Disruption of P450-mediated vitamin E hydroxylase activities alters vitamin E status in tocopherol supplemented mice and reveals extra-hepatic vitamin E metabolism.

The widely conserved preferential accumulation of α-tocopherol (α-TOH) in tissues occurs, in part, from selective postabsorptive catabolism of non-α-TOH forms via the vitamin E-ω-oxidation pathway. We previously showed that global disruption of CYP4F14, the major but not the only mouse TOH-ω-hydroxylase, resulted in hyper-accumulation of γ-TOH in mice fed a soybean oil diet. In the current study, supplementation of Cyp4f14−/− mice with high levels of δ- and γ-TOH exacerbated tissue enrichment of these forms of vitamin E. However, at high dietary levels of TOH, mechanisms other than ω-hydroxylation dominate in resisting diet-induced accumulation of non-α-TOH. These include TOH metabolism via ω-1/ω-2 oxidation and fecal elimination of unmetabolized TOH. The ω-1 and ω-2 fecal metabolites of γ- and α-TOH were observed in human fecal material. Mice lacking all liver microsomal CYP activity due to disruption of cytochrome P450 reductase revealed the presence of extra-hepatic ω-, ω-1, and ω-2 TOH hydroxylase activities. TOH-ω-hydroxylase activity was exhibited by microsomes from mouse and human small intestine; murine activity was entirely due to CYP4F14. These findings shed new light on the role of TOH-ω-hydroxylase activity and other mechanisms in resisting diet-induced accumulation of tissue TOH and further characterize vitamin E metabolism in mice and humans.

following the National Institutes of Health guidelines for laboratory animal use. Cyp4f14 Ϫ / Ϫ mice, which are viable and fertile, and their wild-type littermates, were generated as described previously ( 16 ). At 21 days of age, six mice of each genotype (three males and three females) were weaned onto a modifi ed AIN-93G semipurifi ed diet containing 12 mg/kg ␣ -TOH and supplemented with 800 mg/kg each ␦ -TOH and ␥ -TOH. After 12 weeks, 24 h urine and fecal collections were made as described below. Mice were deeply anesthetized by isofl urane inhalation and exsanguinated via cardiac puncture. If bile was present in the gallbladder, it was collected with a syringe and frozen at Ϫ 80°C until analysis. Heparinized blood was centrifuged at 6,000 g for 5 min, and serum was frozen at Ϫ 80°C until analysis. Liver, kidney, lung, heart, abdominal fat, and brain tissue samples were fl ash frozen in liquid nitrogen for vitamin E quantifi cation by gas chromatography-mass spectrometry (GC-MS) as described previously ( 16 ). Proximal small intestine samples were fl ash frozen for CYP4F14 expression analysis by RT-PCR.

L-Cpr
Ϫ / Ϫ mice, which have trace liver CYP activity by 2 months of age, and their wild-type littermates were generated as described previously ( 17 ). At 2.5 months of age, four male mice of each genotype were fed a modifi ed AIN-93G semipurifi ed diet containing 12 mg/kg ␣ -TOH and supplemented with 800 mg/kg each ␦ -TOH and ␥ -TOH. After 4 weeks, 24 h urine and fecal collections were made as described below. Mice were then euthanized, and plasma and several tissues were collected as described above. Approximately 0.6 g of fresh liver was immediately used for preparation of microsomes as described below.

h urine and fecal collections
Mice were placed individually in wire bottom polycarbonate cages in which urine and fecal pellets were collected separately. Mice were afforded continuous access to food and water, and collections were made over 24 h following a 24 h acclimation period.

Preparation of liver microsome
Microsomes were prepared from fresh mouse liver by standard differential centrifugation as previously described, resuspended in 0.1 M sodium phosphate (NaP) buffer containing 1 mM EDTA and frozen at Ϫ 80°C ( 16 ). Microsomal protein concentrations were determined by a Bradford-based Bio-Rad assay using BSA as the standard.

