A cross-sectional analysis of candidate biomarkers of biological effect in smokers, never-smokers and ex-smokers

Abstract

Context: Biomarkers of biological effect (BOBE) have been proposed as potential tools to assess tobacco product use, toxicity and disease risk.

Objective: To determine if candidate BOBE can distinguish between smokers, never-smokers and former smokers.

Methods: Biomarker levels were compared from 143 smokers, 61 never-smokers and 61 ex-smokers.

Results: In total, 27 candidate biomarkers were assessed, 14 were significantly different between smokers and never-smokers (p < 0.01) and of these 14 biomarkers, 12 were able to distinguish between smokers and former smokers (p < 0.05), which indicates the potential for reversibility.

Conclusions: A total of 12 of 27 BOBE are potentially useful tools for future product assessment.

• Biomarkers
• cancer
• cardiovascular disease
• COPD
• inflammation
• oxidative stress
• smoking
• toxicant

Introduction

In 2001, the US Institute of Medicine (IOM) reported that, since smoking-related diseases were dose-related and because epidemiological studies showed a reduction in the risk of smoking-related diseases following cessation, it might be possible to reduce smoking-related disease risk by developing potential reduced-exposure products (Institute of Medicine, 2001). More recently in the United States, the Family Smoking Prevention and Tobacco Control Act of 2009 granted the US Food and Drug Administration (FDA) the legal authority to regulate the manufacturing, distribution, and marketing of tobacco products, including so-called “modified risk tobacco products” (MRTPs). MRTPs were defined as any tobacco product that is sold or distributed for use to reduce harm or the risk of tobacco-related disease. To be marketed in the United States, the candidate product must meet one of two public health standards: a demonstrated “Modified Risk” claim or a “Special Rule for Certain Products” claim, specifying a reduced-exposure product (in terms of machine-smoked toxicant yield; Food and Drug Administration, 2012). To meet the Modified Risk standard, the applicant must prove with scientific evidence that the product, as actually used by consumers, will (1) significantly reduce harm and the risk of tobacco-related disease to individual users and (2) benefit the health of the population as a whole, taking into account both users and non-users of the product (Food and Drug Administration, 2012). With respect to a Modified Risk claim, the IOM provided an outline of the kind of studies that would be required to provide sufficient scientific evidence of reduced risk. In addition to pre-clinical toxicity testing using in vitro systems and animal models, the IOM proposed the use of biomarkers of exposure to specific tobacco toxicants to establish internal dose to the consumer and suggest their use as disease risk biomarkers, with the caveat that there is currently no evidence that a single or group of constituents is solely responsible for a given disease. Finally, they suggest that “for a biomarker of exposure to be accepted as a biomarker of risk or a surrogate endpoint of disease, there should be a strong biological rational as well as compelling data from clinical and epidemiologic studies” (Institute of Medicine, 2012). Scientists have begun to investigate the potential of biomarkers of exposure as disease risk biomarkers. Most notably, Hecht et al. investigated several biomarkers of exposure in the Shanghai Cohort Study for their potential to predict lung cancer. The authors concluded that total cotinine, total 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL), r-1-,t-2,3,c-4-tetrahydroxy-1,2,3,4-tetrahydrophenanthrene (PheT), and total N′-nitrosonornicotine (NNN) are biomarkers of lung cancer risk and could be included in a future risk prediction algorithm for smokers (Hecht et al., 2013). Biomarkers of exposure alone are unlikely to provide sufficient evidence of disease risk, as there is a lack of mechanistic data to link most of them to established clinical endpoints of disease or pre-disease states (part of the “Biomarker Qualification” process). The IOM defined Biomarker Qualification as an “evidentiary process of linking a biomarker with biological processes and clinical endpoints” (Institute of Medicine, 2010). From theoretical and practical standpoints, we propose that the qualification of a biomarker linking an exposure with “biological processes” is distinct from one that would link exposures or biological processes to “clinical endpoints”. Thus we use the term “biomarkers of biological effects” (BOBE) for the former and “biomarkers of potential harm” for the latter. In addition to the definition offered by the IOM, we further suggest that to qualify for use in MRTP scientific assessment, BOBE should: (1) be measurable in tissues or body fluids that can be obtained by techniques, which are as minimally invasive as possible; (2) show reversibility within a timeframe of less than 6 months after smoking cessation, as an indication of the effect that could be achieved with a suitable candidate MRTP; and (3) have fully validated methods for their measurement, based on existing guidelines (Aggett et al., 2005; Chau et al., 2008; Food and Drug Administration, 2001; Institute of Medicine, 2010, 2012; Lee et al., 2006), in order to reduce inter-laboratory differences in measurement, across studies. Thus, we propose that qualified BOBE would represent acute and sub-chronic response pathways to exposures and that for MRTP scientific assessment, alterations in their concentrations in the direction of those found in studies of smoking cessation would add to the overall “weight of evidence” to evaluate the potential of a MRTP to reduce risk.

In our terminology, we note that none of these BOBE are qualified as a predictive biomarker for a disease endpoint. We propose that they are only qualified against the type of exposure that could be anticipated by use of a MRTP compared to the use of a conventional cigarette. However, BOBE, such as biomarkers of oxidative stress, DNA damage and of inflammation, may measure processes that have themselves been associated with disease endpoints and so alterations in these biomarkers may provide meaningful data for the scientific assessment of MRTPs.

Previously, we have presented cross-sectional data on the ability of candidate BOBE to discriminate between small cohorts of current, former and never-smokers (Lowe et al., 2009). We have subsequently further investigated these biomarkers and others with known association to smoking-related disease mechanisms in a 6-month, single-centre, single-blinded, forced switch clinical study. This manuscript discusses the baseline cross-sectional BOBE analysis from the study, which serves to determine which BOBE have the potential utility to draw meaningful conclusions from a forced product switch from a conventional cigarette to a MRTP.

Materials and methods

Study design

This cohort was formed from the baseline measurements of a larger study which is registered with Current Controlled Trials (number ISRCTN81286286) and the protocol has been published (Shepperd et al., 2013). The study was designed and conducted in accordance with the ethical principles of the Declaration of Helsinki and the International Committee on Harmonization for Good Clinical Practice and German law. The protocol and the informed consent forms were approved by the ethics committee of the Ärtzekammer, Hamburg. All subjects were sourced from the Hamburg area and provided written informed consent prior to the start of the study. The complete study design and associated information is presented in Shepperd et al. (2013). For sample collection, each participant attended the clinic (Momentum Pharma Services, Hamburg, Germany) for a consecutive 2-d residential stay and samples were taken for biomarker analysis on each day. Baseline BOBE data were pooled for these 2 d for statistical analysis.

