Tobacco use is the largest cause of preventable mortality in the U.S. and worldwide. Given the exceptional public health consequences of cigarette smoking, improved success quitting can have a huge positive impact on longevity and quality of life for millions of people. Altered nicotine metabolism, including natural variation in human nicotine-metabolism genes, is associated with differences in smoking-related behaviors. Specifically, we have shown that variation in the FMO3 gene, related to the nicotine N-oxidation metabolism pathway, is associated with differences in the levels of nicotine dependence among smokers. But the mechanism of FMO3’s effect on nicotine dependence is not understood.
The main objective of this project was to explore altered nicotine N-oxidation in a mouse model of nicotine-related phenotypes including nicotine withdrawal. We attempted to manipulate nicotine N-oxidation and nicotine-N-oxide levels in mice by three methods: 1) using an FMO-gene knockout mouse, 2) exposing mice to 3,3’-diindolymethane (DIM), and 3) by administering exogenous nicotine-N-oxide. DIM is a natural derivative of cruciferous vegetables that has been shown to alter FMO activity in humans and rats. However, we found that DIM’s effect on nicotine metabolism is very different in mice than in rats. In rats DIM drastically reduces nicotine N-oxidation without altering another nicotine metabolism pathway, nicotine C-oxidation, but in mouse, DIM did not affect N-oxidation while increasing C-oxidation. This was true in both male and female mice from multiple mouse strains. Because differences in nicotine C-oxidation also impact cigarette consumption in smokers, we chose to pursue further experiments in DIM-fed mice despite this unexpected result.
Nicotine withdrawal was measured in both male and female mice of different strains under different experimental conditions using a variety of phenotypes, including changes in respiration, somatic signs (tremor, paw shaking, etc.), pain sensitivity, locomotion, anxiety, and conditioned place preference. Initial experiments were also performed under most conditions using acute nicotine as well as nicotine withdrawal. Unfortunately, although differences in some phenotypes (notably acute nicotine’s effect on respiration) could be detected between mouse strains, significant differences were not found between FMO knockout mice and sibling controls, or between DIM-fed mice and controls. Exogenous nicotine-N-oxide administration also did not alter nicotine withdrawal phenotypes. Therefore, we conclude that, in light of DIM’s unexpected effects, and our failure to detect significant differences between FMO knockout and control mice, the mouse is not an optimal model to study the effects of altered nicotine N-oxidation using the assays currently in our hands.
Given the exceptional public health consequences of cigarette smoking, improved cessation has a huge positive impact on human longevity and quality of life. Variation in the FMO3 gene, responsible for nicotine N-oxidation, is associated with differences in liability to nicotine dependence among smokers. We sought to explore altered nicotine N-oxidation in a mouse model of nicotine-related phenotypes, including nicotine withdrawal, by manipulating nicotine N-oxidation and nicotine-N-oxide levels by three methods: 1) using an FMO knockout mouse, 2) exposing mice to 3,3’-diindolymethane (DIM), and 3) by administering exogenous nicotine-N-oxide. DIM has been shown to alter FMO activity in humans and rats. However, we found that DIM’s effect on nicotine metabolism is very different in mice than in rats. In mice, DIM did not affect N-oxidation but lead to increased C-oxidation in both male and female mice from multiple mouse strains (C57BL and CD1). Because differences in nicotine C-oxidation also impact cigarette consumption, we nevertheless performed further experiments in DIM-fed mice.
Nicotine withdrawal was measured in both male and female mice of different strains under different experimental conditions using phenotypes including respiration (using full-body plethysmography), somatic signs (tremor, paw shaking, etc.), pain sensitivity (tail-flick), locomotion and anxiety (open-field), and conditioned place preference. Initial experiments were also performed under most conditions using acute nicotine. Unfortunately, although differences in some phenotypes (notably acute nicotine’s effect on respiration) could be detected between mouse strains, significant differences were not found in any phenotype between FMO knockout mice and sibling controls, or between DIM-fed mice and controls. Exogenous nicotine-N-oxide administration also did not alter nicotine withdrawal phenotypes. Therefore, we conclude that, in light of DIM’s unexpected effects, and our failure to detect significant differences between FMO knockout and control mice, mice are not an optimal model to study altered nicotine N-oxidation with the tools and assays currently in our hands.