Preparation of intestinal mucosa microsome
Mice were euthanized by isofl urane exposure, and the upper 12 cm of the small intestine was immediately excised and placed on an ice-cold sheet of glass. The removed piece was immediately fl ushed with ice-cold wash buffer (0.9% NaCl containing 1 mM EDTA and 1 mM PMSF), then cut longitudinally, opened, and rinsed with ice-cold wash buffer to remove intestinal contents. The mucosal cells were gently scraped off with a razor blade. The intestinal mucosal cells were homogenized using a Tefl on/glass homogenizer with ice-cold homogenization buffer [50 mM Tris-HCl, pH 7.4, containing 150 mM potassium chloride (KCl), 1 mM EDTA, 20% glycerol, 1 mM PMSF, and 10% protease inhibitor cocktail] and centrifuged at 10,000 g for 20 min at 4°C. The supernatant was centrifuged at 100,000 g for 1 h at 4°C. The microsomal pellet was resuspended in 0.1 M NaP buffer containing 1 mM EDTA and frozen at Ϫ 80°C. Microsomal protein concentration was determined by a Bradford-based Bio-Rad assay. non-␣ -TOH forms of vitamin E are metabolized to a greater extent than ␣ -TOH, thereby contributing to the ␣ -TOH phenotype ( 15 ).
Using a novel Cyp4f14 knockout mouse model, we recently identifi ed CYP4F14 as the major mouse vitamin E-hydroxylase ( -0, hydroxylation of a terminal methyl group), accounting for 70-90% of whole-body -hydroxy metabolite production ( 16 ). This result demonstrated the existence of other liver vitamin E--hydroxylase enzyme(s) in the mouse. In addition, two novel metabolites were found to be excreted in the feces: 12 ′ -and 11 ′ -OH metabolites of ␦ -and ␥ -TOH, products of -1 and -2 hydroxylation activities, respectively. Interestingly, Cyp4f14 Ϫ / Ϫ mice displayed increased fecal excretion of these novel metabolites as well as increased fecal excretion of unmetabolized TOH. Despite these counterbalancing mechanisms and redundancy in the TOH--hydroxylase pathway, Cyp4f14null mice fed a modest amount of ␥ -TOH in the form of soybean oil accumulated 2-fold more ␥ -TOH than wild-type mice. Although the urine has been previously considered to be the major route of TOH metabolite excretion, we found the fecal route to predominate over the urine.
The central hypothesis of the current work was that postabsorptive catabolism via TOH-hydroxylases constitutes the major limitation on tissue accumulation of non-␣ -TOH. We aimed to investigate whether dietary supplementation with high levels of ␥ -and ␦ -TOH would overcome the counterbalancing effects and result in tissue enrichment above that seen with the previous soybean oil diet. Two experimental models were employed: mice with global disruption of Cyp4f14, and mice with liver-specifi c CPR disruption that therefore lack all hepatic microsomal P450 activity. The L-Cpr model also allowed the opportunity to determine whether tissues other than the liver possess TOH-hydroxylase activity. Additionally, we extended the previous fi ndings concerning novel -1 and -2 hydroxylase activities to determine their potential relevance in humans.

Tocopherol metabolism and tissue accumulation in
Cyp4f14 ؊ / ؊ mice supplemented with ␥ -and ␦ -TOH Use of mice was in accordance with protocols approved by the Cornell Institutional Animal Care and Use Committee and

Statistical analysis
All tocopherols and tocopherol metabolites were log transformed as necessary, and means were compared between wild-type and knockout mice using Student t -test. Additionally, tissue TOH concentrations in Cyp4f14 Ϫ / Ϫ mice were compared using a two-way ANOVA with genotype and gender as main effects, as well as the interaction effect. When the interaction effect was not signifi cant, it was removed from the model. All tests were two-sided, and a P -value < 0.05 was considered statistically signifi cant. Analyses were performed using JMP version 8 (SAS Institute).

