The Journal of Nutrition Vol. 128 No. 2 February 1998, pp. 280-286

The Mongolian Gerbil (Meriones unguiculatus) Is an Appropriate Animal Model for Evaluation of the Conversion of -Carotene to Vitamin A^1,2,3

Christine M. Lee, Janine D. Lederman, Nicolle E. Hofmann, and John W. Erdman Jr.^{*, 4}

Department of Food Science and Human Nutrition, ^* Division of Nutritional Sciences, University of Illinois, Urbana, IL 61801

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Vitamin A (VA) deficiency is the leading cause of blindness in children in developing countries. Dietary intervention with foods rich in provitamin A carotenoids, such as -carotene (C), has been suggested as one solution to this problem. The objective of the two studies described in this paper was to examine the utilization of C as a source of VA at different stages of VA depletion using the Mongolian gerbil as a model. Male 4- to 5-wk-old Mongolian gerbils were fed powdered C-free semipurified diets either with or without VA for 26 d (Study 1), or without VA for 8-10 wk (Study 2). Gerbils were then fed diets with or without VA (20.9 nmol/g diet) and/or C [(67.0 µmol/g diet (Study 1) and 145.9 µmol/g diet (Study 2)] for variable periods. Two (Study 1) or three (Study 2) days before termination of the study, 3-4 gerbils per group were dosed orally with ¹⁴CC. Tissues were evaluated for VA and C content by HPLC. Liver was extracted with and without saponification to evaluate ¹⁴CC and ¹⁴CVA content. The results demonstrate the following: 1) the gerbil is an appropriate animal model to study C utilization; 2) 20.9 nmol VA/g diet is more than sufficient for this species; 3) the daily VA utilization rate for this species is calculated to be 3.1 µg/100 g body weight; 4) a highly bioavailable source of C at a 6:1 weight ratio of C:VA is sufficient to reverse marginal VA status in this model; and 5) a highly bioavailable source of C fed between a 6:1 and 13:1 weight ratio to VA provides equivalent VA status as preformed VA in Mongolian gerbils.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Vitamin A (VA)⁵ deficiency, especially among children, is a major public health concern in many developing countries (Sommer and West 1996). Areas with high rates of VA deficiency are characterized by low consumption of preformed VA (FAO/WHO 1988) and reliance on carotenoids to satisfy the VA requirement. To make public health recommendations regarding the use of carotene-rich foods to reduce VA deficiency, it is important to know the conversion efficiency of -carotene (C) to VA in individuals with low VA status and to be certain that C can be utilized by individuals deficient in VA. Recently, there has been considerable concern regarding the low efficiency with which pro-vitamin A carotenoids are utilized from foods (dePee et al. 1995, Solomons and Bulux 1993, Solomons 1996), especially for those with marginal-to-low VA status (dePee et al. 1995).

Animal models are useful alternatives to human studies when investigating mechanisms of carotene absorption and cleavage (van Vliet 1996). Ferrets (Ribaya-Mercado et al. 1989, White et al. 1993) and preruminant calves (Bierer et al. 1995, Poor et al. 1992) have recently been used for such studies. The Mongolian gerbil has shown promise as an appropriate model to study C metabolism, because this species readily absorbs C intact after a meal containing physiologic levels of C (Pollack et al. 1994).

The purpose of the two studies described here was to use the gerbil model to examine the utilization of C for VA value at different levels of VA depletion. Gerbils were depleted to above and below hepatic stores of 70 nmol/g (20 µg/g), a concentration considered to be the upper level cut-off indicating marginal VA deficiency in humans and in experimental animals such as rats (Olson 1991).

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Animals.  Male 4-wk-old Mongolian gerbils with average weights of 32.7 and 28.0 g for Study 1 and 2, respectively, were obtained from Tumblebrook Farms (Brant Lake, NY) for Study 1 and Harlan Sprague Dawley (Indianapolis, IN) for Study 2; they had been fed a commercial, nonpurified diet (NIH 31) postweaning. To accustom the gerbils to powdered feed, they were fed upon arrival a commercial, nonpurified diet (Purina Mills, St. Louis, MO), which was ground to a fine powder. The gerbils had free access to food and water, and room lighting was on a diurnal cycle (light 0700-1900 h). All animal handling procedures were approved by University of Illinois Laboratory Animal Care Advisory Committee.

