UGT2B10 genotype influences nicotine glucuronidation, oxidation and consumption

Abstract

Background

Tobacco exposure is routinely assessed by quantifying nicotine metabolites in plasma or urine. On average, 80% of nicotine undergoes C-oxidation to cotinine. However, interindividual variation in nicotine glucuronidation is substantial and glucuronidation accounts for from 0 to 40% of total nicotine metabolism. We report here the effect of a polymorphism in a UDP-glucuronsyl transferase, UGT2B10, on nicotine metabolism and consumption.

Methods

Nicotine, cotinine, their N-glucuronide conjugates, and total trans-3'-hydroxycotinine were quantified in the urine (n=327) and plasma (n =115) of smokers. Urinary nicotine N-oxide was quantified in 105 smokers. Nicotine equivalents, the sum of nicotine and all major metabolites, were calculated for each smoker. The relationship of the UGT2B10 Asp67Tyr allele to nicotine equivalents, N-glucuronidation, and C-oxidation was determined.

Results

Individuals heterozygous for the Asp67Tyr allele excreted less nicotine or cotinine as their glucuronide conjugates than wild-type, resulting in a 60% lower ratio of cotinine glucuronide:cotinine, a 50% lower ratio of nicotine glucuronide:nicotine and increased cotinine and trans-3'-hydroxycotinine. Nicotine equivalents, a robust biomarker of nicotine intake, were lower among Asp67Tyr heterozygotes compared to individuals without this allele; 58.2 nmol/ml (95% CI, 48.9 – 68.2) versus 69.2 nmol/ml (95% CI, 64.3 – 74.5).

Conclusions

Individuals heterozygous for UGT2B10 Asp67Tyr consume less nicotine than do wild type smokers. This striking observation suggests that variations in nicotine N-glucuronidation, as reported for nicotine C-oxidation, may influence smoking behavior.

Impact

UGT2B10 genotype influences nicotine metabolism and should be taken into account when characterizing the role of nicotine metabolism on smoking.

Introduction

Tobacco use is associated with cancers of the lung, larynx, nasal cavity, oral cavity, esophagus, liver, pancreas, bladder, cervix, and leukemia (1). Therefore, cancer etiology and epidemiology studies frequently require assessment of tobacco exposure (2–5). Error in estimating tobacco exposure can obscure important observations, while accuracy in assessment is useful to identify tobacco exposure as either a causal or confounding factor. Many approaches have been used to evaluate tobacco exposure broadly including: self-report of smoking history (cigarettes/day, pack-years of smoking, time spent around smoke, etc.), collection of used products (cigarette butts, discarded chew), topography measures (puff volume, frequency), and biomarkers (exhaled carbon monoxide, nicotine metabolites, tobacco specific nitrosamines) (6–11). Selecting a measure(s) to assess exposure is not a trivial choice and depends on the goals of a particular study.

Quantifying one or more nicotine metabolites is an objective measure of current exposure from smoking, smokeless tobacco use, or environmental/secondhand smoke. Plasma or urinary cotinine are the most frequently utilized biomarker of tobacco exposure, other than exhaled carbon monoxide (12,13). The only significant source of cotinine is metabolism from nicotine, and it is highly correlated with nicotine intake (14,15). However, interindividual differences in nicotine metabolism also influence cotinine levels (15,16). The role of ethnicity in interindividual variation is addressed in a recent publication that recommends the use of racial/ethnic specific plasma cotinine levels to distinguish smokers from non-smokers (13). A potentially more robust biomarker of nicotine exposure is nicotine equivalents, defined as the sum of nicotine and its major metabolites in urine (8). Total nicotine (nicotine + nicotine glucuronide), total cotinine (cotinine + cotinine glucuronide) and total trans-3'- hydroxycotinine (trans-3'-hydroxycotinine + its glucuronides) can be assessed with minimal sample preparation (e.g. a single solid phase extraction) by liquid chromatography/mass spectrometry with high throughput capacity (LC/MS) (17). Therefore, assessing nicotine equivalents is feasible in the time typically used to quantify a single nicotine metabolite. Nicotine equivalents directly account for over 80 % of nicotine intake, and since it is a sum of metabolites its value will not change if metabolism shifts from one pathway to another (14,17,18). Nicotine equivalents correlate with cigarettes per day, and have been used to better characterize the link between polymorphisms in nicotinic acetylcholine receptor A subunits, smoking and lung cancer (8,19).

