The Journal of Nutrition Vol. 128 No. 2 February 1998, pp. 271-279

Ferrets (Mustela putoius furo) Inefficiently Convert -Carotene to Vitamin A¹

Janine D. Lederman, Katrina M. Overton, Nicolle E. Hofmann, Billy J. Moore,, Jesse Thornton, and John W. Erdman Jr.²

Department of Food Science and Human Nutrition and Division of Nutritional Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801

	ABSTRACT

Abstract Introduction Methods Results Discussion References

The ferret has recently been used as a model to evaluate the absorption and metabolism of several carotenoids; however, little is known about the vitamin A (VA) requirements of this species or the ability of ferrets to convert dietary -carotene (C) to VA. Three studies were conducted to estimate the daily utilization of VA in ferrets and to determine the effect of prior VA status on the ability of ferrets to utilize C as a source of VA. Weanling male ferrets were fed a pelleted, low carotenoid, semipurified diet either with (+VA) or without VA (VA) for 21- to 35-d prefeeding periods. Upon initiation of the experiments, several ferrets were killed to determine base-line VA status. The remaining ferrets were fed VA, C, or VA and C in pelleted feed (Studies 1-3) or liquid carrier (Study 3) for 16-21 additional days. Hepatic VA and C concentrations were used as the primary indicators of VA status, although serum and adrenal VA and C also were measured. The results showed the following: 1) provision of C at up to a 15:1 weight ratio of C to VA in pelleted feed or liquid carrier was not sufficient to maintain hepatic VA stores after a VA prefeeding period; 2) the daily utilization rate of VA by ferrets ranged from 80 to 171 µg in the three studies; 3) the ferret was confirmed to be a species that has the majority of its serum VA in ester form; and 4) feeding VA diets significantly reduced serum retinyl esters but had less effect on serum retinol. We conclude that although ferrets can convert C to VA, the process is inefficient. The ferret model can be most appropriately used when studying the biological effect of tissue C stores on VA status and is less appropriate for the evaluation of dietary C conversion to VA.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Vitamin A (VA)³ deficiency is an important public health concern, particularly in developing countries where the consumption of preformed VA is low and may constitute only 10-30% of the total VA intake (FAO/WHO 1988). Increasing consumption of carotenoid-rich fruits and vegetables has been suggested as an alternative to the provision of massive oral doses of VA suggested by the World Health Organization (FAO/WHO 1986) to overcome VA deficiency. However, there is concern regarding the low efficiency with which provitamin A carotenoids are utilized from some foods (de Pee et al. 1995, Solomons 1996, Solomons and Bulux 1993). For example, de Pee and co-workers (1995) did not demonstrate improved VA status in breast-feeding Indonesian women with low VA status, who were supplemented daily with carotene-rich, dark-green, leafy vegetables, and suggested that the matrix of these foods may reduce carotenoid utilization.

Development of animal models to study the effects of complex food matrices and the effects of nutritional status of subjects and other interactions on -carotene (C) absorption and metabolism has been a focus in our laboratory. Ferrets, like humans, are able to absorb dietary C intact and accumulate C in tissue and sera. Ferrets also absorb a variety of other carotenoids such as -carotene, lycopene and canthaxanthin (Tang et al. 1993, White et al. 1993a). In addition, ferrets convert C to VA in the intestine (Wang et al. 1991) and in homogenates of liver, lung and adipose tissue (Wang et al. 1992). Tissue distribution of carotenoids in ferrets is similar to that of humans, with liver, adrenal and adipose tissues accumulating substantial amounts of intact carotenoids (Gugger et al. 1992, Ribaya-Mercado et al. 1989 and 1992, Zhou et al. 1996). Therefore use of the ferret as a model for studying human carotenoid metabolism seems well justified.

One difficulty with using this model is that the nutritional requirements (in particular, the VA requirement) of the ferret are not well characterized. Our laboratory has successfully developed a semipurified, pelleted ferret diet that contains 18.9 nmol/g diet (5.4 µg/g diet) VA as crystalline retinyl palmitate (a level that is within the range usually found in commercial feeds; McLain et al. 1988), which results in excellent growth (White et al. 1993b). These studies were designed to achieve the following: 1) yield further information regarding the VA utilization rate of ferrets, 2) determine the effects of prior VA status on the utilization of C for VA, and 3) compare the bioavailability of VA and C when provided in pelleted feed or liquid carrier.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Animals.  Descented male ferrets (n = 47-50), 7-8 wk old with an average weight of 533, 422 and 464 g for Studies 1-3, respectively, were obtained from Marshall Farms (North Rose, NY) and certified to be of good health by a veterinarian. Animals were individually housed in stainless steel rabbit cages in the Edward R. Madigan Laboratory Animal Care Facility at the University of Illinois. Feed intake and body weight were routinely measured. The room was maintained at a constant temperature and room lighting was provided on an automated diurnal cycle of 12 h light and 12 h dark. All animal handling procedures were approved by the University of Illinois Laboratory Animal Care Advisory Committee.

Diets.  The composition of the semipurified pelleted diet used was previously developed in our laboratory (White et al. 1993b), and was shown to support normal growth while depleting tissue stores of carotenoids. The pelleted diets contained either 5.41 µg/g diet VA, provided as crystalline retinyl palmitate (Sigma Chemicals, St. Louis, MO) and/or 35 µg/g (Study 1) or 65 µg/g diet C (Studies 2 and 3) from 10% water-soluble beadlets (a gift from Hoffmann LaRoche, Nutley, NJ). This represented approximately a 6:1 and 12:1 weight ratio of C to VA. The pelleting process used to prepare diets for the first two studies was found to be too harsh and resulted in some destruction of C and VA. For the third study, the slurry of diet ingredients was cooled before extrusion, and during drying of pellets, less heat was used and direct light contact with pellets was avoided to reduce VA and C degradation.