Effect of Cyp4f14 disruption on 24 h TOH and TOH metabolite excretion
Metabolic cages were utilized to obtain 24 h urine and fecal samples from Cyp4f14 +/+ and Cyp4f14 Ϫ / Ϫ mice fed 800 mg/kg of both ␦ -and ␥ -TOH for 12 weeks. In urine, only 3 ′and 5 ′ -carboxychromanol metabolites of ␦ -and ␥ -TOH were detected. Urinary metabolites of ␥ -TOH were reduced by 88% and those of ␦ -TOH were reduced by 77% in Cyp4f14 Ϫ / Ϫ mice compared with their wild-type littermates. Total urinary tocopherol metabolites were reduced by 82% in Cyp4f14 Ϫ / Ϫ mice ( Table 1 ). Analysis of 24 h fecal samples revealed the presence of all six carboxychromanol metabolites, as well as the 13 ′ -OH metabolites of ␥ -and ␦ -TOH, formed via the -oxidation pathway ( Table 2 ). The 13 ′ -COOH metabolite was consistently the predominant -oxidation metabolite in fecal samples. Total -oxidation metabolites of ␥ -TOH were reduced by 79% and those of ␦ -TOH by 89% in Cyp4f14 Ϫ / Ϫ mice. 12 ′ -OH and 11 ′ -OH metabolites of ␥ -and ␦ -TOH were also present, formed via the -1 and -2 oxidation pathways, respectively. These metabolites were also detected in the limited number of gallbladder bile samples that were collected (data not shown).
Twenty-four hour whole-body (urine + fecal) -oxidation metabolites of ␥ -and ␦ -TOH were reduced by 82% in Cyp4f14 Ϫ / Ϫ compared with wild-type mice ( Fig. 1A ). The sum of 24 h -1 and -2 metabolites of ␥ -and ␦ -TOH, although numerically higher in Cyp4f14 Ϫ / Ϫ mice (30%), was not statistically signifi cantly altered by the disruption of Cyp4f14 . As a result, the combination of metabolites from all three pathways was 40% reduced in Cyp4f14 Ϫ / Ϫ mice. Twenty-four hour fecal excretion of unmetabolized TOH was not statistically different between wild-type and Cyp4f14null mice ( Fig. 1B ).

Analysis of TOH and metabolites in 24 h urine and fecal samples
Twenty-four hour fecal collections were homogenized in PBS, and an aliquot was used for analysis. Urine and fecal samples were incubated with ␤ -glucuronidase (800 units for urine or 1,600 units for feces, dissolved in NaP buffer, pH 6.8) and sulfatase (0.4 units for urine or 0.8 units for feces) for 2 h at 37°C. Samples were acidifi ed to pH 2, extracted, and derivatized as previously described ( 16 ). TOH and their metabolites were quantifi ed by GC-MS using d 9 -␣ -TOH and d 9 -␣ -CEHC as internal standards.

Analysis of bile TOH and metabolites in Cyp4f14
+/+ and Cyp4f14 ؊ / ؊ mice Bile samples from Cyp4f14 mice (two wild-type and three knockout) were incubated with enzymes (800 units ␤ -glucuronidase, 0.4 units sulfatase) for 2 h at 37°C. Samples were then acidifi ed and extracted as described above for urine.

Suitability of methods for detection of TOH metabolites in human fecal material
In an exploratory investigation, reference fecal material was obtained from an adult male following 14 days supplementation of either ␥ -TOH or ␣ -TOH (400 mg/kg/day). Samples were processed and analyzed for TOH metabolites as described above for mice.

Tissue TOH status of L-Cpr +/+ and L-Cpr
؊ / ؊ mice supplemented with ␥ -and ␦ -TOH ␣ -TOH concentrations in L-Cpr Ϫ / Ϫ mice were half those of wild-type mice in every tissue analyzed ( Fig. 2C ). ␦ -TOH concentrations in the lung, heart, and fat of L-Cpr Ϫ / Ϫ mice were, on average, 43% those of wild-type mice. No other signifi cant differences in tissue TOH status between genotypes were observed.    with the metabolites formed by liver microsomes from L-Cpr +/+ mice ( Fig. 3A-C ) using GC-MS in SIM mode. The rank order of activity (rate of metabolite production) by substrate for mouse intestinal microsomes was ␥ > ␦ >> ␣ and that for mouse liver was ␥ = ␦ >> ␣ . Mouse intestinal microsomes displayed -hydroxylase activity that was 10% that of mouse liver microsomes for the three substrates.