Selection of study participants

The suitability of participants who gave informed consent was assessed according to the inclusion criteria within 28 d before entering the study and were verified upon arrival at the clinic. The universal inclusion criteria were: weight of at least 52 kg (men) or 45 kg (women); body-mass index (BMI) within the normal range; no relevant clinically abnormal findings detected on physical examination, electrocardiography, lung function tests, or in the medical history, as judged by a medical physician; willingness to refrain from consuming alcohol within 72 h before each in-clinic evaluation; willingness to refrain from consuming and avoid being in the presence of the cooking of grilled, fried or barbequed food for 48 h before each in-clinic evaluation; not being pregnant or breast feeding and no use of any medications that interfere with the cyclo-oxygenase pathway (e.g. non-steroidal anti-inflammatory drugs) within 14 d prior to study commencement.

Additional inclusion criteria for the smoking groups were age between 23 and 55 years; being a regular current smoker of a brand of cigarettes with an ISO tar yield of 6–8 mg and a blend style and mechanics similar to brands sold in Germany; having smoked the same brand for a minimum of 6 months and any brand for at least 5 years before screening; smoking of between 10 and 30 cigarettes per day and having a urinary cotinine level of >100 ng/mL at screening.

Additional inclusion criteria for the ex-smoking group were age between 28 and 55 years; not having smoked for at least 5 years but having been a regular smoker of between 10 and 30 cigarettes per day for at least 5 years previously; having a urinary cotinine level of <30 ng/mL (inclusive of levels 0 and 1 on the NicAlert™ test, Palico, Rotkreuz, Switzerland) at screening and <100 ng/mL (inclusive of levels 0–2 on the NicAlert test) during the study.

Additional inclusion criteria for the never-smoking group were age between 28 and 55 years; never having smoked more than 100 cigarettes during his or her lifetime, and none in the previous 5 years; and having a urinary cotinine level of <30 ng/mL at screening and <100 ng/mL during the study. At study commencement, it became apparent that some non-smokers had urinary cotinine levels higher than the screening criterion of <30 ng/mL (the lower band of the NicAlert test range), most likely due to exposure to passive cigarette smoke. In order to retain study participants and statistical power, the criterion was adjusted to <100 ng/mL, which was still in the “non-smoker” range as described by the NicAlert test.

Cohort demographics

A total of 251 subjects were included in this study and were divided into three groups, current smoker, ex-smokers and never-smokers. The groups were matched as closely as possible for age, gender, ethnicity and BMI (Table 1). To determine the amount the subjects had smoked, pack years were calculated. The mean smoking history was 20.33 ± 11.12 pack years for current smokers and was 13.27 ± 8.94 pack years for ex-smokers (p < 0.001). Lung function was determined by Forced Expiratory Volume (FEV1) using standard testing methods (Miller et al., 2005). The mean FEV1 was 91.94% ± 12.48 for the current smokers, 98.46% ± 13.79 for ex-smokers and was 95.69% ± 12.06 for never-smokers. There was a significant difference between the FEV1 value for current smokers and ex-smokers (p < 0.01) and never-smokers (p < 0.05). There was no significant difference in the FEV1 values for ex-smokers and never-smokers. The cohort demographics are shown in Table 1.

Table 1. Subject demographics and biomarkers of exposure by study group.

	Study group
	(A) Current smoker	(B) Ex-smoker	(C) Never-smoker
Number of subjects (n)	143	61	61
Age (years)
Mean ± SD	38.8 ± 9.1	42.5 ± 7.7	41.8 ± 7.5
Median (min, max)	39 (23,55)	44 (28,55)	43 (28,35)
Gender (n, % of total)
Male	69 (48.25)	31 (50.82)	32 (52.46)
Female	74 (51.75)	30 (49.18)	29 (47.54)
Ethnicity (n, % of total)
Caucasian	143 (100)	61 (100)	59
Afro Caribbean	0 (0)	0 (0)	1 (1.64)
Other	0 (0)	0 (0)	1 (1.64)
BMI (kg/m^2)
Mean ± SD	24.58 ± 2.93	25.21 ± 2.84	25.02 ± 2.69
Median (min, max)	24.40 (18.6, 30.0)	25.40 (19.4, 29.8)	25.0 (18.8, 30.0)
Pack year
Mean ± SD	20.33 ± 11.12	13.27 ± 8.94	0
Median (min, max)	18.13 (2.63, 58.88)	11.00 (2.67, 40.99)	n/a
FEV1 (% predicted value)
Mean ± SD	91.94 ± 12.48	98.46 ± 13.79	95.69 ± 12.06
Median (min, max)	93 (50, 123)	100 (71, 130)	96 (70, 130)
Biomarkers of exposure (mean ± SD, n)
Urinary nicotine equivalents (mg/24 h)	15.039 ± 6.80 (120)^a,b	0.002 ± 0.006 (52)	0.001 ± 0.003 (54)
Urinary total 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol  (NNAL) (ng/24 h)	182.17 ± 13.44 (120)^a,b	5.37 ± 1.35 (52)	5.23 ± 2.59 (55)
Salivary cotinine (ng/mL)	204.87 ± 98.95 (128)^a,b	0.51 ± 0.07 (58)	0.51 ± 0.05 (59)
Exhaled carbon monoxide (CO) (ppm)	19.59 ± 8.00 (128)^a,b	1.33 ± 0.72 (61)	1.30 ± 0.63 (60)

^aBiomarkers of exposure achieving statistical difference relative to never-smokers, p < 0.001.

^bBiomarkers of exposure achieving statistical difference relative to ex-smokers, p < 0.001.

For sample collection, each participant attended the clinic (Momentum Pharma Services, Hamburg, Germany) for a 2-d residential stay and samples were taken for biomarker analysis on each day. Baseline BOBE data were pooled for these 2 d for statistical analysis.

Sample collection and processing

Urine

Prior to the start of a 24-h collection interval, participants emptied their bladder. All urine voided for a complete 24-h period (starting at approximately 19:00 h and ending at the same time on the following day) were collected on the in-clinic evaluation days. Samples were refrigerated at between 2 and 8 °C. For each participant, urine samples collected during a single 24-h interval were pooled together and weighed. At the end of the 24-h period, pooled samples were thoroughly mixed before providing aliquots for urinary biomarker analyses. Urine aliquots were stored in polypropylene tubes at −18 to −25 °C pending analysis, and shipped on dry ice to the analytical laboratory.

Blood-based collections

Blood samples were collected in Monovettes (Sarstedt) using the following anticoagulants for plasma preparation: Ethylenediaminetetraacetic acid (EDTA) for Interleukin 6 (IL-6), Interleukin 8 (IL-8), Monocyte Chemoattractant Protein 1 (MCP-1), ascorbic acid, soluble intercellular adhesion molecule 1 (sICAM-1) and oxidized LDL cholesterol analysis; lithium heparin for leukotriene B4 analysis and citrate for fibrinogen and neutrophil elastase analysis. Blood cells were separated from plasma by centrifugation. Plasma samples for the analysis of total ascorbate and ascorbic acid were prepared in accordance with the acidic deproteinization method described by Lykkesfeldt (2007).