In smokers, nicotine is metabolized by three pathways (Figure 1). The major pathway is CYP2A6-catalyzed 5'-oxidation, which after a second oxidation results in the formation of cotinine. Nicotine may also be metabolized by N-glucuronidation and N-oxidation. On average these two pathways each account for less than 10 % of the nicotine dose (15). However, urinary nicotine N-glucuronide concentrations vary significantly among smokers and can account for as much as 40 % of the total nicotine metabolites excreted (6).

Figure 1.

Nicotine metabolism pathways

Polymorphisms are common in nicotine metabolism genes and their regulatory regions, although only a minority of the known variants have been fully characterized (20). Enzyme amount and relative activity affect the metabolic fate of nicotine. Several polymorphisms in CYP2A6, the hepatic enzyme responsible for oxidation of nicotine to cotinine and a major catalyst of cotinine oxidation to trans-3'-hydroxycotinine, are associated with lower intensity smoking (21–23). For example, CYP2A6 polymorphisms that influence smoking behavior include CYP2A6*4, a deletion allele, CYP2A6*2, an inactivating point mutation, and CYP2A6*9 which alters the TATA box promoter region (21,24,25). Furthermore, low nicotine oxidation, low concentrations of the tobacco carcinogen NNAL and low nicotine equivalents have been observed among Japanese Americans, a population with a high prevalence of CYP2A6 deletion alleles, compared to European Americans (26). UGT2B10 and UGT1A4 catalyze the N-glucuronidation pathways of nicotine metabolism in vitro and the genes encoding these enzymes are also polymorphic (27–30). UGT2B10 is a more efficient catalyst of nicotine N-glucuronidation than is UGT1A4, and it appears to be the key catalyst in vivo (28,31). Recently, we reported that the UGT2B10 polymorphism Asp67Tyr was associated with a 20 % decrease in the excretion of cotinine and nicotine as their glucuronide conjugates in urine of 84 smokers (31). This UGT2B10 variant was characterized in vitro by Chen et al. as having decreased nicotine and cotinine glucuronidation activity (27).

The main goals of the current study were to confirm the relationship between Asp67Tyr and lower nicotine glucuronidation, and to assess if glucuronidation was associated with a shift in nicotine metabolism through other pathways including nicotine-N-oxidation. In addition, the effect of the Asp67Tyr polymorphism on nicotine equivalents as a pathway-independent measure of tobacco exposure was determined. The measure of nicotine equivalents does not routinely include nicotine N-oxide. However, in the present study LC/MS/MS analysis was used to quantify nicotine N-oxide in a subset of subjects.

Methods

Study populations

This research was approved by the University of Minnesota’s Institutional Review Board. A total of 327 adult smokers who were recruited from the Twin Cities metropolitan area and who participated in one of three tobacco research studies conducted at the University of Minnesota were included in this genotype-phenotype investigation. Participants were recruited through advertisements on campus and local media, were screened by telephone, and were willing to participate in a smoking reduction or cessation study. Participants were in good physical and psychiatric health, had no contraindications to nicotine replacement therapy, were not pregnant or nursing, and were not using barbituates or anti-convulsants. For each study, participants were asked to provide a first void morning urine sample while they were smoking as usual, and this sample was collected at least two weeks prior to any study intervention (and before randomization to an intervention group). A blood sample was also collected from participants. Study 1 included 122 participants aged 18–65 who smoked > 10 cigarettes per day and who subsequently participated in a cessation trial of traditional nicotine replacement versus smokeless tobacco. Study 2 included 145 participants aged 18 – 70 who were smokers of 10 – 40 “light” cigarettes (0.7–1.0 mg nicotine/cigarette) per day and who then participated in a study of low and no nicotine cigarettes (19). Study 3 included 84 smokers aged 18 – 70 years who smoked 15 – 45 cigarettes/day and subsequently participated in a smoking reduction trial (6,31). Initial genotype-phenotype analysis of UGT2B10 Asp67Tyr genotype and the percent of nicotine and cotinine excreted as glucuronide conjugates in urine was conducted on study 3 samples and was recently reported (31).

