., 2015). Functional clusters were
annotated manually, guided by overrepresentation analysis
(Varemo et al., 2013). Immune-cell markers were taken from the
literature (Kogel et al., 2016).
2.12. Data availability
The mass spectrometry proteomics data are available from the
ProteomeXchange Consortium via the PRIDE partner repository
(http://www.ebi.ac.uk/pride/archive/)(Vizcaino et al., 2014)
(accession numbers: PXD005958, PXD005959, and PXD006894).
The lipidomics data are available in Supplementary Table 2. The
transcriptomics data have been submitted to ArrayExpress (www.
ebi.ac.uk/arrayexpress) (accession numbers: E-MTAB-5544, E-
MTAB-5545, and E-MTAB-5910). The histopathology data are
available in Supplementary Table 3. The data for the other end-
points are available in Supplementary Table 4.
3. Results
3.1. Exposure and aerosol characterization
The aerosols were generated using commercially available 6-jet
Collison nebulizers (BGI), diluted to the target concentrations using
ltered conditioned air (Table 1), and conveyed to the nose-only
exposure chamber (Fig. 1A) (Phillips et al., 2015). The nal con-
centrations of PG, VG, and nicotine in the aerosols were monitored
up to four times per day by sampling and analysis, demonstrating
that the overall delivery of the constituents was within the target
range (target ±10%) (Fig. 1B, Supplementary Table 1). Particle size
distribution was determined using a spectrophotometric aero-
dynamic particle sizer (Table 3).
Aerosol uptake was conrmed by determining nicotine and
cotinine concentrations in the plasma, and by quantication of
nicotine metabolites present in 24-h urine samples (Fig. 1C and
D,G). In female rats, a higher nicotine concentration was found in
plasma; this is likely explained by the smaller body size but not
proportionately lower minute volumes (for the nicotine-containing
aerosols) (Supplementary Fig. 1). However, levels of the main
nicotine metabolite cotinine (Fig. 1D), and the total amount of the
recovered nicotine metabolites in urine (see below), did not differ
between male and female rats.
PG uptake was demonstrated by measuring plasma concentra-
tion, which showed the same pattern as the concentrations of PG
measured in the test atmospheres (Fig. 1E). The glycerin concen-
trations in the plasma from exposed rats did not correlatewell with
the VG concentrations measured in the aerosols (Fig. 1F), which is
likely due to the presence of a signicant quantity of endogenous
glycerin in plasma, as well as fast-acting glycerin metabolism (Jin
et al., 2014). Of note, for both male and female rats, addition of
nicotine to the PG/VG aerosols resulted in higher plasma glycerin
concentrations, which might be explained by the effects of nicotine
on the global metabolic state of these rats. The total amount of the
recovered nicotine metabolites in urine samples collected over 24 h
(both during and after the exposure) was similar among all
nicotine-exposed groups, indicating a similar uptake by the rats
(Fig. 1G). Some of the differences in the amounts of urinary nicotine
metabolites in the Nicotine þ(PG/VG) groups (Low vs. Medium for
males, and Medium vs. High for females) are likely related to the
technical challenges of urine collection.
Respiratory physiology parameters were measured to assess
potential irritation in the upper respiratory tract caused by the
aerosols (Supplementary Fig. 1). A few changes for the PG/VG
groups were observed, e.g., a lower peak inspiratory ow and
breathing frequency for male rats in the High (PG/VG) group vs.
vehicle, but overall no clear pattern for PG/VG exposure was shown.
More generally, for female rats, addition of nicotine to PG/VG
resulted in increases in the peak inspiratory ow, tidal volume, and
minute volume ewith the resulting minute volumes for the
nicotine-containing aerosols, especially for Nicotine þHigh (PG/
VG), in a comparable range for male and female rats
(Supplementary Fig. 1). Consequently, a higher relative nicotine
uptake per unit of body weight can be expected for female rats, as
was indicated by the determinations of nicotine in plasma (Fig. 1C).
3.2. Body weight and food intake
All animals showed increases in body weight over time
throughout the 90-day exposure period (Supplementary Fig. 2A). At
the end of the exposure period, male rats displayed a slightly
lower body weight in exposure groups containing nicotine
(Supplementary Fig. 2B). In contrast, female rats had a signicantly
higher body weight when exposed to PG/VG containing nicotine.
While a decrease in body weight in males was also noted upon
exposure to nicotine-containing aerosols in our previous 28-day rat
exposure study, the body weights of the female rats did not clearly
differ in that study (Phillips et al., 2015).
Unexpectedly, food consumption in both sexes was higher in the
exposure groups containing nicotine in the PG/VG aerosol
(Supplementary Fig. 2C). Nicotine is reported as the appetite-
suppressing component of smoke that causes a reduction in food
consumption (Grunberg et al., 1986, 1987; Jo et al., 2002). The
higher food consumption noted in this study may be explained by
differences in the study design and the aerosol compositions, and
possibly the complex effects of nicotine on metabolism and stress
response systems (see below). Of note, signicant differences in
food consumption relative to body weight were not observed in any
group compared with the sham group in our previous nicotine/
pyruvate 28-day rat inhalation study (Phillips et al., 2015). Further
studies could determine whether these results are due to nicotine
only, or if other constituents of the aerosols may have an effect on
food consumption.
3.3. Effects on the respiratory system: standard toxicology
We found higher absolute lung (with larynx and trachea)
weights in female rats exposed to the high concentration of PG/VG
plus nicotine compared with the vehicle group (Supplementary
Fig. 3). However, in female rats the statistically signicant effect
in absolute values disappeared after normalization to body weight,
whereas a slight increase in the normalized weight, compared with
the vehicle group, became signicant for the Nicotine þHigh (PG/
VG) group in males. Note that a signicant decrease in normalized
lung weights was also observed for the sham compared with the
Table 3
Particle size distribution (median mass aerodynamic diameter, MMAD). Mean
MMAD values were calculated from 14 determinations with the standard deviation
(SD). The geometrical standard deviation (GSD) was calculated and the ranges are
given.
Mean MMAD [
m
m] SD [
m
m] GSD range
Sham eee
Vehicle 1.4 0.1 1.6e1.7
Low (PG/VG) 1.7 0.1 1.6e1.8
Med (PG/VG) 2.0 0.1 1.7e1.9
High (PG/VG) 2.0 0.1 1.8e2.2
Nic þLow (PG/VG) 1.8 0.1 1.6e1.9
Nic þMed (PG/VG) 2.0 0.1 1.6e1.9
Nic þHigh (PG/VG) 1.9 0.2 1.6e2.1
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332320
vehicle group in male rats, which suggests slight effects of the sa-
line aerosol (vehicle) inhalation on the lung weights.
Histopathological evaluation of the respiratory tract showed
only a low incidence of ndings (i.e., a low number of animals with
a histological response), with minimal severity for a number of
parameters which can be considered as adaptive changes, such as
induced by dehydration (Burger et al., 1989)(Fig. 2AeD,
Supplementary Table 3). These ndings were mainly restricted to
the larynx, with mild ndings in the nicotine-containing PG/VG
groups of basal cell hyperplasia and squamous cell metaplasia.
Inltration of unpigmented macrophages in the lung occurred only
at low levels, with slightly but signicantly lower levels in female
rats exposed to low and medium levels of PG/VG (with and without
nicotine) (Fig. 2E).
To analyze the degree of inammation in the lungs caused by
the inhalation of the aerosols, the number of free lung cells (cells in
BALF) was determined using ow cytometry; only a slightly higher
total cell count was observed in female rats exposed to nicotine-
containing aerosols (Fig. 2F). This was further supported by the
differential cell count (Supplementary Fig. 4), which showed that
the macrophage counts were slightly higher for the female
Nic þHigh (PG/VG) group compared with the vehicle group. No
clear exposure-related changes were detected by multi-analyte
proling of a cytokine panel (Supplementary Fig. 5).
3.4. Effects on the respiratory system: systems toxicology
We completed the assessment of the respiratory effects with a
systems toxicology analysis, which included gene expression (GEX)
and protein expression (PEX) proling of the nasal epithelium and
lung tissue of female rats.
Consistent with the low to absent histopathological ndings in
the nose, GEX proling of nasal epithelium did not show any
differentially expressed genes in the exposure groups compared
with the vehicle control (FDR-adjusted p-value <0.05)
(Supplementary Fig. 6A). PEX proling only showed seven down-
regulated proteins in the High (PG/VG) group compared with
vehicle, which, however, were not identied for the corresponding
Nicotine þHigh (PG/VG) group (Supplementary Fig. 6B).
In the lung, compared with the vehicle control, all exposure
groups displayed only limited effects on the lung transcriptome and
proteome, and no gene/protein was differentially expressed be-
tween the vehicle and sham controls under the applied threshold of
the FDR-adjusted p-value <0.05 (Supplementary Fig. 6C/D). Only
two genes were differentially expressed in the Low PG/PV group:
Alas2 (down) and LOC314492 (up) (immunoglobulin heavy chain
variable region); four genes were down-regulated in the nicotine-
containing Low PG/PV group: Alas2, Hbb, LOC689064 (beta-
globin [Rattus norvegicus]), and LOC287167 (Hba-a1 hemoglobin
alpha, adult chain 1 [Rattus norvegicus]) (Fig. 3A and B). Although
the expression of these genes was not signicantly altered in other
groups, they were quantied with similar high fold-changes. These
changes do not appear to be treatment-specic, and may, for
example, be explained by slight differences in the lung perfusion
(residual blood constituents) between the rats.
The comparisons with the vehicle group did not clearly identify
genes/proteins associated with nicotine exposure. However, the
shape of the volcano plots ewith a tendency toward increased
Fig. 2. Effects on respiratory organs. (AeE) Histopathology of the respiratory tract organs (mean scores ±SEM). Plot titles show the evaluated endpoints. Statistically signicant
differences from the vehicle group are represented by asterisks (*); statistically signicant differences between groups with and without nicotine (at the same PG/VG concentration)
are represented by carets (^)(p-value <0.05). (F) Mean ±SEM absolute BALF cell counts for male and female rats (N ¼10). See Supplementary Fig. 4 for immune-cell type dis-
tributions in BALF. PG, propylene glycol; VG, vegetable glycerin; Nic, nicotine; BALF, bronchoalveolar lavage uid.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 321
Fig. 3. Evaluation of lung effects by systems toxicology. (A) Lung gene expression response in groups of female rats compared with the vehicle group (N ¼6). Volcano plots show
the amplitude (log
2
fold-change, x-axis) and signicance (log10 FDR-adjusted p-value, y-axis) for each quantied transcript. Transcripts with an FDR-adjusted p-value <0.05 are
considered signicant and are shown as yellow (up) and cyan (down) dots above the dotted signicance threshold line at -log
10
(fdr p-value) ¼1.3. (B) Lung protein expression
response for groups of female rats compared with the vehicle group (N ¼6). Volcano plot as in A.(CeD) Volca no plot showing average nicotine effects for (C) gene expression (GEX)
and (D) protein expression (PEX) data. (E) Gene-set analysis (GSA) for GEX data for the average nicotine effect. Set enrichment was assessed by gene-set analysis with the mean of
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332322
levels of signicance for the nicotine containing PG/VG groupsd
suggested a low-level effect on the lung transcriptome and prote-
ome from exposure to PG/VG with added nicotine. Thus, to more
clearly detect these slight effects of nicotine addition, differentially
expressed genes/proteins were identied for the average nicotine
effect by contrasting the three nicotine þPG/VG groups vs. vehicle
with the three PG/VG (without nicotine) groups vs. vehicle (Fig. 3C
and D). This analysis found 97 genes and 6 proteins that were, more
generally, signicantly associated with the added nicotine in the
aerosol (FDR-adjusted p-value <0.05).
