Nutritional status for any specific nutrient depends on nutrient intake and the efficiency of absorption and metabolic utilization, relative to the requirement. Incomplete bioavailability of dietary folate can play a significant role, affecting folate nutritional status (Cuskelly et al. 1996). The development of methods for the synthesis of folates labeled with stable isotopes has provided important tools for the study of folate absorption and metabolism in human beings (Gregory 1989, 1995 and 1997) and has offered new insight into the relationship between diet composition and folate bioavailability (Pfeiffer et al. 1997, Wei et al. 1996). In spite of the power of these methods, their application in evaluating factors affecting folate absorption and postabsorptive metabolism will not reach full potential until additional protocols are evaluated and optimized. Particular needs include the following: 1) synthesis and evaluation of additional isotopically labeled folates to expand the array of potentially usable tracers for in vivo studies; 2) evaluation of the relative merits of urinary excretion of labeled folates vs. determination of short-term labeling of plasma folates for quantification of absorption in bioavailability studies; and 3) determination of the merits of protocols not involving prior administration of nonlabeled folate to achieve "saturation" of subjects' tissues. The main questions regarding the saturation procedure involve the precision of protocols conducted without saturation and whether the high concentration of biliary folate caused by enterohepatic circulation of the saturation doses may alter the rate or extent of absorption of orally administered reference or test doses. Although the saturation procedure involves a nonphysiologic intake of folate before isotopic studies, this procedure is quantitatively useful because it enhances excretion of labeled folates and improves the precision of the protocol; thus, resolution of these issues is required.

The development of improved protocols for the study of folate bioavailability is important in addressing key issues such as the relative bioavailability of dietary folate vs. supplemental folic acid, the efficacy of fortification procedures and the effect of various genetic, age or disease states on aspects of folate absorption and utilization. Such information is essential in evaluating the adequacy of folate intake in reducing the incidence of neural tube defects (CDC 1992, Scott et al. 1995) and the risk of various forms of vascular disease (Boushey et al. 1995, Green and Jacobsen 1995) and certain cancers (Mason 1995).

Quantification of isotopic enrichment by mass spectral methods is facilitated when the mass of the tracer used is at least two units greater than that of the nonlabeled compound. Initial methods of preparing folates with stable isotopes were limited, in terms of in vivo application, because of the single labeling achieved (see, for example, Pastore 1980, Plante et al. 1980). Methods for the synthesis of folates labeled with stable isotopes have been discussed in two reviews (Gregory 1989, Pfeiffer and Gregory 1997). Hachey et al. (1978) first described a method for preparing folic acid labeled with deuterium at the 3 and 5 positions. To permit maximum flexibility in stable isotopic studies of folate bioavailability and metabolism, our laboratory has addressed the synthesis, analysis and application of folates with labeling at the 3 and 5 positions as well as on the glutamate moiety, for example, [3,5-²H₂]folic acid ([²H₂]folic acid),⁴ [3,5-²H₂]folyl polyglutamates and [glu-²H₄]folic acid. Our method of [²H₂]folic acid synthesis is based on catalytic dehalogenation for 3,5 labeling of folic acid (Gregory 1990, Gregory and Toth 1988a) as a somewhat more rapid alternative to the method of Hachey et al. (1978). Analogous 3,5 labeling of pteroic acid permits convenient synthesis of [²H₂]folyl polyglutamates (Pfeiffer and Gregory 1997). [²H₄]Folic acid is prepared by coupling pteroic acid with commercially available L-[²H₄]glutamic acid (Gregory and Toth 1988b). To allow the synthesis of a folate having M+5 labeling (i.e., 5 mass units greater than the nonlabeled form), and facilitated by the commercial availability of L-[¹³C₅]glutamic acid, we recently modified the method for [²H₄]folic acid synthesis and adapted it to the preparation of [glutamate-¹³C₅]folic acid ([¹³C₅]folic acid) (Pfeiffer and Gregory 1997). Although in vivo data suggest equivalent intestinal absorption, metabolism and excretion of [²H₄]folic acid and [²H₂]folic acid, such direct comparisons of [²H₂]folic acid and [¹³C₅]folic acid have not been previously conducted.