Diets.  After the gerbils acclimated to the powdered commercial diet, they were fed a semi-purified diet prepared in our laboratory. This diet has been used successfully in our laboratory for previous feeding studies with gerbils (Pollack et al. 1994). -Carotene (10% water-soluble beadlets) and/or VA, provided as retinyl palmitate (Palmabeads), both generous gifts from Hoffmann La Roche (Nutley, NJ), were added to the appropriate experimental diets. Diets were stored at 4°C until use.

Experimental design. 

Study 1.  After a 6-d acclimation period, gerbils were randomly assigned to receive a semipurified diet that was either VA deficient (VA) (n = 38) or VA sufficient (+VA) (n = 38), containing 20.9 nmol VA/g diet, for 26 d (Fig. 1A). Eight gerbils from each group were killed by cervical dislocation after cardiac puncture (described below) at the end of the prefeeding period (branchpoint) to determine hepatic VA concentration. The remaining gerbils were randomly subdivided into groups and fed powdered diets supplemented with VA (20.9 nmol/g diet) and/or C (67.1 nmol/g diet) for 31 d, representing an ~6:1 weight ratio of C:VA. Forty-eight hours before termination of the study, 3-4 gerbils from each group were given orally 0.18 MBq of ¹⁴CC, specific activity 0.16 TBq/mmol (a gift from Hoffmann La Roche) by gavage, after 12 h of food deprivation.

View larger version (25K): [in this window] [in a new window]   Fig 1. Study design for gerbil Studies 1 (A) and 2 (B). For Study 1, the appropriate diets were supplemented with Vitamin A (VA) or -carotene (C) at 20.9 and 67.1 nmol/g diet, respectively. For Study 2, the appropriate diets were supplemented with VA or C at 20.9 and 145.9 nmol/g diet, respectively.

At the time of killing, blood samples were collected via cardiac puncture in gerbils anesthetized with ketamine hydrochloride/xylazine (95:5, v/v) (Vetlar, Parke-Davis, Morris Plains, NJ, Rompun, Miles Laboratory, Shawnee, KS, respectively) delivered by intramuscular injection, as recommended by the campus veterinarian (0.1 mL/100 g body weight). Liver, kidney and adrenal tissues were removed and weighed. Fecal material was collected postdosing until termination for radiolabel analysis. All samples were stored at 20°C until analysis, which was completed within 1 y.

Gerbils belonging to a specific diet group will be described using the notation, prefeeding diet:experimental diet.

Study 2.  After a 3-d acclimation period, gerbils were fed a semipurified VA-deficient diet for either 8 wk (branchpoint 1), n = 50, or 10 wk (branchpoint 2), n = 50 (Fig. 1B). At the first branchpoint, 10 gerbils were killed and 40 gerbils were randomly divided into four groups and given new diets that were either unsupplemented or supplemented with VA (20.7 nmol/g diet) and/or C (145.9 nmol/g diet) for 5 d. The weight ratio of C:VA was ~13:1. The remaining 50 gerbils continued to consume the VA-deficient diet until the second branchpoint, at which time 10 gerbils were killed and the remaining gerbils were divided and fed diets as described previously for the first branchpoint. Seventy-two hours before termination of the study, 3-4 gerbils per group were dosed with 0.18 MBq of ¹⁴C-labeled C as described previously. At termination, serum, tissue and feces were collected as described above.

All gerbils were fed a VA/C prefeeding diet. Gerbils that were fed the prefeeding diet for 8 and 10 wk will be referred to as branchpoints 1 and 2, respectively, and gerbils that were fed experimental diets for 5 d after the 8- or 10-wk prefeeding period will be referred to as experimental groups from branchpoint 1 and 2, respectively.

Dose preparation and administration.  The ¹⁴CC was purified on a silica gel column (10-SPE, J. T. Baker, Phillipsburg, NJ) with several column volumes of chloroform. Purity was verified by HPLC analysis. The radiolabeled C was added to Ensure⁶ (Ross Laboratories, Columbus, OH) using a minimal amount of methylene chloride that was evaporated before dosing. The oral dose (~500 µL) was administered by gavage at 40°C after 12 h of food deprivation.

Analytical methods

Tissue and serum analysis.  All samples were analyzed for VA, but C was analyzed only for gerbils that were fed C. VA and C were extracted from liver and kidney samples (0.1-0.2 g), whole adrenals (Study 1), pooled adrenals (6 per sample) (Study 2), and serum (200-800 µL) as described previously by our laboratory (Lederman et al. 1998). Extracts were stored at 20°C before HPLC analysis, which was completed within 48 h of extraction.