Gene-set but not protein-set analysis identied several bio sets
possibly affected by nicotine addition to the aerosol (Fig. 3E):
Several up-regulated gene sets reected changes in the metabolism
of xenobiotics (e.g., Xenobioticsand Metabolism of xenobiotics
by P450), and the overwhelming majority of down-regulated gene
sets were T-cell-related (e.g., TCR-a(ctivation) pathwayand T-
helper pathway).
Xenobiotic metabolism enzymes up-regulated by added nico-
tine included Cyp1a1 and Fmo3 (Fig. 3F). Cyp1a1, which was only
quantied by transcriptomics, was previously identied as a
nicotine-responsive gene in a 28-day rat inhalation study (Phillips
et al., 2015). Flavin-containing monooxygenase 3 (Fmo3) can
catalyze the oxidation of nicotine to nicotine N
0
-oxide (Hukkanen
et al., 2005).
To follow up on the observed down-regulation of T-cell-related
gene sets, we also evaluated the expression of immune-cell marker
genes in the lung tissue (Fig. 3G). Of the cell types evaluated
(neutrophils, myeloid cells, B-cells, and T-cells), T-cell markers
showed the clearest down-regulation in the nicotine-containing
exposed groups, with signicant down-regulation of the T-cell re-
ceptor components Cd3e and Cd3g for the average nicotine effect
(FDR-adjusted p-value <0.05).
In summary, while PG/VG exposure displayed very limited ef-
fects on the lung transcriptome and proteome, a few molecular
effects of exposure to the nicotine-containing PG/VG aerosols could
be identied; these included the up-regulation of the xenobiotic
metabolism enzymes Cyp1a1 and Fmo3 and the down-regulation
of T-cell-related transcripts.
3.5. Effects on hematology and blood clinical chemistry
Except for a slightly higher mean corpuscular hemoglobin
concentration (MCHC) in the blood of male rats exposed to the
highest concentration of PG/VG plus nicotine compared with PG/
VG exposure without nicotine, the red blood cell evaluation of male
rats revealed no changes in response to exposure compared with
the vehicle-exposed group (Supplementary Fig. 7). Female rats had
a mild but statistically signicant reduction in total erythrocyte
counts and a lower mean corpuscular hemoglobin concentration,
but a higher mean corpuscular volume after exposure to the
highest PG/VG concentration containing nicotine (Supplementary
Fig. 7). Similarly, in a 28-day inhalation toxicity study on nicotine
and pyruvic acid, no change in red blood cell parameters was
detected when male animals were exposed to three concentrations
of nicotine compared with a sham group, but female rats displayed
higher mean corpuscular volumes (Phillips et al., 2015). Of note,
such changes in these red blood cell parameters can occur sec-
ondary to acute stress effects (see below) (Everds et al., 2013).
The analysis of other blood chemistry parameters revealed
lower total protein concentrations (female rats only), lower total
cholesterol, and lower glucose concentrations in rats exposed to all
PG/VG concentrations containing nicotine when compared with
the vehicle-exposed groups or to the respective PG/VG-exposed
groups without nicotine (Fig. 4 AeB, Supplementary Fig. 7D). In
female rats, triglyceride levels also tended to be lower in the
nicotine-containing PG/VG versus the PG/VG only groups, but sig-
nicance was only reached for Nicotine þHigh (PG/VG) versus High
(PG/VG), and triglyceride levels displayed variability between the
groups in male rats, with signicantly increased levels in the
vehicle versus sham group (Fig. 4 C). Other changes found only in
the female rats exposed to PG/VG plus nicotine included lower
creatinine and calcium concentrations (Supplementary Fig. 7). The
effects of nicotine on cholesterol and glucose were similar to the
ndings from a previous study (Phillips et al., 2015), which, in both
sexes, also showed lower cholesterol and glucose concentrations
upon exposure to nicotine-containing aerosols compared with
sham. However, lower calcium concentrations were not observed
in the nicotine-exposed female groups within the shorter 28-day
time frame of that study.
In line with the global down-regulation of clinical parameters in
the PG/VG nicotine-exposed groups, many lipid classes were down-
regulated in the plasma of female rats (Fig. 4 D).
3.6. Effects in the liver: standard toxicology
We found higher relative and absolute liver weights in male and
female rats exposed to all concentrations of PG/VG with nicotine
compared with the vehicle-exposed group (Fig. 5A, Supplementary
Table 4). Similar results (higher absolute liver weights in females)
were obtained in a previously reported 28-day rat inhalation study
on nicotine and pyruvic acid (Phillips et al., 2015).
The exposure to nicotine resulted in increased activities of the
two liver enzymes alanine aminotransferase and alkaline phos-
phatase in female rats, and a tendency for an increased activity of
these enzymes in male rats (Fig. 5B and C). However, note that
alanine aminotransferase activity also displayed test-item inde-
pendent changes, with signicant differences between the sham
and vehicle groups, suggesting that the small relative changes be-
tween the groups are to be interpreted with caution. Moreover, ALP
is an indicator of cholestasis in rats and humans which, in this
study, could be secondary to the diffuse swelling of hepatocytes,
causing compression of bile canaliculi and ducts.
A more severe occurrence of hepatocyte vacuolation (maximum
severity 1.7 out of 5) was generally observed in liver sections from
female and male rats in the nicotine-containing PG/VG groups
compared with the vehicle-exposed group (Fig. 5 DeF). To deter-
mine the glycogen content of hepatocytes in the liver sections,
periodic acid-Schiff (PAS) staining after diastase digestion was
performed (Supplementary Fig. 8). With the potential caveat that
glycogen staining can be sensitive to the xation method (Horobin,
2014; Zakout et al., 2010), the results suggested that glycogen
storage was not the main reason for cytoplasmic vacuolation, as
after diastase digestion only a partial reduction in the intensity of
the PAS staining was observed. Moreover, the distribution of the
PAS staining was centro-lobular, and not pan-lobular as it would be
with cytoplasmic vacuolation. In line with our ndings, in a
the absolute log2 fold-change of the average nicotine effect as the gene-set statistic (from mSigDB c2.cp collection). Gene permutation (Q1) and sample permutation (Q2) were used
to assess statistical signicance. Gene sets with an adjusted p-value <0.05 for both Q1 and Q2 were considered and represented in a bar chart, with the length of the bars reecting
the mean fold-change of the gene set. (F) Gene expression (GEX) and protein expression (PEX) proles for Cyp1a1 and Fmo3. Differences compared with the vehicle group with a
raw p-value <0.05 are marked (*). Note that the average nicotine effect for these changes is signicant (FDR-adjusted p-value <0.05) and that Cyp1a1 was only detected by GEX. (G)
Gene expression proles for a panel of immune-cell markers. Fold-changes of the average nicotine effect and the group comparisons with the vehicle group are color-coded, and
statistical signicance is marked (adjusted p-value, x ¼<0.05, * ¼<0.01).
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 323
previous 28-day rat inhalation study with aerosols of nicotine and
pyruvic acid (Phillips et al., 2015), cytoplasmic vacuolation of he-
patocytes was observed for most rats of the study with increased
levels for rats exposed to nicotine and sodium pyruvate. However,
in this study, the increase in vacuolation could be associated with
an altered intracytoplasmatic deposition of glycogen.
3.7. Effects on the liver: systems toxicology
We leveraged the liver systems toxicology data to characterize
the nicotine exposure-related effects in more detail. Gene and
protein expression proling did not show signicant effects for the
PG/VG exposure groups compared with vehicle exposure; also, no
gene or protein was differentially expressed between the sham and
vehicle groups (Supplementary Fig. 9). However, a number of genes
and proteins demonstrated signicant differential expression be-
tween the nicotine PG/VG and vehicle exposure groups; there were
three, four, and six differentially expressed proteins for the low,
medium, and high PG/VG groups with nicotine vs. vehicle groups.
Of note, the volcano plots indicated an overall stronger effect for
PG/VG groups with nicotine than for the PG/VG groups without
nicotine on the group comparison statistics; however, these effects
of added nicotine mostly did not reach signicance at the 5% FDR
level.
To increase the sensitivity of the analysis, the average effect of
added nicotine was analyzed across all PG/VG concentrations: 290
genes and 218 proteins were signicantly associated with nicotine
in the aerosols (Fig. 6A). Gene-set and protein-set analysis revealed
several biological bio sets possibly affected by the added nicotine
(Fig. 6B/C). Up-regulated categories included cholesterol biosyn-
thesis, xenobiotic metabolism, and lipid metabolism. Down-
regulated categories included cytokine- and chemokine-,
extracellular matrix-, and nitrogen metabolism-related processes.
To complement this analysis, we compared the proteins and genes
differentially affected by nicotine in the PG/VG aerosols (Fig. 6D). Of
the 218 differentially expressed proteins, 28 (approximately 13%)
were also differentially expressed at the gene level. While no bio-
logical function was enriched for the 9 shared down-regulated
proteins, the 19 shared up-regulated proteins were enriched for
biological functions in the metabolism of lipids and xenobiotics.
From these gene and protein expression data, nicotine-driven lipid
and xenobiotic metabolism emerged as the most clearly affected
biological functions.
To further expand this integrative analysis and to evaluate the
nicotine addition-related changes in more detail, we applied a
joined functional association clustering approach to the proteomics
and transcriptomics data (Titz et al., 2015a, 2015b). To do so, we rst
identied clusters of functionally associated proteins and genes
affected by nicotine-containing PG/VG aerosol exposure in the
liver; subsequently, identied functional gene and protein clusters
that shared a signicant number of components were re-clustered
and annotated. The main four identied clusters were enriched for
fatty acid metabolism,xenobiotic, steroid, retinol, and bile acid
metabolism,TCA cycle, pyruvate, and gluconeogenesis, and
cholesterol biosynthesisfunctions (Fig. 6E). Thus, the identied
clusters were in overall agreement with the gene-set analysis and
signicant gene/protein overlap results.