There are few single factors with a more profound negative impact on human longevity than tobacco use. Smoking status is the single best predictor of human mortality, with the average moderate smoker’s life ending twelve years early. Conversely, smoking cessation provides immediate and long-term health benefits that increase lifespan.
Heritable differences in nicotine metabolism influence smoking phenotypes among dependent smokers. The enzyme CYP2A6 is responsible for the majority of nicotine metabolism (70-80%) in most people, and variation in CYP2A6 is associated with cigarette consumption. Smokers with reduced CYP2A6 function metabolize nicotine more slowly, which delays the onset of the nicotine withdrawal and allows them to wait longer between cigarettes. Nicotine is metabolized by two additional pathways: glucuronidation by the UGT2B10 enzyme, and N-oxidation by the Flavin-containing monooxygenase FMO3. A common FMO3 haplotype that leads to reduced in vivo nicotine N-oxidation is also associated with altered liability to nicotine dependence. In a large sample of smokers seeking to quit, we previously found a significant association between FMO3 genotype and the Fagerström Test of Nicotine Dependence (FTND), the most commonly used instrument for measuring nicotine dependence. But unlike CYP2A6, FMO3 genotype was not significantly associated with cigarette consumption.
The main objective of this work was to explore altered nicotine N-oxidation in a mouse model of nicotine-related phenotypes. To that end, we attempted to manipulate nicotine N-oxidation and nicotine-N-oxide levels through both genetic and pharmacological means, namely: 1) using an FMO-gene knockout mouse, 2) by exposing mice to 3,3’-diindolymethane (DIM), and 3) by administering exogenous nicotine-N-oxide. DIM, a natural derivative of cruciferous vegetables, has been shown to alter FMO activity in humans and rats. Human subjects administered a diet rich diet in Indole-3-carbonyl (I3C), the precursor to DIM, had reduced FMO3 activity as determined by the ratios of trimethylamine to trimethylamine-N-oxide in their urine. FMO1 protein levels were also significantly decreased in the livers of rats fed I3C or DIM, and liver microsomes prepared from these animals showed a dramatic decrease in N-oxidation of FMO substrates including nicotine. The inhibition was more pronounced for DIM than for I3C, and appeared to be specific to N-oxidation. Importantly, nicotine C-oxidation was unaffected by DIM treatment in rats.
Unfortunately, we discovered that DIM’s effects on nicotine metabolism differed significantly between mice and rats. Furthermore, we did not detect statistically significant differences between FMO knockout mice and controls. Therefore, we conclude that mouse may not be the optimal model to study the influence of altered nicotine N-oxidation on smoking behaviors using the reagents and assays currently at our disposal.
All protocols were approved by the Washington University Institutional Animal Care and Use Committee. Mice were approximately 8-weeks old at the initiation of all experiments (i.e. DIM feeding, pump insertion etc., depending on protocol). C57BL/6 and C3H mice were purchased from Jackson Laboratories. CD1 mice were purchased from Charles River. FMO1/2/4 knockout mice created in the C57BL/6 background were received from the Shephard lab at King’s College London and maintained in Washington University Medical School animal facilities. To produce FMO1/2/4 -/- knockout mice and FMO1/2/4 +/+ control littermates, FMO1/2/4 -/- males were mated to C57BL/6 females purchased from Jackson labs and FMO1/2/4 +/- offspring were crossed and subsequent progeny genotyped for experiments.
Diluted pup tail snip preps were used as template in PCR reactions using two primer sets, to detect either the intact FMO1 gene or an inserted cassette in the knockout construct:
For the purpose of planning experiments, the sexes of young pups were sometimes confirmed using primers that detect the Y-chromosome, with positive control primers:
Reaction conditions were 95°C for 5min, followed by 30 cycles of 95°C for 30 s, 59°C for 30 s and 72°C for 30 s, and a final extension at 72°C for 7 min. Amplification products were resolved by electrophoresis through a 2% (w/v) agarose gel and visualized by staining with ethidium bromide.