Erythrocyte lysate

Whole blood was collected in Monovettes (Sarstedt) using EDTA as an anticoagulant. Blood was centrifuged at 800 × g for 10 min at 4 °C, and the plasma and white buffy layers were removed using a pipette. The remaining erythrocytes were lysed in four times their volume of ice-cold HPLC-grade water, and then centrifuged at 10 000 × g for 15 min at 4 °C. The supernatant (erythrocyte lysate) was recovered and stored for enzyme activity assays and malondialdehyde analysis at −80 °C.

Bioanalytical methods

Biomarker analysis was performed by Celerion (Lincoln, NE), Quest Diagnostics (Valencia, USA and Heston, UK), Eurofins (Breda, Netherlands) and Covance Laboratories (Harrogate, UK). The methodology and lower limit of quantification (LLOQ) are shown in Table 2.

Table 2. Biomarker analytical methods.

Biomarker	Matrix	Method	LLOQ
Nicotine equivalents		LC-MS/MS; Xu et al., 2004
Nicotine	Urine		10 ng/mL
Cotinine	Urine		10 ng/mL
Trans-3′-Hydroxycotinine	Urine		10 ng/mL
Nicotine-N-Glucuronide	Urine		10 ng/mL
Cotinine-N-Glucuronide	Urine		20 ng/mL
Trans-3′-Hydroxycotinine-O-Glucuronide	Urine		50 ng/mL
Total NNAL	Urine	LC-MS/MS; Kavvadias et al., 2009	5 pg/mL
Cotinine	Saliva	NicAlert™	1 ng/mL
CO	Exhaled breath	Electrochemical sensor	1 ppm
8-iso-PGF2α Type III	Urine	LC-MS/MS	25 pg/mL
8-iso-PGF2α Type VI	Urine	LC-MS/MS	25 pg/mL
11-dehydrothromboxane B2	Urine	LC-MS/MS	20 pg/mL
White blood cell (WBC) count	Blood	Flow cytometry	n/a
Neutrophil count	Blood	Flow cytometry	n/a
Monocyte count	Blood	Flow cytometry	n/a
High-density lipoprotein (HDL)	Serum	Enzyme activity via autoanalyser	3 mg/dL
Low-density lipoprotein (LDL)	Serum	Derived via Friedewald's equation	n/a
Oxidized low-density lipoprotein (oxLDL)	Plasma	ELISA	0.0018 IU/L
Total cholesterol	Serum	Enzyme activity via autoanalyser	25 mg/dL
High-sensitivity C-reactive protein (hsCRP)	Serum	Turbidity	0.2 mg/L
Triglycerides	Serum	Enzyme activity via autoanalyser	20 mg/dL
Fibrinogen	Plasma	Coagulation	80 mg/dL
Total antioxidant capacity (TAC)	Serum	Colorimetric	0.5 mmol/L
Superoxide dismutase (SOD) activity	Erythrocytes	Enzyme activity	0.025 U/mL
Glutathione reductase activity	Erythrocytes	Enzyme activity	20 nmol/min/mL
Glutathione peroxidase (GPx) activity	Erythrocytes	Enzyme activity	50 nmol/min/mL
Catalase activity	Erythrocytes	Enzyme activity	2 nmol/min/mL
Malondialdehyde	Erythrocytes	LC-MS/MS	1.00 μmol/L
8-Oxo-7,8-dihydro-2′′-deoxyguanosine (8-oxodG)	Urine	LC-MS/MS	0.2 ng/mL
Cis-thymidine glycol	Urine	LC-MS/MS	0.2 ng/mL
Ascorbic acid	Plasma	LC-MS/MS	5.00 μmol/L
Dehydroascorbic acid	Plasma	LC-MS/MS	0.500 μmol/L
Monocyte chemotactic protein 1 (MCP-1)	Plasma	Mulitplexed ELISA	3.3 pg/mL
Neutrophil elastase	Plasma	ELISA	15.60 ng/mL
Leukotriene B4 (LTB4)	Plasma	ELISA	10.29 pg/mL
Soluble intercellular adhesion molecule 1 (sICAM-1)	Plasma	ELISA	15.63 pg/mL
Interleukin 6 (IL-6)	Plasma	Mulitplexed ELISA	3 pg/mL
Interleukin 8 (IL-8)	Plasma	Mulitplexed ELISA	3 pg/mL

Urinary 8-iso-PGF2α Type III, 8-iso-PGF2α Type VI and 11-dehydrothromboxane B2 were analysed at Celerion (Lincoln, NE) by LC-MS/MS. Briefly for 8-iso-PGF2α type III and VI, urine samples supplemented with a stable label internal standard were extracted and concentrated via solid phase extraction and affinity separation with a polymeric mixed mode anion exchange sorbent. Sample extracts were injected for chromatographic separation on a Thermo Electron Corporation, Kromasil® C18 column with a shallow gradient mobile phase delivery system. Mobile phase A was 90% acetonitrile, 5% methanol and 5% ammonium acetate (500 mM). Mobile phase B was 30 mM ammonium acetate pH 4.0. Detection of negative ions was monitored in a multiple reaction monitoring mode on an AB/MDS Sciex API 4000 mass spectrometer. Ionization was achieved in APCI mode.

For 11-dehydrothromboxane B2, urine samples, supplemented with a stable label internal standard were extracted and concentrated via solid phase extraction, polymeric mixed mode anion exchange sorbent. Sample extracts were injected for chromatographic separation on a Waters ACQUITY UPLCTM BEH, C18 column with a shallow gradient mobile phase delivery system. Mobile phase A was 90% acetonitrile, 5% methanol and 5% ammonium acetate (500 mM). Mobile phase B was 30 mM ammonium acetate pH 4.0. Detection of negative ions was monitored in a multiple reaction monitoring mode on an AB/MDS Sciex API 5000 mass spectrometer. Ionization was achieved in APCI mode. Methods were developed and validated based on the good laboratory practice (GLP) principles described in 21 CFR Part 58 (FDA, 2013) and the Guidance for Industry – Bioanalytical Method Validation (FDA, 2001).

8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) and cis-thymidine glycol levels were determined by LC-MS/MS with solid phase extraction, using an API 4000™ (AB SCIEX, Framingham, MA) at Covance Laboratories, Harrogate, UK. With respect to cis-thymidine glycol, internal standard solution (20 µL, cis-thymidine glycol-d4, 500 ng/mL) and ammonium sulphate buffer (750 µL, 10 mM, pH 9.0) were added to thawed urine samples (250 µL). The samples were vortex mixed and subjected to solid phase extraction (Bond Elut 96 PBA 100 mg). The resulting eluates were evaporated to dryness under a stream of nitrogen and the residues were dissolved in acetonitrile: water: acetic acid (150 µL, 80: 20: 0.1, v/v/v) and centrifuged. The final extracts were submitted for LC-MS/MS analysis. Chromatographic separation was achieved within 5 min by an isocratic mobile phase containing acetonitrile: water: acetic acid (80:20:0.1, v/v/v), flowing through Phenomenex, Luna NH2, 3 µm, 150 × 3 mm (id) analytical column, at a flow rate of 0.5 mL/min. Mass spectrometric detection was by selected reactant monitoring (cis-Thymidine glycol m/z 274.9 → 116.0, IS m/z 278.9 → 119.9). Concentrations of cis-thymidine glycol in calibration standards and QC samples were determined using least squares linear regression with the reciprocal of the concentration (1/×) as weighting.

