Inhibition of benzo[a]pyrene-induced DNA adduct in buccal cells of smokers by black raspberry lozenges

Abstract

Using LC–MS/MS analysis we previously showed for the first time (Carcinogenesis 43:746–753, 2022) that levels of DNA damage induced by benzo[a]pyrene (B[a]P), an oral carcinogen and tobacco smoke (TS) constituent, were significantly higher in buccal cells of smokers than those in nonsmokers; these results suggest the potential contribution of B[a]P in the development of oral squamous cell carcinoma (OSCC) in humans. Treating cancers, including OSCC, at late stages, even with improved targeted therapies, continues to be a major challenge. Thus interception/prevention remains a preferable approach for OSCC management and control. In previous preclinical studies, we and others demonstrated the protective effects of black raspberry (BRB) against carcinogen-induced DNA damage and OSCC. Thus, to translate preclinical findings, we tested the hypothesis in a Phase 0 clinical study that BRB administration reduces DNA damage induced by B[a]P in the buccal cells of smokers. After enrolling 27 smokers, baseline buccal cells were collected before the administration of BRB lozenges (5/day for 8 weeks, 1 gm BRB powder/lozenge) at baseline, at the middle and the end of BRB administration. The last samples were collected 4 weeks after BRB cessation (washout period). B[a]P-induced DNA damage (BPDE-N²-dG) was evaluated by LC–MS/MS. BRB administration resulted in a significant reduction in DNA damage: 26.3% at the midpoint (P = .01506) compared to baseline, 36.1% at the end of BRB administration (P = .00355), and 16.6% after BRB cessation (P = .007586). Our results suggest the potential benefits of BRB as a chemopreventive agent against the development of TS-initiated OSCC.

In a clinical study, we demonstrated that black raspberry (BRB) inhibited benzo[a ]pyrene-induced DNA damage in smokers’ buccal epithelial cells. The results support the potential benefits of BRB as a chemopreventive agent against the development of tobacco smoke-initiated OSCC.

Graphical Abstract

Graphical Abstract.

Introduction

Worldwide, head and neck squamous cell carcinoma (HNSCC) is the sixth most common malignancy; the most predominant type of this disease is oral squamous cell carcinoma (OSCC) (1). Major causative agents in the development of HNSCC include tobacco, alcohol, oncogenic HPV and betel quid (1–5). Although over 60 carcinogens were identified in tobacco smoke (TS) (6), it remains unclear which compounds in TS contribute to human HNSCC development. However, studies reported in the literature indicate that enzymatic bioactivation of certain classes of chemical carcinogens, such as polycyclic aromatic hydrocarbons (PAHs) and tobacco-specific nitrosamines (TSNA), generates active metabolites that can damage DNA, leading to OSCC development (4–6). In addition, human oral epithelium possesses all the enzymes associated with nitrosamine metabolism (7), and PAHs can be bioactivated by human oral mucosal explants; these results are consistent with the presence of functional cytochrome P-450 1A1 and 1B1 (CYP1A1/1B1) (8).

PAHs are products of incomplete combustion or pyrolysis processes, and they are present in cigarette smoking, grilled and smoked food, and vehicle-generated exhausts (9). The environmental PAH carcinogen, benzo[a]pyrene (B[a]P), has been a model compound that has been extensively used to study the biological mechanism(s) of action of PAHs and to assess their exposure (10). B[a]P was classified by the International Agency for Research on Cancer as Group 1 (known human carcinogen) (11). Feeding B[a]P in the diet for two years induced papillomas and carcinomas of the tongue in mice (4,12). Furthermore, B[a]P induced both papillomas and carcinomas in the cheek pouch of hamsters (4). B[a]P can be metabolically activated by CYP1A1/1B1 and epoxide hydrolase that can lead to the formation of B[a]P-dihydrodiol (B[a]P-DHD) intermediate and its ultimate carcinogen 7,8-dihydroxy-anti-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (anti-BPDE) which reacts with DNA predominately at the N²-position of guanine to form BPDE-N²-dG adduct (Fig. 1) (4). A minor adduct with N⁶-position of adenine was previously reported in animal models to form N⁶-deoxyadenosine adduct, but BPDE-N²-dG remains the major adduct reported (13). Since the formation of DNA adducts from PAHs is a critical step in the multi-step process of carcinogenesis, adduct formation can serve as a valid biomarker to assess cancer risk. It has been reported that the major DNA adducts resulting from exposure to the ultimate tumorigenic metabolite of B[a]P, BPDE, were detected in human oral buccal cells (14,15). However, the methods (immunohistochemistry, ELISA) used in these previous studies to detect this adduct suffer from cross-reactivity with adducts other than B[a]P-DNA adduct, cannot provide unequivocal information on the structure of the adduct and are at best semiquantitative. In fact, using Linear Quadruple Ion Trap Mass Spectrometry, Bessette et al. could not identify this adduct in buccal cells in a pilot study of six smoking subjects (16). Thus, to avoid such limitations associated with previously reported methods, in a recent study using a quantitative LC-MS/MS method, we demonstrated that BPDE-N²-dG adducts derived from B[a]P in buccal cells of smokers were significantly higher than those detected in nonsmokers (17).