All mice were fed Purina 5058 mouse chow. For DIM experiments, mice were fed modified LabDiet 5058 containing 0.25% i.e. 2500 ppm DIM purchased from Purina TestDiet, always beginning 2 weeks prior to further experimental manipulations such as pump insertions.
Microsome preparation and incubations
Microsomes were prepared from flash-frozen tissues. All experiments included at least 5 samples per data point. Tissues in ice-cold potassium phosphate buffer pH 7.4 were homogenized 6x with a glass-teflon homogenizer, centrifuged 20 min at 9,000 g. The S9 fraction was centrifuged 60 min at 100,000 g to obtain a microsomal pellet which was resuspended, further homogenized 6x, centrifuged another 60 min at 100,000 g and finally resuspended and homogenized 6x in potassium phosphate buffer containing 20% glycerol for storage at -80°C. Prior to freezing, protein concentrations were determined by the Bradford Protein assay. All incubations were carried out in 96-well PCR plates in 50ml total volume at 37ºC for 10 minutes in tricine buffer containing 1 mM NADPH and 100mM nicotine. Reactions were quenched by adding 10ml 15% zinc sulfate containing 200ng/ml of the internal standard, (d3)-nicotine-N-oxide. Samples were centrifuged and supernatants removed for LC/MS analysis.
Measurement of nicotine and nicotine metabolites
Nicotine and nicotine metabolites were analyzed by liquid chromatography tandem mass spectrometry in the laboratory of Evan Kharasch at Washington University. Nicotine and metabolites were quantified using the ratio of metabolite to internal standard. LC-MS\MS analyses were performed on an Applied Biosystems Sciex API 4000 Q-TRAP triple quadrupole mass spectrometer equipped with an electrospray source. The HPLC system consisted of two LC 20AC pumps with a CTO-20A column oven, SIL-20A autosampler, DGU-20A3 degasser, FCF-11AL valve, and a CBM 20A controller (Shimadzu). The chromatographic separation was performed on an Waters 150 x 2.1, 3.5μm xBridge column with a pre-column inline 0.2 μm filter. The injection volume was 5 μl and the oven temperature was 40 °C. Mobile phase (0.3 ml/min) was (A) 4.5 mM ammonium acetate pH 4.0 and (B) 4.5 mM ammonium acetate in acetonitrile using the following program: 1% B for 4 minutes, linear gradient to 25% B between 4 and 5 minutes, held at 25% B until 6.0 minutes, and then re-equilibrated to initial conditions between 6.5 and 8.5 minutes. The instrument was operated in positive-ion mode at 450 °C with an ion spray voltage of 5500 V, entrance potential of 10 V and exit potential of 22 V. The curtain gas was set at 20, ion source gas 1 at 30, and ion source gas 2 at 40. Transitions monitored for nicotine were m/z 163→130, for nicotine iminium m/z 161→130, for cotinine, m/z 177→98, for nicotine-N-oxide m/z 179→132, and for deuterated (d3)-nicotine-N-oxide, m/z 182→132.