With respect to 8-oxodG, internal standard solution (20 µL, 250 ng/mL) and acetic acid (500 µL, 0.5% v/v) were added to thawed urine samples (0.5 mL). The samples were vortex mixed and subjected to solid phase extraction. The resulting eluates were evaporated to dryness under a stream of nitrogen and the residues were dissolved in acetonitrile: water: formic acid (150 µL, 75:25:0.1 v:v:v) and centrifuged. The final extracts were submitted for LC-MS/MS analysis. Chromatographic separation was achieved within 5 min by an isocratic mobile phase containing acetonitrile: water: formic acid (75: 25: 0.1, v/v/v) flowing through Phenomenex, Luna NH2, 3 µm, 150 × 3 mm (id) analytical column, at a flow rate of 0.5 mL/min. Mass spectrometric detection was by selected reactant monitoring (c8-hydroxy-2′-deoxyguanosine m/z 284.2 → 168.2, 15N5-8-hydroxy-2′-deoxyguanosine m/z 289.1 → 173.2). Concentrations of 8-oxodG in calibration standards and QC samples were determined using least squares linear regression with the reciprocal of the concentration (1/×) as weighting.

White blood cell (WBC), neutrophil and monocyte counts from whole blood were determined using flow cytometry on an automated haematology analyser (XE-5000™, Sysmex, Kobe, Japan). Serum high-density lipoprotein (HDL), total cholesterol and triglycerides were determined by enzyme activity using an automated analyser (AU5400 Chemistry Analyser, Olympus, Tokyo, Japan). Serum low-density lipoprotein (LDL) was calculated using the Friedewald equation. Serum high-sensitivity C-reactive protein (hsCRP) was measured by turbidity assay with a nephelometer (Dade Behring BNII, Siemens Healthcare, Erlangen, Germany). Serum fibrinogen was determined using the coagulation assay and Multifibren U Reagent (Siemens Healthcare Diagnostics, Erlangen, Germany) on an automated analyser (Behring Coagulation System®, Siemens Healthcare Diagnostics). Serum total antioxidant capacity (TAC) was determined using a commercial colourimetric kit (Randox Antrim, UK) on a chemistry analyser (ROCHE Modular P800, Basel, Switzerland). A multiplexed enzyme-linked immunosorbent assay (ELISA) system (Luminex™ 100, Luminex, Austin, TX) with a commercial kit (Fluorokine™ MAP Human Cytokine Custom Premix Kit, R&D Systems, Minneapolis, MN) was used to determine plasma levels of interleukin 6 (IL-6), interleukin 8 (IL-8) and monocyte chemotactic protein 1 (MCP-1). An automated ELISA system (BEP 2000, Siemens Healthcare) and commercial kits were used to measure plasma levels of oxidized low-density lipoprotein (oxLDL; MP Products, Amersfoort, Netherlands) and neutrophil elastase (Biovendor, Brno, Czech Republic). Plasma levels of soluble intercellular adhesion molecule 1 (sICAM-1) were determined by ELISA using a commercial assay (Custom Human 1-Plex Assay, Aushon BioSystems, Billerica, MA) on an automated analyser (SearchLight® Plus CCD Imaging and Analysis System, Aushon BioSystems). Plasma leukotriene B4 (LTB4) was determined by ELISA using a commercial assay (Parameter™, R&D Systems, Minneapolis, MN) and plate reader (GENios Pro™ Reader, Tecan, Männedorf, Switzerland). Plasma ascorbic acid and dehydroascorbic acid levels were determined by LC-MS/MS using a proprietary method conducted at Eurofins (Breda, Netherlands). The method was developed and validated in accordance with Guidance for Industry – Bioanalytical Method Validation (FDA, 2001). Briefly, the assay for the determination of total ascorbic acid is based upon the reduction of ascorbic acid oxidation products into ascorbic acid using dithiothreitol as a reducing agent, followed by protein precipitation and LC-MS/MS detection using an API 4000™ (AB SCIEX, Framingham, MA). To determine the concentration of dehydroascorbic acid, ascorbic acid concentration was subtracted from the concentration of total ascorbic acid.

Erythrocyte malondialdehyde concentrations were determined by a proprietary LC-MS/MS method developed at Eurofins (Breda, Netherlands). The method was developed and validated in accordance with Guidance for Industry – Bioanalytical Method Validation (FDA, 2001). Samples were extracted by protein precipitation and then derivatised by the addition of thio-barbituric-acid. The derivatised samples were extracted, internal standard added and analysed by HPLC equipped with an API 4000™ (AB SCIEX, Framingham, MA). Positive ions were monitored in the MRM mode. Erythrocyte levels of superoxide dismutase (SOD), glutathione reductase, glutathione peroxidase (GPx) and catalase activity were determined using commercially available kits (Cayman Chemical, Ann Arbor, MI) and a plate reader (SpectraMax® Plus384, Molecular Devices, Sunnyvale, CA). In brief, blood collections were performed using ethylenediaminetetraacetic acid as an anticoagulant. Whole blood samples were then centrifuged and erythrocytes lysed in water and enzyme activity determined. Haemoglobin levels in the lysates were determined by the cyanmethaemoglobin method using an automated analyser (ADVIA® 2120i, Siemens Healthcare, Erlangen, Germany). Levels were reported as the enzyme activity per µmol of haemoglobin. All erythrocyte samples were stored at −80 °C and analysed within 1 month of preparation.

Data presentation and statistical analysis

After protocol deviations were removed, biomarker results from the two individual days were averaged for each subject and results are presented as group mean, standard deviation (SD), median, minimum (min) and maximum (max). Descriptive statistics were used to summarize the demographic characteristics separately for each group. After testing for equal variances (Levene’s test for homogeneity), differences between days were tested using a paired t-test or Wilcoxon signed rank sum test accordingly. Repeated measures ANOVA was used to test for group differences. Age and gender were included in the model as covariates and Tukey’s test was used to account for multiplicity. To correct for violations of the regression assumptions, the natural logarithm was used to transform the biomarker values. In cases where data transformation was not effective, Kruskal–Wallis test was used to test group differences. The significance level was set at 5%.