Liquid carrier preparation. 

Study 3.  A liquid carrier for VA and C was provided to two groups in Study 3. The liquid carrier was a nondairy creamer (Coffee-Mate, General Mills, Minneapolis, MN) with negligible C or VA content. The carrier contained 25 g/100 g fat (dry basis), which should be sufficient for optimal C absorption. The carrier was prepared daily as a 10% solution (wt/wt) of Coffee-Mate in water. -Carotene (10% cold water-soluble beadlets) or retinyl palmitate (Palmabeads, a gift from Hoffmann-LaRoche) was added to warm Coffee-Mate, sonicated and stirred until completely dispersed. The C from the 10% water-soluble beadlets is highly bioavailable to ferrets (White et al. 1993a).

Experimental design. 

Study 1.  (Fig. 1A) Upon initiation of the study, four ferrets were killed by severing the brachial vessels between the pectoralis major and latissimus dorsi after cardiac puncture (described below) to determine baseline VA and C concentrations in tissue and serum. The remaining ferrets (n = 42) were divided into groups and fed either a +VA or VA pelleted diet for 35 d. Ferrets had free access to food and water. Seven ferrets from each group were then killed to determine the extent of VA accumulation or depletion. The remaining ferrets were fed diets containing either VA (18.9 nmol/g diet) or C (65.2 nmol/g diet) for 16 d. Forty-eight hours before termination of the experiment, several ferrets from each treatment group were administered 6.63 Bq ¹⁴C-C (gift from Hoffmann LaRoche) orally in a 1.0 mL dose of Ensure⁴ (Ross Laboratories, Columbus, OH), after overnight food deprivation. Feces were collected postdosing until termination of the experiment.

View larger version (21K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Fig 1. Study design for Study 1 (A), Study 2 (B) and Study 3 (C). All ferrets were fed a semipurified pelleted diet. The appropriate diets were supplemented with 5.41 µg/g diet vitamin A (VA) and 35 µg/g diet -carotene (C) for Study 1 or 5.41 µg/g diet VA and 65 µg/g diet for Studies 2 and 3.

Blood samples were collected at the termination of the studies by cardiac puncture of ferrets under ketamine hydrochloride/xylazine (10:1 v/v) anesthesia (Vetallar, Parke-Davis, Morris Plains, NJ; Rompun, Miles Laboratory, Shawnee, KS, respectively) delivered by intramuscular injection (~0.30 mL/kg body weight). Tissues were removed, rinsed, blotted and weighed. Serum and tissues were stored at 20°C until analysis.

Study 2.  (Fig. 1B) The design of Study 2 was similar to that of Study 1 with the following exceptions: 1) the length of the VA or +VA pelleted feeding was 33 d; 2) the experimental feeding was 21 d; 3) some ferrets were fed both VA and C during the experimental period; 4) C concentration in the diet was 121 nmol/g diet instead of 65.2 nmol/g diet; and 5) no radioactive dose was administered.

Study 3.  (Fig. 1C) Upon initiation of the study, seven ferrets were killed to determine base-line VA and C levels in tissue and serum. Ferrets were fed a VA-deficient, semipurified, pelleted diet for 21 d at which time seven ferrets were killed to determine the extent of liver VA depletion. The remaining animals were assigned to consume one of five diets for 16 d as follows: 1) VA/+C pelleted feed, 2) VA/C pelleted feed with a +C liquid carrier, 3) +VA/C pelleted feed, 4) VA/C pelleted feed with a +VA liquid carrier, or 5) VA/C pelleted feed. Ferrets not receiving C or VA in liquid carrier received 1 mL placebo carrier. Groups were adjusted to ensure similar mean body weights for all groups.

Ferrets received C at approximately a 12:1 weight ratio of C to VA from pelleted feed or liquid carrier. Liquid carrier treatments provided C or VA in single, daily 1-mL oral supplement with the C or VA content equivalent to the average quantity of C or VA consumed the previous day by ferrets fed the corresponding C or VA pelleted diets. The liquid carrier was administered (by 1-mL pipetman) in the early afternoon after 4-5 h of food and water deprivation. Food and water were then returned to the ferrets. At the termination of the study, serum and tissues were removed as described previously.

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

Radiolabel dose preparation. 

(Study 1).  The radiolabeled dose was prepared from purified 10, 10, 11, 11 ¹⁴C-C with a specific activity of 5.26 MBq/mg (a gift from Hoffmann-LaRoche). The ¹⁴C-C was dissolved in 500 µL hexane and purified by elution from a silica gel column (10-SPE, J.T. Baker, Phillipsburg, NJ) with 10 mL hexane followed by HPLC analysis. The ¹⁴C-C was added to Ensure and heated to 40°C under nitrogen to evaporate the hexane. -Carotene beadlets (10% water-soluble beadlets, Hoffmann LaRoche) were dissolved in water and added to Ensure at a concentration of 7.45 nmol/L. The mixture was stored overnight at 20°C and warmed to 25°C before administration. The adult enteral formula Ensure was chosen on the basis of its negligible carotenoid content (Bowen et al. 1988) and ease of administration. The formula was not expected to have confounding effects on the absorption or metabolism of the C.

