Pyridoxine-5--D-glucoside (PN-glucoside)⁴ is a major naturally occurring form of vitamin B-6 in fruits, vegetables and cereal-grains (Gregory and Ink 1987, Gregory and Sartain 1991, Kabir et al. 1983). Approximately half of the dietary vitamin B-6 intake of Americans is from plant sources (Kant and Block 1990), although this varies widely according to individual food selection patterns. The bioavailability of PN-glucoside as a source of vitamin B-6 depends primarily on the extent of in vivo enzymatic hydrolysis, which is the rate-limiting step in the metabolic utilization of PN-glucoside (Ink et al. 1986, Trumbo and Gregory 1988, Trumbo et al. 1988). Isotopic studies comparing the metabolism of PN-glucoside administered orally or by injection have indicated that most but not all of the hydrolysis occurs in the intestine in rats and humans (Gregory et al. 1991, Trumbo and Gregory 1988). PN-glucoside and pyridoxine (PN) have been found to undergo similar extents of intestinal absorption when administered orally in rats (Ink et al. 1986).

With respect to its utilization in vitamin B-6 metabolism, the bioavailability of PN-glucoside in rats has been found to be ~25% relative to PN in a metabolic study with radiolabeled compounds (Ink et al. 1986) and in a bioassay with nonlabeled compounds (Trumbo et al. 1988). Using a dual-label stable isotopic study, we later found that [²H₂ ]PN-glucoside exhibited ~58% bioavailability in men, relative to simultaneously administered [²H₅ ]PN (Gregory et al. 1991). The relative bioavailability of PN-glucoside varies among species, apparently as a function of differences in intestinal pyridoxine-5--D-glucoside hydrolase activities (Banks and Gregory 1994).

In addition to the role of PN-glucoside as a source of partially available dietary vitamin B-6, PN- glucoside is also a naturally occurring antagonist that weakly inhibits the metabolic utilization of simultaneously ingested PN. In radioisotopic studies with rats, we have found that PN-glucoside partially interferes with the metabolism of co-ingested PN (Gilbert and Gregory 1992), and that this apparent competitive effect of PN-glucoside is transient and consistent with the short in vivo residence time of intact PN-glucoside (Nakano and Gregory 1995a). This interaction was confirmed on a more long-term basis with rats by chronically feeding PN-glucoside in the presence or absence of PN (Nakano and Gregory 1995b), which indicated the nutritional significance of this phenomenon. In vitro studies with isolated rat hepatocytes indicated that PN-glucoside and PN compete for the same transport system and that PN-glucoside competitively inhibits the uptake of PN by hepatocytes (Zhang et al. 1993). These in vitro observations and their correlation with the time course of in vivo observations provide evidence of the mechanism of this antagonistic effect.

The objectives of the present studies with human subjects were as follows: 1 ) to reevaluate the bioavailability of PN-glucoside relative to PN by administering deuterium-labeled PN or PN-glucoside singly to humans to eliminate the possibility of interactions that may occur in a dual-label protocol as employed previously; and 2 ) to determine the extent of any in vivo antagonism of PN-glucoside on the metabolic utilization of PN in humans, as seen previously in rats, using quantities of administered PN-glucoside within the range encountered in typical diets. These studies were conducted with a protocol that permitted examination of the time course of the metabolic utilization of PN and PN-glucoside and any antagonistic effect of PN-glucoside.

SUBJECTS AND METHODS

Trials 1 and 2 involved the administration of a single dose of [²H₂ ]PN or [²H₂ ]PN-glucoside to assess the bioavailability of deuterium-labeled PN-glucoside compared with deuterium-labeled PN. In Trials 3 and 4, [²H₂ ]PN was administered along with two amounts of nonlabeled PN-glucoside to evaluate the dose dependence of the inhibitory effect of PN-glucoside on co-ingested PN in humans. Trial 1 served as a reference in the assessment of [²H₂ ]PN-glucoside bioavailability and as a trial without PN-glucoside in the assessment of the antagonistic effects of PN-glucoside. The doses used in the four trials are summarized in Table 1 and described in more detail below.