Measurement of DIM
DIM concentrations in mouse plasma, liver and brain were determined in the laboratory of Stephen Hecht at U. Minnesota. Liver and brain tissues were first homogenized with a glass-teflon homogenizer in ice-cold potassium phosphate buffer pH 7.4. Analyses were performed on a Thermo Fisher Scientific TSQ Quantum Discovery Max instrument in the positive ion mode with N2 as the nebulizing and drying gas. MS parameters: spray voltage 3.2 kV; sheath gas 25; capillary temperature 250°C; collision energy 17 V; scan width 0.05 amu; Q2 gas pressure 1.0 mTorr; source CID 9 V; tube lens offset 104 V; Q1 0.2 amu and Q3 0.7 FWHM. Mass spectrometry data were acquired and processed by Xcalibur software version 1.4. 8μL of the sample were injected from an autosampler using an Agilent 1100 capillary LC system equipped with a 5 μm, 150 × 0.5 mm Zorbax SB-C18 column. The flow rate was 15 μL/min for the first 3 minutes then 10 μL/min with a gradient from 60% methanol in 15 mmol/L NH4OAc to 100% methanol in 8 minutes and held for an additional 29 minutes. The mass transitions monitored were m/z 247.12 → 130.07 for DIM and m/z 249.12 → 132.07 for [2H2]DIM. Quantitation was done by comparing the MS peak area ratio of DIM to that of [2H2]DIM using a calibration curve prepared by plotting the MS peak area ratio of DIM to [2H2]DIM against their concentration ratios using standard mixes containing constant levels of [2H2]DIM and varying levels of DIM.
Nicotine and nicotine-N-oxide administration and withdrawal
Solutions of nicotine and/or nicotine-N-oxide for acute or chronic administration were prepared in sterile saline. Nicotine was prepared from nicotine hydrogen tartate salt with concentrations according to the free base. All acute injections were subcutaneous. For chronic administration control saline, nicotine and or nicotine-N-oxide were loaded into Alzet osmotic pumps, models 1002 or 2002 (for 2 weeks), or 1004 or 2004 (for 4 weeks) in a sterile hood and weighed prior to and following filling to confirm they were correctly filled. Mice were weighed 1-2 days prior to pump insertion in order to calculate the correct nicotine and N-oxide concentrations to ensure the desired dose. For pump insertion, mice were anesthetized with 2-3% isoflurane, an area of the back shaved with electric clippers and prepped with betadine. A small skin incision was made with a scalpel, large enough to permit formation of a pocket to insert the osmotic pump. The skin was closed with surgical staples.
Spontaneous nicotine withdrawal was initiated by again anesthetizing mice with isoflurane, making a second incision, removing the pump, and reclosing the incision with staples. Experiments were performed 8-12 hours following spontaneous nicotine withdrawal depending on the assay. To initiate precipitated withdrawal, mice were injected with 3mg/kg mecamylamine and assayed beginning 30 minutes later.
Respiratory parameters were measured in a Buxco Small Animal Whole Body Plethysmograph. Mice were acclimated to the plethysmograph chamber for one hour prior to each experiment. Baseline measurements were taken for 10 minutes prior to experimental manipulation. They were then removed from the chambers, injected subcutaneously with sterile saline, mecamylamine or nicotine, returned to the chambers and parameters measured for thirty or sixty minutes. For certain acute nicotine experiments, mice were used in three separate experiments on three different days. In those cases, half of the mice in each strain/sex group were injected sequentially with 0.35, 0.00, and 0.70 mg/kg nicotine, while the other half were injected with 0.00, 0.70 and 0.35 mg/kg nicotine. Four parameters were typically measured: 1) frequency, i.e. number of breaths per minute, 2) minute volume, total volume inhaled per minute, 3) tidal volume, volume per breath, and 4) inspiratory time, seconds per breath. Data was analyzed as averages over ten-minute periods post-injection.
Time to first rear
During acute nicotine experiments performed in the plethysmograph, sedation was also measured using the time to first rear following nicotine injection. Mice that appeared sedated lay flat on the floor of the chamber within one minute following injection. The time each mouse first rose with both front paws off the floor was recorded as the time of first rearing, with the time of injection as time zero. Time of first rear was recorded as zero for mice that did not lay flat in the chamber within one minute following injection.
Somatic signs of nicotine withdrawal were recorded 12 hours following pump removal and spontaneous nicotine withdrawal, or thirty minutes following precipitated withdrawal with mecamylamine. Mice were placed in a 15x15 cm vented Plexiglass box and videoed for 30 minutes. Videos were later reviewed by scorers blind to genotype or experimental condition and a tally taken of rears, and somatic signs of nicotine withdrawal, i.e. body tremors, head shakes, paw shakes were tallied. The occurrence of ptosis and piloerection were also noted. Somatic signs were analyzed separately and as a combined score.