Nicotine and Metabolite analysis

Urine was aliquoted and stored at − 20°C until the time of analysis. Urinary nicotine and nicotine metabolite concentrations were analyzed by GC/MS and were previously reported (6). Free nicotine, total nicotine (free nicotine + N-glucuronide), free cotinine, and total cotinine (free cotinine + N-glucuronide) and total trans-3'-hydroxycotinine (trans-3'-hydroxycotinine + its glucuronides) were quantified. Analyses were performed in duplicate for each urine sample and repeated if values differed by > 10 %, and average duplicate error was 3 %. Repeat analysis was required on fewer than 10 % of the samples; no samples were excluded due to measurement error. Nicotine-N-oxide was analyzed in urine of study 1 samples by LC/MS. In brief, urine (20 µl) with added [d₃-methyl]-nicotine N-oxide (Toronto Research Chemicals (North York ON, Canada) as an internal standard was processed by solid phase extraction with an Oasis MCX column eluted with 400 µl 2 % ammonium hydroxide in methanol. LC/MS/MS was performed using a Finnigan Discovery triple quadrupole mass spectrometer (Thermo Electron, San Jose, CA) with an ESI source in positive ion mode. Selective reaction monitoring was performed of the mass transitions for d0-nicotine N-oxide (m/z 179→130 and m/z 179→117).

Plasma was separated from blood and stored at −20°C for study 1 samples. Plasma free cotinine, total cotinine (free + N-glucuronide), and total trans-3'-hydroxycotinine were quantified by GC/MS as performed for urinary metabolites except for the addition of an initial solid phase extraction using an Oasis MCX column (Waters Corporation, Milford, MA) (32). Analyses were performed in duplicate and repeated if values differed by > 10 %.

UGT2B10 Genotyping

DNA was isolated from blood using the DNeasy kit (Qiagen, Valencia, CA). PCR-restriction fragment length polymorphism (RFLP) analysis was performed targeting UGT2B10 at the codon 67 position with HinfI digestion as previously described (Chen et al 2008). A second RFLP for the SNP rs7657958 that is linked with UGT2B10 Asp67Tyr in Caucasians (haplotype C) was performed in a subset of samples (N = 107, 32 %) (29). Genotyping results were concordant between the two assays, as well as with DNA sequence analysis of 10 samples.

Statistical analysis

Statistical analyses were conducted using SAS (SAS Institute, Cary, NC) and Excel (Microsoft, Redmond, WA). Wilcoxon two-sided t-approximation statistics were calculated to compare means of continuous variables that did not have a normal distribution, and a p-value less than 0.05 was considered significant. Spearman nonparametric correlation was used to assess univariate correlations. A general linear model was used to compare phenotypes by UGT2B10 genotype, and to adjust for nicotine equivalents. Individuals with the highest and lowest 1 % of urinary nicotine equivalents were excluded from analyses, < 9 nmol/ml or > 236 nmol/ml. Nicotine and cotinine N-glucuronidation were assessed as the percent of the analyte that was present as its N-glucuronide and as the square-root transformed ratio of glucuronide conjugate to free analyte. The square-root transformed ratio of total trans-3'-hydroxycotinine to free cotinine in plasma or total trans-3'-hydroxycotinine to free cotinine in urine was used as an estimate of C-oxidation.