Analytical methods

Diet analysis.  All analysis was performed under yellow lighting. The pelleted diets were analyzed for VA and C content before feeding. Pelleted feed was ground to a fine powder in a blender, mixed thoroughly and further ground by hand with a mortar and pestle. Triplicate samples (0.36-0.43 g) were weighed and 6 mL absolute ethanol containing BHT (1 g/L) was added. Samples were saponified by addition of 2 mL saturated KOH and incubated (70°C) in a water bath for 30 min. After cooling to room temperature, 2 mL distilled water was added and samples were extracted three times with 10 mL hexane. The combined hexane layers were completely evaporated under vacuum in a Speedvac AS160 (Savant Instruments, Farmington, NY) and the residues analyzed by HPLC for C and VA content.

Tissue and serum VA and C.  Duplicate liver samples (0.20-0.25 g) or both adrenal glands were homogenized (Polytron 10/35, Brinkman Instrument, Westbury, NY) in 5 mL absolute ethanol containing BHT (1g/L). Samples were saponified by addition of 1 mL saturated KOH and incubated (70°C) in a water bath for 30 min. After samples were cooled to room temperature, 2 mL distilled water as added. Samples were extracted three times with 8 mL hexane and the combined hexane layers were evaporated to dryness. Echinenone (a gift from Hoffmann LaRoche) was used as internal standard for quantification of tissue C.

-apo-8-Carotenic acid ethyl ester (Sigma Chemical) was added to serum as internal standard. Duplicate aliquots (0.5-1.5 mL) of serum were denatured by the addition of an equal volume of absolute ethanol containing BHT (1 g/L) and extracted three times with up to 4 mL hexane. The combined hexane layers were evaporated completely in a Speedvac and residues were analyzed by HPLC for C and VA.

Serum retinol and retinyl-ester analysis. 

(Studies 1 and 3).  Eight retinyl esters (RE) were synthesized as standards to establish HPLC retention times (all materials were purchased from Sigma Chemical). Five RE (retinyl palmitate, retinyl stearate, retinyl oleate, retinyl linoleate and retinyl myristate) were synthesized by a condensation reaction of retinol with the respective ester anhydrides in triethylamine (Hubbard et al. 1990). Arachidonate and pentadecanoate esters were synthesized by the transesterification of the methyl ester with retinyl acetate (Futterman and Andrews 1964). All standard solutions were purified on open silica columns by elution with hexane containing 50 mL/L ethyl acetate. Aliquots (0.5-2.0 mL) of serum were denatured by the addition of an equal volume of absolute ethanol containing BHT (1 g/L). Serum was extracted as described previously. Indentification of retinyl esters was based upon relative retention times to retinyl palmitate⁵ and absorbance at 325 nm. Retinyl palmitate (Sigma Chemical) and all-trans retinol (ROH) (Sigma Chemical) standard curves were constructed and used to quantitate serum ROH and RE concentrations. Because all RE have similar extinction coefficients, the retinyl palmitate standard curve was used for all RE quantifications.

Radiolabel distribution in saponified and nonsaponified liver. 

(Study 1).  For saponified liver analysis, procedures were conducted as above. Nonsaponified liver samples (0.3-0.4 g) were homogenized on ice with 3.1 mL 50 mmol/L Tris buffer. Collagenase solution (0.14 mL) [50 g/L collagenase type IV (Sigma Chemical) in PBS, pH 7.4] was added to the homogenate, vortexed and incubated at 37°C. After incubation, 0.14 mL protease solution [5 mg/mL protease type XXV (Sigma Chemical) in 50 mmol/L Tris buffer] was added and incubated at 37°C for 30 min. Samples were then cooled on ice and 3.7 mL ethanol containing BHT (1.0 g/L) was added (Peng et al. 1993). Samples were extracted three times with 8 mL hexane. Combined hexane layers were evaporated to dryness under vacuum and residues analyzed by HPLC.

For both saponified and nonsaponified liver, sample extracts were injected onto an HPLC column (no. 58298, Supelco, Bellefonte, PA) and fractions collected at 30-s intervals (Ultrorac II 2070, Pharmacia Biotech, Piscataway, NJ). The mobile phase was evaporated, and the residue was transferred to a scintillation vial containing 10 mL scintillation cocktail (Budgetsolve, Research Products International, Mt. Prospect, IL). Radioactivity of samples was obtained by liquid scintillation counting (Beckman LS 9000, Beckman Instruments, Fullerton, CA). Radioactivity eluting at times corresponding to ROH, RE and C was verified to be associated with those compounds by a Waters 991 Photodiode Array (PDA) detector (Millipore, Bedford, MA).

Radiolabel analyses.  Feces were dried at 50°C and dry weights recorded. Samples were ground with a mortar and pestle to a uniform consistency and stored at 20°C until analysis. Fecal (n = 5) samples were analyzed for each ferret receiving the radiolabeled dose. To prepare stock solutions, dried, ground feces (0.2 g) were solubilized with 1 mL Solvable (Dupont/NEN Research Products, Boston, MA) and incubated at 50°C in a water bath for 3 h. Samples were allowed to cool and were decolorized by the addition of 30% hydrogen peroxide. Hydrogen peroxide was added in 0.1-mL aliquots to avoid bubbling and allowed to sit overnight. Scintillation cocktail was added to aliquots of the stock fecal solution and radioactivity of the samples was obtained by liquid scintillation counting.

Serum aliquots (20 µL) were counted directly with 10 mL scintillation cocktail. Radioactivity in tissue was determined by solubilization and decolorization (when necessary), as described above, of small tissue samples (0.05-0.2 g).

Carrier analysis. 