Figure 1.

Metabolic activation of B[a]P to the corresponding diol epoxide BPDE leads to the formation of BPDE-N²-dG adduct

Treating cancers, including HNSCC at advanced stages remains a significant challenge even with improved targeted therapies, making prevention and early detection desirable strategies to manage and control this disease (4,5). Diets rich in fruits and vegetables have the potential to reduce the risk of cancer. This observation may reflect, in part, the high content of constituents that have the ability to inhibit carcinogen initiation and progression (5). Among several sources of phytochemicals, black raspberry (BRB) has shown great promise in cancer prevention in both preclinical and clinical studies (4). BRB is known to affect various biological activities, such as a modifier of Phase I, Phase II and DNA repair enzyme activities (18). BRB has also been shown to inhibit carcinogen-induced DNA damage in oral tissues and OSCC in preclinical animal models (4). As BRB is approximately 90% water, freeze-drying concentrates BRB components approximately 10-fold (19). Thus, in this Phase 0 clinical trial, we examined the effects of freeze-dried BRB on the levels of DNA damage induced by B[a]P in buccal cells of current smokers in an attempt to assess the potential cancer prevention effects of an oral lozenge that can deliver BRB.

Materials and methods

Chemicals and enzymes

Chemicals and enzymes, including protease K, RNase A, and 8-hydroxyquinoline, were purchased from Sigma-Aldrich (St. Louis, MO). [¹⁵N]-dG was obtained from Spectra Stable Isotopes (Columbia, MD). BPDE-N²-dG and its [N¹⁵]-labeled internal standard were synthesized using our previously published microwave-supported synthesis (17). The crude reaction product was partially purified using a C-18 Sep-PAK cartridge. The methanolic eluate was concentrated and reconstituted in 2 ml of MeOH and purified by HPLC to remove unreacted dG.

Black raspberry [BRB] lozenges as the test agent

As we reported (20), each participant received a dose of 5 g freeze-dried BRB per day in the form of dissolvable slow-release BRB lozenges (each lozenge contains 1 g BRB freeze-dried powder). Each participant took 1 lozenge each time, five times/day and did not eat or drink 30 min following lozenge use. In a previous Phase 0 clinical trial, it was reported that a comparable dose regimen in the form of troches has been used (21). All lozenges were prepared under GMP from a single harvest year of black raspberry (Rubus occidentalis of the Munger cultivar), which was obtained by BerriHealth, Corvallis, OR, from farms located in Corbett and Sandy, OR. Each lozenge contains Inert ingredients and binders, including 150 mg organic honey crystal, 150 mg organic rice extract, cellulose, and inorganic elements (copper, magnesium). Lozenges possess a dark reddish-purple color with a total mass of 1.3 g each; the shape is round (7.43 mm × 22.79 mm). These lozenges were analyzed by Eurofins Food Chemistry Testing (Madison, WI) for compositions of constituents, including vitamins, minerals, anthocyanosides, free anthocyanidins, ellagic acid and quercetin dihydrate. The total free anthocyanidins and anthocyanosides in our formulation is 41.8 mg/serving size based on the Certificate of Analysis; further details have been recently reported by us (20).

Demographics, inclusion and exclusion criteria of participants

As we recently reported (20), the participants (21–75-year-old) were recruited from a tobacco user research registry which was maintained at the Penn State College of Medicine in Hershey, Pennsylvania. Our study was designed in a way that every participant serves as his (or her) own internal control and assists by using an internal control for those who, by virtue of their medical conditions, need to take medications such as antiinflammatory agents on a daily basis. The possible effects of daily meloxicam and low-dose ASA for these participants are already recorded on their baseline levels. Since we are looking at changes to the baseline levels, those participants who take daily medications should not be a problem. In this study, 27 smokers (10 males, 17 females) were included, as shown in the demographic data summarized in Table 1.

Table 1.

Demographic characteristics of smokers enrolled in the Phase 0 clinical trial.

Demographics
Gender, n (%)
Male	10 (37)
Female	17 (63)
Age, mean (SD)	44.7 (11.6)
Smoking Characteristics
Cigarettes per day, mean (SD)	19.4 (8.4)
Cigarette flavor, n (%)
Menthol	17 (63)
Nonmenthol	10 (37)

Study design

The overall design of this study (depicted in Fig. 2) was approved by the Penn State University Institutional Review Board (IRB#13621). Furthermore, this Phase 0 clinical trial is registered at clinicaltrials.gov (NCT04372914). Urine samples were collected at baseline one week after being enrolled (week 1), followed by mid-BRB (week 5), end-BRB (week 9) and a washout period (week 13). These four urine collections were more than sufficient for cotinine and creatinine analysis. However, to ensure an adequate amount of DNA from buccal cells for the BPDE-N²-dG analysis, we conducted two sample collections (1 week apart) for each collection period as follows: baseline (weeks 0, 1), during the middle of BRB administration (weeks 4, 5), at the end of BRB administration (weeks 8, 9), wash out period after BRB cessation (weeks 12, 13). Participants were instructed by our clinical coordinator to consume 5 BRB lozenges per day for 8 weeks and not to swallow or crush them. All of the collected samples were aliquoted and kept in a −80°C freezer.

Figure 2.

The study design of the Phase 0 trial demonstrates the duration of BRB supplementation and the time points for the collection of buccal cells and urine of smokers

Buccal cell collection methods

The procedure for buccal cell collection was recently reported (20). In brief, participants rinsed their mouths twice with water at the beginning of the trial. Then, they brush the inside of the mouth up and down on each side of the cheek for a duration of one minute using a soft bristle toothbrush. Approximately 20 ml of saline was used for rinsing (SALJET; Winchester Laboratories LLC; St. Charles, IL) for two minutes before collection using a 50 ml centrifuge tube.

DNA isolation

A phenol-chloroform extraction method was used for DNA isolation from oral buccal cells (22). Buccal cells were first suspended in 200 µl alkaline lysis buffer followed by the addition of proteinase K. These buccal cells were kept in an incubator for 3–4 h under 58°C; then 300 µl of Phenol/chloroform/isoamyl alcohol mix (25:24:1, pH 6.7) was added to the solution. After centrifugation for 5 min at 13 000 rpm, the top aqueous layer was collected, followed by adding sodium acetate buffer (to a final concentration of 0.3 M, pH 5.2) and ethanol to precipitate DNA. The resulting pellet was further washed with 70% ethanol before being stored at −20°C.

DNA hydrolysis, deoxyribonucleosides analysis, and purification

The procedure of DNA hydrolysis has been previously described (20). Briefly, 300 pg of [¹⁵N]-labeled BPDE-N²-dG adducts were added to 10 μg of DNA prior to enzymatic digestion. We hydrolyzed DNA in the presence of MgCl₂ and DNase I for 1.5 h at 37°C, followed by another incubation with enzymes including nuclease P1, phosphodiesterase and alkaline phosphatase. A small portion of the DNA hydrolysate was used to examine for the completion of hydrolysis by HPLC. We partially purified the remaining hydrolysate using an Oasis HLB column (1 ml, 30 mg; Waters Ltd.).

LC-MS/MS analysis

The detailed method for the analysis of BPDE-N²-dG using a Sciex QTRAP 6500 + mass spectrometry was previously reported (17). Briefly, the mass spectrometry is coupled with a Sciex EXion HPLC separation system. The analyte was separated on a 1.7 µm Acquity UPLC BEH C18 analytical column (2.1 × 50 mm, Waters, Ireland). We conducted a gradient elution by a flow rate of 0.3 ml/min with the following conditions: initially, we used 40% solvent B (containing 0.1% formic acid in methanol) and 60% solvent A (containing 0.1% formic acid in water), then a linear gradient to 100% solvent B in 2 min was obtained and kept for two additional minutes before going back to initial conditions.

The decluster potential was 70 V for BPDE-N²-dG and [¹⁵N]-BPDE-N²-dG; the entrance potential was 10 V for BPDE-N²-dG and [¹⁵N]-BPDE-N²-dG; the collision energy was 17 V for BPDE-N²-dG; the collision cell exit potential (CXP) was 24 V for BPDE-N²-dG and [¹⁵N]-BPDE-N²-dG. Other parameters used are: the curtain gas (CUR): 35 L/h, the collision gas (CAD): medium, the ionSpray voltage: 5500 V, the temperature: 400°C, gas 1: 30 L/h, and gas 2: 15 L/h. Multiple reaction monitoring mode (MRM) was used to quantify BPDE-N²-dG as well as its internal standards [¹⁵N]-BPDE-N²-dG with the transitions of m/z 570 → 454 for BPDE-N²-dG, 575 → 459 for [¹⁵N]-BPDE-N²-dG. All peaks were integrated and quantified by Sciex OS 1.5 software.

Creatinine analysis

Using an established assay (23), creatinine levels in the urine were analyzed spectrophotometrically following a reaction between creatinine and alkaline picrate. A volume of 50 µl of each sample was mixed with 200 μl of 0.12% of picric acid in 0.15 N sodium hydroxide in a 96 well plate. Samples were incubated at room temperature for a period of 30 min, followed by the measurement with a microplate reader (Biotek Synergy HTX) at the absorbance at 490 nm. Creatinine concentration in each sample was determined using a creatinine standard curve (0.01–0.1 mg/ml).

Cotinine analysis

The cotinine levels in the urine were determined using a commercial competitive enzyme-linked immunoassay (ELISA) kit (Calbiotech, El Cajon, CA) following the manufacturer’s instruction. Briefly, samples were added in duplicate to wells coated with an antibody to cotinine. Enzyme conjugate was added, and the plates were incubated for an hour at room temperature. All wells were washed. After the addition of a substrate reagent, the plate was further incubated for 20 min. Subsequently, we added the stop solution and measured the absorbance at 450 nm. The concentration of cotinine was then calculated based on the standard curve generated from the manufacture provided standards.

Statistical analysis

This is a cohort design in which each participant is measured at four time points: Baseline (weeks 0 and 1), during the middle of BRB administration (weeks 4 and 5), at the end of BRB administration (weeks 8 and 9), wash out period after BRB cessation (weeks 12 and 13). Each participant (at baseline) serves as his or her own control at later time points for greater statistical power. For all endpoints, their distribution at baseline, mid-BRB, end-BRB, and washout periods was first inspected using boxplots; cube or square root transformation was then applied to each endpoint so that its distribution is more symmetric and closer to a normal distribution. The change of these endpoints at each later period, respectively, from their corresponding value at the baseline was calculated in the transformed scale. The t-test and the Wilcoxon signed-rank test were used to test whether each change deviates from 0 in mean or median.

Sample size calculation

All of the 27 participants enrolled in this Phase 0 clinical trial came from a larger project studying the impact of BRB lozenges on biomarkers in the urine and buccal cells of active smokers. In the original design of this larger project, the tobacco-specific nitrosamine Nʹ-Nitrosonornicotine-releasing adduct (4-hydroxy-1-(3-pyridyl)-1-butanone [HPB]) was proposed as the primary biomarker. A sample size of 47 smokers was required. However, the biomarker examined in the current study (BPDE-N²-dG) is considered a secondary biomarker. Although no power analysis was conducted for the secondary biomarker in the current manuscript, it does not affect the validity of the results (24).