For comparative studies of oral bioavailability or fractional absorption of many nutrients, administration of simultaneous oral and intravenous doses followed by evaluation of kinetics of plasma isotopic enrichment allows the most direct determination. Such an approach could not be conducted with labeled folates initially because of limits of detection, although current gas chromatography mass spectrometry (GCMS) instruments permit this analysis. If total clearance of a drug (or labeled nutrient) administered by oral and intravenous routes is equivalent, then comparison of area under the curve provides a direct indication of fractional absorption or bioavailability (Ritschel 1992, Rowland and Tozer 1989). We were greatly interested in plasma folate kinetics as a possibly useful index of folate bioavailability in stable isotopic protocols.

In this paper, we report studies relevant to the design of future in vivo protocols using folates labeled with stable isotopes. These studies were conducted to evaluate the following: 1) the metabolic equivalence of [¹³C₅]folic acid and [²H₂]folic acid when administered orally, and 2) the short-term kinetics of plasma folate enrichment for studies of folate bioavailability in human subjects. In addition, the merits of isotopic protocols conducted without prior administration of nonlabeled folic acid to subjects (i.e., "folate saturation") are considered.

SUBJECTS AND METHODS

Experimental protocols with human subjectsoverview.

Subjects were free-living and during the trial periods (1 d predose, 1 d postdose) were advised to consume a self-selected diet that avoided foods high in folate. Two trials were conducted in these studies, one with oral doses of labeled compounds and one with oral and intravenous administration of labeled folates, as described in detail below.

For 24 h before administration of the labeled compounds (d 1), beginning at 0700 h, subjects collected all urine into opaque 2-L plastic bottles each containing 3 g solid sodium ascorbate. In addition, 500 mg of dithiothreitol was added to the urine before subsampling for further protection against folate oxidation. Urine was refrigerated to minimize bacterial growth and subsamples frozen daily after saturation with nitrogen gas. The predose urine collection allowed the assessment of any isotopic carry-over between the two trials. After administration of the test doses on d 2 of each trial, subjects collected all urine for the ensuing 24-h period.

The isotopic distribution of total urinary folate (i.e., ¹H, ¹³C₅ and ²H₂) was determined after affinity chromatographic isolation of total folate, cleavage of the 9C-10N bond, HPLC isolation of the p-aminobenzoylglutamate (pABG) fragment and derivatization using trifluoroacetic anhydride and trifluoroethanol by gas chromatography-mass spectrometry (GCMS) in the electron-capture negative chemical ionization mode (Gregory and Toth 1988b, Toth and Gregory 1988). Peak areas at charge-to-mass ratios (m/z) of 426, 428 and 431, corresponding to the molecular ions from nonlabeled (¹H), ²H₂ and ¹³C₅ isotopomers, respectively, were determined in this analysis. The molar ratios of ¹³C₅/¹H and ²H₂/¹H species in samples were then calculated for the observed ratios of GCMS peak areas by using simultaneous equations that corrected for the natural abundance of the stable isotopes. Total urinary excretion of [¹³C₅] and [²H₂]folates was calculated from the daily urine volume, total folate concentration and the ratios of ¹³C₅ and ²H₂ to nonlabeled folate. Urinary excretion ratios were calculated as follows: (% of [¹³C₅]folate dose excreted/24 h)/(% of [²H₂]folate dose excreted/24 h).

RESULTS

Table 1. Urinary folate excretion by human subjects during 24-h predose and 24-h postdose periods following administration of [13C5] and [2H2]folic acid orally (Trial A) or by oral and intravenous routes (Trial B)1 [View Table]

Table 2. Molar ratios and excretion of urinary folate isotopomers over 24-h postdose period by human subjects after administration of [13C5] and [2H2]folic acid orally (Trial A) or by oral and intravenous (IV) routes (Trial B)1 [View Table]

To facilitate a comparison of the excretion behavior between the two trials involving oral vs. intravenous administration of the reference dose of [²H₂]folic acid, we also expressed molar ratios as urinary ¹³C₅/²H₂ excretion ratios (% of dose of ¹³C₅ folates/% of dose of ²H₂ folates), which adjusted excretion data for the size of the doses. For Trial A, this excretion ratio was 0.96 ± 0.055. A ratio of 1.0 would indicate equivalent in vivo behavior of the [¹³C₅] and [²H₂] forms of folic acid administered. Within the precision of this study, there was no apparent difference in absorption or in vivo behavior of these two forms of labeled folic acid (Table 2).