Either -apo-8 carotenoic methyl ester (Fluka, Ronkonkoma, NY) or echinenone (a gift from Hoffmann La Roche) was added as internal standard to all samples analyzed for C. Retinyl palmitate (Sigma Chemical, St. Louis, MO) was added to serum samples as an internal standard for retinol.

HPLC.  Samples analyzed for C were reconstituted in methylene chloride and analyzed as described by Lederman et al. (1998).

Vitamin A samples were reconstituted in either methylene chloride or a 2:3 (mobile phase/methylene chloride) mix before HPLC analysis. Elution of VA from a Supelco LC-18 column (no. 58298 Supelco, Bellefonte, PA) was monitored at 325 nm on a system consisting of a Dynamax model SD-200 pump (Rainin Instrument, Woburn, MA), Waters 486 tunable absorbance detector (Millipore, Bedford, MA), and Dynamax HPLC Methods Manager integrator (Rainin Instrument), with methanol/acetonitrile/chloroform (47:47:6, v/v/v) mobile phase, at a flow rate of 1.5 mL/min.

Radioisotope analysis.  The distribution of radiolabel in saponified and nonsaponified liver tissue was analyzed. For nonsaponified liver analysis, liver samples from each group were pooled (total of 0.20-0.35 g), ground with 7 g anhydrous sodium sulfate by using a mortar and pestle, and extracted six times, alternating between 15 mL methylene chloride and 15 mL ethyl acetate. The combined solvent extract was reduced to 3 mL in a Rotoevaporator (Brinkmann Instruments, Westbury, NY); the remaining solvent was evaporated to dryness under vacuum in a Speedvac (Savant Instruments, Farmington, NY). For saponified liver analysis, samples were prepared as described previously for C and VA analysis. Sample residues were reconstituted in methylene chloride and analyzed as described previously by our laboratory (Lederman et al. 1998).

To determine fecal excretion of the ¹⁴CC dose, triplicate samples of dried, ground feces were weighed into scintillation vials, and 0.75 mL Solvable (Dupont/NEN Research Products, Boston, MA) was added; samples were incubated in a water bath at 50°C for 6 h and cooled to room temperature. Three aliquots of 0.2 mL of 30% hydrogen peroxide were added to samples at 12-h intervals for decolorization. Scintillation cocktail was added to the decolorized samples, samples were vortexed and radioactivity measured by liquid scintillation counting.

The total ¹⁴C content in tissues was measured by adding 0.5 mL Solvable to small tissue samples. Samples were incubated in a water bath at 50°C for 3 h and cooled to room temperature. Scintillation cocktail was added, and radioactivity was obtained as previously described. Serum radioactivity was measured directly by liquid scintillation counting. Total serum radioactivity was determined on the basis of the assumption that serum accounts for 5.46 g/100 g total body weight (Lombardi and Oler 1967).

Calculation of VA utilization.  By monitoring the change in hepatic VA stores from gerbils (Study 2) from the start of the VA/C diet to after 10 wk of consuming the diet, we calculated the daily hepatic loss of VA/100 g body weight.

Statistical analysis.  Differences between groups were determined using one-way ANOVA and Fisher's protected least significant difference analysis (StatView 4.5, Abacus Concepts, Berkley, CA). Differences were considered significant at P < 0.05. Values shown represent group means ± SD. Some data were log transformed before statistical analysis as indicated in table footnotes.

	RESULTS

Abstract Introduction Methods Results Discussion References

Study 1.  There were no differences in final body weights among groups (data not shown), and no signs of VA deficiency were observed. After the prefeeding period, gerbils had hepatic VA stores of 1.8 µmol (0.68 µmol/g or 195.6 µg/g) and 0.6 µmol (0.25 µmol/g or 71.9 µg/g) in those fed the +VA and VA diets, respectively. Gerbils fed the +VA:+VA/+C and +VA:+VA/C diets had significantly higher hepatic VA stores compared with those fed only the +VA prefeeding diet (P < 0.05) (Table 1). However, hepatic VA stores were slightly lower in gerbils fed the +VA:VA/+C diet compared with the +VA branchpoint (1.5 vs. 1.8 µmol, P = 0.1). -G e r b i l s  f e d  ;ms V A : ;pl V A / ;pl ;gb C,  ;ms V A : ;pl V A / ;ms ;gb C,  o r- VA:VA/+C had significantly greater hepatic VA stores than the VA branchpoint. Regardless of prior VA status, the experimental diet that contained both VA and C resulted in the greatest increase of VA stores compared with the respective +VA or VA branchpoint, with the VA/+C diet having the least effect. Vitamin A stores in kidney (1.9-3.0 nmol) and adrenal (1.5-2.0 nmol) tissues were not significantly different among groups. Vitamin A concentration in serum ranged from 1.4 to 2.0 µmol/L. Gerbils that were fed +VA:VA/+C had significantly higher serum VA concentrations than all other groups.