Within the fatty acid metabolismcluster, the majority of
genes/proteins (at least 9 out of 17) were involved in lipid beta-
oxidation: Acss2, Eci1, Cpt2, Hadha, Hadhb, Acad11, Acadl, Ech1,
and Ehhadh. This included the core enzymes that mediate the lipid
oxidation and cleavage cycles; for example, acyl-coA hydrogenases
(Acad11 and Acadl) catalyze the initial oxidation step, and Hadha
and Hadhb, the components of the mitochondrial tri-functional
Fig. 4. Blood clinical chemistry and serum lipidomics. (A) Total cholesterol, (B) glucose, and (C) triglycerides (mean ±SEM). Statistically signicant differences from the vehicle
group are represented by asterisks (*); statistically signicant differences between groups with and without nicotine (at the same PG/VG concentration) are represented by carets (^)
(p-value <0.05). (D) Plasma lipid changes measured by lipidomics. Log
2
fold-change of lipid classes compared with vehicle group or for nicotine effect is color-coded, and sta-
tistically signicant changes are marked. TAG, triacylglycerol; SM, sphingomyelin; PI, phosphatidylinositol; PE, phosphatidylethanolamine; PC, phosphatidylcholine; LPE; lyso-
phosphatidylethanolamine; LPC, lyso-phosphatidylcholine; DAG, diacylglycerol; CE, cholesteryl esters.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332324
enzyme, catalyze the subsequent hydratase, dehydrogenase, and
thiolase steps during the beta-oxidation cycles. This cluster also
included two key enzymes involved in ketone body formation: 3-
hydroxy-3-methylglutaryl-CoA synthase 2 (Hmgcs2) and type 1
3-hydroxybutyrate dehydrogenase (Bdh1). Ketone bodies include
acetoacetate, beta-hydroxybutyrate, and acetone, and are produced
from fatty acids in the liver, especially during fasting and extensive
prolonged exercise (McGarry and Foster, 1980).
The observed up-regulation of genes/proteins by nicotine
exposure in the TCA cycle, pyruvate, and gluconeogenesiscluster
is likely functionally linked to the shift toward lipid oxidation and
ketogenesis (see Discussion). This included protein up-regulation of
phosphoenolpyruvate carboxykinase 1 (Pck1), the key enzyme of
gluconeogenesis, and serine dehydratase (Sds) and glutamic-
Fig. 5. Effects on the liver. (A) Liver weight normalized to body weight for male and female rats (mean ±SEM, N ¼10). Statistically signicant differences from the vehicle group
are represented by asterisks (*); statistically signicant differences between groups with and without nicotine (at the same PG/VG concentration) are represented by carets (^)(p-
value <0.05). (BeC) Liver enzyme activities measured in blood (mean ±SEM): (B) Alkaline phosphatase and (C) alanine aminotransferase. (D) Liver vacuolation assessed by
histopathology (mean ±SEM). (EeF) Microscopy images of hepatocyte vacuolation. Representative examples of H&E staining results are shown for a female (E) and a male (F) rat.
Periodic acid-Schiff (PAS) and PAS-D staining results are shown in Supplementary Fig. 8.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 325
Fig. 6. Evaluation of liver effects using systems toxicology analysis. (A) Volcano plot showing average nicotine addition effect for gene expression (GEX, left) and protein
expression (PEX, right) data for the liver. (B) Gene-set analysis (GSA) for GEX data for the average nicotine effect. Setenrichment was assessed by gene-set analysis with the mean of
the absolute log2 fold-change of the average nicotine effect as the gene set statistic (from mSigDB c2.cp collection). Gene permutation (Q1) and sample permutation (Q2) were used
to assess statistical signicance. Gene sets with an adjusted p-value <0.05 for both Q1 and Q2 were considered and represented as a bar chart, with the length of the bars reecting
the mean fold-change of the gene set. (C) As in B, but for PEX data. (D) Venn diagram for the unique and shared genes (GEX) and proteins (PEX) affected by nicotine. Gene-set
overrepresentation results for the 19 signicantly up-regulated proteins/genes shared between PEX and GEX (FDR-adjusted p-value <0.05). Note that no gene set was
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332326
oxaloacetic transaminase 1 (Got1), enzymes that convert amino
acids into substrates for gluconeogenesis (Yang et al., 2011). While
the aforementioned up-regulated Hmgcs2 enzyme channels
Acetyl-CoA into the ketone-body synthesis pathway, its isoform
Hmgcs1 channels this molecular building block into the cholesterol
biosynthesis pathway (Hegardt, 1999). Added nicotine increased
the expression of Hmgcs1, together with other genes/proteins in
the cholesterol biosynthesiscluster. Finally, the xenobiotic, ste-
roid, retinol, and bile acid metabolismcluster contains both en-
zymes more closely associated with xenobiotic metabolism
processes, e.g. Cyp1a1 and Ephx, and enzymes that contribute to
the synthesis of endogenous metabolites such as steroids (e.g.,
Cyp3a18, Cyp17a1) and bile acids (e.g., Cyp7a1).
The most down-regulated gene sets in the liver transcriptomics
data included chemokine receptors bind chemokinesfrom the
Reactome database and cytokine-cytokine receptor interaction
from the KEGG database (Fig. 6B), but these were not covered by
the analyses focused on overlapping effects between the prote-
omics and transcriptomics data. Thus, to complement the other
analyses, we investigated the expression proles of these gene sets
in more detail (Fig. 6F). These down-regulated genes included the
chemokines Cxcl9 and Cxcl14 and the interferon gamma receptor 2
(Ifngr2). Of potential interest in our context, Cxcl14 knockout mice
have been reported to demonstrate a metabolic phenotype, and
Cxcl14 has been proposed as a regulator of glucose metabolism
(Hara and Tanegashima, 2012).
Finally, we evaluated whether these nicotine exposure-related
gene/protein adaptations resulted in changed lipid proles in the
liver (Fig. 6GeI). As for the other tissues, for lipidomics, we had only
included samples from rats exposed to High PG/VG with and
without nicotine (and the control groups). Only one lipid, Lac-
Cer(d18:1/24:1), was signicantly down-regulated in the High
Nic þPG/VG vs. vehicle comparison, and two lipids, PI 18:1/18:1
and PC 16:0/18:1, were signicantly up-regulated in the compari-
son of High PG/PV with nicotine vs. PG/PV without nicotine
(Nicotine) (FDR-adjusted p-value <0.05) (Fig. 6G). Next, we
focused on the differential abundance of conjugated different fatty
acids (FA) (aggregated over the different lipid classes) (Fig. 6H): FA
22:5, FA 22:4, FA 20:1, FA 18:1, and FA 16:0 (number of carbons:
number of double bonds) were signicantly more abundant
following exposure to nicotine-containing aerosol (FDR-adjusted p-
value <0.05), but no chain-length- or desaturation-dependent
pattern could be identied. In addition, for the lipid classes
(Fig. 6I), only diacylglycerols (DAGs) showed a signicant increase
following exposure to nicotine-containing aerosol, and LacCer
decreased signicantly in the Nic-High PG/PV group vs. vehicle.
Overall, the hepatic response to nicotine exposure suggested a
shift toward lipid oxidation, gluconeogenesis, ketone body forma-
tion, and cholesterol biosynthesis. However, the effects on the
steady-state lipid levels in the liver were limited and did not reveal
a clear pattern.
3.8. Systemic stress responses
Previously, systemic effects in toxicology studies have been
linked to generalized stress responses (Everds et al., 2013),
including in nicotine exposure studies (Phillips et al., 2015). It has
been proposed that some of these changes are linked to the acti-
vation of the hypothalamic-pituitary-adrenal (HPA) axis, which
involves corticosterone, adrenocorticotropic hormone (ACTH), and
corticotrophin-releasing hormone (CRH). For example, the sys-
temic response of rats to acute and chronic immobilization stress
includes metabolic changes such as a decrease in the triacylglycerol
concentration in plasma (Ricart-Jan et al., 2002).
A decrease in thymus weight has been proposed as a sensitive
indicator of systemic stress (Everds et al., 2013). In the current
study, nicotine exposure, especially for female rats, but also for the
Nicotine þLow (PG/VG) male group, resulted in a signicant
decrease in thymus weight (Supplementary Fig. 10A). In addition,
an increase in adrenal gland weight is commonly observed with
stress. The female rats in particular, but also the male rats, showed a
signicant increase in adrenal gland weights in our study. The left
adrenal weights normalized to body weight are shown in
Supplementary Fig. 10B.
In agreement with a previous nicotine inhalation study (Phillips
et al., 2015), these observations suggest a nicotine exposure-related
stress response that is especially prominent in female rats ewhich
might be due to the higher plasma nicotine concentrations, and
thus higher systemic exposure, achieved under these inhalation
conditions in the female rats (see 3.1).
4. Discussion
The inhalation toxicology of PG/VG aerosols with and without
nicotine was investigated in a 90-day rat inhalation study according
to OECD TG 413 (OECD, 2009b), in combination with additional
molecular endpoints to allow for a systems toxicology assessment.
Aerosol generation and delivery. The aerosols were adminis-
tered via the inhalation route using nose-only exposure to ensure
reproducibility of aerosol uptake. Nose-only exposure minimizes
aerosol deposition on the animal fur, which could be subsequently
taken up by grooming and contribute as an oral uptake or trans-
dermal component to the total nicotine uptake. Aerosols were
generated with commercially available 6-jet Collison nebulizers
that function by applying pressure to force liquid solutions through
small apertures resulting in the production of a ne aerosol. This
mode of aerosol generation and exposure has been previously
validated as being appropriate for studying nicotine exposure via
inhalation (Shao et al., 2013), and was applied to nicotine aerosol
generation in our prior 28-day inhalation study where stable
aerosols were generated through the study conduct (Phillips et al.,
2015). This setup allowed for exact control of the exposure condi-
tions, including the reproducible exposure to three concentrations
of the PG/VG (þ/nicotine) mixtures, which was conrmed both
by analyzing the constituents in the aerosol and after the uptake
biomarkers in the plasma/urine (Fig. 1). In future studies, the ability
to control inhalation-route aerosol delivery of mixtures with pre-
cision could, for example, support comparative assessments of
aerosols containing different avor mixtures. With its experimental
design, our study extends and complements previous toxicity
studies for PG and VG, which assessed other modes of exposure
(e.g., RTECS, 1985), inhalation of the individual compounds (e.g.,
Renne et al., 1992; Suber et al., 1989; Werley et al., 2011), or PG/VG
in ECIG liquids (e.g., Werley et al., 2016).
signicantly enriched for the 9 shared down-regulated proteins/genes. (E) Gene expression (GEX) and protein expression (PEX) proles for functional clusters shared between GEX
and PEX data. See methods for details on the identication of the six marked functional clusters. Fold-changes of the average nicotine effect and the group comparisons with the
vehicle group are color-coded, and statistical signicance is marked (adjusted p-value, x ¼<0.05, * ¼<0.01). (F) Gene expression proles for the chemokine receptors bind
chemokines [REACTOME]and cytokine-cytokine receptor interaction [KEGG]gene sets. Note that the respective proteins were not quantied, otherwise representation as in E.