Results

Biomarkers of exposure

To confirm smoking status a select number of tobacco specific biomarkers of exposure were measured (Table 1). The mean urinary nicotine equivalents and total NNAL were significantly higher in current smokers 15.039 ± 6.80 mg/24 h and 182.17 ± 13.44 ng/24 h, respectively, than in ex-smokers 0.002 ± 0.006 mg/24 h and 5.37 ± 1.35 ng/24 h, respectively, and never-smokers 0.001 ± 0.003 mg/24 h and 5.23 ± 2.59 ng/24 h, respectively. The mean salivary cotinine was 204.87 ± 98.95 ng/mL for the current smokers, 0.51 ± 0.07 ng/mL for ex-smokers and was 0.51 ± 0.05 ng/mL for never-smokers. The mean exhaled carbon monoxide was 19.59 ± 8.00 ppm for the current smokers, 1.33 ± 0.72 ppm for ex-smokers and was 1.30 ± 0.63 ppm for never-smokers. All biomarkers of exposure achieved significant statistical difference relative to ex-smokes and never-smokers (p < 0.001). These data confirm the self-reported smoking status of the three groups.

Biomarkers of biological effect

Of the 27 candidate BOBE assessed in this study 14 biomarkers were found to be statistically significantly different between current smokers and never-smokers groups as shown in Table 3.

Table 3. Biomarkers of biological effect achieving statistical difference relative to never-smokers.

							p Values
Biomarker	Study group	n	Mean ± SD	Median	Min	Max	A versus B	A versus C	B versus C
8-Iso-PGF2α Type III	(A) Current smoker	120	245.72 ± 115.70	220.72	69.84	748.16	<0.0001	<0.0001	NS
(ng/24 h)	(B) Ex-smoker	52	176.27 ± 68.87	162.18	77.10	437.56
	(C) Never-smoker	55	166.84 ± 63.26	154.86	75.06	371.64
11-Dehydrothromboxane B2	(A) Current smoker	119	799.73 ± 354.15	753.33	170.23	2402.53	NS	<0.001	NS
(ng/24 h)	(B) Ex-smoker	51	711.72 ± 344.46	665.40	140.88	1691.68
	(C) Never-smoker	55	616.63 ± 337.98	577.77	76.76	2326.17
WBC count (×10^−9/L)	(A) Current smoker	127	7.32 ± 1.73	7.15	4.00	11.50	<0.0001	<0.0001	NS
	(B) Ex-smoker	60	5.78 ± 1.03	5.65	3.80	8.80
	(C) Never-smoker	60	5.58 ± 1.10	5.50	3.80	9.00
Neutrophil count (×10^−9/L)	(A) Current smoker	127	4.04 ± 1.36	3.79	1.80	7.52	<0.0001	<0.0001	NS
	(B) Ex-smoker	60	3.13 ± 0.70	3.09	1.91	4.96
	(C) Never-smoker	60	3.02 ± 0.79	3.01	1.75	5.42
Monocyte Count (×10^−9/L)	(A) Current smoker	127	0.62 ± 0.182	0.58	0.27	1.44	<0.0001	<0.0001	NS
	(B) Ex-smoker	60	0.52 ± 0.121	0.50	0.33	0.84
	(C) Never-smoker	60	0.48 ± 0.123	0.49	0.25	0.77
HDL (mg/dL)	(A) Current smoker	127	48.84 ± 13.43	47.37	18.56	89.13	<0.01	<0.0001	NS
	(B) Ex-smoker	60	56.79 ± 18.27	54.53	28.42	134.57
	(C) Never-smoker	60	57.28 ± 16.27	57.14	29.39	96.10
oxLDL (IU/L)	(A) Current smoker	127	73.55 ± 25.04	69.50	28.00	134.50	<0.0001	<0.01	<0.05
	(B) Ex-smoker	60	58.41 ± 15.75	55.50	26.00	109.50
	(C) Never-smoker	60	65.73 ± 19.63	65.75	32.50	130.50
Total antioxidant capacity	(A) Current smoker	127	1.55 ± 0.10	1.53	1.35	1.80	<0.05	<0.0001	<0.05
(mmol/L)	(B) Ex-smoker	60	1.58 ± 0.09	1.57	1.40	1.81
	(C) Never-smoker	60	1.63 ± 0.12	1.62	1.28	1.90
Catalase activity	(A) Current smoker	127	1759.35 ± 1018.11	1358.75	525.05	4166.70	<0.05	<0.0001	<0.0001
(nmol/µmol/min)	(B) Ex-smoker	60	1477.29 ± 577.09	1310.95	511.70	3055.55
	(C) Never-smoker	60	943.92 ± 292.09	934.95	536.65	2577.9
Malondialdehyde	(A) Current smoker	127	118.03 ± 68.44	101.75	34.65	592.68	NS	<0.0001	<0.05
(nmol/mmol)	(B) Ex-smoker	59	96.12 ± 83.22	81.60	25.80	531.25
	(C) Never-smoker	59	67.44 ± 46.39	57.23	29.63	285.63
Ascorbic acid (µmol/L)	(A) Current smoker	127	53.38 ± 19.71	56.45	2.50	101.83	<0.05	<0.01	NS
	(B) Ex-smoker	59	58.99 ± 16.98	56.80	15.17	96.59
	(C) Never-smoker	59	59.72 ± 13.49	57.63	23.24	98.98
MCP-1 (pg/mL)	(A) Current smoker	127	134.60 ± 34.18	133.00	65.00	238.50	<0.0001	<0.0001	NS
	(B) Ex-smoker	60	107.88 ± 30.60	103.75	61.00	279.50
	(C) Never-smoker	60	105.06 ± 26.02	104.75	42.00	151.00
Neutrophil elastase (ng/mL)	(A) Current smoker	127	30.45 ± 15.84	30.50	8.00	102.50	<0.05	<0.0001	NS
	(B) Ex-smoker	60	21.13 ± 15.12	17.50	8.00	86.50
	(C) Never-smoker	60	19.39 ± 10.68	19.25	8.00	43.50
LTB4 (pg/mL)^a	(A) Current smoker	127	21.21 ± 44.43	5.15	5.15	299.79	<0.0001	<0.0001	NS
	(B) Ex-smoker	60	34.66 ± 34.19	27.82	5.15	176.95
	(C) Never-smoker	60	68.27 ± 203.61	23.59	5.15	1331.34

NS, Not significant.

^aLn transformation required.

The total amount of urinary 8-iso-PGF2α Type III and 11-dehydrothromboxane B2 excreted in 24 h were higher in current smokers than in never-smokers (p < 0.001). Total WBC, neutrophil and monocyte count were all also determined to be higher in current smokers than in never-smokers (p < 0.0001). In the current smoker group, the levels of serum HDL were determined to be lower than in the never-smoker group (p < 0.0001) and the levels of plasma oxidized LDL cholesterol were determined to be higher than in the never-smoker group (p = 0.01). As an overall indication of antioxidant status the TAC in serum was determined to be lower in current smokers than in never-smokers (p < 0.0001). Catalase activity was found to be significantly higher in current smokers compared to never-smokers (p < 0.0001). As an indicator of lipid peroxidation, the erythrocyte malondialdehyde concentrations were determined to be higher in current smokers than in never-smokers (p < 0.0001). In the current smoker group, plasma levels of the anti-oxidant ascorbic acid were observed to be lower than in the never-smoker group (p < 0.01). Plasma levels of the inflammatory mediators MCP-1 and neutrophil elastase were higher in current smokers than in never-smokers (p < 0.0001). The plasma levels of the eicosanoid inflammatory mediator LTB4 were determined to be lower in current smokers than in never-smokers (p < 0.0001).