Altered nociception was measured in mice using the tail-flick assay. Baseline measurements were taken the day prior and then approximately 12 hours following pump removal and spontaneous nicotine withdrawal, or prior to and then beginning thirty minutes following precipitated withdrawal with mecamylamine. Each data point was the average of three measurements for each mouse, baseline and post-withdrawal.
Each mouse was scruffed, its tail immersed in a 50°C water bath, and the time measured until it withdrew its tail from the bath.
Locomotion and anxiety were measured in an open field assay. Mice were placed in the center of a 25x25cm open field 8-12 hours following pump removal and spontaneous nicotine withdrawal, or thirty minutes following precipitated withdrawal with mecamylamine. The mice were recorded for thirty minutes and their movement was analyzed using Noldus EthoVision XT software to determine total locomotion and time spent in the center of the field (approximately 25% of the total area) versus the periphery.
Conditioned place preference
Mice were trained in an unbiased, balanced three-compartment conditioning apparatus. The apparatus consisted of two large square chambers connected by a small vestibule that could be closed off on both sides. In each apparatus, one large chamber was decorated with vertical black and white stripes, the other with horizontal black and white stripes. On pre-conditioning day, mice were allowed free access to all three chambers for 20 minutes. Time spent in each compartment was recorded with a video camera and analyzed with Noldus EthoVision XT software.
For nicotine conditioning, mice were injected subcutaneously with saline and placed in the randomized non-conditioning side of the apparatus with the vestibule closed for 20 minutes in the morning, and at least 4 hours later, injected with nicotine and placed in the conditioning side of the apparatus for 20 minutes in the afternoon. Training occurred on two consecutive days. To test for nicotine place preference on the third day, mice were again allowed free access to the three compartments and recorded on video. Scores were calculated by subtracting the time spent in the nicotine-paired compartment, post-conditioning minus pre-conditioning.
For conditioning with spontaneous nicotine withdrawal, osmotic pumps were removed after pre-conditioning in the apparatus. 12 hours later, mice were placed in one side of the apparatus (the conditioning side) for 50 minutes. To test for place preference, 5 days later, mice were again allowed free access to the three compartments, recorded on video and analyzed with Noldus EthoVision XT software. Scores were calculated by subtracting the time spent in the nicotine-withdrawal paired compartment, post-conditioning minus pre-conditioning.
Respiratory parameters were measured and analyzed using FinePointe software. Open-field and conditioned place preference results were measured and analyzed using Noldus EthoVision XT software. Additional statistical tests were performed using the statistical software package R.
Nicotine metabolism in FMO1/2/4 knockout DIM-fed mice
We had previously prepared liver and brain microsomes from male and female FMO1/2/4 knockout mice, as well as from heterozygous and homozygous control littermates. Nicotine-N-oxide formation was detected in all incubations, but N-oxidation activity in knockout microsomes was only 3-5% and 1-2% of the activity detected in control liver and brain microsomes respectively, for both males and females. FMO1/2/4+/- heterozygous livers had 40-50% as much N-oxidation activity as FMO1/2/4+/+ liver microsomes.
In order to determine the effects of DIM on hepatic nicotine metabolism, liver microsomes were prepared from male and female C57BL/6J and CD1 mice fed standard rodent chow or chow containing DIM for two weeks. Contrary to our expectations based on prior published evidence from rats, nicotine N-oxidation activity was not significantly reduced in DIM-fed mouse liver microsomes of either sex or strain (Figure 1). However, nicotine C-oxidation (measured by nicotine iminium ion) was increased (Figure 2).
Figure 1. Nicotine iminium produced by liver microsomes; each column n=5
Figure 2. Nicotine-N-oxide produced by liver microsomes; each column n=5.