Table 1. Summary of doses of vitamin B-6 compounds administered to human subjects1 [View Table]

In Trial 1, subjects were administered an oral dose of 5 µmol [²H₂ ]PN and 1.93 µmol nonlabeled PN (total of 6.93 µmol PN) in 50 mL of apple juice. In Trial 2, subjects were given a dose of 8.62 µmol [²H₂ ]PN-glucoside and 3.33 µmol nonlabeled PN-glucoside. For Trial 3, PN-glucoside was provided at a relative molar concentration of 15% of total vitamin B-6 by the administration of 5 µmol [²H₂ ]PN and 1.29 µmol nonlabeled PN, along with 1.11 µmol nonlabeled PN-glucoside. In Trial 4, PN-glucoside constituted 40% of the total vitamin B-6 (molar basis) in a dose comprised of 5 µmol [²H₂ ]PN and 3.33 µmol nonlabeled PN-glucoside. Calculation of all doses was based on an assumed 58% biological equivalence of PN-glucoside and PN to maintain total ingested available vitamin B-6 of 6.93 µmol. This dose was considered reasonable and relevant to normal human diets in view of the Recommended Dietary Allowances for men and women of 2.0 and 1.6 mg, corresponding to 11.8 and 9.47 µmol, respectively (Food and Nutrition Board 1989). This dosage regimen was developed to provide the same amount of biologically available vitamin B-6 in each trial on the basis of the assumption of 58% relative molar bioavailability of PN-glucoside (Gregory et al. 1991).

During each trial, the subjects were asked to collect all urine for the 24-h predose period beginning the morning before administration of labeled compounds. This sample was useful as an indicator of vitamin B-6 nutritional status by measurement of 4PA concentration. All urine samples were collected in 1000-mL amber-color polyethylene bottles and kept refrigerated. Between 0730 to 0800 h the next morning, before consumption of any food, subjects were given their respective oral dose of vitamin B-6 (Table 1), which was mixed in 50 mL of pasteurized apple juice. After consuming this portion of juice, a second 50-mL portion of juice containing no added vitamin B-6 compounds was used to rinse the cup and then consumed by the subject. Urine collections were continued for the following 48-h postdose period. During the first 24 h postdose, urine from each 8-h interval was kept separately; urine was collected for the entire 24-h period on the second day postdose. After measuring urine volume, urine samples were divided into six portions and stored at 20°C until analysis.

Mass spectral analysis of deuterium-labeled 4-pyridoxic acid. The isolation and purification of 4PA from urine samples were performed according to a modification of the method of Hachey et al. (1985) as described by Gregory et al. (1991). Urine (20 mL) was acidified to pH 2.0 with HCl and applied to a 20 cm × 0.7 cm i.d. column packed with Bio-Rad AG 50W-X8 resin (100-200 mesh, H⁺ form, Bio-Rad Laboratories, Richmond, CA) that had been equilibrated with 0.1 mol/L acetic acid. The column was washed with 20 mL of 0.1 mol/L acetic acid and then 4PA was eluted with 0.2 mol/L triethylamine. The fractions containing 4PA were pooled, then evaporated to dryness at 60°C under a stream of nitrogen gas. Final purification was achieved by a reversed-phase HPLC (Partisil 10 ODS-3 Magnum 9 column, 9 mm i.d. × 25 cm, Whatman, Clifton, NJ) with 0.1 mol/L formic acid, pH 2.0, containing 3% (v/v) acetonitrile as a mobile phase. The peak corresponding to 4PA was collected manually and evaporated to dryness. Samples were prepared for gas chromatography-mass spectrometry (GCMS) analysis by conversion of the isolated 4PA to the 3-O-acetyl-4-pyridoxic acid lactone derivative as described by Hachey et al. (1985).

GCMS analysis was performed in the electron capture negative ionization mode with selected-ion monitoring at m/z 207 and 209 for nonlabeled and [²H₂ ] forms (Gregory et al. 1991). GCMS analysis was performed by Metabolic Solutions (Merrimack, NH).