View this table: [in this window] [in a new window]   Table 1. Tissue vitamin A (VA) stores and serum VA concentrations from gerbils after 26 d of consuming either a +VA or VA diet (branchpoints) and after an additional 31 d of consuming an experimental diet containing VA and/or -carotene (C) (Study 1)1

No C was detected in gerbils fed diets without C. After the experimental period, gerbils fed the VA prefeeding diet had higher liver stores of C than gerbils fed the +VA prefeeding diet, for each of the corresponding experimental diets (Table 2). Kidney C stores ranged from 67.1 to 79.0 pmol and were not different among groups. Adrenal C concentrations ranged from 2.2 to 3.9 nmol/g. Gerbils fed +VA:+VA/+C had significantly higher adrenal C concentrations than any other group.

View this table: [in this window] [in a new window]   Table 2. Liver and kidney -carotene (C) stores and adrenal C concentrations from gerbils fed either C or C and Vitamin A (VA) for 31 d after 26 d of consuming either a +VA or VA prefeeding diet (Study 1)1,2

The percentage of recovery of the radiolabeled dose from liver and feces was variable and ranged from 4.2 to 8.9 and 36.7 to 59.1%, respectively; there were no significant differences among groups (data not shown). Photodiode array (PDA) analysis of saponified and nonsaponified liver samples showed that the radioactivity associated with either VA or C was associated primarily with VA for all groups, demonstrating that all groups were able to absorb the radiolabeled dose and convert most of the ¹⁴CC to vitamin A and store the ¹⁴CVA the liver, primarily as retinyl esters (data not shown).

Study 2.  There were no significant differences in final body weights for any group (data not shown) and no signs of VA deficiency were observed. After the VA-deficient diet was consumed for 8 and 10 wk, liver VA stores were 0.21 µmol (79.9 nmol/g, 22.9 µg/g) and 0.18 µmol (62.5 nmol/g, 17.9 µg/g), respectively. Gerbils fed VA diets for 8 wk reached marginally deficient VA status; after 10 wk, liver VA concentrations were considered deficient, although these two levels did not differ significantly from each other. Gerbils fed experimental diets that contained either VA or C had significantly higher liver VA stores compared with their respective branchpoint (Table 3). Feeding both VA and C resulted in the greatest increase in liver VA stores, with levels reached that were significantly higher than any other group within the same branchpoint.

View this table: [in this window] [in a new window]   Table 3. Liver and kidney Vitamin A (VA) stores, and adrenal and serum VA concentrations from gerbils fed VA and or -carotene (C) after either 8 (branchpoint 1) or 10 wk (branchpoint 2) of consuming a VA/C diet (Study 2)1

We calculated the daily VA utilization rate for Mongolian gerbils to be 3.1 µg VA/100 g body weight. Assuming that an adult gerbil weighing 75 g consumes 6 g diet/d, the dietary VA levels required to match the utilization rate would be 1.44 nmol VA/g diet (0.41 µg VA/g diet).

Serum VA was not different between branchpoints 1 and 2, between groups receiving the same experimental diets or between experimental groups within each branchpoint (Table 3). However, the serum VA for gerbils within branchpoint 1 was significantly higher than experimental groups from the first branchpoint that were fed +VA/C, VA/+C or +VA/+C. The serum VA concentration for branchpoint 2 was also significantly higher than that for experimental groups fed VA/+C or +VA/+C from branchpoint 2.

Kidney VA stores ranged from 2.05 to 2.94 nmol and were not significantly different among groups. Adrenal VA concentrations ranged from 14.19 to 30.63 nmol/g for branchpoint 1 and 11.67 to 20.57 nmol/g for branchpoint 2. For both branchpoints, adrenal VA concentrations were numerically highest in the group fed the +VA/C experimental diets and lowest in the group fed the VA/+C diet.

-Carotene liver stores and serum concentrations ranged from 25.92 to 40.64 nmol and 34.68 to 51.06 nmol/L, respectively, and there were no significant differences among groups (Table 4). Kidney C stores ranged from 6.90 to 11.63 pmol. Kidney C stores in gerbils fed the VA/+C experimental diet were significantly lower than in those fed +VA/+C for groups within branchpoint 1; however, for branchpoint 2, the reverse was true.