To confirm that the effects of DIM on nicotine metabolism observed in microsomes correspond with similar effects in vivo, we also measured plasma levels of nicotine metabolites in CD1 male mice subcutaneously injected with nicotine. CD1 males were chosen because of the large change in nicotine C-oxidation activity observed in liver microsomes from these mice (>19 fold increase, Figure 1) relative to the more modest increases observed in the other three sex/strain groups (2.4-3.5 fold). Cotinine was measured because nicotine iminium ion is rapidly converted to cotinine in vivo. Consistent with the microsome incubations, plasma levels of cotinine were significantly greater in DIM-fed animals 10 and 20 minutes following nicotine injection (Figure 3), and the plasma nicotine was significantly reduced at 10 minutes (Figure 4). Nicotine-N-oxide plasma levels were not significantly different.
Figure 3. each data point n=5.
Figure 4. each data point n=5.
These results are included in a manuscript prepared for publication. Because variation in nicotine C-oxidation is also associated with altered smoking behaviors, we chose to continue with several further experiments in DIM-fed mice.
Respiration in FMO1/2/4 knockout and DIM-fed mice
Both acute nicotine and nicotine withdrawal alter respiratory parameters in human subjects, but the effect of nicotine withdrawal on respiration is little studied in mice due to the challenges of measuring respiratory metrics compared to phenotypes such as heart-rate, blood-pressure or locomotion. Unrestrained whole-body plethysmography offers a precise, non-invasive method to measure respiratory phenotypes in experimental animals, avoiding the stresses or anesthesia common to traditional plethysmography techniques.
We began by examined respiratory phenotypes in mice of different strains acutely exposed to different doses of nicotine. Male and female mice of two inbred strains, C57BL/6J and C3H/HeJ, were injected subcutaneously with two doses of nicotine, 0.35 or 0.70 mg/kg, and compared to saline-injected controls for thirty minute. 0.35 and 0.70 mg/kg were chosen based on initial experiments measuring respiration in male and female C57BL/6J mice injected with saline and various doses of nicotine up to 1 mg/kg. Grieder and colleagues (2017) have also shown that 0.35 mg/kg nicotine is reinforcing in C57BL mice. We also performed experiments comparing subcutaneous and intraperitoneal injection (with 0.35 mg/kg nicotine in C57BL males) and found no significant difference in respiratory phenotypes.
We found injection with both 0.35 and 0.70 mg/kg nicotine significantly reduced respiratory frequency in C57BL mice but only 0.70 mg/kg nicotine significantly affected the parameter among C3H mice (Figure 5 & 6). Similarly, both 0.35 and 0.70 mg/kg nicotine significantly affected minute volume, tidal volume, and inspiratory time among C57 mice, while only 0.70 mg/kg affected these parameters among C3H mice. For C57 mice, the higher dose also had larger effects on all respiratory parameters than the lower dose.
Figure 5. Average respiratory frequency (breaths per minute) during three ten minute intervals post-injection. Each data point represents 8 male and 8 female mice. Figure 6. Average respiratory frequency (breaths per minute) during the first ten minutes post-injection. Each data point represents 8 male and 8 female mice.
None of the four respiratory parameters were significantly associated with sex for either strain under any condition. For both mouse strains, frequency was highly correlated with minute volume and inspiratory time (R2>0.90 for minute volume and >0.77 for inspiratory time), and minute volume and inspiratory time were highly correlated (R2>0.64). The exceptional parameter was tidal volume; among the C57 mice, across the three intervals, tidal volume was moderately correlated with the other three parameters (R2>0.47), but among the C3H mice, tidal volume was poorly correlated with minute volume (R2=0.11) and not correlated with frequency (R2=0.005) or inspiratory time (R2=0.006).
Many of the C57 and C3H mice injected with nicotine appeared sedated within one minute after injection. In order to quantify this effect, we measured the time at which mice first rose on their hind legs (time until first rear) following the sedated period after injection, sometimes exceeding thirty minutes. Not surprisingly, this measure paralleled the differences observed in respiratory parameters.
To determine whether the observed strain differences were related to nicotine metabolism, C57BL/6J and C3H/HeJ mice were also injected subcutaneously with 0.35 mg/ml nicotine, and plasma nicotine measurements were made. Although plasma nicotine concentrations differed by sex, there was no difference in concentrations between strains. These data are included in a manuscript submitted for publication.