Results

A total of 327 participants were included in the genotype-phenotype analyses, which included samples collected at baseline while participants were smoking as usual from 3 studies. Study participants were 87 % European American, 8 % African American and the remainder were Asian, American Indian or mixed race. The mean ages +/− SD by study were 1) 43 +/− 11.6 years; 2) 41 +/− 14 years; and 3) 46 +/− 10.6 years. Diary recorded cigarettes per day at baseline were mean +/ SD: 1) 21 +/− 6.5; 2) 21 +/− 8.8; 3) 26 +/− 6.9. No effect of UGT2B10 Asp67Tyr genotype on cigarettes per day was observed. The mean +/− SD (wild type, Asp67Tyr heterozygotes) by study were 1) 20.2 +/− 6.1, 22.8 +/− 7.7; 2) 20.7 +/− 8.7, 22.3 +/− 9.1; 3) 26.2 +/− 7.1, 25.6 +/− 7.0. Total allele frequency of the UGT2B10 Asp67Tyr variant was 10.6 %, and this was consistent with Hardy-Weinberg equilibrium. Correspondingly individuals who were heterozygous for the Asp67Tyr allele represented 20 % of participants: study 1, 17.4 %; study 2, 18.3 %; and study 3, 22.2 %. The percent of cotinine excreted as a glucuronide conjugate in first morning urine was significantly decreased among individuals who were heterozygous for the Asp67Tyr allele compared to individuals without this allele in each of the three studies. Subsequently, analyses were conducted on the combined data unless indicated.

The distribution of urinary nicotine and its metabolites by UGT2B10 Asp67Tyr genotype is presented in Table 1. Analytes are presented as molar percentages of the sum of nicotine and its metabolites in urine, and reflect the relative contribution of different pathways to the fate of nicotine. The molar percentages of nicotine and cotinine N-glucuronides in urine were 30 % lower for Asp67Tyr heterozygous individuals than for wild-type, 4.2 % versus 6.1 % (p = 0.003) and 17.6 % versus 24.2 % (p < 0.0001) respectively. Similarly, the fractions of nicotine and cotinine excreted as their glucuronide conjugates, ie. percent glucuronidation, were lower among Asp67Tyr heterozygotes than wild-type; 29.5 % versus 38.2 % (p = 0.007) and 51.3 % versus 62 % (p < 0.0001). As might be expected, a decrease in the molar percent of cotinine glucuronide was accompanied by an increase in excretion of free cotinine, 16 % versus 13.4 % (p = 0.003) for Asp67Tyr heterozygotes compared to wild-type. Moreover, a significant increase was observed in the molar percent of total trans-3'-hydroxycotinine in urine of Asp67Tyr heterozygotes compared to wild-type, 51.1 % versus 44.3 %, p = 0.001. In summary, the Asp67Tyr allele was associated with relatively lower urinary abundance of nicotine and cotinine N-glucuronides, higher free cotinine, and higher total trans-3'-hydroxycotinine.

Table 1.

Distribution of urinary metabolites by genotype as a molar percentage of nicotine and its metabolites^*

Analyte	Wild-type Mean percent (SD) (n = 263)	UGT2B10 Asp67Tyr† Mean percent (SD) (n=62)	p-value
Nicotine	11.8 (9.5)	11.8 (8.5)	0.75
Nicotine glucuronide	6.1 (5.1)	4.2 (3.3)	0.003
Cotinine	13.4 (6.8)	16.0 (7.4)	0.003
Cotinine glucuronide	24.2 (10.7)	17.6 (8.0)	< 0.0001
Trans-3'-hydroxycotinine	44.3 (16.7)	51.1 (14.6)	0.001
Percent nicotine glucuronidation‡	38.2 (23.6)	29.5 (18.8)	0.007
Percent cotinine glucuronidation‡	62.8 (18.2)	51.3 (18.5)	< 0.0001

^*Molar percentage = (analyte / Σ nicotine + cotinine + trans-3'-hydroxycotinine + glucuronides)*100.

^†UGT2B10 Asp67Tyr heterozygous individuals

^‡Percent of analyte excreted as its glucuronide conjugate

There were 3 individuals whose DNA yielded a restriction digest pattern that would be consistent with homozygosity for UGT2B10 Asp67Tyr. No urine samples were available for one individual who appeared homozygous, another individual excreted a low percentage of cotinine as its glucuronide conjugate (7.1 %) and had undetectable nicotine glucuronidation, and the third individual had a normal-to-high glucuronidation phenotype (73 % cotinine glucuronidation, 67 % nicotine glucuronidation). DNA sequence analysis of UGT2B10 is planned for these samples, and at this time they were excluded from the analyses which were focused on characterizing heterozygous carriers of UGT2B10 Asp67Tyr.