(G) Lipidomics response proles in liver. Volcano plots (as in Fig. 3A) represent the group differences from vehicle exposure and the average nicotine effect on the liver lipidome. (H)
Effects on concentrations of the different (conjugated) fatty acids with different length and degree of desaturation (carbon chain length: number of double bonds). Heatmap
representation as in Fig. 6E. (I) Effects on lipid class concentrations as in Fig. 4D.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 327
The PG/VG mixture was assessed at three test atmosphere target
concentrations in this study: PG (mg/L)/VG (mg/L): 0.174/0.210
(low), 0.520/0.630 (medium), and 1.520/1.890 (high). These con-
centrations correspond approximately to a daily delivered dose of
50 (PG)/60 (VG) mg/kg (low), 150/181 (medium), and 438/544 mg/
kg (high) (formula adapted from (Alexander et al., 2008): D ¼
(C RMV d)/BW), where D is the dose (mg/kg), C is the con-
centration of constituent in aerosol (mg/L), RMV is the respiratory
minute volume (for rats, 0.2 L/min was used), d is the duration of
exposure (min), and BW is body weight (kg) (for rats, 0.25 kg was
used). Calculating the human equivalent dose (HED) based on the
body surface area by dividing the rat dose by a factor of 6.2 (CDER,
2005) yields a maximum HED of approximately 71 mg/kg for PG
and 88 mg/kg for VG, or a total uptake of 4.3 g/day of PG and 5.3 g/
day of VG for a 60-kg adult human. The delivered quantities exceed
normal daily use of e-liquids: The majority of human users
consume less than 4 mL e-liquid per day (Action on Smoking and
Health, 2016) with variable ratios of PG and VG (Peace et al.,
2016); 4 mL of a 100% (v/v) PG solution correspond to 4.2 g PG,
and 4 mL of a 100% (v/v) VG solution correspond to 5.0 g VG.
The added nicotine concentration in the test atmosphere was
targeted to be 23
m
g/L (corresponding to approximately 6.6 mg/kg/
day according to (Alexander et al., 2008)). The daily inhaled dose
was close to the maximum tolerated dose (MTD) of 7.5 mg/kg for
acute administration and approximately half of the 80% mortality
dose (15 mg/kg) determined by single intratracheal instillation
(Phillips et al., 2015). Calculating the human equivalent dose (HED)
based on the body surface area by dividing the rat dose by a factor
of 6.2 (CDER, 2005) yields a HED of 1.1 mg/kg, corresponding to a
daily nicotine dose of 66 mg for a 60-kg adult human. As discussed
above, the majority of human users consume less than 4 mL e-
liquid per day (Action on Smoking and Health, 2016), and the Eu-
ropean Tobacco Products Directive (TPD Article 20, 3. B) caps the
maximum nicotine concentration in an e-liquid at 20 mg/mL
(European Parliament and Council, 2014). Based on this, a
maximum of 80 mg inhaled nicotine per day is to be expected, close
to (121% of) the calculated HED for a 60-kg adult human. At these
nicotine concentrations, a mild effect of nicotine was expected
based on our previous study (Phillips et al., 2015).
Respiratory effects. At the three tested concentrations, PG/VG
without nicotine did not show any clear effects on the measured
endpoints in the respiratory system, including histopathology, lung
inammation, and the molecular responses measured by tran-
scriptomics and proteomics. In a published 90-day rat inhalation
study (for 6 h/day, 6 days/week), 1.1 mg/L and 2.2 mg/L PG in the
aerosol (317 mg/kg and 634 mg/kg daily delivered dose, according
to (Alexander et al., 2008), see above) caused an increase in the
numbers of goblet cells or increased mucin production in the nose,
and 2.2 mg/L PG (634 mg/kg daily delivered dose) caused nasal
hemorrhage, possibly as a dehydration effect (Suber et al., 1989). In
another rat inhalation study for PG, bleeding around the nose was
observed after an acute (7-day) exposure, and squamous meta-
plasia in the larynx was observed after 28 days of PG exposure
(Werley et al., 2011), which was only observed in the presence of
nicotine in our study (see below). However, Werley et al. used a
different exposure regimen, and the NOEL for PG in the 28-day
study was estimated at 30 mg/L PG for 12 min per day (288 mg/
kg daily delivered dose, according to (Alexander et al., 2008), see
above). In a 13-week rat inhalation study with glycerol, Renne et al.
mainly observed mild squamous metaplasia of the epithelium lin-
ing of the epiglottis for 0.662 mg/L glycerol (6 h/day, 5 days/week)
(191 mg/kg daily delivered dose, according to (Alexander et al.,
2008), see above), which was considered an adaptive response to
the mild irritant effects of the aerosol. With this, the lack of effects
of the PG/VG aerosol (or mild effects that did not reach
signicance), up to a daily delivered dose of 438 mg/kg PG and
544 mg/kg VG, is overall in line with the observation of only mild
irritation effects in previous studies of the individual compounds.
The addition of nicotine to the PG/VG aerosol resulted in only a
low incidence and mild severity of histopathological ndings in the
upper respiratory tract (Fig. 2). These ndings were mainly local-
ized in the larynx, but also included slightly higher counts of free
lung cells for female rats exposed to nicotine-containing aerosols
(Fig. 2F). Squamous metaplasia of the larynx is among the most
sensitive endpoints for inhalation studies, reacts to a wide range of
conditions, including dehydration by low-humidity air, and has
been identied as a sensitive adaptive rather than toxicologically
relevant response (Burger et al., 1989; Osimitz et al., 2007). These
mild effects following exposure to the nicotine-containing aerosols
should be placed in context by contrasting the ndings of this study
with those for cigarette smoke (CS) from the 3R4F reference ciga-
rette in two previously published 90-day rat inhalation studies
with a comparable exposure regimen (6 h per day, 5 days per week)
(Kogel et al., 2016; Oviedo et al., 2016; Wong et al., 2016). In these
studies, exposure to mainstream 3R4F smoke at a concentration of
23
m
g/L nicotine caused severe squamous cell hyperplasia-
metaplasia in the larynx and severe lung inammation, including
an increase in free lung cells (up to >5-fold higher free cell numbers
for 3R4F vs. sham exposure) and a dramatic increase in BALF neu-
trophils, as well as several signicantly affected cytokines in the
lavage uid.
Similarly, while CS from 3R4F clearly affected the lung tran-
scriptome in these previous studies (e.g., >1200 signicantly
differentially regulated transcripts in female rats exposed to smoke
from 3R4F (23
m
g/L nicotine) vs. Sham (Kogel et al., 2016)), only very
few signicantly differentially expressed transcripts could be
identied in the current study (Fig. 3). However, the aggregated
analysis of the nicotine exposure effects across all exposure groups
allowed for identication of the low-level effects of nicotine: These
included up-regulation of the transcripts for xenobiotic enzymes
(Cyp1a1 and Fmo3) and down-regulation of T-cell-associated
genes. This is consistent with a previous rat inhalation study, where
the up-regulation of Cyp1a1 following nicotine aerosol exposure
had been identied as the clearest concentration-dependent effect
of nicotine in the lung tissue (Phillips et al., 2015). To our knowl-
edge, Fmo3 up-regulation has not been reported as a marker of
nicotine exposure in rat lungs. Fmo3 is known as the enzyme that
catalyzes nicotine-N
0
-oxide formation in the liver (Yildiz, 2004),
and has been hypothesized to contribute to the metabolism of
nicotine to nicotine-N
0
-oxide in rat lung microvascular endothelial
cells (Ochiai et al., 2006). Thus, Fmo3 induction is likely to be
functionally associated with the xenobiotic response of the lung
tissue to nicotine exposure. T-cell-associated genes were down-
regulated following nicotine exposure in lung tissue (Fig. 3). This
is in line with observations in our previous nicotine aerosol expo-
sure study (Phillips et al., 2015) and reproduces a generally
observed exposure effect for nicotine-containing aerosols in the
lung (e.g., Kogel et al., 2016). This observation could either reect
the down-regulation of T-cell receptor components or a more
general decrease in T-cell numbers in the lung following nicotine
exposure. Possibly, this nding is linked to the previously noted
anti-inammatory effects of nicotine in the lung immune system,
including its suppression of T-cell responses (e.g., Singh et al., 2000;
Sopori et al., 1998).
Hematology and clinical chemistry. While hematology and
clinical chemistry did not show any signicant and consistent ef-
fects for PG/VG exposure, the exposure to nicotine-containing
aerosols compared with vehicle exposure (and comparing PG/VG
aerosols at corresponding concentrations with and without nico-
tine), especially in female rats, resulted in a signicant reduction in
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332328
blood glucose and lipid levels, including cholesterol and tri-
acylglycerols (Fig. 4). Similar effects of nicotine-containing aerosols
have been observed in previous rat inhalation studies (Oviedo et al.,
2016; Phillips et al., 2015; Wong et al., 2016). Unexpectedly, expo-
sure to nicotine-containing PG/VG aerosols also resulted in
increased body weight in female rats and in increased food con-
sumption in both male and female rats (Supplementary Fig. 2), but
these effects were not observed in Phillips et al. (2015) and,
generally, nicotine is more commonly reported to exert a sup-
pressive effect on food consumption and weight gain in the liter-
ature (e.g Grunberg et al., 1986).
Liver effects. In line with the other endpoints, the PG/VG
aerosols did not show clear and signicant effects on liver-related
endpoints. However, rats exposed to PG/VG aerosols with added
nicotine exhibited several hepatic effects, including increased liver
weight, increased liver enzyme activities in plasma, and increased
vacuolation of liver cells compared with vehicle exposure (Fig. 5).
Similar effects of nicotine on the liver have been reported in a
previous nicotine inhalation study (Phillips et al., 2015) and were,
more generally, also observed in previous rat inhalation studies for
other nicotine-containing aerosols, including CS (Oviedo et al.,
2016; Wong et al., 2016). Hepatocyte vacuolization could be
considered an adverse effect as it is consistently seen after mild and
subacute liver injury. However, several lines of evidence point to
the possibility that hepatocyte swelling with nonlipidic vacuoliza-
tion may reect a cellular adaptation, rather than a degenerative
change. Observations in animal and human livers suggest vacuo-
lated hepatocytes observed during liver injury are cells adaptively
altered to resist further insult. In particular, Nayak et al. suggest
hepatocyte vacuolization may reect an adaptive cellular response
(Nayak et al., 1996). For example, morphological and biochemical
investigations have shown that cytoplasmic vacuolation of hepa-
tocytes following low doses of CCl4 was due to excess accumulation
of glycogen. Low dose CCl4 exposed cells lacked features of
degeneration or regeneration, and were much less susceptible to
injury by larger subsequent CCl4 doses, as assessed by structural
and serum enzyme analyses (Nayak et al., 1996). In our previous
studies, we have also demonstrated that both severity score and
incidence of nicotine-induced cytoplasmic vacuolization in liver
cells reverted back to baseline levels after the 42 days recovery
period (Wong et al., 2016). Reversibility is an important factor in the
holistic interpretation of toxicity studies and given that hepatocyte
vacuolization induced by high dose of nicotine has been shown to
be readily and completely reversible on cessation of treatment this
would indicate a lower level of concern (Lewis et al., 2002).