(Study 3).  Triplicate 1-mL aliquots of the C and VA doses in liquid carrier were added to 3 mL absolute ethanol containing BHT (1g/L). Samples were saponified by the addition of 1.5 mL saturated KOH and incubated (70°C) in a water bath for 30 min. After the samples were cooled to room temperature, 2 mL of distilled water was added. The mixture was extracted four times with 6 mL hexane. The C and VA content of the liquid carriers was determined by HPLC.

HPLC.  Separate HPLC systems were used for C and VA analyses. A Milton-Roy Constametric III Pump (Riveria Beach, FL), Rheodyne 7125 (Cotati, CA) sample injection valve, BioRad (Richmond, CA) 1790 programmable variable wavelength detector and Shimadzu (Columbia, MD) CR601 integrator were used for C analysis. Carotenoids were separated using a Vydac C-18 (No. 201TP54) analytical column protected by an Upchurch (Oak Harbor, WA) precolumn packed with ODS C-18. The isocratic elution of carotenoids with methanol/acetonitrile/water (88:9:3, v/v/v) at a flow rate of 2 mL/min was monitored at 452 nm.

Analysis of VA was conducted on a system consisting of a Tracor (Austin, TX) pump, Rheodyne 7125 injection valve, Waters (Millipore) 486 tunable absorbance detector, and Shimadzu CR601 integrator. A Supelco LC-18 (No. 58298) column protected by an Upchurch precolumn packed with ODS C-18 was used for ROH and RE separation. Isocratic elution of ROH with methanol/acetonitrile/chloroform (47:47:6, v/v/v at a flow rate of 1.5 mL/min was detected at 325 nm. Serum RE were eluted isocratically with acetonitrile/methylene chloride (80:20, v/v with 0.5 g/L ammonium acetate added at a flow rate of 1.5 mL/min and detected at 325 nm (DeRuyter and DeLeenheer 1978, Furr 1990, Ross 1981).

Statistical analysis.  Data were compared using one-way ANOVA and Fisher's protected least significant difference analysis (StatView 512⁺, BrainPower, Calabasas, CA). Differences were considered significant at P < 0.05. Values shown represent group means ± SD. Some values were log transformed before ANOVA as indicated in table footnotes.

	RESULTS

Abstract Introduction Methods Results Discussion References

Diet analysis.  Analysis of the pelleted diets used in Studies 1 and 2 revealed substantial losses (17-55%) of C and VA. The actual concentrations of the diets were 15.7 nmol/g VA and 28.9 nmol/g C (Study 1) and 28.9 nmol/g VA and 79.5 nmol/g C (Study 2). Analysis of the improved, pelleted diets fed in Study 3 showed a high percentage of recovery of both C (96%) and VA (94.5%). Similarly, analysis of the prepared oral doses showed 100% recovery of the added C and VA. Therefore the actual weight ratio of C to VA provided in the diets was ~3.5:1, 15:1, and 12:1 in Studies 1, 2 and 3, respectively.

Study 1

Growth.  The pelleted diets provided for adequate and generally uniform growth for all groups. There were no significant differences in body weight among groups at the beginning of the experiment (P > 0.08). At the termination of the study, the +VA:VA/+C-fed group had significantly lower body weights (P < 0.05) than other groups, which may be explained by the slightly (but not significantly) lower initial body weight of this group. All groups grew well and did not exhibit any signs of VA deficiency (data not shown).

Tissue VA.  Upon trial initiation, base-line total liver VA stores were 17.5 µmol (Table 1). After a 35-d prefeeding period, total liver VA stores were higher in +VA-fed ferrets and lower for ferrets fed VA diets compared with base-line values. Ferrets fed the VA prefeeding diet and the +VA/C experimental diet had significantly greater liver VA stores compared with liver stores of ferrets fed only the VA prefeeding diet. Ferrets that were fed VA:VA/+C diets had significantly lower hepatic VA stores compared with those that received only the VA prefeeding diet. Significantly lower adrenal VA concentrations were observed in ferrets fed +VA:VA/+C and VA:VA/+C diets compared with stores from ferrets fed only the +VA prefeeding diet or the +VA:+VA/C diet (Table 1).

View this table: [in this window] [in a new window]   Table 1. Tissue and serum vitamin A (VA) levels from ferrets that were fed either a VA-sufficient (+VA) or VA-deficient (VA) diet for 35 d and then fed a diet with either VA or -carotene (C) for 16 d (Study 1)1,2

An estimate of the daily utilization of VA was calculated by subtracting total hepatic VA stores at trial initiation from VA stores after the VA prefeeding period and dividing by the number of days fed. The estimated utilization rate was 0.28 µmol/d (~80 µg/d).

Serum.  Retinol (ROH), RE (sum of retinyl esters: oleate, palmitate, stearate, linoleate, myristate, pentadecanoate and heptadecanoate), and total VA (ROH + RE) in serum are presented in Table 1. Serum RE and total VA were both significantly lower in the VA:VA/+C group and in the ferrets that were fed only the VA prefeeding diet than all other groups. Figure 2 shows a typical serum RE profile for ferrets fed +VA:+VA/C (left panel), showing high levels of retinyl stearate and palmitate, and ferrets fed VA:VA/+C (right panel), showing reduced levels of retinyl stearate and palmitate.

View larger version (24K): [in this window] [in a new window]   Fig 2. Serum retinyl ester profile of ferrets receiving either +VA:+VA/C (prefeeding diet:experimental diet) (left panel), or VA:VA/+C (right panel). The chromatogram shows retinol (Rol) and esters: retinyl linoleate, oleate, palmitate and stearate (RL, RO, RP and RS, respectively). Diets: +VA, vitamin A-containing diet; VA, vitamin A-free diet; +C, -carotene-containing diet; C, -carotene-free diet.