In addition to C-oxidation and N-glucuronidation, nicotine may be metabolized by N-oxidation. The product of N-oxidation, nicotine-N-oxide, though typically a minor metabolite, may be increased in individuals with decreased metabolism through other pathways (33). Therefore, we evaluated the influence of UGT2B10 genotype and N-glucuronidation and C-oxidation phenotype on nicotine-N-oxide abundance in urine. Urinary nicotine-N-oxide was quantified in study 1 participants (N = 108). The concentration of nicotine N-oxide excreted by smokers who were heterozygous or wild type for the UGT2B10 Asp67Tyr allele was not significantly different, 4.1 versus 3.8 nmol/ml, respectively (p=.34). The molar percent of nicotine N-oxide excreted by these two groups was also not different, 6.27 ± 2.83 (range 1.11 to 12.7) and 5.53 ± 3.16 (range 0.76 – 11.8) respectively. Likewise, the excretion of nicotine-N-oxide did not differ significantly by N-glucuronidation phenotype, the urinary ratio of cotinine glucuronide to free cotinine (Table 2). The phenotype was categorized as low, high, or average, based on the lowest and highest quartiles of the metabolite ratio distribution and the interquartile range. The mean nicotine-N-oxide abundances were 5.5 % , 6.1 % and 6.8 % for high, average, and low N-glucuronidation phenotype, respectively; and the Chi-square test for trend was p = 0.055. In contrast, the molar percent of nicotine N-oxide excreted increased as the ratio of plasma trans-3'-hydroxycotinine to free cotinine ratio decreased (Table 2). That is, nicotine N-oxide excretion was significantly related to C-oxidation phenotype. The abundance of nicotine-N-oxide increased with successively lower trans-3'-hydroxycotinine:free cotinine ratios, means: 4.3 %, 6.1 %, and 8.7% for high, average, and low C-oxidation phenotype groups respectively (p < 0.005 between each group).

Table 2.

Nicotine-N-oxide as a molar percentage of nicotine and its metabolites in urine by C-oxidation and glucuronidation phenotype (Study 1)

Phenotype*	Quartile ratio values	N	Molar % of Nicotine-N-oxide Mean (95 % CI)
Cotinine glucuronide/cotinine†
Low	< 0.89	27	6.8 (5.7 – 7.9)
Average	0.89 – 1.5	54	6.1 (5.4 – 6.9)
High	> 1.5	27	5.5 (4.4 – 6.6)
Trans-3'-hydroxycotinine/cotinine‡
Low	< 0.45	25	8.7 (7.7 – 9.7)
Average	0.45 – 0.69	52	6.1 (5.4 – 6.8)
High	> 0.69	25	4.3 (3.3 – 5.3)

^*Phenotype was assessed by metabolite ratios: low = lowest quartile, average = interquartile range, high = highest quartile.

^†Cotinine glucuronide:free cotinine quantified in first morning urine. Nicotine-N-oxide means did not differ by phenotype. Chi-square test for trend: p-value 0.055.

^‡Ratio of total trans-3'-hydroxycotinine: free cotinine quantified in plasma. Mean nicotine-N-oxide values differed significantly between each phenotypic group with p-values < 0.005.

The effect of the UGT2B10 Asp67Tyr allele on plasma concentrations of nicotine metabolites was evaluated in study 1 participants (Table 3). Free plasma cotinine was higher among Asp67Tyr heterozygotes than wild-type individuals, 1.31 nmol/ml (231 ng/ml) versus 1.09 nmol/ml (192 ng/ml), p = 0.03. No difference in plasma cotinine glucuronide concentration was detected, 0.13 nnmol/ml versus 0.15 nmol/ml, for Asp67Tyr heterozygotes and for wild-type individuals respectively. Since the percent of conjugated cotinine present in plasma is low, the overall mean was 10 %, it was difficult to quantify relatively small differences in cotinine N-glucuronide in plasma. Total trans-3'-hydroxycotinine was higher in the plasma of Asp67Tyr heterozygotes compared to wild-type, 0.46 nmol/ml versus 0.39 nmol/ml, p = 0.04. Therefore, the effect of the Asp67Tyr allele on free cotinine and total trans-3'-hydroxycotinine was consistent in urine and plasma; both were higher among heterozygotes than wild-type.