View this table: [in this window] [in a new window]   Table 4. Liver and kidney -carotene (C) stores and adrenal and serum C concentrations from gerbils fed Vitamin A (VA) and or C after either 8 (branchpoint 1) or 10 wk (branchpoint 2) of consuming a VA/C diet (Study 2)1,2

The average percentage of the radiolabeled dose recovered from liver ranged from 3.7 to 16.1% for the various groups (Table 5). Within branchpoints, gerbils with C in their experimental diet had less ¹⁴C in the liver than gerbils that did not have C in their diet, although this relationship was significant only (P < 0.05) for branchpoint 1 (P > 0.2 for branchpoint 2). Recovery of ¹⁴C from liver was less in groups from animals fed the prefeeding diet for 10 wk for all experimental diets, but the difference was significant only for the C experimental diets (P < 0.05). PDA analysis of the radiolabeled liver tissue showed that the radioactivity associated with either VA or C was associated predominantly with retinol and retinyl esters for all groups, with minimal radioactivity associated with C (Fig. 2). The amount of radioactivity associated with VA was much less in the gerbils that were fed the prefeeding diet for 10 wk rather than 8 wk, and gerbils fed C in their experimental diets had less ¹⁴CVA than those gerbils that were not fed C. The amount of radioactivity associated with VA per gram liver tissue for each group is presented in Figure 3. Gerbils fed experimental diets without C had significantly more ¹⁴CVA in their livers than those that were fed C, for both the 8- and 10-wk prefeeding groups. Total radioactivity recovered from the liver that was associated with VA ranged from 26.0 to 30.1% for groups not fed C, and 17.1 to 20.5% for groups fed C and was lower for the 10-wk prefeeding period for each of the respective diet groups (data not shown). The percentage of the radioactivity recovered from the liver that was associated with C ranged from 0.39 to 3.5 for groups not fed C and 0.57 to 5.3 for groups fed C and was higher for the 10-wk prefeeding period for each of the respective diet groups (data not shown).

View this table: [in this window] [in a new window]   Table 5. The percentage of recovery of 14C from a 0.18 MBq 14C--carotene dose in serum and tissues from groups of gerbils fed Vitamin A (VA) and or -carotene (C) after either 8 (branchpoint 1) or 10 wk (branchpoint 2) of consuming a VA/C diet (Study 2)1

View larger version (22K): [in this window] [in a new window]   Fig 2. Radioactivity of HPLC fractions collected from a lipid extract of saponified liver tissue from gerbil Study 2. Panel A shows groups that were fed a Vitamin A/-carotene (VA/C) diet for 8 wk and then fed a VA/C, +VA/C, VA/+C, or +VA/+C diet for an addition 5 d. Panel B shows groups that were fed a VA/C diet for 10 wk and then fed a VA/C, +VA/C, VA/+C, or +VA/+C diet for an addition 5 d. All gerbils received a 0.18 MBq dose of 14CC 72 h before termination of the study (dpmq). Photodiode array (PDA) analysis verified peak 1 as retinol and peak 2 as C. Diets: +VA, vitamin A-containing diet; VA, vitamin A-free diet; +C, -carotene-containing diet; C, -carotene-free diet.

View larger version (23K): [in this window] [in a new window]   Fig 3. Liver radioactivity associated with retinol following saponification of liver tissues from gerbils that were prefed a Vitamin A (VA) and -carotene (C) free diet (VA/C) for either 8 or 10 wk and then fed experimental diets (VA/C, +VA/C, VA/+C or +VA/+C) for 5 d. Groups on the horizontal axis are labeled with length of prefeeding:experimental diet. All gerbils received a 0.18 MBq dose of 14CC 72 h before termination of the study. Bars represent group means ± SD, n = 3 or 4. Bars with no letters in common are significantly different, P < 0.05.

Recovery of ¹⁴C from feces was highly variable and ranged from 25.9 to 62.2% (Table 5). Fecal recovery from gerbils fed the +VA/C experimental diet was significantly less than that for the +VA/+C diet within branchpoint 1, whereas there were no significant differences within branchpoint 2. Fecal recovery of ¹⁴C was significantly less in the gerbils fed the prefeeding diet for 10 wk compared with 8 wk for all of the experimental diet groups except +VA/C.

The percentage of recovery of the radioactive dose from serum ranged from 0 to 1.03 (Table 5). Within branchpoint 2, serum radioactivity was significantly higher in the group that received the VA/C experimental diet. No radioactivity was detected in the serum of gerbils from groups receiving C in their experimental diets.