Radiolabel analysis.  Radioactivity measured in feces was highly variable, ranging from 4.1 to 63.5% recovery, and was not significantly different among treatments (data not shown). No significant differences were seen between groups in radioactivity recovered in liver and adrenal tissue; however, recovery of the ¹⁴C dose from serum of the ferrets fed VA:+VA/C was significantly higher than recovery form the +VA:VA/+C group (Table 2).

View this table: [in this window] [in a new window]   Table 2. Radioactivity of serum, liver and adrenals from ferrets that received 6.63 Bq 14C--carotene (C) 2 d before termination in Study 11,2

The distribution of radioactivity was compared between saponified and nonsaponified liver samples. The results of one ferret per treatment are shown in Figure 3. After saponification, radioactivity eluting with the major liver retinyl esters (palmitate and stearate) decreased as the radioactivity associated with ROH increased. Ferrets in all treatment groups converted ¹⁴C-C to VA except for those that received VA pelleted diets during the prefeeding period followed by VA/+C diets in the experimental feeding period. This group had some unmetabolized ¹⁴C-C in liver (reflected by radioactivity in the peak corresponding to elution of C + retinyl palmitate after saponification).

View larger version (29K): [in this window] [in a new window]   Fig 3. The distribution of radioactivity in saponified and nonsaponified liver tissue after a 6.63 Bq 14C--carotene (C) dose from one ferret per diet group in Study 1 showing that the ferrets were able to convert the C to vitamin A (VA). Diet treatments are indicated by prefeeding diet:experimental diet. The prefeeding diet was fed for 35 d and the experimental diet for 16. The radiolabeled dose was given orally to ferrets 2 d before termination of the study. Photodiode array (PDA) analysis verified peaks 1, 2 and 3 to be retinol, retinyl palmitate +-carotene and retinyl stearate, respectively.

Study 2

View this table: [in this window] [in a new window]   Table 3. Liver Vitamin A (VA) of ferrets that were fed either a VA-sufficient (+VA) or VA-deficient (VA) prefeeding diet for 33 d followed by an experimental diet containing VA and/or -carotene (C) for 21 d (Study 2)1,2

Study 3

Hepatic total VA stores of ferrets receiving +VA/C pelleted feed:placebo carrier or VA/C pelleted feed:+VA liquid carrier were significantly higher than in ferrets after the VA prefeeding period (branchpoint) and were significantly greater than in all other experimental groups, but were not significantly different from one another (Table 4). Provision of C alone resulted in significantly lower liver VA concentrations, but not total stores, compared with values after the VA prefeeding period. Baseline values of both liver VA concentration and total stores were significantly higher than those of any of the experimental groups.

View this table: [in this window] [in a new window]   Table 4. Liver and adrenal Vitamin A (VA) of ferrets that were fed a VA-deficient (VA) diet for 21 d followed by a diet that contained either VA or -carotene (C) in either a pelleted form or a liquid carrier for 16 d (Study 3)1,2

Adrenal VA stores and concentrations were significantly higher in the experimental groups that were fed VA than in all other experimental groups (Table 4).

As expected, no C was detected in tissues of ferrets fed diets without C. Surprisingly, liver and adrenal C stores were significantly greater in ferrets receiving C from pelleted feed than in ferrets receiving C in a liquid carrier (Table 5). Similarly, serum C concentrations were significantly higher in the pelleted group than in the carrier-fed group.

View this table: [in this window] [in a new window]   Table 5. Liver and adrenal -carotene (C) stores and serum C concentration in ferrets fed a vitamin A (VA) deficient diet for 21 d followed by a diet containing C in either liquid or pelleted form for 16 d (Study 3)1,2

Following the 21-d VA prefeeding period, serum ROH, RE and total VA concentrations were significantly lower than base-line values (Table 6). Serum ROH concentrations were significantly lower in the group of ferrets fed VA/C pelleted:placebo carrier than in all other groups. Serum RE concentration in ferrets fed VA/C pelleted:placebo carrier was 4.4% of base-line values; however, serum ROH concentrations were 36.7% of base-line values. After the 16-d experimental feeding period, total serum VA (ROH + RE) and RE for both groups fed VA were significantly greater than all for other treatment groups, but not different from one another. Ferrets fed VA/+C pelleted:placebo carrier had significantly higher total serum VA and RE concentrations than ferrets fed VA/C pelleted:+C carrier.

View this table: [in this window] [in a new window]   Table 6. Serum retinol, retinyl esters (RE) and total vitamin A (VA) in ferrets that were fed a VA-deficient diet for 21 d followed by an experimental diet containing VA and/or -carotene (C) in either a liquid of pelleted form for 16 d (Study 3)1,2

	DISCUSSION

Abstract Introduction Methods Results Discussion References

The incidence of hypovitaminosis A in many countries around the world is quite high and leads to increased morbidity and mortality among children (Sommer et al. 1984, Sommer and West 1996). Populations in which VA deficiency is prevalent invariably rely on carotenoid-containing foods for the majority of their VA intake (FAO/WHO 1988). Solomons and Bulux (1993) and Erdman et al. (1993) enumerated a variety of factors that contribute to the poor bioconversion of plant carotenoids to VA in humans. Among those factors are complex food matrices, intestinal parasites and low fat intakes. The VA status of the individual also may affect their ability to utilize carotenoids.