The ¹³C₅/²H₂ molar ratios for urinary folates (29.1 ± 17.0) exhibited a high degree of variability among subjects but tended to exceed the ratio of 4.58 of the administered folates (Table 2). These ratios for excreted and administered folates were not significantly different largely because high between-subject variability and low subject number yielded low statistical power. Examination of the data indicated that this variability in the ¹³C₅/²H₂ ratio was due mainly to the variation in excretion of [²H₂]folates derived from the intravenously injected [²H₂]folic acid.

As in Trial A, the relative extent of 24-h postdose excretion in Trial B was small (2.42 ± 0.88% of dose for [¹³C₅]folate and 1.29 ± 0.94% of dose for [²H₂]folate; Table 2). The urinary excretion ratio (% of dose of ¹³C₅ folates/% of dose of ²H₂ folates) exceeded 1.0 in all subjects, indicating greater relative excretion of the orally administered [¹³C₅]folate (Table 2). Although the mean excretion ratio of Trial B was not significantly >1.0 and the difference between excretion ratios of Trial A and Trial B did not reach significance (P = 0.19), a paired t test of excretion values (% of dose) for [¹³C₅] and [²H₂]folates indicated a significant difference in Trial B. These results indicated greater urinary excretion of folates derived from the orally administered tracer than from the injected folate in this protocol. Differences in urinary excretion (% of dose) between trials were small for each form of labeled folate and were not significant.

DISCUSSION

To our knowledge, short-term kinetics of plasma folate have not been examined previously using stable-isotope labeling. Similar studies of plasma or serum folate concentrations after oral and parenteral doses of nonlabeled 5-formyl-tetrahydrofolate (i.e., folinic acid, leucovorin) have been reported, generally with much higher doses (Anderson et al. 1992, Bunni et al. 1989, Wolfram et al. 1990, Zittoun et al. 1993); plasma kinetics after an intramuscular dose and oral doses of folic acid also have been reported (Loew et al. 1987, Menke et al. 1993, Schuster et al. 1993). In addition, Bailey et al. (1988) evaluated serum folate area under the curve 8 h postdose as an indicator of folate bioavailability and found that large doses (1.13 µmol, 500 µg) were required to elicit a consistent response. Such studies of short-term changes in plasma folate concentration or urinary excretion provide no information about the in vivo fate of the administered dose because the plasma and urinary folate pool contains both newly absorbed folates as well as those derived from previous dietary intake and turnover of body folate pools.

In Trial B, the oral dose ([¹³C₅]folic acid) yielded a minimal response in plasma relative to the injected dose of [²H₂]folic acid. This could be due to incomplete absorption, although previous studies involving intragastric administration of [³H]folates in food-deprived rats indicated that intestinal absorption was nearly complete (Bhandari and Gregory 1992). In addition, the apparent bioavailability of folic acid in an oral supplement was estimated to be 76-97% in humans (Schuster et al. 1993). A more probable explanation of the limited short-term appearance of orally administered [¹³C₅]folate in plasma is the avid uptake of newly absorbed folates by the liver and the active enterohepatic circulation. Steinberg (1984) reported that 10-20% of reduced folates (derived from mucosal reduction of absorbed folic acid) in portal blood is taken up by the liver during the first pass after absorption in rats, whereas hepatic first-pass uptake of unchanged folic acid is much greater. Steinberg (1979) also estimated that approximately half of the folates reaching peripheral tissues had passed through the enterohepatic circulation. This study provides direct kinetic evidence of a first-pass effect in humans. Because of this phenomenon, we believe that studies involving serial measurements of folate enrichment in plasma are not conducive to evaluation of folate bioavailability in foods or test doses containing folates labeled with stable isotopes in nutritionally relevant amounts.