We performed further plethysmograph experiments measuring the effects of acute nicotine, both 0.35 and 0.70 mg/kg, on respiration in male and female FMO1/2/4 knockout mice and littermate controls, and found no significant difference between genotypes. This result was not altogether surprising given that nicotine N-oxidation is not the primary nicotine metabolism pathway and FMO3 genotype is associated with liability to nicotine dependence rather than the acute effects of nicotine in smokers.
We have also hypothesized that the metabolite nicotine-N-oxide itself might mediate the effects of altered FMO activity on smoking behavior, perhaps by regulating the activity of nicotinic acetylcholine receptors. Therefore, we further tested the respiratory effects of nicotine with or without the addition of nicotine-N-oxide (70mg/kg) in C57BL male and female mice, but again found no significant difference.
Finally, we were most interested to measure the effects of nicotine withdrawal on respiratory measures with the aim of assaying the relationship between N-oxidation activity and withdrawal. Therefore, we first measured the respiratory effects of precipitated nicotine withdrawal with mecamylamine in C57BL male and female mice implanted with osmotic pumps after receiving 36mg/kg per day for 14 days. Unfortunately, we did not see a significant difference in any respiratory parameter between mecamylamine and saline-injected mice (Figure 7) and due to limited access to the plethysmograph we were unable to perform further experiments.
Figure 7. Average respiratory frequency (breaths per minute) during six ten minute intervals following injection with mecamylamine or saline. Each line represents 6 mice/
Somatic signs of withdrawal and altered nociception
Often in concert with other nicotine withdrawal experiments, we measured somatic signs of withdrawal, i.e. rearing (withdrawal is usually associated with reduced rearing), body tremors, head shakes, paw shakes etc. We attempted to measure somatic signs in experiments including precipitated and spontaneously withdrawn FMO1/2/4 knockouts and control littermates, and in spontaneously withdrawn CD1 mice, DIM-fed CD1 mice, and CD1 mice administered nicotine-N-oxide in addition to nicotine. Although we had success in initial experiments with precipitated nicotine withdrawal in CD1 mice, we found this phenotype especially fragile and did not observe significant withdrawal-associated differences in somatic signs in other experiments. In conjunction with these experiments, we also measured differences in nociception using the tail-flip assay, and similarly, we saw differences in precipitated nicotine withdrawn CD1 mice, but could not demonstrate differences in nociception related to any other variable (i.e. genotype, DIM or nicotine-N-oxide.
Open field assay
To measure the effects of nicotine withdrawal on locomotion and anxiety under different conditions, we videoed withdrawing mice in an open field and analyzed their movements. All mice were withdrawn from nicotine after receiving 36 mg/kg per day nicotine or control saline for four weeks by osmotic pump. Experiments included comparing FMO1/2/4 knockouts to control littermates, both precipitated (mecamylamine) and spontaneous (pump removal) withdrawal; comparing spontaneous withdrawal from nicotine to withdrawal from nicotine plus 70 mg/kg per day nicotine-N-oxide in CD1 mice; and spontaneous withdrawal in DIM-fed versus control CD1 mice (Figures 8 & 9). Unfortunately, we did not find a significant difference in total locomotion or the ratio of time spent in the center versus periphery of the field (a measure of anxiety behavior) under any experimental conditions.
Figures 8 & 9. Distance traveled (cm) and percent of time spent in the center of an open field, for spontaneously withdrawn mice administered 1) control saline, 2) nicotine, or 3) nicotine + nicotine-N-oxide, and 4) DIM-fed mice administered control saline, or 5) nicotine. Each column represents 8 mice.