Table 3.

Plasma nicotine metabolites by UGT2B10 genotype (Study 1)

Analytes	Wild-type (n = 95) Mean (SD)	Asp67Tyr*(n = 20) Mean (SD)	Range
Free cotinine (nmol/ml)	1.09 (0.49)	1.31 (0.40) †	0.19 – 2.75
Cotinine glucuronide (nmol/ml)	0.15 (0.14)	0.13 (0.14)	0.00 – 0.54
Percent cotinine glucuronide	11.3 (9.7)	8.72 (9.0)	0.0 – 36.4
Total Trans-3'-hydroxycotinine (nmol/ml)	0.39 (0.32)	0.46 (0.24) ‡	0.04 – 1.92
Total Trans-3'-hydroxycotinine: free cotinine	0.38 (0.28)	0.37 (0.18)	0.07 – 1.29

^*UGT2B10 heterozygous individuals

^†p = 0.03

^‡p = 0.04

The relationship between Asp67Tyr genotype and phenotype assessed as metabolite ratios is presented in Table 4. UGT2B10 Asp67Tyr genotype and N-glucuronidation ratios, as a biomarker of phenotype, were strongly associated. The urinary ratio of cotinine glucuronide to cotinine was significantly lower among Asp67Tyr heterozygotes than wild-type individuals, mean (95 % CI): 1.26 (1.00 – 1.56) versus 2.03 (1.86 – 2.20), p < 0.0001. The ratio of urinary nicotine glucuronide to nicotine was also significantly lower among Asp67Tyr heterozygotes than wild-type individuals (p 0.005). The metabolite ratios in Table 4 were adjusted for nicotine equivalents. However, the relationship between Asp67Tyr and low N-glucuronidation was evident whether or not the glucuronide ratio was adjusted for nicotine equivalents. No statistical difference was observed in the urinary ratio of total trans-3'-hydroxycotinine:free cotinine, between Asp67Tyr variant and wild type individuals. A higher ratio of total trans-3'- hydroxycotinine:total cotinine was observed for Asp67Tyr heterozygotes mean (95 % CI): 1.73 (1.46 – 2.03) versus 1.29 (1.18 – 1.41), for wild type (p = 0.048).

Table 4.

Urinary nicotine equivalents and metabolite ratios by UGT2B10 genotype: geometric means and 95 % confidence intervals

UGT2B10 Genotype	Cotinine glucuronidation ratio*	Nicotine glucuronidation ratio*	Total trans-3'-hydroxycotinine: free cotinine†	Nicotine equivalents‡
Wild-type (N = 264)	2.03 (1.86 – 2.20)	0.77 (0.64 – 0.89)	3.80 (3.96 – 4.15)	69.2 (64.3 – 74.5)
Asp67Tyrδ(N = 63)	1.26 (1.00 – 1.56)	0.41 (0.25 – 0.62)	3.96 (3.35 – 4.64)	58.2 (48.9 – 68.2)
	p < 0.0001§	p = 0.005§	p = 0.715§	p = 0.048

^*Glucuronide conjugate / unconjugated analyte

^†Total trans-3'-hydroxycotinine / total cotinine was higher among Asp67Tyr heterozygotes compared to wild-type: 1.73 (1.46 – 2.03) versus 1.29 (1.18 – 1.41), p = 0.003, adjusted for nicotine equivalents

^‡Sum of nicotine, cotinine, trans-3'-hydroxycotinine, and their respective glucuronides in nmol/ml, quantified in first morning urine. Urinary nicotine-N-oxide was quantified in a subset of individuals (n = 108, study 1) and did not differ by genotype: mean, 4.03 +/− 2.8 nmol/ml; range, 0.08 – 14.5 nmol/ml.