From the 14 BOBE found to be statistically significantly different between the current smokers and never-smokers groups, 12 of the biomarkers were also found to be statistically significantly different between current smokers and ex-smokers groups. The biomarkers that achieved a significant difference between current smokers and ex-smokers were urinary 8-iso-PGF2α Type III (p < 0.0001), total WBC, neutrophil and monocyte count (p < 0.0001), serum HDL (p < 0.01), plasma oxidized LDL cholesterol (p < 0.0001), serum TAC (p < 0.05), erythrocytes catalase enzyme activity (p < 0.05), plasma ascorbic acid (p < 0.05), MCP-1 (p < 0.0001), neutrophil elastase (p < 0.05) and LTB4 (p < 0.0001). The mean levels of 11-dehydrothromboxane B2 in current smokers were higher than in the ex-smoker group. However, due to the variation within the groups this was not statistically significant.

None of the other 13 candidate BOBE assessed in this study were found to be statistically significantly different between current smokers and never-smokers groups and are shown below in Table 4 for reference purposes. The plasma levels IL-6, IL-8 and TNF-α were below the limit of quantification for the assays used in this study.

Table 4. Biomarkers of biological effect not achieving statistical difference relative to never-smokers.

							p Values
Biomarker	Study group	n	Mean ± SD	Median	Min	Max	A versus B	A versus C	B versus C
8-Iso-PGF2α Type VI	(A) Current smoker	120	1536.45 ± 636.39	1403.67	351.83	3889.19	NS	NS	NS
(ng/24 h)	(B) Ex-smoker	52	1589.06 ± 585.87	1549.89	441.51	3169.80
	(C) Never-smoker	55	1615.47 ± 836.67	1382.58	489.94	4492.22
hsCRP (males, mg/L)	(A) Current smoker	63	2.083 ± 5.628	0.450	0.100	41.350	NS	NS	NS
	(B) Ex-smoker	30	0.985 ± 1.059	0.500	0.100	4.050
	(C) Never-smoker	31	1.195 ± 1.218	0.750	0.100	4.000
Total cholesterol	(A) Current smoker	127	190.69 ± 38.02	186.20	103.25	301.82	NS	NS	NS
(mg/dL)	(B) Ex-smoker	60	201.47 ± 36.95	196.44	129.35	286.55
	(C) Never-smoker	60	199.36 ± 48.14	200.12	101.90	373.75
LDL (mg/dL)	(A) Current smoker	126	118.28 ± 34.30	113.30	51.04	213.65	NS	NS	NS
	(B) Ex-smoker	60	122.04 ± 29.05	118.14	65.93	181.56
	(C) Never-smoker	60	118.66 ± 38.78	117.46	48.72	243.23
Triglycerides (mg/dL)	(A) Current smoker	127	121.40 ± 67.02	103.63	39.86	590.32	NS	NS	NS
	(B) Ex-smoker	60	113.06 ± 58.12	97.65	41.19	329.48
	(C) Never-smoker	60	116.11 ± 66.12	92.56	42.07	402.55
Fibrinogen (mg/dL)	(A) Current smoker	127	292.40 ± 55.78	285.00	195.00	495.00	NS	NS	NS
	(B) Ex-smoker	60	288.00 ± 44.76	277.50	210.00	395.00
	(C) Never-smoker	60	292.67 ± 45.92	290.00	210.00	455.00
SOD activity	(A) Current smoker	127	4.31 ± 1.02	4.05	2.75	6.85	<0.0001	NS	<0.0001
(IU/µmol)	(B) Ex-smoker	60	3.66 ± 0.59	3.58	2.45	4.90
	(C) Never-smoker	60	4.48 ± 0.93	4.15	3.10	7.30
Glutathione reductase	(A) Current smoker	127	30.64 ± 15.17	28.25	0.20	73.75	<0.0001	NS	NS
activity (nmol/	(B) Ex-smoker	60	16.47 ± 11.48	14.73	0.15	53.90
µmol/min)	(C) Never-smoker	60	34.35 ± 65.14	24.75	0.18	517.05
GPx activity	(A) Current smoker	127	323.62 ± 73.60	315.90	172.95	559.95	NS	NS	NS
(nmol/µmol/min)	(B) Ex-smoker	60	346.82 ± 82.48	330.73	218.45	618.05
	(C) Never-smoker	60	341.86 ± 72.62	335.81	228.50	588.45
8-OxodG (ng/24 h)	(A) Current smoker	120	4912.63 ± 2270.72	4581.12	819.67	14297.43	NS	NS	NS
	(B) Ex-smoker	52	4993.84 ± 2500.68	4832.49	1399.02	17074.38
	(C) Never-smoker	56	4913.08 ± 2025.30	4733.56	1540.35	11282.71
Cis-thymidine glycol	(A) Current smoker	120	2612.18 ± 1204.74	2406.26	446.94	8037.57	NS	NS	NS
(ng/24 h)	(B) Ex-smoker	52	2521.80 ± 1263.84	2442.97	0.00	7830.68
	(C) Never-smoker	56	2378.13 ± 910.18	2350.89	776.04	4663.68
Dehydroascorbic acid	(A) Current smoker	127	25.52 ± 4.33	25.82	12.74	36.00	NS	NS	NS
(µmol/L)	(B) Ex-smoker	59	25.65 ± 5.48	26.08	10.52	37.24
	(C) Never-smoker	59	27.34 ± 5.16	27.81	16.10	43.08
sICAM-1 (pg/mL)	(A) Current smoker	127	339.55 ± 92.01	336.32	132.68	649.69	<0.01	NS	NS
	(B) Ex-smoker	60	287.65 ± 94.33	279.21	148.01	596.06
	(C) Never-smoker	60	320.28 ± 103.11	306.48	143.17	557.41

Discussion

The BOBEs measured in this study have been associated with key disease processes including those of oxidative stress, inflammation and oxidatively damaged DNA, which are known to contribute to the development of smoking-related diseases.