Animal models are necessary to adequately study the factors that affect the bioconversion of carotenoids from foods to VA (Van Vliet 1996). An appropriate animal model should closely mimic the human uptake, absorption and metabolism of VA and C. The ferret model has been evaluated for uptake and tissue distribution of a variety of carotenoids (Ribaya-Mercado et al. 1989 and 1992, Tang et al. 1993 and 1995, Wang et al. 1991 and 1992, White et al. 1993a and 1993b, Zhou et al. 1996). Most notably, Wang et al. (1991) demonstrated that incubation of ferret intestinal homogenates with C resulted in the formation of substantial quantities of retinoids and -apo-carotenals. Moreover, perfusion of the upper portion of the small intestine with radiolabeled C resulted in the appearance of C and these same metabolic products in mesenteric lymph and portal circulation. A notable exception to similarities between ferrets and humans is substantially elevated serum VA in ferrets, particularly the retinyl ester contration (Ribaya-Mercado et al. 1992).

This work was designed to compare the utilization of C and VA in pelleted diets or liquid carrier in ferrets with differing VA status. Hepatic VA stores were the major indicator of utilization. Concentrations of C (liver, serum and adrenal), VA (serum and adrenal) and the presence of radiolabeled C and metabolites (Study 1) also were monitored. A secondary objective was to use hepatic losses of VA in ferrets fed VA-deficient diets to estimate the daily VA utilization rate of this species. Third, the effect of VA depletion on the serum total VA and retinyl ester concentrations was studied.

Ferrets with marginal VA status were unable to utilize dietary C at a rate sufficient to replenish needed liver VA stores. In humans, liver VA concentrations <70 µmol (20 µg)/g generally are considered marginally deficient, whereas concentrations <17.5 µmol (5 µg)/g are considered deficient (Olson 1991). Ferrets fed VA diets throughout Study 1 reached liver VA concentrations as low as 6.6 µg/g yet did not convert dietary C to VA in quantities large enough to increase or maintain liver VA stores. The actual efficiency with which ferrets convert C to VA is not known; however, it is evident that C provided at ~ 5:1 (Study 1), ~ 15:1 (Study 2) or ~ 12:1 (Study 3) weight ratios was not adequate to produce hepatic VA stores equivalent to those resulting from VA feeding even though the ratios were higher than the estimated 6:1 conversion efficiency of C to VA required to meet the human RDA (NRC 1989) for Studies 2 and 3. In contrast, in Study 1, ferrets converted ¹⁴C-C to VA when C was provided in the liquid carrier Ensure in a single dose. Determination of hepatic ¹⁴C-C and ¹⁴C-VA metaolites indicated that all groups of ferrets efficiently converted ¹⁴C-C to ¹⁴C-ROH and RE except for ferrets fed VA/+C diets after a VA prefeeding period (Fig. 3), suggesting inefficient or impaired conversion with diets devoid of VA.

Neither the method of delivery (pelleted feed or liquid carrier in Study 3) nor the improvement of pelleting procedures (Study 3) improved the utilization of C as measured by changes in hepatic total VA stores. For Study 3, liver VA concentrations did not differ in ferrets fed VA in pelleted feed or liquid carrier. Similarly, liver VA stores were the same in ferrets fed C from either source but were significantly lower than in ferrets fed VA. Somewhat contradictory is the fact that C concentrations in liver, adrenal and serum were greater in ferrets fed pelleted C compared with ferrets fed C in liquid carrier. This probably occurred because the relative percentage of bioavailibility of VA does not change over a broad range of dose levels in many species (Olson 1991), whereas the bioavailibility of C decreases with increased dose level (Brubacher and Weisner 1985). Another explanation may be that the VA or C from pelleted diet was fed for ~20 h/d, whereas the oral dose was provided once daily. Thus, the single daily dose of C from the oral dose may have exceeded the ferret's absorptive capacity for C from that "meal."

Liver VA stores for the groups of ferrets fed C in either pelleted or liquid form from Study 3 were similar to those of ferrets that consumed the VA/C pelleted diet with the placebo carrier. Similarly, Ribaya-Mercado et al. (1992) reported no difference in concentrations of liver RE and ROH in ferrets fed 80 µg of C/g diet (compared with our 65 µg/g) for 3 wk compared with control-fed ferrets. The average intake of C from the supplemented diets in that study was ~15.5 mg/d (compared with ~3.8 mg/d in Study 3).

Many investigators have studied the effects of VA deficiency on the utilization of dietary C. Villard and Bates (1986) suggested that low tissue VA increased dioxygenase activity in rats. Similarly, Van Vilet et al. (1992) demonstrated increased 15,15 dioxygenase activity in vitro in intestinal homogenates from VA-deficient hamsters. As VA status of the hamsters and rats decreased, 15,15 dioxygenase activity increased as indicated by the increased production of VA from C. Feeding VA depletion diets to weanling rats for 3 or 5 wk followed by oral supplementation with VA or C in oil (emulsion-based solutions) or diet indicated that VA repletion (judged by growth rate and tissue VA accumulation) was best from the emulsion, poor from the diet and improved with severity of VA depletion (Grolier et al. 1995). Upregulation of C cleavage has not been studied in an animal model that absorbs intact C at physiologic doses; however, it seems logical that a similar mechanism may be involved in the conversion of C to VA in the ferret.

Wang and co-workers (1992) demonstrated in vivo conversion of C to VA (ROH, RE) and several -apo-carotenoids after intestinal perfusion of ferret small intestine with a C in micellar solution. Although these investigators did not specifically measure conversion efficiency, Wang (personal communication) estimated that the conversion of C to VA in ferrets was ~ 12:1 under the optimal conditions of the perfusion assays. The current studies suggest that the conversion efficiency of C to VA is < 15:1 when C is provided in food or a liquid supplement. Thus, although this species can utilize C for VA value, the ferret is among species that are poor converters.