The percentage of recovery of ¹⁴C from kidneys ranged from 0.02 to 0.24 (Table 5). Within branchpoint 1, recovery from the VA/C group was significantly higher than that for either of the C-fed groups. Recovery from animals fed the VA/C diet from branchpoint 2 was significantly higher than that for the other groups within the branchpoint.

The percentage of recovery of the ¹⁴C from lung tissue ranged from 1.44 to 3.39 for groups not fed C, and 0.35 to 0.88 for C-fed groups (Table 5). Recovery was significantly higher in the groups not fed C compared with that for groups fed C within each branchpoint.

Recovery of the radioactive dose from adrenals was < 0.001% of the initial dose (data not shown).

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Vitamin A deficiency is a public health concern in developing countries and has been associated with increased respiratory infection, risk of measles, incidence of diarrhea and decreased immune response (Fawzi et al. 1995, Sommer and West 1996). Dietary intervention with foods rich in provitamin A carotenoids has been suggested as a long-term solution to this problem. However, the efficacy of C utilization from some foods has been questioned (dePee et al. 1995, Solomons and Bulux 1993, Solomons 1996) because of the limited ability of provitamin A carotenoids to increase the VA status of individuals suffering from clinical and subclinical VA deficiency.

Rats have been used extensively (Biesalski and Weiser 1993, Grolier et al. 1995) to evaluate the efficiency of C utilization by monitoring changes in liver VA stores. The rat intestinal mucosal cells effectively cleave C. However, this species does not absorb physiologic doses of C intact. Because humans both cleave C to form VA and absorb a variety of carotenoids (including C) intact, the rat is not the most appropriate model for C utilization studies.

The ferret is a model that does absorb C intact; however, studies in this laboratory (Lederman et al. 1998) have shown that the efficiency of conversion of a highly bioavailable source of C to VA was poorer than 15:1, much lower than the expected conversion in humans.

Although little is known about VA metabolism in gerbils, we have shown that this species does absorb small doses of C intact (Pollack et al. 1994). Other studies in this laboratory evaluated the uptake of VA and C into intestinal mucosal cells by using brush border membrane vesicles isolated from normal rats and from gerbils of differing VA status (Moore et al. 1996). The gerbil brush border membrane vesicles used in the gerbil studies were isolated from intestinal segments from gerbils in Study 2 described in this paper. Moore et al. (1996) showed that brush border membrane vesicle uptake of retinol was not affected by VA status; however, uptake of C was significantly lower in the more deficient animals, which had hepatic VA concentrations of 62.5 nmol/g, suggesting that low VA status may reduce the intestinal uptake of C.

The purpose of these studies was to evaluate the effect of declining VA status on the ability to utilize a highly bioavailable source of C for VA in the Mongolian gerbil. Although Studies 1 and 2 had similar study designs, there were differences in the prefeeding diets, the length of feeding and the quantity of C in the diets. Despite these differences, the results of these studies together yield insight into the utilization of C by gerbils at differing VA status ranging from adequate to approaching deficiency.

In Study 1, gerbils that consumed the +VA:+VA/+C and +VA:+VA/C diets had significantly higher hepatic VA stores than the +VA branchpoint, but their stores were lower (P = 0.1) after the VA/+C experimental diet. However, gerbils fed the VA prefeeding diet had significantly higher hepatic VA stores after all three experimental diets (Table 1). These results demonstrate the following: 1) a dietary level of 20.9 nmol VA/g diet is more than sufficient to increase hepatic VA stores in either VA-sufficient or moderately VA-sufficient gerbils; 2) feeding C at a 6:1 weight ratio of C:VA to gerbils with sufficient VA status provides enough VA value to maintain, but not increase hepatic VA stores; and 3) as VA status declines to a moderately sufficient level, this amount of C is sufficient to increase hepatic VA stores, but not to the level of preformed VA. These conclusions are based on feeding highly bioavailable, commercial C beadlets. If C is derived from natural sources, from which bioavailability would probably be lower, the weight ratio of C:VA to achieve equivalent VA status might be greater than 6:1.

In Study 2, gerbils were fed VA depletion diets for 8 or 10 wk, which reduced hepatic VA stores to 79.9 nmol/g and 62.5 nmol/g, respectively. Using 70 nmol/g as a cut-off level, below which a deficient VA status is indicated (Olson 1991), gerbils fed a VA-deficient diet for 8 and 10 wk achieved marginally deficient and deficient status, respectively. The results of hepatic VA analysis (Table 3) showed that feeding a highly bioavailable source of C at a weight ratio of ~13:1 (C:VA) for 5 d was sufficient to more than double hepatic VA stores to a level equivalent to feeding preformed VA. It is possible that lower ratios of C:VA would also have been effective in improving VA status of marginally deficient gerbils.