Biomarkers of oxidative stress

Elevations in levels of urinary and plasma isoprostane have been previously summarized in smokers compared to non-smokers (Lowe et al., 2013), and furthermore significant reductions in isoprostane levels were observed following smoking cessation (Chehne et al., 2002; Flores et al., 2004; Morrow et al., 1995; Oguogho et al., 2000). Given the utility of this biomarker for distinguishing smoking status, demonstrating reversibility upon smoking cessation and the association with smoking-related diseases (e.g. cardiovascular disease and COPD) via oxidative stress-mediated mechanisms (Morrow, 2005; Rahman, 2005), F2 isoprostanes appear to be promising candidate biomarkers to measure short-term changes in systemic oxidative burden in studies involving a switch from a conventional cigarette to a candidate modified risk tobacco product. In this study, smokers had significantly higher baseline levels of 8-iso-PGF2α Type III compared to both former and never-smokers, whereas no significant difference was evident with 8-iso-PGF2α Type VI. There are numerous studies in the literature which utilize ELISA methods for the quantification of 8-isoprostanes. While the use of ELISA has advantages in terms of throughput and cost, reports have suggested that commercial ELISA kits suffer from poor specificity and correlations between mass spectrometry based methods and ELISA are poor (Smith et al., 2011; Tacconelli et al., 2010). The potential impact of such cross-reactivity is highlighted in the current study, as the variability for 8-iso-PGF2α Type VI was substantially higher than for 8-iso-PGF2α Type III, indicating that levels of 8-iso-PGF2α Type VI in the general population are quite variable, irrespective of smoking status. It is possible that by using ELISA-based assay kits with poor specificity, the potential of 8-iso-PGF2α Type III as a BOBE for smoking-related oxidative stress could be diminished, if certain ELISAs are also sensitive to 8-iso-PGF2α Type VI. Given previous data from the literature and the data from this study, the use of urinary 8-iso-PGF2α Type III in future product assessment studies is recommended. Urinary 8-iso-PGF2α Type VI was not able to discriminate smoking status in this study and hence its use could be limited to very large cohorts with much greater statistical power.

Ascorbic acid levels have been used widely as a biomarker of general oxidative stress, and studies show that levels of ascorbic acid are generally lower in smokers compared with non-smokers and that levels appear to correlate to some extent with the number of cigarettes smoked per day (Lowe et al., 2013). Baseline data from this study supports previously published cross-sectional data in that ascorbic acid levels in smokers were significantly lower than both former and never-smokers, suggesting elevated levels of general oxidative stress in the smoker group.

With respect to dehydroascorbic acid (oxidized ascorbic acid), its potential utility in product assessment studies has been described (Lowe et al., 2013). In this study a large number of individuals returned negative values (as the levels are derived from a subtraction of ascorbic acid from total ascorbate, as described by Lykkesfeldt (2007), and hence for statistical purposes were assigned a value of zero. There was no significant difference between the study groups for this biomarker, in contrast to published data (Chávez et al., 2007).

Other oxidative stress-related BOBEs included erythrocyte catalase, glutathione reductase and glutathione peroxidase activity, erythrocyte malondialdehyde levels and plasma dehydroascorbic acid. These BOBE have known mechanistic links to oxidative stress, which are in turn associated with the development of cardiovascular disease and COPD (Bonomini et al., 2008; Frikke-Schmidt & Lykkesfeldt, 2009; Li et al., 2013; Lubos et al., 2011; Mak, 2008; Rahman & Adcock, 2006). Despite previously reported differences in erythrocyte enzyme activity and lipid peroxidation; of which malondialdehyde is an indicator, between smokers and non-smokers (Ho et al., 2005; Lowe et al., 2013; Orhan et al., 2005; Pannuru et al., 2011) all of these biomarkers exhibited considerable heterogeneity in the data making interpretation difficult. These data support the conclusions of Bogdanska et al. (2003) that healthy individuals exhibit significant heterogeneity with the activities of these enzymes, possibly due to individual circumstances at the time of sampling, measurement variability and genetic variability inherent in the general population. Hence, these BOBEs are unlikely to be useful for future product assessment studies without extensive research to better understand the inter-individual variability and analytical method standardization.

TAC is a measurement of the buffering potential of a biofluid against oxidation. In general, TAC is decreased in conditions associated with oxidative stress (Young, 2001), and some studies show that TAC is able to discriminate smoking status (Bloomer, 2007; Buico et al., 2009; Lowe et al., 2013), and demonstrate reversibility upon smoking cessation (Petruzzelli et al., 2000). In this study, baseline levels of TAC were significantly lower in smokers compared to both former and never-smokers, which is in agreement with published data. As suggested by Young (2001), TAC data should be considered alongside other endpoints of oxidative stress. In this case, the TAC data supports that of urinary 8-iso-PGF2α Type III and plasma ascorbic acid in suggesting that smokers experience elevated levels of general oxidative stress.

Inflammatory biomarkers

Elevations in WBC have been associated with smoking and smoking-related disease (Fröhlich et al., 2003; Lao et al., 2009), and decreases have been reported following smoking cessation (Blann et al., 1997; Hammett et al., 2007; Lao et al., 2009). In this study, WBC was significantly higher in smokers compared to former and never-smokers at baseline in agreement with published literature. This indicates clear discrimination of smoking status and the potential for reversibility following a product switch, and hence WBC is potentially useful for future product assessment studies. The monocyte and neutrophil counts followed a very similar pattern to WBC and hence could also be potentially useful BOBEs.

11-Dehydrothromboxane B2 is a urinary metabolite of the potent platelet agonist and vasoconstrictor; thromboxane A2, and is associated with platelet activation, and smoking status (Frost-Pineda et al., 2011; Lowe et al., 2009; Nowak et al., 1987; Wennmalm et al., 1991). Furthermore, significant reductions in the level of 11-dehydrothromboxane B2 were reported following smoking cessation (Rångemark et al., 1993; Saareks et al., 2001) and following a switch to an electrically heated tobacco product (Roethig et al., 2008). In this study, 11-dehydrothromboxane B2 levels in smokers were significantly higher at baseline than never-smokers, but not in former smokers in agreement with our previous data (Lowe et al., 2009). The lack of discrimination between smokers and former smokers is likely to be due to inherent variability in the general population caused by factors other than gender and BMI (Liu et al., 2011) and hence, there was a lack of statistical power in this study to overcome this. Given the utility of 11-dehydrothromboxane B2 in other studies to differentiate smoking status, demonstrate reversibility following smoking cessation and following a switch to a tobacco product with reduced machine-smoked toxicant yields, this biomarker could be included in future product assessment studies.

Recruitment of inflammatory cells from the vascular system is an essential part of the damage-repair process in endothelial dysfunction and tissue injury, and elevations in sICAM-1 and MCP-1 have implications for the development of atherosclerosis (Deo et al., 2004; Gross et al., 2012; Piemonti et al., 2009). With respect to atherosclerosis, leucocyte trafficking is controlled by interactions between the endothelium and leukocytes via chemotactic signalling and subsequent binding of leukocytes to adhesion molecules on the endothelial layer (Palmer et al., 2002; Scott et al., 2000). MCP-1 is such a chemotactic factor which targets monocytes. There is surprisingly little data in the literature with respect to studies which have investigated MCP-1 levels in the plasma/sera of healthy individuals who smoke. A small study by Garlichs et al. (2009) reported elevated plasma MCP-1 in smokers compared to non-smokers; however, it did not reach statistical significance possibly due to the sample size. Animal studies have shown a relationship between elevated plasma MCP-1 and smoke exposure (Yuan et al., 2007) and elevated MCP-1 is associated with peripheral artery disease and increased carotid intima-media thickness (Kusano et al., 2004; Larsson et al., 2005; Satiroglu et al., 2011). In this study, baseline levels of MCP-1 were significantly higher in smokers compared to former and never-smokers indicating that this BOBE could be potentially useful in future product assessment studies.