TOH--hydroxylase activity in human intestine microsomes and human liver microsomes
HLM and HIM displayed -hydroxylase activity toward ␦ -, ␥ -, or ␣ -TOH ( Fig. 4 ). The rank order of activity (rate of metabolite production) by substrate for HLM was ␥ > ␦ >> ␣ and that for HIM was ␥ = ␦ >> ␣ . HIM displayed -hydroxylase activity that was 20-30% that of HLM for the three substrates.

DISCUSSION
We previously reported that CYP4F14 is the major vitamin E--hydroxylase in the mouse and that its disruption in mice fed a soybean oil diet containing modest amounts of ␥ -TOH resulted in elevated tissue levels of ␥ -TOH relative to wild-type mice ( 16 ). The accumulation of ␥ -TOH occurred despite two counterbalancing mechanisms, namely, i ) higher excretion of unmetabolized ␥ -TOH and ii ) elevated excretion of novel -1 and -2 metabolites of ␥ -TOH. The previous studies also revealed the presence in mice of -hydroxylase activity not mediated by CYP4F14. The current work aimed to further characterize the role of vitamin E--hydroxylase in limiting tissue accumulation of TOH. Specifi cally, we investigated whether dietary supplementation with high levels of ␥ -and ␦ -TOH would overcome the counterbalancing effects and result in tissue enrichment above that seen with the previous soybean oil diet. Two experimental models were employed: mice with global disruption of Cyp4f14 and mice with liver-specifi c in tissue ␥ -and ␦ -TOH occurred in the absence of any overt detrimental effects. The combination of supplementation and Cyp4f14 disruption resulted in a 14-fold increase (186 nmol/g) in liver ␥ -TOH ( Fig. 1C ), illustrating the potential of this model to study the biological consequences of tissue enrichment with specifi c forms of vitamin E. Interestingly, female mice accumulated signifi cantly more vitamin E in many tissues compared with their male counterparts, irrespective of genotype. Due to the small sample size when segregated by gender, statistical analysis of metabolic cage data was not possible, although it appears that females neither metabolize vitamin E nor excrete unmetabolized TOH to a lesser extent than males. The reason for this gender effect and its consequences are unknown but warrants further investigation.
Unlike any other tissue in Cyp4f14 Ϫ / Ϫ and wild-type mice, adipose tissue displayed higher levels of ␦ -TOH compared disruption of Cpr , both supplemented with high levels of ␥ -and ␦ -TOH. Supplementation with ␥ -and ␦ -TOH resulted in substantial increases in tissue levels of these forms in both wild-type and Cyp4f14 knockout mice in comparison to mice previously fed an unsupplemented soybean oil diet ( 16 ). The increases in tissue levels (7-to 25-fold) were similar or greater than the increase in dietary levels (5-to 10-fold). Although the fold-difference in tissue ␥ -and ␦ -TOH between knockout and wild-type mice was similar to that previously observed in unsupplemented mice, the absolute difference in tissue levels was substantially greater. For example, the effect of Cyp4f14 disruption resulted in an increase of 14 nmol/g ␥ -TOH in the liver of unsupplemented mice ( 16 ) but an increase of 100 nmol/g in liver of supplemented mice ( Fig. 1C ), despite being a 2-fold difference between genotypes in both cases. The dramatic elevation Supplemented mice of both genotypes excreted substantial quantities of unmetabolized TOH in the feces, which constituted a signifi cant mechanism of resistance to tissue TOH accumulation. Whether this resulted from decreased intestinal absorption or increased biliary secretion of tocopherols is unknown. In the supplemented mice, fecal TOH elimination was a more important means of disposal of dietary TOH than was metabolic elimination. In wild-type mice, fecal TOH elimination was fi ve times that of whole-body metabolic disposal, whereas in knockout mice, fecal TOH disposal was 10-fold that of metabolic disposal. In this respect, fecal elimination of unmetabolized TOH served as a high-capacity mechanism of resisting tissue accumulation of dietary TOH under conditions of supplementation. These fi ndings suggest that supplementation was not able to completely overcome the mechanisms of resistance of tissue accumulation constituted by vitamin E metabolism and fecal TOH excretion.