Our results concerning serum retinoid profile analysis support Ribaya-Mercado et al. (1992) and Tang et al. (1993) who reported high RE concentrations, in particular retinyl palmitate and retinyl stearate, in ferret serum. We found retinyl esters to be the predominant form of VA in ferret serum (51-94% of total serum VA, in Studies 1 and 3). Ferrets with the greatest liver VA stores had the highest percentage of serum RE compared with ferrets of lower VA status. Ribaya-Mercado et al. (1992) reported that serum RE represented the majority (92-94%) of the total serum VA before, after and without (control) C supplementation of healthy, VA-sufficient ferrets. Provision of ¹⁴CC to ferrets via an intestinal perfusion system resulted in 36 ± 6% of the absorbed ¹⁴CC in portal vein blood associated with retinyl esters and 8 ± 2% associated with ROH (Wang et al. 1992).

High serum RE concentrations are common in carnivores (Schweigert et al. 1988); however, noncarnivores normally have < 5% of total serum VA as RE. However, elevated serum RE clearly are normal for the ferret and do not indicate a toxic VA state. With hepatic VA stores of 3.5 and 4.9 µmol, as seen in Studies 1 and 3, respectively (levels considered marginally deficient in humans; Olson, 1991), total RE still accounted for 65-67% of total serum VA. Similarly, measurable concentrations of plasma RE were observed in dogs fed VA-restricted diets for 1 y (Wilson et al. 1987). Fasting plasma RE concentrations in non-VA-supplemented, healthy humans range from 0-0.34 µmol/L (Bankson et al. 1986). Humans administered pharmacologic doses of VA had peak plasma RE concentrations of 8 µmol/L (Tang and Russell 1991). Satisfactory plasma total VA concentrations for humans range from 1.0 to 2.8 µmol/L. Clearly, ferrets have high serum RE concentrations (0.83-8.61 µmol/L in Study 3) relative to humans, indicating differences between ferret and human VA metabolism.

We conclude the following: 1) the efficiency with which C is converted to VA by ferrets under conditions of marginal or more deficient VA status is poorer than 15:1; 2) the in vivo conversion of dietary C to VA and the subsequent accumulation of VA in tissue occurred less efficiently when VA status was compromised; and 3) the ferret has a low capacity to derive VA from C. Thus, it appears that the ferret is not an appropriate model for evaluating the ability of carotenoids to increase VA status in VA-deficient populations. However, this species readily absorbs C intact, and studies involving the effects of absorbed C on metabolism should be quite appropriate.

A set of studies similar to those described above with ferrets was performed in our laboratory with the Mongolian gerbil (Lee et al. 1998), a species that absorbs physiologic doses of C intact (Pollack et al. 1994). It was found (Lee et al. 1998) that the gerbil was much more like the human in its ability to utilize dietary C for VA. The gerbil is thus the species of choice for C utilization studies.

It is estimated that the daily utilization rate of VA by ferrets approaches 175 µg (estimate ranged from 80 to 171 µg in 3 studies). Feeding VA at a level of 5.41 nmol/g diet as VA resulted in appropriate growth and hepatic stores. In the two studies (1 and 3) in which retinyl esters were monitored, ferrets began with very high serum levels of retinyl esters, especially palmitate and stearate. VA depletion substantially reduced retinyl esters but had little effect on retinol concentration.

These studies provide information regarding the utilization of C for VA stores in the ferret, an approximate utilization rate of VA for this species and data on retinyl ester changes in serum after partial depletion of VA.