Results

The results of this study on the effects of BRB on buccal cell levels of BPDE-N²-dG and urinary levels of cotinine and creatinine in smokers, as well as the average number of cigarettes per day and age, are provided in Supplementary Table S1. We determined the compliance based on the following equation, and the participants deemed compliant based on a report that if their rates were 80% or more (25): (number of lozenges provided − number of lozenges returned)/(number of lozenges instructed to take per day × number of days until return to the study) = rate of compliance. Based on this approach, the compliance rate was more than 97%.

Analysis of BPDE-N2-dG levels in smokers’ buccal cells

Representative chromatograms of the LC-MS/MS analyses of BPDE-N²-dG and [¹⁵N]-BPDE-N²-dG are shown in Fig. 3. The area of the peak, which co-eluted with the internal standard [¹⁵N]-BPDE-N²-dG (retention time = 3.223 min), was used to quantitate the level of BPDE-N²-dG. The baseline level of BPDE-N²-dG (mean ± SD) in the buccal cells of participants before the consumption of BRB was determined to be 22.44 ± 25.58 BPDE-dG/10⁸ dG (Table 2). We found that at the middle of BRB administration, at the end of BRB administration and at the washout period, the levels of BPDE-N²-dG in buccal cells (16.54 ± 24.92, 14.35 ± 17.74. and 18.72 ± 28.50 BPDE-dG/10⁸ dG, respectively) were significantly lower than baseline (P = .01506, .00355 and .007586, respectively); the median levels at each time point were 14.36 (baseline), 10.63 (middle of BRB administration), 6.74 (end of BRB administration) and 8.96 (washout) BPDE-N²-dG/10⁸ dG, respectively. In Fig. 4, the levels of buccal cell BPDE-N²-dG were plotted after cube root transformation.

Figure 3.

Representative chromatograms of BPDE-N²-dG obtained from stable isotope dilution HPLC-MS/MS analysis of DNA adducts isolated from human oral buccal cells

Table 2.

Levels of BPDE-N²-dG/10⁸dG before, during and after BRB cessation.

Subjects visit	Mean (Median) (BPDE-N 2 -dG/10 8 dG)	SD	P-value (compared with baseline) a
Baseline	22.44 (14.36)	25.58
Middle of BRB	16.54 (10.63)	24.92	.01506
End of BRB	14.35 (6.74)	17.74	.00355
Washout	18.72 (8.96)	28.50	.007586

^aAnalyzed by Wilcoxon signed-rank exact test.

Figure 4.

The effects of BRB on BPDE-N²-dG levels in buccal cells of 27 smokers after cube root transformation. The area of the peak, which co-eluted with the internal standard [¹⁵N]-BPDE-N²-dG (retention time = 3.223 min, cf.Fig. 3), was used to quantify the levels of BPDE-N²-dG. The baseline level of BPDE-N²-dG (mean ± SD) in buccal cells of all participants prior to the consumption of BRB was determined and found to be statistically significant compared to those at the middle of BRB administration (P = .015) at the end of BRB administration (P = .0035) and at the washout period (P = .0075). In this Figure, the levels of buccal cell BPDE-N²-dG were plotted after cube root transformation

The effects of BRB on BPDE-N2-dG levels in buccal cells of the individual subjects

As described above, we showed that BRB inhibited the formation of BPDE-N²-dG adduct in the buccal cells of smokers. However, considering the interindividual differences among smokers regarding the uptake and the metabolic activation of B[a]P leading to the formation of BPDE-N²-dG adduct and the further impact of BRB on Phase I and Phase II and DNA repair enzyme activities (4, 18, 28–30), it was essential to determine BRB effects on each of the 27 smokers enrolled in this study. The results are shown in Fig. 5 (spaghetti plot graph), and as expected, BRB had varied effects among individual subjects; briefly, the levels of BPDE-N²-dG adduct decreased in 21 subjects but increased in 6 subjects.

Figure 5.

The effects of BRB on BPDE-N²-dG levels in buccal cells of the individual participants

The effects of BRB on BPDE-N2-dG levels in buccal cells of males and females

We next examined the effects of BRB on adduct levels among smokers of different genders. As shown in Table 1, we have enrolled 10 males and 17 females in our trial. The results demonstrate that BRB significantly reduced adduct levels in males but not in females (Supplementary Fig. S1). The effect of BRB on racial differences cannot be derived from our study since 96% of smokers are white.

Measurement of urinary levels of cotinine and creatinine

Urinary levels of cotinine ranged from 590 to 20 420 ng per mg creatinine in smokers enrolled in this trial throughout the duration of this study. Cotinine levels were normalized by dividing with the corresponding levels of creatinine. Our results indicate that the consumption of BRB had no effect on urinary cotinine levels of our participants (Fig. 6a), but we found that urinary cotinine levels in each smoker differed across the various time points over the course of this trial (Fig. 6b; spaghetti plot graph). We further examined whether the overall cotinine levels had an influence on how BRB affected adduct levels and found that there are three subjects, among the 27 smokers enrolled in this trial, that had very low levels of cotinine. Thus, by excluding these three subjects in our analysis, we found that BRB remains to significantly reduces the adduct levels in the remaining 24 subjects (Supplementary Fig. S2).

Figure 6.

The effect of BRB on (a) levels of urinary cotinine in smokers (ng/mg creatinine) and (b) on individual smokers. The baseline value represents the levels of urinary cotinine in smokers prior to the administration of BRB

Discussion

The first site of exposure to cigarette smoke constituents (including B[a]P) is the oral cavity, and tobacco smoking is causatively associated with several diseases, including HNSCC (4,5). The oral epithelial cells can be collected noninvasively to assess the extent of DNA damage in smokers. DNA damage is a reliable biomarker for both exposure and cancer risk (26). LC-MS/MS-based methods have been considered superior to others reported in the literature for detection, structural confirmation and accurate quantification of DNA damage (10,27). In a recent study, we used an LC-MS/MS method to detect BPDE-N²-dG adducts in buccal cells, and the levels were significantly higher in smokers than in nonsmokers (17). In a more recent study using LC-MS/MS methods, we showed that 8-hydroxy-2ʹ-deoxyguanosine (8-oxodG) in buccal cells and urine of smokers was significantly reduced by BRB (20). However, the effect of BRB on B[a]P-induced covalent adducts (e.g. BPDE-N²-dG) in buccal cells of smokers has never been reported, and thus this is the focus of the current clinical study.

All of the data presented from this study, including the DNA adducts, the nicotine metabolite (cotinine), and creatinine levels, demonstrated appreciable interpatient variations as manifested by large standard deviations. The cotinine level variations most likely reflect dietary differences, as higher protein consumption is positively associated with cotinine. These findings are consistent with the out-bred human population and the extensive interpatient variability in drug/carcinogen metabolizing enzymes (28). This established inter-participant high variability was the basis for obtaining baseline samples and employing every participant as their own internal control.

Using BRB at the dose of 5 g/day for 8 weeks resulted in a significant reduction of BPDE-N²-dG in smokers’ buccal cells during and at the end of BRB administration; the inhibition of this DNA damage remains after BRB cessation (washout period). This persistence of BRB’s protective effects may reflect increased detoxification capacities in oral keratinocytes following lozenge use. A similar dose of BRB was used in our previous clinical trial (20) as well as in a previous clinical trial reported in the literature (21), and even a much higher dose was used by others (29). Administration of BRB by oral and topical applications in previous clinical trials did not produce any toxicity; however, some patients experienced mild disturbances of the gastrointestinal tract that resolved in a few days (30). To our knowledge, the toxicity of individual constituents identified in BRB has not been reported and considering the relatively low toxicity of BRB (31), its utilization in future trials would appear ideal as a whole-food-based approach for cancer interception/prevention.

Numerous sources of phytochemicals have been proposed for cancer prevention (4,32). Literature data strongly support the notion that the effects of BRB as an inhibitor of Phase I and inducers of Phase II enzymes may, in part, account for its inhibition of DNA damage, mutagenesis and carcinogenesis (4). In fact, in a previous report, we showed that BRB extract enhanced the levels of GSH in oral cancer cells and in oral leukoplakia cells, thereby increasing the detoxification of diol epoxide derived from dibenzo[a,l[pyrene, a potent tobacco carcinogen (33). As the most abundant intracellular nonprotein thiol, GSH is an essential cofactor for the detoxification of reactive electrophilic metabolites, such as those derived from B[a]P that can damage DNA. However, oral administration of GSH directly may not be a suitable approach due to its poor bioavailability and short half-life (two minutes) in the blood (34). Thus, the administration of BRB provides a plausible approach to increasing cellular levels of GSH. It appears that anthocyanins are the primary components that can account for the chemopreventive efficacy of BRB (4). Anthocyanins are known to induce Phase II enzymes in cultured cells by activating glutathione-related enzymes (glutathione S-transferase, glutathione reductase, glutathione peroxidase and glutamate-cysteine ligase) (35,36), which is consistent with our previous finding demonstrating that BRB increased the protein expression of glutamate-cysteine ligase catalytic subunit (GCLC) (33). In addition, administration of dietary BRB has also been reported to induce glutathione S-transferase activity in the rat (37).

Cells have developed elaborate DNA repair machinery to cope with DNA damage. Previous studies that examined the effects of BRB on DNA repair are rather limited (38,39). Previously, we employed an established DNA repair model system, and our results showed that pretreatment of HeLa cells with BRB extract enhanced DNA repair capacity of two structurally different nucleotide excision repair substrates, including the BPDE-N²-dG adduct (40). TS and environmental pollutants such as B[a]P can lead to the formation of bulky adducts (e.g. BPDE-N²-dG) and if not repaired can lead to mutations that often drive malignant transformation.

More than 70% of nicotine is metabolically converted into cotinine, and the latter is the most widely used biomarker of TS exposure (41). The ratio of urinary cotinine (30–550 ng) per mg creatinine has been shown to differentiate smokers from nonsmokers, and in the current study, the levels of cotinine detected in our participants are consistent with those reported in smokers (42–44). In the present study, BRB consumption had no effect on urinary cotinine levels and likely on the amount of nicotine uptake by our participants.

BRB demonstrates numerous cancer-preventive effects, including its role as a powerful scavenger of reactive oxygen species and an inhibitor of biomarkers of oxidative stress (8-oxodG) in buccal cells, as reported by us (20). Our results show an oral cavity-directed formulation (lozenge) demonstrated positive chemopreventive effects in the mouth. BRB use as a whole food additive to reduce levels of other forms of DNA damage (e.g. BPDE-N²-dG), as shown in the present study, could benefit high-risk populations such as smokers who are unable to quit and nonsmokers or former smokers who are exposed to environmental carcinogens. Systemic administration, however, is not likely to result in the same high intraoral BRB levels achieved by lozenge use. The results of the present study cannot be widely extrapolated to predict the effects of BRB on DNA damage (BPDE-N²-dG) among smokers of different racial backgrounds; this is a limitation of our study. However, we showed that BRB significantly reduced adduct levels in males. It did not exhibit an effect in females. The incidence and mortality of HNSCC are higher in men than women, and Black/African Americans have a higher incidence of advanced disease and poorer survival than Whites (45–47). Therefore, it is plausible that the dose of BRB required to inhibit the various DNA damages in the buccal cells of smokers may vary drastically based on gender and race. Future clinical trials that incorporate these key demographics should provide important insights to answer some of these limitations.

Author contributions

K.-M.C.: Formal analysis, investigation, methodology, writing—original draft. Y.-W.S.: Formal analysis, investigation, methodology. N.M.K.: Resources, project administration, writing—review and editing. L.R.: Formal analysis, methodology, writing—review and editing. D.S.: Formal analysis. J.L.: Conceptualization, software, formal analysis, funding acquisition, methodology, writing—original draft, project administration, writing—review and editing. J.K. and S.A.: Synthesis and purification of the DNA adduct derived from benzo[a]pyrene. G.S.: Conceptualization, writing—original and editing. S.R.M.: Conceptualization, formal analysis, funding acquisition, methodology, writing—original draft, project administration, writing—review and editing. K.E.-B.: Conceptualization, software, formal analysis, funding acquisition, methodology, writing—original draft, project administration, writing—review and editing.

Ethical statement

The Ethics Committee approval and study protocols were obtained from the Pennsylvania State University Investigational Review Board (IRB#13621).

Conflict of interest

None declared.

Funding

This work was supported by the National Cancer Institute (CA173465).

Data availability

The data generated in this Phase 0 Clinical Trial are available within the contents of this article and its Supplementary Table S1 and Supplementary Figs S1 and S2.