Nicotine and nicotine-withdrawal conditioned place preference (CPP)
CPP experiments were also conducted, often in concert with measurements of altered nociception and somatic signs of withdrawal or with open field experiments. Initial experiments were performed conditioning with acute nicotine in male and female FMO1/2/4 knockouts and littermate controls (as well as +/- heterozygous littermates) and found that the mice would condition to a relatively high dose of nicotine (0.70 mg/kg), but we did not see a significant difference by genotype. This was not a surprise given that we had not seen a significant difference in other phenotypes (i.e. sedation and respiration) by FMO1/2/4 genotype with acute nicotine.
We further performed CPP experiments conditioning with spontaneous withdrawal (see methods) in FMO1/2/4 knockouts versus littermate controls (in the C57BL background) after receiving 36 mg/kg per day of nicotine for four weeks, and also in CD1 mice after receiving 36 mg/kg per day nicotine, or control saline, or nicotine together with 70 mg/kg per day nicotine-N-oxide for four weeks. Unfortunately, we did not observe conditioning with spontaneous nicotine withdrawal in either mouse strain under any conditions.
The primary aims of the experiments described in this report were to demonstrate how nicotine N-oxidation could be manipulated genetically or pharmacologically in mice, and investigate how altered nicotine N-oxidation affects nicotine-related phenotypes using the mouse as a model. This line of investigation was inspired by provocative results from a human genetic study of smoking behavior. The most common reduced-function FMO3 allele is also associated with significantly reduced nicotine dependence among treatment-seeking smokers. This was a surprising finding given that N-oxidation by FMO3 is responsible for less than 7% of nicotine metabolism in most subjects. Furthermore, the influence of FMO3 genotype on smoking behavior is unique in that it affects urgency to smoke (time to first cigarette upon arising) independent of consumption levels, indicating that it acts through a novel mechanism. Therefore, we formulated two hypotheses to explain this association: 1) FMO3 activity in the brain significantly determines local nicotine clearance, influencing dependence, or 2) nicotine-N-oxide itself is pharmacologically active and the concentration of this metabolite influences dependence. Unlike CYP2A6, FMO enzymes are expressed and active in the human brain and we have demonstrated nicotine N-oxidation in human and mouse brain microsomes. But nicotine-N-oxide also modulates the activity of α4β2 nicotinic acetylcholine receptors in vivo. Therefore, we sought a model to distinguish between these mechanisms, and mouse was the obvious first choice given the available FMO knockout and the potential of pharmacologically suppressing FMO activity with DIM.
However, there are numerous reasons why the association between FMO3 genotype and nicotine dependence might not translate into a murine system. Firstly, animal model measures of nicotine-related phenotypes are never a perfect analog of human smoking. For example, due to the limitations of our methods, mice received nicotine passively and continuously via an implanted pump, unlike nicotine consumed via cigarette smoking, which is self-administered, pulsatile and inhaled through the lungs. It is also possible that, relevant to one of our hypotheses, nicotine-N-oxide does not regulate α4β2 nicotinic acetylcholine receptor activity in mice in the same way it does in humans, or that the differing roles and expression of acetylcholine receptor subtypes in the mouse CNS reduces the effects of nicotine-N-oxide on mouse behavior. Finally, as we discovered, FMO enzyme expression is not regulated in mouse exactly as it is in humans and rats. Unlike in rats, DIM does not affect FMO activity, but rather enhances nicotine C-oxidation.
Although we are disappointed at our inability to demonstrate differences in nicotine-related phenotypes by manipulating FMO activity and nicotine-N-oxide in mouse, we remain optimistic about FMO3 as a target for smoking cessation therapy. The association between FMO3 and liability to nicotine dependence remains intriguing, and so does the possibility of regulating FMO activity in humans using DIM. Although indirect evidence indicates FMO3 activity is repressed by DIM consumption in humans, the effects of DIM on nicotine metabolism in humans remains poorly studied. Notably, our work in mouse demonstrates that DIM may enhance CYP-mediated nicotine metabolism in some species (an unexpected result based on rat). If this is the case in humans, DIM might be expected to increase nicotine withdrawal and cigarette consumption in current smokers, an effect that could complicate its use as an adjunct to smoking cessation therapy.
We do not have any current plans to submit grant applications related to nicotine N-oxidation or DIM in mouse.