^δUGT2B10 Asp67Tyr heterozygous individuals

^§Adjusted for nicotine equivalents

To investigate the stability of the urinary glucuronide ratio and the effect of nicotine dose, we evaluated the ratio among individuals who decreased their smoking. The ratio of cotinine glucuronide:cotinine was assessed in a group of individuals (n= 27) at baseline and after 12 weeks at which point they had reduced their nicotine intake by 60 %, as assessed by nicotine equivalents. The ratio of cotinine glucuronide:cotinine decreased by 40 % as they decreased their nicotine intake, mean (95 % CI): 2.59 (2.05 – 3.24) at baseline versus 1.51 (1.04 – 2.10) at week 12, p = 0.008. However, when the ratio was adjusted for nicotine equivalents this difference disappeared, mean (95 % CI): 1.99 (1.49 – 2.52) and 2.07 (1.58 – 2.63) for baseline and week 12 respectively, p = 0.81. A decrease of similar magnitude was observed in the urinary ratio of total trans-3'-hydroxycotinine: free cotinine as individuals decreased their nicotine intake, and adjusting for nicotine equivalents eliminated this difference.

During the evaluation of urinary nicotine metabolites, it became clear that nicotine equivalents were lower among Asp67Tyr individuals than for wild-type (Table 4). Nicotine equivalents were decreased 15 % for Asp67Tyr heterozygotes relative to wild type, mean (95 % CI): 58.2 nmol/ml (48.9 – 68.2) versus 69.2 nmol/ml (64.3 – 74.5) respectively, p = 0.048. Nicotine equivalents were also lower among Asp67Tyr individuals compared to wild-type when a nested comparison was used to take into account the effect of study group, and there was no difference in nicotine equivalents for Asp67Tyr individuals by study. Nicotine equivalents accounts for greater than 80 % of nicotine intake and therefore this biomarker is useful to estimate nicotine intake during ad libitum smoking. The only significant pathway of nicotine metabolism not accounted for by nicotine equivalents is N-oxidation. However, as discussed above the excretion of nicotine N-oxide was not influenced by N-glucuronidation phenotype (Table 2). Moreover, we demonstrated in study 1 that there is no substantial shift to nicotine-N-oxidation among Asp67Tyr heterozygotes.

Discussion

Interindividual variation in glucuronidation of nicotine and cotinine is substantial and the contribution of glucuronidation to total nicotine metabolism ranges from < 1 – 40 % based on nicotine and its metabolites recovered in smokers’ urine (6). On average, an estimated 25 % of nicotine is metabolized to N-glucuronide conjugates of nicotine or cotinine (15). Variation in N-glucuronidation influences the distribution of nicotine metabolites and therefore may affect nicotine exposure assessment using nicotine metabolite biomarkers. We report here that smokers who are heterozygous for the Asp67Tyr allele compared to individuals without this allele have decreased nicotine glucuronidation, increased nicotine C-oxidation and decreased nicotine consumption as measured by excreted nicotine equivalents, a robust biomarker of nicotine intake. The observed increase in nicotine glucuronidation is consistent with our previous report and predicted based on in vitro studies (27,31). The parallel increase in nicotine C-oxidation as measured by significant increases in the levels of both free cotinine and total trans-3'-hydroxycotinine, in plasma and urine suggests that nicotine glucuronidation may play a more important role in circulating nicotine and cotinine levels then previously appreciated. The observed decrease in total nicotine consumption is consistent with an influence of decreased glucuronidation on circulating nicotine levels. The striking observation that smokers who are UGT2B10 Asp67Tyr heterozygotes consume less nicotine suggests that variations in nicotine N-glucuronidation, as with C-oxidation (21), may influence nicotine consumption by smokers.

UGT2B10 is the most efficient catalyst of nicotine and cotinine N-glucuronidation in vitro followed by UGT1A4 (28,34). Both UGT2B10 and UGT1A4 are polymorphic, and for UGT2B10 there are 3 missense single nucleotide polymorphisms (SNPs) and over 100 synonymous SNPs reported in the NCBI SNP database. The UGT2B10 Asp67Tyr allele was identified by Chen et al. because it was linked to a haplotype that was associated with decreased N-glucuronidation of the tobacco carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanal (NNAL) by human liver microsomes (29). Subsequently, a 20 – 30 % decrease in nicotine and cotinine glucuronidation was observed among human liver microsomes that were heterozygous for the Asp67Tyr allele compared to wild-type (27).