Mice exhibit multiple hepatic TOH-hydroxylase activities in the absence of CYP4F14 that infl uence diet-induced tissue levels of TOH. Therefore, we utilized L-Cpr Ϫ / Ϫ mice, in which all hepatic microsomal CYP activity was absent. If vitamin E-metabolizing capacity was restricted to the liver, with ␥ -TOH, despite similar levels of these forms in the diet. One possible explanation for this fi nding is that adipose tissue may have vitamin E-metabolizing capability such that ␥ -TOH is being metabolized to a greater extent than ␦ -TOH. The concentration of ␦ -TOH was not affected by the disruption of Cyp4f14 , indicating that any vitamin E-metabolizing capability in the adipose tissue is through a CYP4F14-independent mechanism. We observed higher levels of unmetabolized fecal ␥ -TOH than ␦ -TOH in both Cyp4f14 Ϫ / Ϫ and wild-type mice, which could also be contributing to higher ␦ -TOH levels in adipose tissue.
Cyp4f14 disruption in supplemented mice resulted in substantial loss of whole-body vitamin E--hydroxylase activity (80%), the magnitude of which was similar to that previously observed in unsupplemented mice ( 16 ). Although not statistically signifi cant, the higher level of 11 ′and 12 ′ -OH metabolites excreted in the supplemented Cyp4f14 Ϫ / Ϫ mice of the current study resulted in attenuation of the overall metabolic defi cit, reducing the wholebody decrement in vitamin E metabolic activity to 40%. The counterbalancing effect of these alternative pathways remained an important mechanism in limiting tissue TOH accumulation in the context of supplementation. -hydroxylase activity toward several forms of vitamin E at levels approximately 10% that of the liver. TOH-hydroxylase activity present in the intestine was shown to be mediated entirely by CYP4F14, as Cyp4f14 Ϫ / Ϫ mice showed a complete absence of this activity. We conclude that mouse intestine, unlike the liver, has only one enzyme capable of TOH--hydroxylase activity. We confi rmed the presence of vitamin E hydroxylase activity in human small intestinal mucosa, validating the mouse as a model to investigate the role of the intestine in fi rst-pass vitamin E metabolism. The contribution of the intestine to whole-body vitamin E metabolic activity is at present unknown but could be investigated given the availability of intestinespecifi c Cpr knockout mice ( 20 ).
The -1 and -2 hydroxylations of ␦ -and ␥ -TOH represented quantitatively important mechanisms of vitamin E metabolism, accounting for up to 30% of whole-body vitamin E metabolites in wild-type mice. The relevance of this fi nding to human vitamin E metabolism was demonstrated by the identifi cation of all -0, -1, and -2 hydroxy metabolites of ␥ -TOH in human feces. We additionally identifi ed several fecal -hydroxy metabolites and both -1 and then these mice should completely lack the ability to metabolize vitamin E. However, in L-Cpr Ϫ / Ϫ mice supplemented with ␥ -and ␦ -TOH, whole-body metabolic capacity was only reduced by 70%, clearly demonstrating the presence of extra-hepatic , -1, and -2 hydroxylation activities. This is the fi rst report to show the existence of extra-hepatic vitamin E metabolic activity.
Despite signifi cant reductions in vitamin E metabolism in L-Cpr Ϫ / Ϫ mice supplemented with ␥ -and ␦ -TOH, tissue levels of all three TOH were similar to or actually lower than those of wild-type mice. This fi nding may have resulted from a reduced effi ciency of TOH absorption secondary to the inability to synthesize bile acids. Liver-specifi c disruption of Cpr has been shown to cause 90% reduced bile acid production due to the disruption of CYP7A1 in the liver, the rate-limiting step of neutral bile acid biosynthetic pathway ( 18 ). This feature additionally complicates the interpretation of the 70% reduction in whole-body in vivo metabolic capacity and could underestimate the magnitude of extra-hepatic metabolic activity.
Shin et al. reported Cyp4f14 expression in the small intestine ( 19 ); therefore, we evaluated that tissue and found