Transcriptomics analysis of the liver showed that these changes
were associated with alterations in the metabolic state of the liver
(see below). In addition, PG/VG with nicotine exposure resulted in
down-regulation of chemokine and cytokine gene sets in the liver
(Fig. 6), including the chemokines Cxcl9 and Cxcl14, and the pro-
lactin (Prlr), Ifngr2, and oncostatin M (Osmr) receptors. Down-
regulation of cytokine and interferon signaling gene sets by nico-
tine exposure has also been observed in (Phillips et al., 2015) and
may, together with the aforementioned down-regulation of T-cell
markers in the lung, be related to the effect of nicotine on the
immune system (Han and Lau, 2014).
General metabolic shift. Exposure to the nicotine-containing
aerosols resulted in a general shift of the metabolic state, most
prominently in female rats ewhich might be due to the higher
plasma nicotine concentrations, and thus higher systemic expo-
sure, achieved under these inhalation conditions in the female rats
(see 3.1). Nicotine-exposed female rats displayed increased body
weights that were associated with increased food consumption. At
the same time, levels of glucose and a broad range of lipid classes
such as triacylglycerols and cholesterol were reduced in blood.
Systems toxicology investigations indicated that nicotine exposure
also affected metabolic pathways the in liver, including up-
regulation of fatty acid beta-oxidation, cholesterol synthesis,
gluconeogenesis, and ketone body formation pathways. Similar
effects of nicotine exposure were also observed in a previous
nicotine/pyruvate inhalation study (Phillips et al., 2015). In that
study, the metabolic effects of the nicotine-containing aerosols,
which were also most pronounced in female rats, included a
reduction in blood glucose, cholesterol, and other plasma lipid
levels. Overall consistent with our results, effects on liver meta-
bolism pathways in that study included up-regulation of choles-
terol biosynthesis, oxidative phosphorylation, and
gluconeogenesis, as well as effects on lipid and steroid metabolism.
Similar metabolic effects were also observed in previous CS expo-
sure studies, and have been attributed to a nicotine effect, poten-
tially conjoined with a systemic stress response (Everds et al., 2013;
Wong et al., 2016) (see below). Consistent with our results, Golli
et al. observed decreased cholesterol plasma levels and increased
liver gluconeogenesis enzymes (Pck1) for rats exposed intra-
peritoneally to e-liquids with nicotine for 28 days (Golli et al.,
2016). However, other responses to nicotine exposure were
different in that study, including the observed effects on food
consumption, body weight gain, and glucose plasma levels. In
addition, several non-inhalation nicotine exposure studies found
increases in lipid and cholesterol levels in plasma (Adluri et al.,
2008; Ashakumary and Vijayammal, 1997; Chattopadhyay and
Chattopadhyay, 2008; Valenca et al., 2008). Overall, we can
conclude that nicotine exposure affects lipid and glucose meta-
bolism in a complex manner. However, the resulting (and experi-
mentally observed) physiological states are likely dependent on the
route, concentration, and duration of exposure, among other fac-
tors. Possibly contributing to the differences in the nal induced
metabolic states are the complex effects of nicotine on the
cholinergic nervous system, interlinked with the physiological
response to general exposure stress, as discussed in the next
section.
Systemic stress. Nicotine exerts complex effects on the stress
response of animals (al'Absi et al., 2013; Picciotto et al., 2002).
Nicotine has been reported to affect different stress regulation
systems, most prominently the hypothalamicepituitaryeadrenal
(HPA) axis (Faraday et al., 2005; Matta et al., 1998; Rhodes et al.,
2001; Rohleder and Kirschbaum, 2006). Likewise, cold stress or
heat stress has been shown to increase the blood glucagon level in
rats, concomitant with increases in free fatty acids and glycerin
(Kuroshima et al., 1981). While free fatty acids were not measured
in the current study, the observed elevated glycerin levels in plasma
are consistent with such a metabolic stress response. Moreover, in
healthy non-smokers, nicotine infusion was observed to increase
serum levels of noradrenaline, adrenaline, glycerol, and free fatty
acids, while the plasma levels of glucagon, insulin, glucose, pyru-
vate, lactate, and cortisol did not signicantly change (Andersson
et al., 1993). In our study, consistent with an effect of the
nicotine-containing aerosols on these stress systems, we observed
several clinical responses that have been identied previously as
stress response phenotypes (Everds et al., 2013; Selye, 1936),
including a decrease in thymus weight and an increase in adrenal
gland weight (Supplementary Fig. 10). Notably, complex in-
teractions of nicotine and other sources of stress have been
observed (e.g., Chen et al., 2008; Cheng et al., 2005; Faraday et al.,
1999), which could have further modulated (induced or sup-
pressed) the stress-related effects in our study, in which the rats
were concurrently exposed to the stress of the nose-only exposure
procedure and nicotine-containing aerosols. For example, the HPA
axis is functionally interlinked with adipose tissue metabolism via
leptin (Spinedi and Gaillard, 1998), and glucocorticoids produced
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 329
by the adrenal cortex affect metabolism (e.g., increased gluconeo-
genesis in the liver) and modulate the activity of the immune
system (Nicolaides et al., 2015; Rose and Herzig, 2013). Therefore,
we cannot exclude the possibility that some of the observed
metabolic and immune-related effects of nicotine exposure may be
explained by the complex modulation of stress response pathways
by nicotine.
Limitations of the study. A general challenge in toxicological
assessment studies is to judge the possible implications, especially
of observed mild effectsdhere, for example, the histopathological
effects observed for the nicotine-containing aerosols on the larynx
and free lung cells. Above, we contrasted the effects of the tested
aerosols with the much more pronounced effects observed
following exposure to CS in a previous study. However, to further
improve comparability of the effect sizes in future studiesde.g.,
between ECIG aerosols and CSdit would be benecial to include
exposure to CS as a reference. In particular, systems toxicology
endpoints would benet from such a positive controlfor the
induced exposure effects. Using a Collison nebulizer for aerosol
generation, the current study assessed the effects of the unaltered
main ECIG liquid components. However, depending on the ECIG
design used for aerosol generation, overheating of the liquid may
occur, which can then, for example, result in the formation of al-
dehydes (Ahmad et al., 2016; Bekki et al., 2014; Farsalinos et al.,
2015). Thus, for a given ECIG/e-liquid combination, beyond the
assessment of the pure compounds, it will be pertinent to also
specically assess to what extent thermal decomposition products
are generated and how they affect the toxicity of the generated
aerosol. Another deviation of our nebulizer-generated test
aerosol from genuine ECIG aerosols produced by the dedicated
electronic nicotine delivery systems is the larger particle size,
which may result at least in part from hygroscopic growth due to
the dilution, and from differences in the measurement methods
(Fuoco et al., 2014; Lerner et al., 2015a; Manigrasso et al., 2015), or
the lack of trace metal (such as copper) contamination from
heater elements (Lerner et al., 2015a). Thus, the present paper re-
ports the basic inhalation toxicity of the e-liquids as the rst part of
a multi-stage assessment approach (see introduction). To fully
assess the risk of ECIG aerosols, additional studies will be required,
which would also include the contributions from the delivery
devices.
5. Conclusion
A 90-day rat inhalation study according to OECD TG 413 (OECD,
2009b) for the toxicological assessment of nebulized PG/VG aero-
sols with and without nicotine was conducted. Standard toxico-
logical endpoints were complemented with systems toxicological
analyses using transcriptomics, proteomics, and lipidomics of lung
tissue, liver tissue, and serum. Both standard and systems toxi-
cology endpoints demonstrated very limited biological effects of
PG/VG aerosol with no signs of toxicity. Systems toxicology ana-
lyses detected biological effects of nicotine exposure, which
included up-regulation of the xenobiotic-metabolizing enzymes
Cyp1a1 and Fmo3 in the lung and metabolic effects, likely inter-
linked with a generalized stress response to nicotine present in the
exposure aerosols.
Altogether, under the conditions of this 90-day SD rat study,
several nicotine related responses have been observed but taking in
to account the overall weight of evidence no adverse effects were
observed for PG/VG/nicotine up to 438/544/6.7 mg/kg/day, since
the vast majority of the effects observed are considered to be
adaptive changes caused by the nicotine levels and they have been
shown to be reversible on cessation of treatment.
Competing interests and funding statements
The work reported in this publication was funded solely by
Philip Morris International (PMI). All authors are (or were) em-
ployees of PMI R&D or worked for PMI R&D under contractual
agreements.
Authorscontributions
BP, BT, UK, DS, SL, DS, ME, SK, EV, MCP, JH, and PV contributed to
the conception or design of the work. BP, UK, DS, JH, CN, EG, AE, and
NVI contributed to the data collection. BP, BT, UK, DS, PL, YX, DS, CN,
WS, FM, and PV contributed to the data analysis and interpretation.
BP, BT, UK, and PV drafted the article. All authors critically revised
and approved the article.
Acknowledgements
The authors would like to thank the study team, especially
acknowledging the technical assistance and support of Sophie Di-
jon, Dariusz Peric, Karine Baumer, David Bornand, Remi Dulize, and
Sophie Scheuner. The authors thank Jan Verbeeck for support with
the exposure setup, Vincenzo Belcastro for the transcriptomics data
submission, and Sam Ansari for supporting the sample manage-
ment. Lipidomics data were generated by Zora Biosciences (Espoo,
Finland). The authors thank Ansgar Buettner (Histovia GmbH,
Overath, Germany) and Klaus Weber (AnaPath GmbH, Liestal,
Switzerland) for the histopathological evaluations.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.fct.2017.09.001.
Transparency document
Transparency document related to this article can be found
online at https://doi.org/10.1016/j.fct.2017.09.001.
References
ACGIH, 1989. American conference of governmental industrial hygienists. Threshold
Limit Values and Biological Exposure Studies for 7988-89. Cincinnati.
Action on Smoking and Health, 2016. Use of Electronic Cigarettes (Vapourisers)
Among Adults in Great Britain. Ash. Action on Smoking and Health.
Adluri, R.S., Nagarajan, D., Periyaswamy, V., Venugopal, P.M., 2008. Dose-response
effect of ferulic acid against nicotine-induced tissue damage and altered lipid
levels in experimental rats: a pathohistological evaluation. Fundam. Clin.
Pharmacol. 22, 557e567.
Ahmad, E.-H., Salman, R., El-Hage, R., Talih, S., Malek, N., Baalbaki, R.,
Karaoghlanian, N., Nakkash, R., Shihadeh, A., Saliba, N.A., 2016. Nicotine and
carbonyl emissions from popular electronic cigarette products: correlation to
liquid composition and design characteristics. Nicotine Tob. Res. ntw 280.
al'Absi, M., Nakajima, M., Grabowski, J., 2013. Stress response dysregulation and
stress-induced analgesia in nicotine dependent men and women. Biol. Psychol.
93, 1e8.
Alexander, D.J., Collins, C.J., Coombs, D.W., Gilkison, I.S., Hardy, C.J., Healey, G.,
Karantabias, G., Johnson, N., Karlsson, A., Kilgour, J.D., McDonald, P., 2008. As-
sociation of Inhalation Toxicologists (AIT) working party recommendation for
standard delivered dose calculation and expression in non-clinical aerosol
inhalation toxicology studies with pharmaceuticals. Inhal. Toxicol. 20,
1179 e118 9 .