In this study we corroborated our prior finding that UGT2B10 Asp67Tyr is associated with decreased nicotine and cotinine N-glucuronidation in vivo (31). Significantly less nicotine and cotinine were excreted as their glucuronide conjugates by individuals who were heterozygous for the Asp67Tyr allele compared to wild-type; the mean values were 22 % and 17 % lower for the Asp67Tyr group, respectively. In addition, UGT2B10 genotype clearly differentiated glucuronidation phenotypes, assessed as the ratio of urinary cotinine glucuronide to cotinine or nicotine glucuronide to nicotine. The glucuronide ratio for cotinine was 63 % lower and for nicotine was 53 % lower for Asp67Tyr heterozygotes compared to wild-type (Table 4).

The distribution of urinary metabolites was distinct for individuals with the Asp67Tyr allele compared to wild-type as these individuals had a lower fraction of nicotine and cotinine N-glucuronides, but higher free cotinine and trans-3'-hydroxycotinine. Hence, low glucuronidation phenotype is compensated by an increase in the clearance of cotinine through C-oxidation. However, there was no significant increase in nicotine N-oxidation, the 3^rd nicotine metabolism pathway (Figure 1). Higher plasma concentrations of free cotinine and trans-3'-hydroxycotinine were also observed in Asp67Tyr heterozygotes. Overall, we observed that the Asp67Tyr allele affects concentrations of both plasma and urinary nicotine metabolites and that the effect is consistent with a decrease in N-glucuronidation.

The dynamic nature of metabolite pathways may influence both biomarkers of nicotine exposure and phenotypic measures of nicotine metabolism. Specifically, when consuming similar amounts of nicotine, smokers with decreased nicotine glucuronidation will typically have higher plasma cotinine concentrations relative to smokers with average levels of glucuronidation. Therefore, UGT activity is one of several factors contributing to plasma cotinine levels. However, since both total trans-3'-hydroxycotinine and free cotinine increased when N-glucuronidation decreased, the C-oxidation ratio (total:free) was not significantly affected by UGT activity and variations in this ratio likely still reflect variations in P450 2A6 activity.

We report here that urinary ratios of nicotine metabolites, both the ratio of cotinine glucuronide:cotinine and trans-3'-hydroxycotinine:cotinine (total:total or total:free), are influenced by nicotine dose. Among individuals who significantly reduced their nicotine intake in a 12 week smoking reduction study (study 3), urinary metabolite ratios were significantly lower when the individuals were smoking less. This difference disappeared after adjusting for nicotine equivalents. Therefore, nicotine equivalents used as a measure of nicotine dose allow one to compare nicotine metabolism phenotypes across different levels of tobacco consumption. In contrast, adjusting for nicotine equivalents had no effect or actually increased the strength of the reported associations between UGT2B10 genotype and nicotine metabolite ratios (Table 4).

The most significant finding of this study was that individuals who were heterozygous for the Asp67Tyr allele had lower nicotine equivalents. When stratified by UGT2B10 genotype, nicotine equivalents were 58.2 among heterozygotes compared to 69.2 nmol/ml for individuals who did not have an Asp67Tyr allele (p < 0.05). The only pathway of nicotine metabolism not routinely included in the measure of nicotine equivalents is nicotine N-oxidation (8). We saw no difference in the level of nicotine N-oxide excreted by Asp67Tyr UGT2B10 heterozygotes compared to wild type smokers. Since the study participants were smoking ad libitum the observed difference in nicotine equivalents represents a moderate but significant decrease in nicotine consumption that is likely attributable to the Asp67Tyr polymorphism in UGT2B10. Nicotine equivalents were 84 % of that for wild-type, a change that is similar in magnitude to what has been reported for the influence of CYP2A6 genotype on smoking. In two recent relatively large studies of smoking and CYP2A6 genotype, the percent decreases in reported cigarettes per day was 24 % (35) and 17 % (36) in individuals defined as slow nicotine inactivators. This is the first report demonstrating that a polymorphism in UGT2B10 may influence smoking behavior. In different populations, the frequency of polymorphisms in CYP2A6 and in UGT2B10 may influence the relative contribution of these alleles to variation in nicotine metabolism and potentially to tobacco consumption.