Levels of sICAM-1 rise in a dose-dependent manner following smoke exposure (Bermudez et al., 2002) and they lower rapidly, within 4 weeks of smoking cessation, which has been shown to be sustained for 52 weeks (Bermudez et al., 2002; Palmer et al., 2002; Scott et al., 2000) and can be reduced to that of never-smoker levels (Bermudez et al., 2002). In this study, initial baseline comparisons of smokers compared to former and never-smokers showed no significant difference in levels of plasma sICAM-1, in contrast to published literature and the reason for this is unclear, especially since these studies utilized immunoassay techniques. Comparisons of day 1 versus day 2 data showed a significant difference between days in the smoker group (p < 0.05), however, the absolute values were not vastly different. There was no significant difference between these time-points for the former and never-smokers, indicating that the analytical method was unlikely to be a major factor (data not shown). Given the current literature reports on sICAM-1 demonstrating its ability to distinguish smoking status and demonstrate reversibility upon smoking cessation, this biomarker could be included in future product assessment studies for further analysis.

Neutrophil elastase and leukotriene B4 also demonstrated a significant increase and decrease, respectively, in smokers, compared to former and never-smokers indicating the potential for future use. However, given that the absolute levels were relatively low in plasma, the heterogeneity in the data would make the assessment of product-related effects difficult, thus these BOBEs are unlikely to be useful for future studies.

Lipid biomarkers

Elevated levels of LDL cholesterol and triglycerides, and reduced levels of HDL cholesterol following smoking have been implicated as a causative factor in the development and progression of cardiovascular disease. (Chelland Campbell et al., 2008). Numerous studies have reported significant differences between smokers and non-smokers and a reversal of the pro-disease lipid profile mentioned above, in people who have quit smoking (Chelland Campbell et al., 2008; Hata & Nakajima, 2000; Ohsawa et al., 2005) or switched to an electronically heated tobacco product (Roethig et al., 2008). In this study, levels of HDL were significantly lower in current smokers compared to both former and never-smokers at baseline, in agreement with most published literature. Given the ability of HDL to discriminate smoking status in most literature reports, demonstrate reversibility following smoking cessation and following exposure to tobacco products with reduced machine-smoked toxicant yields, this biomarker should be included in future product assessment studies. With respect to total cholesterol, LDL and triglycerides, there was no significant difference between smokers and non-smokers in agreement with our previous data (Lowe et al., 2009), thus the use of these BOBEs in future short term studies is of questionable value. Elevated oxLDL has been heavily implicated as a contributing factor in the development of atherosclerosis (Boullier et al., 2001; Holvoet, 2004; Kita et al., 2001). Studies in comparing smokers and non-smokers have shown conflicting data. Smokers exhibited significantly higher (2.3-fold) 3-N-T levels in LDL apoprotein than non-smokers (Yamaguchi et al., 2005) and elevated oxLDL levels have been reported in smokers in other studies (Scheffler et al., 1990, 1992; Yoshida et al., 2004). In contrast, studies have reported no difference between smokers and non-smokers (Hininger et al., 1997; Marangon et al., 1997). In this study, baseline data showed that smokers had significantly higher levels of oxLDL compared to both former and never-smokers, indicating that this biomarker has potential for future studies.

All other BOBEs not discussed here did not show any significant difference between smokers and non-smokers at baseline, and hence their use for future short-term product comparison studies is questionable. Eleven BOBEs, which were able to distinguish between current smokers and former smokers, were in general agreement with published literature. Although the association between plasma MCP-1 levels and atherosclerosis is not novel, there is currently very limited data available on plasma MCP-1 levels in healthy current and former smokers in a cohort of this size. Hence, we propose that plasma MCP-1 should be further scrutinized in future longitudinal studies to confirm its utility for product assessment studies.

The current data show BOBE levels at a single baseline time-point. To better understand the robustness of these responses, longitudinal analysis of each BOBE using each subject as their own control would provide further insight into the usefulness of these BOBEs for product assessment. Indeed, it is possible that some BOBEs that we have identified as potentially useful, may not be robust enough when data from multiple time-points are included, and vice versa for BOBEs where we have questioned their usefulness. Dietary and metabolic influences on these BOBEs are not to be underestimated, and may have contributed to the heterogeneity observed with some BOBEs in this study; especially those BOBEs which change in shorter timeframes (e.g. oxidative stress biomarkers). Without doing specific studies on the influence of known or hypothesized dietary and/or metabolic interactions with candidate BOBEs, analysis at multiple time-points over a longer period of time would help to give confidence on whether or not BOBEs are affected by “general dietary and/or metabolic variation” throughout the life of the study. Where dietary and/or metabolic interactions are known, they should be accounted for in the study design via inclusion/exclusion criteria if appropriate (e.g. known polymorphisms of key genes known to affect the BOBE, or restriction of certain food groups), or by tracking dietary intake including it in the data analysis as a co-variate.

Conclusions

The data presented in this study serves to determine areas of focus for longitudinal analysis of BOBEs in current smokers following a forced switch from conventional cigarettes to a MRTP or other nicotine delivery devices. The value of BOBE for tobacco/nicotine delivery product assessment is still very much open to debate given the lack of studies to relate these endpoints to disease risk and benefits to human health (i.e. it is currently difficult to prove that changes in BOBE levels directly relate to pathological/risk reducing outcomes), and such studies are very much needed. Furthermore, studies which investigate correlations between biomarkers of exposure and BOBE may provide insights into which specific, or classes of, smoke toxicants drive adverse biological responses associated with disease development pathways. Robustness assessments on candidate BOBEs over multiple time-points are also needed to determine reliability, especially given that in most cases, confounding factors, which may affect BOBE levels, have not been fully characterized. Currently, the US FDA draft guidance for industry has stated that MRTPs may enter inter-state commerce if they satisfy an “Exposure Modification Order”, which is defined as “a modified risk tobacco product that reduces or eliminates exposure to a substance and for which the available scientific evidence suggests that a measurable and substantial reduction in morbidity and mortality is reasonably likely to be demonstrated in future studies” (Food and Drug Administration, 2012). The study of BOBEs in product assessment studies may serve to provide regulators with the necessary information to satisfy this level of regulation, in conjunction with smoke chemistry, biomarker of exposure, in vitro and potentially animal data to determine the impact of a candidate MRTP at an individual level.

Declaration of interest

The authors declare no conflict of interest outside of the funding source. This study was funded by British American Tobacco (Investments) Ltd.