Andersson, K., Eneroth, P., Arner, P., 1993. Changes in circulating lipid and carbo-
hydrate metabolites following systemic nicotine treatment in healthy men. Int.
J. Obes. Relat. metabolic Disord. J. Int. Assoc. Study Obes. 17, 675e680.
Ansari, S., Baumer, K., Boue, S., Dijon, S., Dulize, R., Ekroos, K., Elamin, A., Foong, C.,
Guedj, E., Hoeng, J., Ivanov, N.V., Krishnan, S., Leroy, P., Martin, F., Merg, C.,
Peck, M.J., Peitsch, M.C., Phillips, B., Schlage, W.K., Schneider, T., Talikka, M.,
Titz, B., Vanscheeuwijck, P., Veljkovic, E., Vihervaara, T., Vuillaume, G.,
Woon, C.Q., 2016. Comprehensive systems biology analysis of a 7-month ciga-
rette smoke inhalation study in C57BL/6 mice. Sci. data 3, 150077.
Ashakumary, L., Vijayammal, P.L., 1997. Effect of nicotine on lipoprotein metabolism
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332330
in rats. Lipids 32, 311e315.
Bekki, K., Uchiyama, S., Ohta, K., Inaba, Y., Nakagome, H., Kunugita, N., 2014.
Carbonyl compounds generated from electronic cigarettes. Int. J. Environ. Res.
Public Health 11, 11192e11200.
Bolstad, B., Bolstad, M.B., BiocGenerics, I., biocViews Microarray, O., 2013. Package
affyPLM.
Breland, A., Soule, E., Lopez, A., Ramoa, C., El-Hellani, A., Eissenberg, T., 2017. Elec-
tronic cigarettes: what are they and what do they do? Ann. N. Y. Acad. Sci.1394,
5e30.
Brown, C.J., Cheng, J.M., 2014. Electronic cigarettes: product characterisation and
design considerations. Tob. Control 23 (Suppl. 2) ii4e10.
Burger, G.T., Renne, R.A., Sagartz, J.W., Ayres, P.H., Coggins, C.R., Mosberg, A.T.,
Hayes, A.W., 1989. Histologic changes in the respiratory tract induced by
inhalation of xenobiotics: physiologic adaptation or toxicity? Toxicol. Appl.
Pharmacol. 101, 521e542.
CDER (Center for Drug Evaluation and Research), 2005. Guidance for Industry
Estimating the Maximum Safe Starting Dose in Initial Clinical Trials for Ther-
apeutics in Adult Healthy Volunteers. Accessed from. https://www.fda.gov/
downloads/drugs/guidances/ucm078932.pdf.
C.F.R, 2014. Code of Federal Regulations 21CFR 182.1320, April, 1990. Ofce of the
Federal Register, National Archives and Records Administration, Washington,
D.C.
Cannon, W.C., Blanton, E.F., McDonald, K.E., 1983. The ow-past chamber: an
improved nose-only exposure system for rodents. Am. Industrial Hyg. Assoc. J.
44, 923e928.
Chattopadhyay, K., Chattopadhyay, B.D., 2008. Effect of nicotine on lipid prole,
peroxidation &antioxidant enzymes in female rats with restricted dietary
protein. Indian J. Med. Res. 127, 571e576.
Chen, H., Fu, Y., Sharp, B.M., 2008. Chronic nicotine self-administration augments
hypothalamic-pituitary-adrenal responses to mild acute stress. Neuro-
psychopharmacol. ofcial Publ. Am. Coll. Neuropsychopharmacol. 33, 721e730 .
Cheng, S.Y., Glazkova, D., Serova, L., Sabban, E.L., 2005. Effect of prolonged nicotine
infusion on response of rat catecholamine biosynthetic enzymes to restraint
and cold stress. Pharmacol. Biochem. Behav. 82, 559e568.
Chowdhury, P., Rayford, P.L., Chang, L.W., 1992. Induction of pancreatic acinar pa-
thology via inhalation of nicotine. In: Proceedings of the Society for Experi-
mental Biology and Medicine, vol. 201. Society for Experimental Biology and
Medicine, pp. 159e164.
Dai, M., Wang, P., Boyd, A.D., Kostov, G., Athey, B., Jones, E.G., Bunney, W.E.,
Myers, R.M., Speed, T.P., Akil, H., 2005. Evolving gene/transcript denitions
signicantly alter the interpretation of GeneChip data. Nucleic acids Res. 33
e175ee175.
El-Hellani, A., Salman, R., El-Hage, R., Talih, S., Malek, N., Baalbaki, R.,
Karaoghlanian, N., Nakkash, R., Shihadeh, A., Saliba, N.A., 2016. Nicotine and
carbonyl emissions from popular electronic cigarette products: correlation to
liquid composition and design characteristics. Nicotine Tob. Res. ofcial J. Soc.
Res. Nicotine Tob. 1e
9. http://dx.doi.org/10.1093/ntr/ntw280.
European Parliament and Council, 2014. Directive 2014/40/EU of the European
parliament and of the Council of 3 April 2014 on the approximation of the laws,
regulations and administrative provisions of the Member States concerning the
manufacture, presentation and sale of tobacco and related products and
repealing Directive 2001/37/EC Text with EEA relevance, in: T.E.P.A.O.T.
COUNCIL (Ed.), DIRECTIVE 2014/40/EU.
Everds, N.E., Snyder, P.W., Bailey, K.L., Bolon, B., Creasy, D.M., Foley, G.L., Rosol, T.J.,
Sellers, T., 2013. Interpreting stress responses during routine toxicity studies: a
review of the biology, impact, and assessment. Toxicol. Pathol. 41, 560e614.
Faraday, M.M., Blakeman, K.H., Grunberg, N.E., 2005. Strain and sex alter effects of
stress and nicotine on feeding, body weight, and HPA axis hormones. Phar-
macol. Biochem. Behav. 80, 577e589.
Faraday, M.M., O'Donoghue, V.A., Grunberg, N.E., 1999. Effects of nicotine and stress
on startle amplitude and sensory gating depend on rat strain and sex. Phar-
macol. Biochem. Behav. 62, 273e284.
Farsalinos, K.E., Voudris, V., Poulas, K., 2015. E-cigarettes generate high levels of
aldehydes only in dry puffconditions. Addiction 110, 1352e1356.
Flora, J.W., Meruva, N., Huang, C.B., Wilkinson, C.T., Ballentine, R., Smith, D.C.,
Werley, M.S., McKinney, W.J., 2016. Characterization of potential impurities and
degradation products in electronic cigarette formulations and aerosols. Regul.
Toxicol. Pharmacol. RTP 74, 1e11.
Fuoco, F.C., Buonanno, G., Stabile, L., Vigo, P., 2014. Inuential parameters on particle
concentration and size distribution in the mainstream of e-cigarettes. Barking,
Essex : 1987 Environ. Pollut. 184, 523e529.
Gentleman, R.C., Carey, V.J., Bates, D.M., Bolstad, B., Dettling, M., Dudoit, S., Ellis, B.,
Gautier, L., Ge, Y., Gentry, J., Hornik, K., Hothorn, T., Huber, W., Iacus, S.,
Irizarry, R., Leisch, F., Li, C., Maechler, M., Rossini, A.J., Sawitzki, G., Smith, C.,
Smyth, G., Tierney, L., Yang, J.Y., Zhang, J., 2004. Bioconductor: open software
development for computational biology and bioinformatics. Genome Biol. 5,
R80.
Golli, N.E., Dkhili, H., Dallagi, Y., Rahali, D., Lasram, M., Bini-Dhouib, I., Lebret, M.,
Rosa, J.-P., El Fazaa, S., Allal-El Asmi, M., 2016. Comparison between electronic
cigarette rell liquid and nicotine on metabolic parameters in rats. Life Sci. 146,
131e138.
Goniewicz, M.L., Knysak, J., Gawron, M., Kosmider, L., Sobczak, A., Kurek, J.,
Prokopowicz, A., Jablonska-Czapla, M., Rosik-Dulewska, C., Havel, C.,
Jacob 3rd, P., Benowitz, N., 2014. Levels of selected carcinogens and toxicants in
vapour from electronic cigarettes. Tob. Control 23, 133e139.
Grunberg, N.E., Bowen, D.J., Winders, S.E., 1986. Effects of nicotine on body weight
and food consumption in female rats. Psychopharmacology 90, 101e105.
Grunberg, N.E., Winders, S.E., Popp, K.A., 1987. Sex differences in nicotine's effects
on consummatory behavior and body weight in rats. Psychopharmacology 91,
221e225.
Han, Y., Lau, Y.-l., 2014. Nicotine, an anti-inammation molecule. Inamm. Cell
Signal. 1.
Hara, T., Tanegashima, K., 2012. Pleiotropic functions of the CXC-type chemokine
CXCL14 in mammals. J. Biochem. 151, 469e476 .
Hecht, S.S., Carmella, S.G., Kotandeniya, D., Pillsbury, M.E., Chen, M., Ransom, B.W.,
Vogel, R.I., Thompson, E., Murphy, S.E., Hatsukami, D.K., 2015. Evaluation of
toxicant and carcinogen metabolites in the urine of e-cigarette users versus
cigarette smokers. Nicotine Tob. Res. ofcial J. Soc. Res. Nicotine Tob. 17,
704e709.
Hegardt, F.G., 1999. Mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase: a
control enzyme in ketogenesis. Biochem. J. 338 (Pt 3), 569e582.
Hiemstra, P.S., Bals, R., 2016. Basic science of electronic cigarettes: assessment in cell
culture and in vivo models. Respir. Res. 17, 127.
Horobin, R.W., 2014. Histochemistry: an Explanatory Outline of Histochemistry and
Biophysical Staining. Elsevier.
Hukkanen, J., Jacob 3rd, P., Benowitz, N.L., 2005. Metabolism and disposition ki-
netics of nicotine. Pharmacol. Rev. 57, 79e115 .
Iskandar, A.R., Gonzalez-Suarez, I., Majeed, S., Marescotti, D., Sewer, A., Xiang, Y.,
Leroy, P., Guedj, E., Mathis, C., Schaller, J.P., Vanscheeuwijck, P., Frentzel, S.,
Martin, F., Ivanov, N.V., Peitsch, M.C., Hoeng, J., 2016. A framework for in vitro
systems toxicology assessment of e-liquids. Toxicol. Mech. Methods 26,
389e413.
Jensen, R.P., Luo, W., Pankow, J.F., Strongin, R.M., Peyton, D.H., 2015. Hidden form-
aldehyde in e-cigarette aerosols. N. Engl. J. Med. 372, 392e394.
Jin, E.S., Sherry, A.D., Malloy, C.R., 2014. Interaction between the pentose phosphate
pathway and gluconeogenesis from glycerol in the liver. J. Biol. Chem. 289,
32593e32603.
Jo, Y.H., Talmage, D.A., Role, L.W., 2002. Nicotinic receptor-mediated effects on
appetite and food intake. J. Neurobiol. 53, 618e632.