In Study 2, the percentage of the administered ¹⁴C recovered in both feces and tissues was higher in groups fed the VA depletion diet for 8 wk compared with 10 wk (Table 5). The fecal results demonstrate that the efficiency of absorption of the label by intestinal mucosal cells was higher in the more deficient gerbils, supporting upregulation of C absorption. The liver results suggest that, in the more deficient gerbils, more of the absorbed radiolabel had been metabolized to fulfill VA needs of tissues. Within the liver, ¹⁴CVA made up almost all of the hepatic VA plus C radioactivity, although ~70% of the ¹⁴C recovered was not associated with either VA or C, presumably reflecting metabolic products of either VA or C.

As mentioned above, Moore et al. (1996) found that the uptake of C into the membranes of intestinal cells, using a brush border membrane vesicle model, was decreased when the gerbil became deficient in VA. However, the fecal recovery of the radiolabeled dose in Study 2 suggests increased C absorption when gerbils are VA deficient. Although the uptake of C into the intestinal mucosal cells is decreased, it is possible that the absorption from the intestinal cells into the lymphatic circulation is increased, resulting in a net increase in C absorbtion.

It is interesting to note that for both the 8- and 10-wk depletion periods, both recovery of ¹⁴C from tissues (Table 5) and hepatic ¹⁴C associated with VA and C (Figs. 2 and 3) were lower in groups fed experimental diets containing C than in groups fed diets without C. A clear explanation of these results is not obvious, but the differences may be due to the timing of the ¹⁴CC dose. One explanation may be that the experimental diets were fed 2 d before and immediately after dosing, possibly resulting in competition for mucosal cell uptake of the ¹⁴CC and nonlabeled C in the animals that had C in their diet. Even though the gerbils were deprived of food for 12 h before dosing, the intestinal mucosal cells of the gerbils fed C may have been "saturated" with C, thus decreasing absorption of the radiolabeled dose; immediate refeeding of C would result in a dilution of the ¹⁴CC with cold C, further decreasing absorption of the radiolabeled dose. The increased fecal recovery of the ¹⁴C radiolabel from groups fed C (Table 5) supports this explanation. This trend was not seen in Study 1, probably due to the higher VA status of those gerbils.

There were significant differences in serum VA between groups in both studies. However, because all group means are well above 0.7 µmol/L, a value that indicates marginal VA deficiency (Olson 1991), these differences are not considered to be physiologically important.

Little is known about the VA requirement of gerbils. Previously, the recommended level of VA for a gerbil was a range of 19-34 nmol VA/g diet (NRC 1978), which corresponded to the level used in these studies (20.9 nmol VA/g diet). However, the recommendation was reduced recently to 2.5 nmol VA/g diet (NRC 1995). We calculated the daily VA utilization rate for the Mongolian gerbil to be 3.1 µg VA/100 g body weight. Assuming that an adult gerbil weighing 75 g consumes 6 g diet/d the dietary VA levels required to match the utilization rate would be 1.44 nmol VA/g (0.41 µg VA/g) diet. Because this level assumes 100% utilization efficiency from diet, which cannot be achieved, the recent recommendation of 2.5 nmol VA/g diet (NRC 1995) appears to be appropriate.

The gerbil appears to be a good animal model for evaluation of C bioavailability and metabolism for a number of reasons. This species absorbs C intact when it is fed at low dietary levels (Pollack et al. 1994), and, as shown here, the C is converted to VA for use in metabolism and storage. The conversion efficiency weight ratio established by the Food and Nutrition Board (NRC 1989) for C:VA is 6:1 for humans, assuming a mixed dietary source of C and normal VA status. For gerbils fed a highly bioavailable source of VA, a C:VA weight ratio between 6:1 and 13:1 is adequate, although 6:1 certainly prevents VA deficiency under the conditions used in these two studies. The efficiency of absorption and conversion of C to VA appears to increase as the gerbil becomes more depleted. These observations are similar to those expected in humans. For humans, the efficiency of utilization of VA is high and is maintained over a large range of VA status (Olson 1991), but C utilization declines markedly at higher concentrations in the diet. Moreover, as humans approach marginal-to-deficient VA status, it is expected that there is an upregulation of the efficiency of C utilization as VA.

The results described in this paper with the gerbil model appear to be more reflective of human physiology than that of either the rat, which does not absorb physiologic doses of C intact, or the ferret, which is a poor converter of C to VA, suggesting that the gerbil is a more appropriate animal model for additional evaluations of C bioavailability and metabolism.