To facilitate quantitative analysis, 4PA standards were prepared as 4PA lactone from nonlabeled PN and PN-glucoside, and [²H₂ ]PN and [²H₂ ]PN-glucoside according to Gregory and Kirk (1977) except that PN-glucoside was hydrolyzed to PN by -glucosidase (200 U/74 mg, Boehringer Mannheim, Indianapolis, IN) before the reaction with manganese dioxide. All 4PA standards were purified by reversed-phase HPLC (Partisil 10 ODS-3 Magnum 9 column, 9 mm i.d. × 25 cm, Whatman, Clifton, NJ) with 0.1 mol/L formic acid, pH 2.0, containing 3% (v/v) acetonitrile as a mobile phase, and prepared for GCMS as described above. Simultaneous equations were prepared to relate the ratio of observed GCMS peak areas to actual molar ratios of isotopomers. The urinary excretion of each isotopic form of 4PA ([¹H] and [²H₂ ]) was calculated using the isotope ratios determined by GCMS and the concentration of total urinary 4PA determined by HPLC. Excretion data were also normalized as percentage of dose. To facilitate comparison of excretion of [²H₂ ]4PA over the 48-h postdose period in which the time interval of urine collection differed (i.e., three 8-h and one 24-h collections), excretion data were also expressed as relative rate (percentage of dose excreted per hour). Cumulative excretion over the 48-h postdose period was determined as the sum of [²H₂ ]4PA excretion values in all collection periods.

RESULTS

Table 2. Urinary excretion of total 4-pyridoxic acid at various time intervals in collected urine before or after test doses by human subjects1 [View Table]

Relative bioavailability of [²H₂ ]pyridoxine-5--D-glucoside and [²H₂ ]pyridoxine.

Table 3. Molar ratio of deuterium-labeled ( [2H2] )-4-pyridoxic acid and nonlabeled 4-pyridoxic acid at various time intervals in collected urine after dose of either [2H2] pyridoxine or [2H2]pyridoxine glucoside by human subjects1,2 [View Table]

Table 4. Total 48-h excretion of [2H2]4-pyridoxic acid, expressed as percentage of oral dose of [2H2]pyridoxine or [2H2]pyridoxine-5--D-glucoside by human subjects1,2 [View Table]

Influence of pyridoxine-5--D-glucoside on metabolism of [²H₂ ]pyridoxine.

DISCUSSION

PN-glucoside accounts for 5-80% of the naturally occurring total vitamin B-6 in fruits, vegetables and cereal-grains (Gregory and Ink 1987, Gregory and Sartain 1991, Kabir et al. 1983). Although other glycosylated forms of vitamin B-6 have been identified and may be quantified, the pyridoxine-5--D-glucoside derivative used in this study is the predominant form. Analysis of composites from studies involving typical mixed diets in the U.S. have indicated that glycosylated vitamin B-6 frequently constitutes ~15% of the total vitamin B-6 (Andon et al. 1989, Gregory et al. 1991). Within any particular diet, the actual percentage of PN-glucoside may vary widely depending on the selection of food items constituting the major sources of vitamin B-6. The quantities and relative percentages of PN-glucoside in doses used in these studies were chosen to be relevant to typical dietary intakes.

This study indicated a relative bioavailability of PN-glucoside of ~50% in humans on the basis of a protocol that eliminated any possible interactive effects between PN-glucoside and other forms of vitamin B-6. This confirms and extends the previous study in which PN-glucoside bioavailability was found to be ~58% by using a protocol in which PN-glucoside and PN were administered simultaneously (Gregory et al. 1991).

It should be noted that the subjects of this study were selected on the basis of adequate vitamin B-6 nutritional status. We have recently reported that the activity of jejunal cytosolic pyridoxine--D-glucoside hydrolase in rat intestine increases as a function of the extent of vitamin B-6 deficiency (Banks et al. 1994, Nakano and Gregory 1995b). Pyridoxine--D-glucoside hydrolase has been identified as a distinct, highly specific and apparently novel species of soluble -glucosidase in mammalian intestinal mucosa (Nakano et al. 1995), with a function as the major site of postabsorptive hydrolysis of dietary PN-glucoside. Whether such an induction of pyridoxine--D-glucoside hydrolase occurs in humans during inadequate vitamin B-6 status and whether the bioavailability of dietary PN-glucoside would be increased in vitamin B-6 deficiency in humans have not been determined and will be examined in future studies.

With respect to the previously reported metabolic antagonism between PN-glucoside and PN (Gilbert and Gregory 1992, Nakano and Gregory 1995a, Zhang et al. 1993), the present study has indicated that this effect is kinetically identifiable in humans. Considering that the peak excretion of [²H₂ ]4PA was significantly depressed only at the 40% level of PN-glucoside and that the net excretion of [²H₂ ]4PA was not significantly altered by PN-glucoside, we conclude that the antagonistic effect of PN-glucoside on PN utilization in humans consuming typical diets is probably small. This finding is in contrast to observations in rats, in which effects on 4PA excretion from a given dose of radiolabeled PN and on the in vivo pattern of labeled vitamin B-6 metabolites were much more pronounced (Gilbert and Gregory 1992, Nakano and Gregory 1995a) and in a study involving chronic feeding of PN-glucoside in rats (Nakano and Gregory 1995b).

We hypothesize that the differences in bioavailability and in the antagonistic effect of PN-glucoside are due to the difference between humans and rats in intestinal mucosal cytosolic pyridoxine--D-glucoside hydrolase activity, which has been found to be much higher in humans than in rats (Trumbo et al. 1990). In contrast to the ~50% bioavailability observed in humans, rats exhibit ~25% bioavailability as described previously. Assuming effective absorption of PN-glucoside in each species, this suggests that the concentration of intact PN-glucoside to which liver cells are exposed after absorption would be greater in rats than in humans in such isotopic protocols. Thus, the inhibitory action of PN-glucoside on the hepatic cellular uptake of other forms of vitamin B-6 would probably be subject to greater inhibition in rats than in humans. Whether the antagonistic effect of PN-glucoside may also involve competitive inhibition of the renal reabsorption of PN has not been determined.

Several recent studies provide relevant information regarding the interpretation of this study. Hansen et al. (1996) conducted a study in which adult women were fed diets of equivalent total vitamin B-6 content but containing either high (27%) or low (9%) percentages of PN-glucoside from food sources in a cross-over design. During the 18-d period of the high PN-glucoside diet, the subjects exhibited a reduction in urinary excretion of 4PA and total vitamin B-6 as well as plasma PLP values equivalent to a 15-18% reduction in intake of available vitamin B-6. This observation is fully consistent with our findings regarding the ~50% bioavailability of PN-glucoside in the present study. Hansen et al. (1996) also reported an unexpected significant reduction in erythrocyte PLP concentration when feeding the high PN-glucoside diet. Hansen et al. (1996) hypothesized that this effect of diet on erythrocyte PLP might have been due to an antagonistic effect of PN-glucoside as seen in isolated liver cells (Zhang et al. 1993).

Cheng and Trumbo (1993) reported that the bioavailability of PN-glucoside in pregnant rats was nearly equivalent to that of PN. Although their study did not include nonpregnant controls, this suggests that pregnancy may influence the bioavailability of PN-glucoside and should be examined in women during pregnancy. In the present study, there was no significant difference between the relative bioavailability of PN-glucoside between male and nonpregnant female subjects.

Although there is no evidence that pyridoxine--D-glucosides occur naturally in foods, this isomer has been found to be capable of being taken up by isolated liver cells (Joseph et al. 1996) and to exhibit greater in vivo utilization in rats as a source of available vitamin B-6 than that of the -D-glucoside (Tsuge et al. 1996). Tsuge et al. (1996) also found that pyridoxine--D-glucoside undergoes intestinal absorption largely intact, without hydrolysis, which is consistent with its limited bioavailability in rats observed previously in our laboratory.

In summary, the results of this study provide new quantitative information that extends our understanding of the role of PN-glucoside in human nutrition. PN-glucoside serves as a source of nutritionally available vitamin B-6, with ~50% bioavailability relative to PN. PN-glucoside also exerts an antagonistic effect analogous to that observed in rats, although of lesser magnitude. In this regard, PN-glucoside reduces the rate of PN utilization of co-ingested PN, although it does not alter the overall extent of utilization. This application of stable isotopic techniques has yielded data relevant to assessing the adequacy of various diets in meeting nutritional requirements for vitamin B-6 in humans.