Kogel, U., Schlage, W.K., Martin, F., Xiang, Y., Ansari, S., Leroy, P., Vanscheeuwijck, P.,
Gebel, S., Buettner, A., Wyss, C., Esposito, M., Hoeng, J., Peitsch, M.C., 2014. A 28-
day rat inhalation study with an integrated molecular toxicology endpoint
demonstrates reduced exposure effects for a prototypic modied risk tobacco
product compared with conventional cigarettes. Food Chem. Toxicol. Int. J. Publ.
Br. Industrial Biol. Res. Assoc. 68, 204e217.
Kogel, U., Titz, B., Schlage, W.K., Nury, C., Martin, F., Oviedo, A., Lebrun, S., Elamin, A.,
Guedj, E., Trivedi, K., Ivanov, N.V., Vanscheeuwijck, P., Peitsch, M.C., Hoeng, J.,
2016. Evaluation of the Tobacco Heating System 2.2. Part 7: systems toxico-
logical assessment of a mentholated version revealed reduced cellular and
molecular exposure effects compared with mentholated and non-mentholated
cigarette smoke. Regul. Toxicol. Pharmacol. RTP 81 (Suppl. 2), S123eS138.
Kuroshima, A., Yahata, T., Ohno, T., 1981. Changes in plasma glucagon levels to
stressful environmental temperatures. Jpn. J. physiology 31, 43e52.
Lerner, C.A., Sundar, I.K., Watson, R.M., Elder, A., Jones, R., Done, D., Kurtzman, R.,
Ossip, D.J., Robinson, R., McIntosh, S., Rahman, I., 2015a. Environmental health
hazards of e-cigarettes and their components: oxidants and copper in e-ciga-
rette aerosols. Barking, Essex : 1987 Environ. Pollut. 198, 100e107.
Lerner, C.A., Sundar, I.K., Yao, H., Gerloff, J., Ossip, D.J., McIntosh, S., Robinson, R.,
Rahman, I., 2015b. Vapors produced by electronic cigarettes and e-juices with
avorings induce toxicity, oxidative stress, and inammatory response in lung
epithelial cells and in mouse lung. PLoS One 10 e0116732.
Lewis, R.W., Billington, R., Debryune, E., Gamer, A., Lang, B., Carpanini, F., 2002.
Recognition of adverse and nonadverse effects in toxicity studies. Toxicol.
Pathol. 30, 66e74.
Liberzon, A., Subramanian, A., Pinchback, R., Thorvaldsdottir, H., Tamayo, P.,
Mesirov, J.P., 2011. Molecular signatures database (MSigDB) 3.0. Bioinformatics
27, 1739e174 0 .
Majeed, S., Frentzel, S., Wagner, S., Kuehn, D., Leroy, P., Guy, P.A., Knorr, A., Hoeng, J.,
Peitsch, M.C., 2014. Characterization of the Vitrocell(R) 24/48 in vitro aerosol
exposure system using mainstream cigarette smoke. Chem. Central J. 8, 62.
Manigrasso, M., Buonanno, G., Fuoco, F.C., Stabile, L., Avino, P., 2015. Aerosol
deposition doses in the human respiratory tree of electronic cigarette smokers.
Barking, Essex : 1987 Environ. Pollut. 196, 257e
267.
Marco, E., Grimalt, J.O., 2015. A rapid method for the chromatographic analysis of
volatile organic compounds in exhaled breath of tobacco cigarette and elec-
tronic cigarette smokers. J. Chromatogr. A 1410, 51e59.
Margham, J., McAdam, K., Forster, M., Liu, C., Wright, C., Mariner, D., Proctor, C.,
2016. Chemical composition of aerosol from an e-cigarette: a quantitative
comparison with cigarette smoke. Chem. Res. Toxicol. 29, 1662e1678.
Matta, S.G., Fu, Y., Valentine, J.D., Sharp, B.M., 1998. Response of the hypothalamo-
pituitary-adrenal axis to nicotine. Psychoneuroendocrinology 23, 103e113 .
McCall, M.N., Bolstad, B.M., Irizarry, R.A., 2010. Frozen robust multiarray analysis
(fRMA). Biostatistics 11, 242e253.
McGarry, J.D., Foster, D.W., 1980. Regulation of hepatic fatty acid oxidation and
ketone body production. Annu. Rev. Biochem. 49, 395e420.
McNeill, A., Brose, L., Calder, R., Hitchman, S., Hajek, P., McRobbie, H., 2015. E-cig-
arettes: an Evidence Update. A report commissioned by Public Health England.
Public Health England.
NACLAR, N.a.c.f.l.a.r., 2004. Guidelines on the Care and Use of Animals for Scientic
Purposes. National advisory committee for laboratory animal research,
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332 331
Singapore.
Nayak, N.C., Sathar, S.A., Mughal, S., Duttagupta, S., Mathur, M., Chopra, P.,1996. The
nature and signicance of liver cell vacuolation following hepatocellular
injuryean analysis based on observations on rats rendered tolerant to hepa-
totoxic damage. Virchows Archiv Int. J. pathology 428, 353e365.
Nicolaides, N.C., Kyratzi, E., Lamprokostopoulou, A., Chrousos, G.P., Charmandari, E.,
2015. Stress, the stress system and the role of glucocorticoids. Neuro-
immunomodulation 22, 6e19.
Ochiai, Y., Sakurai, E., Nomura, A., Itoh, K., Tanaka, Y., 2006. Metabolism of nicotine
in rat lung microvascular endothelial cells. J. Pharm. Pharmacol. 58, 403e407.
OECD, 2009a. OECD Principles of Good Laboratory Practice (As Revised in 1997).
OECD, 2009b. Test No. 413: Subchronic Inhalation Toxicity: 90-day Study. OECD
Publishing, Paris.
Osimitz, T.G., Droege, W., Finch, J.M., 2007. Toxicologic signicance of histologic
change in the larynx of the rat following inhalation exposure: a critical review.
Toxicol. Appl. Pharmacol. 225, 229e237.
Oviedo, A., Lebrun, S., Kogel, U., Ho, J., Tan, W.T., Titz, B., Leroy, P., Vuillaume, G.,
Bera, M., Martin, F., Rodrigo, G., Esposito, M., Dempsey, R., Ivanov, N.V., Hoeng, J.,
Peitsch, M.C., Vanscheeuwijck, P., 2016. Evaluation of the Tobacco Heating
System 2.2. Part 6: 90-day OECD 413 rat inhalation study with systems toxi-
cology endpoints demonstrates reduced exposure effects of a mentholated
version compared with mentholated and non-mentholated cigarette smoke.
Regul. Toxicol. Pharmacol. RTP 81 (Suppl. 2), S93eS122.
Peace, M.R., Baird, T.R., Smith, N., Wolf, C.E., Poklis, J.L., Poklis, A., 2016. Concen-
tration of nicotine and glycols in 27 electronic cigarette formulations. J. Anal.
Toxicol. 40, 403e407.
Phillips, B., Esposito, M., Verbeeck, J., Boue, S., Iskandar, A., Vuillaume, G., Leroy, P.,
Krishnan, S., Kogel, U., Utan, A., Schlage, W.K., Bera, M., Veljkovic, E., Hoeng, J.,
Peitsch, M.C., Vanscheeuwijck, P., 2015. Toxicity of aerosols of nicotine and
pyruvic acid (separate and combined) in Sprague-Dawley rats in a 28-day OECD
412 inhalation study and assessment of systems toxicology. Inhal. Toxicol. 27,
405e431.
Picciotto, M.R., Brunzell, D.H., Caldarone, B.J., 2002. Effect of nicotine and nicotinic
receptors on anxiety and depression. Neuroreport 13, 1097e110 6 .
R Development Core Team, 2007. R: a Language and Environment for Statistical
Computing.
Renne, R., Wehner, A., Greenspan, B., Deford, H., Ragan, H., Westerberg, R.,
Buschbom, R., Burger, G., Hayes, A., Suber, R., 1992. 2-Week and 13-week
inhalation studies of aerosolized glycerol in rats. Inhal. Toxicol. 4, 95e111.
Rhodes, M.E., O'Toole, S.M., Czambel, R.K., Rubin, R.T., 2001. Male-female differences
in rat hypothalamic-pituitary-adrenal axis responses to nicotine stimulation.
Brain Res. Bull. 54, 681e688.
Ricart-Jan, D., Rodr, V., Benavides, A., Peinado-Onsurbe, J., Llobera, M., 2002.
Immobilization stress alters intermediate metabolism and circulating lipopro-
teins in the rat. Metabolism 51, 925e931.
Rohleder, N., Kirschbaum, C., 2006. The hypothalamic-pituitary-adrenal (HPA) axis
in habitual smokers. Int. J. Psychophysiol. ofcial J. Int. Organ. Psychophysiol.
59, 236e243.
Rose, A.J., Herzig, S., 2013. Metabolic control through glucocorticoid hormones: an
update. Mol. Cell. Endocrinol. 380, 65e78.
RTECS, 1985. Registry of toxtc effects of chemtcal substances (RTECS) propanedlol
TY2000 000 Edited by R L Lewis and D Sweet Publicatlon No 86-103 US
Department of Health and Human Services, Washington, DC.
Salturk, Z., Cakir, C., Sunnetci, G., Atar, Y., Kumral, T.L., Yildirim, G., Berkiten, G.,
Uyar, Y., 2015. Effects of electronic nicotine delivery system on larynx: experi-
mental study. J. voice ofcial J. Voice Found. 29, 560e563.
Selye, H., 1936. Thymus and adrenals in the response of the organism to injuries and
intoxications. Br. J. Exp. Pathology 17, 234.
Shao, X.M., Xu, B., Liang, J., Xie, X.S., Zhu, Y., Feldman, J.L., 2013. Nicotine delivery to
rats via lung alveolar region-targeted aerosol technology produces blood
pharmacokinetics resembling human smoking. Nicotine Tob. Res. ofcial J. Soc.
Res. Nicotine Tob. 15, 1248e1258.
Singh, S.P., Kalra, R., Puttfarcken, P., Kozak, A., Tesfaigzi, J., Sopori, M.L., 2000. Acute
and chronic nicotine exposures modulate the immune system through different
pathways. Toxicol. Appl. Pharmacol. 164, 65e72.
Smyth, G.K., 2004. Linear models and empirical bayes methods for assessing dif-
ferential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3.
Article3.
Sopori, M.L., Kozak, W., Savage, S.M., Geng, Y., Kluger, M.J., 1998. Nicotine-induced
Modulation of T Cell Function, Drugs of Abuse, Immunomodulation, and Aids.
Springer, pp. 279e289.
Spinedi, E., Gaillard, R.C., 1998. A regulatory loop between the hypothalamo-
pituitary-adrenal (HPA) axis and circulating leptin: a physiological role of
ACTH. Endocrinology 139, 4016e4020.
Suber, R.L., Deskin, R., Nikiforov, I., Fouillet, X., Coggins, C.R.,1989. Subchronic nose-
only inhalation study of propylene glycol in Sprague-Dawley rats. Food Chem.
Toxicol. Int. J. Publ. Br. Industrial Biol. Res. Assoc. 27, 573e583.
Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., Huerta-Cepas, J.,
Simonovic, M., Roth, A., Santos, A., Tsafou, K.P., Kuhn, M., Bork, P., Jensen, L.J.,
von Mering, C., 2015. STRING v10: protein-protein interaction networks, inte-
grated over the tree of life. Nucleic Acids Res. 43, D447eD452.
Titz, B., Schneider, T., Elamin, A., Martin, F., Dijon, S., Ivanov, N., Hoeng, J., Peitsch, M.,
2015a. Analysis of Proteomic Data for Toxicological Applications, Computational
Systems Toxicology. Springer, New York, pp. 257e284.
Titz, B., Sewer, A., Schneider, T., Elamin, A., Martin, F., Dijon, S., Luettich, K., Guedj, E.,
Vuillaume, G., Ivanov, N.V., Peck, M.J., Chaudhary, N.I., Hoeng, J., Peitsch, M.C.,
2015b. Alterations in the sputum proteome and transcriptome in smokers and
early-stage COPD subjects. J. Proteomics 128, 306e320.
Valenca, S.S., Gouveia, L., Pimenta, W.A., Porto, L.C., 2008. Effects of oral nicotine on
rat liver stereology. Int. J. Morphol. 26, 1013e1022.
Varemo, L., Nielsen, J., Nookaew, I., 2013. Enriching the gene set analysis of genome-
wide data by incorporating directionality of gene expression and combining
statistical hypotheses and methods. Nucleic Acids Res. 41, 4378e4391.
Vizcaino, J.A., Deutsch, E.W., Wang, R., Csordas, A., Reisinger, F., Rios, D., Dianes, J.A.,
Sun, Z., Farrah, T., Bandeira, N., Binz, P.A., Xenarios, I., Eisenacher, M., Mayer, G.,
Gatto, L., Campos, A., Chalkley, R.J., Kraus, H.J., Albar, J.P., Martinez-Bartolome, S.,
Apweiler, R., Omenn, G.S., Martens, L., Jones, A.R., Hermjakob, H., 2014. Pro-
teomeXchange provides globally coordinated proteomics data submission and
dissemination. Nat. Biotechnol. 32, 223e226.
Waldum, H.L., Nilsen, O.G., Nilsen, T., Rorvik, H., Syversen, V., Sanvik, A.K.,
Haugen, O.A., Torp, S.H., Brenna, E., 1996. Long-term effects of inhaled nicotine.
Life Sci. 58, 1339e1346.
Werley, M.S., Jerome, A.M., Oldham, M.J., 2014. Toxicological evaluation of aerosols
of a tobacco extract formulation and nicotine formulation in acute and short-
term inhalation studies. Inhal. Toxicol. 26, 207e221.
Werley, M.S., Kirkpatrick, D.J., Oldham, M.J., Jerome, A.M., Langston, T.B., Lilly, P.D.,
Smith, D.C., McKinney Jr., W.J., 2016. Toxicological assessment of a prototype e-
cigaret device and three avor formulations: a 90-day inhalation study in rats.
Inhal. Toxicol. 28, 22e38.
Werley, M.S., McDonald, P., Lilly, P., Kirkpatrick, D., Wallery, J., Byron, P., Venitz, J.,
2011. Non-clinical safety and pharmacokinetic evaluations of propylene glycol
aerosol in Sprague-Dawley rats and Beagle dogs. Toxicology 287, 76e90.
Wong, E.T., Kogel, U., Veljkovic, E., Martin, F., Xiang, Y., Boue, S., Vuillaume, G.,
Leroy, P., Guedj, E., Rodrigo, G., Ivanov, N.V., Hoeng, J., Peitsch, M.C.,
Vanscheeuwijck, P., 2016. Evaluation of the Tobacco Heating System 2.2. Part 4:
90-day OECD 413 rat inhalation study with systems toxicology endpoints
demonstrates reduced exposure effects compared with cigarette smoke. Regul.
Toxicol. Pharmacol. RTP 81 (Suppl. 2), S59eS81.
Yang, J., MacDougall, M.L., McDowell, M.T., Xi, L., Wei, R., Zavadoski, W.J.,
Molloy, M.P., Baker, J.D., Kuhn, M., Cabrera, O., Treadway, J.L., 2011. Polyomic
proling reveals signicant hepatic metabolic alterations in glucagon-receptor
(GCGR) knockout mice: implications on anti-glucagon therapies for diabetes.
BMC Genomics 12, 281.
Yildiz, D., 2004. Nicotine, its metabolism and an overview of its biological effects.
Toxicon 43, 619e632.
Zakout, Y.M.-A., Salih, M.M., Ahmed, H., 2010. The effect of xatives and tempera-
ture on the quality of glycogen demonstration. Biotech. Histochem. 85, 93e98.
B. Phillips et al. / Food and Chemical Toxicology 109 (2017) 315e332332
... Affected homeostasis and increased proliferation may result in DNA damage accumulation and pose potential risks in long-term exposure [139]. Glycerine is also an ingredient in the liquid used in electronic cigarettes, with no toxicologically relevant effects [144]. ...
... To our knowledge, there is no study investigating the impact of regular use on the oral tissues and properties of saliva. Propylene glycol is also an ingredient in the liquid used in electronic cigarettes, with no toxicologically relevant effects observed [144]. ...
Article
Full-text available
Mouthwashes are a very popular additional oral hygiene element and there are plenty of individual products, whose compositions are in a state of flux. The aim of our study was to investigate the compositions of mouthwashes and their functions, as well as to discuss their effectiveness in preventing and curing oral diseases and side effects. We searched for mouthwashes available on the market in Poland. We identified 241 individual mouthwash products. The extraction of compositions was performed and functions of the ingredients were assessed. Then, analysis was performed. The evaluation revealed that there are plenty of ingredients, but a typical mouthwash is a water–glycerine mixture and consists of additional sweetener, surfactant, preservative, and some colourant and flavouring agent, as well as usually having two oral health substances, anticaries sodium fluoride and antimicrobial essential oils. The effectiveness or side effects of several substances of mouthwashes were thoroughly discussed. We recommend not multiplying individual mouthwash products and their ingredients beyond medical or pharmaceutical necessity, especially without scientific proof.
... About 170 million pounds of PG is consumed in human foods (Center for the Evaluation of Risks to Human Reproduction 2004). Thus, it is generally recognized as a safe food additive (Phillips et al. 2017). Besides, PG has numerous other applications in industry, including as a humectant, moisturizer, non-toxic antifreeze, and carrier in fragrance oils (Gaworski et al. 2010;Binks et al. 2013). ...
... Processed cereal-based foods and other infant foods should be free from chemical and biological hazards (Ojuri et al. 2018). Currently, PG is generally recognized as a safe food additive, which is oversupplied in Chinese market (Phillips et al. 2017;Tao et al. 2020;Zhao et al. 2020). PG has very low toxicity and is considered to be a nontoxic or non-carcinogenic compound for adults (Zaripov et al. 2020). ...
Article
Full-text available
Propylene glycol (PG) is widely used in the foods, pharmaceuticals, oil industry, animal feed, cosmetics and other industries. Because of the existence of a chiral carbon center, PG forms R (Rectus)- and S (Sinister)-enantiomers. Currently, the toxicity study of its R-, S-enantiomers is still very scarce. In this study, we have assessed the developmental toxicity and neurotoxicity of the R-, S-, and RS-PG enantiomers in zebrafish larvae. We found that exposure to R-, S-, and RS-PG enantiomers did not significantly affect the basic developmental endpoints of embryos or larvae (i.e., embryonic movement, hatching, mortality, malformation, heartbeat, body length), indicating that R-, S-, and RS-PG exposures did not exhibit the basic developmental toxicity in zebrafish larvae. The toxicity of three enantiomers was lower than that of ethanol, and there was no significant difference between them. However, R-, S-, and RS-PG exposures with high doses could significantly change the eye diameter and locomotor activity of larval zebrafish, indicating that R-, S-, and RS-PG enantiomers of high doses could potentially exhibit the neurotoxicity and ocular developmental toxicity in zebrafish larvae. Therefore, the potential neurotoxicity and ocular developmental toxicity of R-, S-, and RS-PG enantiomers for infants and toddlers should be considered.
... A higher percentage of propylene glycol seems enhancing flavour and strengthening the so-called "throat hit", whereas a higher percentage of vegetable glycerine may increase vapor production [45]. Vegetable glycerine exposure has been associated with irritation of eyes, lungs, and oesophagus mucosa [46]. Likewise, its higher boiling point requires the heating element to reach higher temperatures, resulting in a greater risk of toxicants emission [47]. ...
Article
Full-text available
The use of electronic cigarettes (e-cigarette) and vaping devices started as a potential aid for cessation and reducing the harmful consequences of cigarette smoking, mainly in the adult population. Today e-cigarette use is highly increasing in vulnerable populations, especially young and pregnant women, due to the misconception of its harmless use. Despite the growing acknowledgment in e-cigarette as a potential harmful device, and due to mixed information found concerning its beneficial aid for smokers, along with an insufficient clinical study done in human models, it is important to further evaluate the possible benefits and risks of non-combusting, vaping nicotine or non-nicotine delivery devices. In this review we tried to summarize the latest updated information found in the literature, concentrating mainly in the variety of adverse effects of e-cigarette use and its contribution for recent and future health concerns.
... PG/VG vehicle was only shown to be toxic at higher doses in vitro, which was enhanced by additional flavouring chemicals (35). In a rodent model, there was no evidence of toxicity induced by PG/VG aerosols at different concentrations compared with saline (36). Furthermore, additional chemicals other than flavouring may exist in the e-fluids (37), as the labels do not always accurately reflect the chemical composition in e-fluid products (38). ...
Article
Full-text available
Long term e-cigarette vaping induces inflammation, which is largely nicotine independent. High-fat diet (HFD) consumption is anoter cause of systemic low-grade inflammation. The likelihood of using e-cigarettes as a weight control strategy is concomitant with the increase in obesity. In Australia, only nicotine-free e-fluid is legal for sale. Therefore, this study aimed to investigate how nicotine-free e-cigarette vapour exposure affects inflammatory responses in mice with long term HFD consumption. Mice were fed a HFD for 16 weeks, while in the last 6 weeks, half of the chow and HFD groups were exposed to nicotine-free e-vapour, while the other half to ambient air. Serum, lung, liver and epididymal fat were collected to measure inflammatory markers. While both e-vapour exposure and HFD consumption independently increased serum IFN-γ, CX3CL1, IL-10, CCL20, CCL12, and CCL5 levels, the levels of IFN-γ, CX3CL1, and IL-10 were higher in mice exposed to e-vapour than HFD. The mRNA expression pattern in the epididymal fat mirrors that in the serum, suggesting the circulating inflammatory response to e-vapour is from the fat tissue. Of the upregulated cytokines in serum, none were found to change in the lungs. The anti-inflammatory cytokine IL-10 was increased by combining e-vapour and HFD in the liver. We conclude that short-term nicotine-free e-vapour is more potent than long term HFD consumption in causing systemic inflammation. Future studies will be needed to examine the long-term health impact of nicotine-free e-cigarettes.