Bioavailability is operationally defined in pharmacokinetic methods as the ratio of area under the curve derived from an oral (i.e., extravascular) dose to the area under the curve derived from an intravenous reference dose (Rowland and Tozer 1989). This definition is valid only if the total clearance is constant, i.e., independent of the route of administration, which is not the case for folates. Although this phenomenon does not invalidate the plasma kinetic approach entirely, the use of a control trial involving an oral reference dose would be required. It should also be noted that the route of administration of the labeled folates affects the extent of short-term excretion patterns in this protocol without saturation to an even greater extent than seen in trials using saturation (Gregory et al. 1992). The ¹³C₅/²H₂ urinary excretion ratio of Trial B, which involved oral [¹³C₅] and injected [²H₂]folic acid, was quite variable (6.49 ± 3.81). Although this ratio did not significantly exceed 1.0 with this degree of variability and low number of subjects, a paired t test indicated a significant difference in urinary excretion (% of dose) for [¹³C₅] and [²H₂]folates in this trial. This phenomenon (greater percentage of urinary excretion derived from the orally administered form of folic acid) was observed in a similar study conducted without saturation (Pfeiffer et al. 1997). It was also observed in a protocol using prior saturation when comparing injected labeled [²H₂]folic acid and orally administered [²H₄]folic acid, although the excretion ratio was ~1.0 for trials in which [²H₄]tetrahydrofolates (e.g., 5-methyl, 5-formyl, and 10-formyl) were administered (Gregory et al. 1992). These findings are consistent with those of Steinberg (1979) who reported a greater hepatic first-pass uptake of nonreduced folic acid from portal blood of rats, relative to 5-methyl-tetrahydrofolate. In view of these previous results and the elevated urinary excretion ratio observed in Trial B of this study, indicating a strong and consistent tendency for greater short-term urinary excretion of labeled folates derived from the oral versus the intravenous dose, users of protocols such as these are urged to use care in the design of studies by incorporating appropriate trials to control for such effects.

In this study, the urinary excretion of intact labeled folates was found to be ~2% of the administered dose for each tracer over 24 h postdose. Complete recovery of the isotopically labeled dose as intact urinary folate is not expected as a result of extensive in vivo retention, enterohepatic circulation, catabolism and fecal excretion. In this protocol, absolute recoveries are not required because the response is determined relative to the urinary excretion of a reference dose (i.e., [²H₂]folate) injected immediately after consumption of the oral [¹³C₅]folic acid dose. Many of our previous stable-isotopic protocols for the study of folate bioavailability have involved prior "saturation" of subjects by daily administration of supplemental nonlabeled folic acid before the study and between trials. Such a saturation regimen is intended to improve precision through increasing folate excretion and reducing variability in excretion, while also increasing the rate of excretion of the labeled folate(s). The disadvantage of saturation procedures is that normal absorption and metabolic disposition may not occur because of the resulting large biliary secretion of folate into the intestine, reduced tissue retention and enhanced urinary excretion of folates.

In this study, which did not involve folate saturation, high subject-to-subject variations in absolute urinary folate excretion as well as in isotopic excretion ratios (as indicators of relative bioavailability) constituted limitations of this protocol. We used a small number of subjects because this study was intended only to evaluate several procedural variables. Shortly after this study, we conducted a series of seven similar dual-label trials to evaluate the absorption of folate from supplements and fortified cereal-grain foods without prior saturation (Pfeiffer et al. 1997). Although this study provided evidence of effective absorption of folic acid under these conditions, the within-group variations inherent in this protocol reduced the statistical power of the study, thus limiting the ability to detect small differences among treatments when using 14 subjects.

In summary, the results of this study provide further understanding of various details of protocols for the investigation of folate absorption using stable-isotopic methods. We have shown that [¹³C₅]folic acid and [²H₂]folic acid exhibit highly similar in vivo behavior when administered by the same route and thus are well suited for use in vivo. We also have shown that measurement of plasma kinetics is feasible for short-term evaluation of factors affecting folate absorption, although this technique would be insensitive due to hepatic retention and enterohepatic circulation. This study has provided support for the feasibility of conducting short-term studies of absorption without the use of prior saturation with nonlabeled folate. However, in view of the undesirably low precision seen with this type of protocol, we recommend that protocols involving multiple doses of labeled folate(s) be considered. These will be evaluated in our future research.