	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

• Bankson D. D., Russell R. M., Sadowski J. A. Determination of retinyl esters and retinol in serum or plasma by normal phase liquid chromatography: methods and applications. Clin. Chem. 1986; 32:35-40 [Abstract/Free Full Text]
• Bowen P. E., Mobarhan S., Henderson C., Stacewicz-Sapuntzakis M., Friedman H., Kaiser R. Hypocarotenemia in patients fed enterally with commercial liquid diets. J. Paren. Enteral Nutr. 1988; 12:484-489
• Brubacher G. B., Weisner, The vitamin A activity of beta-carotene. Int. J. Vitam. Nutr. Res. 1985; 55:5-15 [Medline]
• de Pee S., West C. E., Muhilal, Karyadi D., Hautvast G.A.J. Lack of improvement in vitamin A status with increased consumption of dark-green leafy vegetables. Lancet 1995; 346:75-81 [Medline]
• DeRuyter M. G., DeLeenheer A. P. Simultaneous determination of retinol and retinyl esters in serum or plasma by reversed-phase-high-performance-liquid-chromatography. Clin. Chem. 1978; 24:1920-1923 [Abstract/Free Full Text]
• Erdman J. W. Jr., Bierer T. J., Gugger E. T. Absorption and transport of carotenoids. Ann. N.Y. Acad. Sci. 1993; 691:76-85 [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.
• Furr H. C. Reversed-phase high-performance liquid chromatography of retinyl esters. Methods Enzymol. 1990; 189:85-94 [Medline]
• Futterman S., Andrews S. J. Metabolism of the retina. J. Biol. Chem. 1964; 239:81-84 ^{[Free Full Text]}
• Grolier P., Agoudavi S., Azais-Braesco V. Comparative bioavailability of diet-, oil- and emulsion-based preparations of vitamin A and beta-carotene in the rat. Nutr. Res. 1995; 15:1507-1516
• Gugger E. T., Bierer T. L., Henze T. M., White W. S., Erdman, J. W. Jr. Beta-carotene uptake and tissue distribution in ferrets (Mustela putoius furo). J. Nutr. 1992; 122:115-119
• Hubbard R., Brown P. K., Bownds D. Methodology of vitamin A and visual pigments. Vitamin A Group 1990; 243:615-653
• Lee C. M., Lederman J. D., Hofmann N. E., Erdman J. W. Jr. The Mongolian gerbil (Meriones unguiculatus) is an appropriate animal model for the evaluation on the conversion of -carotene to vitamin A. J. Nutr. 1998; 128:280-286 [Abstract/Free Full Text]
• McLain, D. E., Thomas, J. A. & Fox, J. G. (1988) Biology and Diseases of the Ferret. Lea & Febinger, Philadelphia, PA.
• National Research Council (1989) Recommended Dietary Allowances, 10th 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
• Peng Y. M., Peng Y. S., Lin Y. A nonsaponification method for the determination of carotenoids, retinoids, and tocopherols in solid human tissue. Cancer Epidemiol. Biomarkers Prev. 1993; 2:139-144 [Abstract]
• Ribaya-Mercado J. D., Fox J. G., Rosenblad W. D., Blanco M. C., Russell R. M. -Carotene, retinol, and retinyl ester concentrations in serum and selected tissues of ferrets fed -carotene. J. Nutr. 1992; 122:1898-1903
• Ribaya-Mercado J. D., Holmgreen S. C., Fox J. G., Russell R. M. Dietary -carotene absorption and metabolism in ferrets and rates. J. Nutr. 1989; 119:665-668
• Ross C. A. Separation of long-chain fatty acid esters of retinol by high-performance liquid chromatography. Anal. Biochem. 1981; 115:324-330 ^[Medline]
• Schweigert F. J. Insensitivity of dogs to the effects of non specific bound vitamin A in plasma. Int. J. Vitam. Nutr. Res. 1988; 58:23-25 [Medline]
• Solomons N. W., Bulux J. Plant sources of provitamin A and human nutriture. Nutr. Rev. 1993; 51:199-204 [Medline]
• Solomons N. W. Plant sources of vitamin A and human nutrition: renewed strategies. Nutr. Rev. 1996; 54:89-93 ^[Medline]
• Sommer A., Katz J., Tarawoto I. Increased risk of respiratory disease and diarrhea in children with preexisting mild vitamin A deficiency. Am. J. Clin. Nutr. 1984; 40:1090-1098 [Abstract/Free Full Text]
• Sommer, A. & West, K. P., Jr. (1996) Vitamin A Deficiency: Health, Survival and Vision. Oxford University Press, New York, NY.
• Tang G., Dolnikowski G. G., Blanco M. C., Fox J. G., Russell R. M. Serum carotenoids and retinoids in ferrets fed canthaxanthin. J. Nutr. Biochem. 1993; 4:58-63
• Tang G., Russell R. M. Formation of all-trans-retinoic acid from all-trans retinyl palmitate in humans. J. Nutr. Biochem. 1991; 2:210-213
• Tang G., Blanco M. C., Fox J. G., Russell R. M. Supplementing ferrets with canthaxanthin, other carotenoids, vitamin A and vitamin E. J. Nutr. 1995; 125:1945-1951
• Van Vliet, T. (1996) Absorption of -carotene and other carotenoids in humans and animal models. Eur. J. Clin. Nutr. 50 (suppl. 3): 532-537.
• Van Vliet T., Van Schaik F., Van Den Berg H. Beta-carotene metabolism: the enzymatic cleavage to retinal. Neth. J. Nutr. 1992; 53:186-190
• Villard V., Bates C. J. Carotene dioxygenase (EC 1.13.11.21) activity in rat intestine: effects of vitamin A deficiency and of pregnancy. Br. J. Nutr. 1986; 56:115-122 ^[Medline]
• Wang, X. D., Krinsky, N. I., Marini, R. P., Tang, G., Yu, J., Hurley, R., Fox, J. G. & Russell, R. M. (1992) Intestinal uptake and lymphatic absorption of -carotene in ferrets: a model for human -carotene metabolism. Am. J. Physiol. G263: 480-486.
• Wang, X. D., Tang, G. W., Fox, J. G., Krinsky, N. I. & Russell, R. M. 1991) Enzymatic conversion of -carotene into beta-apo-carotenals and retinoids by human, monkey, ferret, and rat tissue. Arch. Biochem. Biophys. 285: 8-16.
• White W. S., Peck K. M., Bierer T. L., Gugger E. T., Erdman J. W. Jr. Interactive effects of oral -carotene and canthaxanthin in ferrets. J. Nutr. 1993a; 123:1405-1413
• White W. S., Peck K. M., Ulman E. A., Erdman J. W. Jr. The ferret as a model for the evaluation of the bioavailabilities of all-trans -carotene and its isomers. J. Nutr. 1993b; 123:1129-1139
• Wilson, D. E., Hejazi, J., Elstad, N. L., Chan, I. F., Gleeson, J. M. & Iverius, P. H. (1987) Novel aspects of vitamin A metabolism in the dog: distribution of lipoprotein retinyl esters in vitamin A- deprived and cholesterol fed animals. Biochem. Biophys. Acta 922: 247-258.
• Zhou J-R., Gugger E. T., Erdman J. W. Jr. The crystalline form of carotenes and the food matrix in carrot root decrease the relative bioavailability of beta and alpha carotene in the ferret model. J. Am. Coll. Nutr. 1996; 15:84-91 [Abstract]

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