	ACKNOWLEDGMENTS

The authors acknowledge Tazima Smith for her assistance with the radiolabel analysis.

	FOOTNOTES

Manuscript received 21 April 1997. Initial reviews completed 19 June 1997. Revision accepted 20 October 1997.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Bierer T. L., Merchen N. R., Erdman J. W. Jr. Comparative absorption and transport of five common carotenoids in preruminant calves. J. Nutr. 1995; 125:1569-1577
• Biesalski, H. K. & Weiser, H. (1993) -Carotene supplements cannot meet all vitamin A requirements of vitamin A-deficient rats. In: Carotenoids in Human Health, (Coufield, L. M., Krinski, N. I. & Olson, J. A., eds.). Academy of Science, New York, NY.
• dePee S., West C. E., Muhilal, Karyadi D., Hautvast J.G.A. Lack of improvement in vitamin A status with increased consumption of dark-green leafy vegetables. Lancet 1995; 346:75-81 ^[Medline]
• FAO/WHO Expert Group (1988) Requirements of Vitamin A, Iron, Folate, and Vitamin B-12. FAO Food and Nutrition Series no. 23, Rome, Italy.
• Fawzi W. W., Herrera M. G., Willett W. C., Nestel P., Amin A. E., Mohamed K. A. Dietary vitamin A intake and the incidence of diarrhea and respiratory infection among Sudanese children. J. Nutr. 1995; 125:1211-1221
• Grolier P., Agoudave S., Azais-Braesco V. Comparative bioavailability of diet-, oil- and emulsion-based preparations of vitamin A and -carotene in the rat. Nutr. Res. 1995; 15:1507-1516
• Lederman J. D., Overton K. M., Hofmann N. E., Moore B. J., Thornton J., Erdman J. W. Jr. Ferrets (Mustela putoius furo) inefficiently convert beta carotene to vitamin A. J. Nutr. 1998; 128:271-279 [Abstract/Free Full Text]
• Lombardi B., Oler A. Choline deficiency fatty liver protein synthesis and release. Lab. Invest. 1967; 17:308-321 [Medline]
• Moore A. C., Gugger E. T., Erdman J. W. Jr. Brush border membrane vesicles from rats and gerbils can be utilized to evaluate the intestinal uptake of all-trans and 9-cis -carotene. J. Nutr. 1996; 126:2904-2912
• National Research Council (1978) Nutrient Requirements of Laboratory Animals, 3rd ed. National Academy Press, Washington, DC.
• National Research Council (1989) Recommended Dietary Allowances, 10th ed. National Academy Press, Washington, DC.
• National Research Council (1995) Nutrient Requirements of Laboratory Animals, 4th ed. National Academy Press, Washington, DC.
• Olson, J. A. (1991) Vitamin A. In: Handbook of Vitamins. Machlin, L. J., ed., pp. 1-57. Marcel Dekker, New York, NY.
• Pollack J., Campbell J. M., Potter S. M., Erdman J. W. Jr. Mongolian gerbils (Meriones unguiculatus) absorb -carotene intact from a test meal. J. Nutr. 1994; 124:869-873
• Poor C. L., Bierer T. L., Merchen N. R., Fahey G. D. Jr., Murphy M. R., Erdman J. W. Jr. Evaluation of the preruminant calf as a model for the study of human carotenoid metabolism. J. Nutr. 1992; 122:262-268
• Ribaya-Mercado J. D., Holmgren S. C., Fox J. G., Russel R. M. Dietary -carotene absorption and metabolism in ferrets and rats. J. Nutr. 1989; 119:665-668
• Solomons N. W. Plant sources of vitamin A and human nutrition: renewed strategies. Nutr. Rev. 1996; 54:89-91 ^[Medline]
• Solomons N. W., Bulux J. Plant sources of provitamin A and human nutriture. Nutr. Rev. 1993; 51:199-204 [Medline]
• Sommer, A. & West, K. P., Jr. (1996) Vitamin A Deficiency: Health, Survival, and Vision. Oxford University Press, New York, NY.
• van Vliet T. Absorption of -carotene and other carotenoids in humans and animal models. Eur. J. Clin. Nutr. 1996; 50:S32-S37
• White W. S., Peck K. M., Ulman E. A., Erdman J. W. Jr. The ferret as a model of evaluation of the bioavailabilities of all-trans--carotene and its isomers. J. Nutr. 1993; 123:1129-1139

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences