α-Retinol and 3, 4-didehydroretinol support growth in rats when fed at equimolar amounts and α-retinol is not toxic after short-term pharmacological doses

Abstract

Dietary α-carotene is found in orange and purple-orange carrots. Upon α-carotene’s central cleavage in the intestine, α-retinal and retinal are formed and reduced to α-retinol and retinol. Previous reports suggested that α-retinol has 2% biopotency of all-trans-retinyl acetate due in part to its inability to bind to retinol-binding protein. The current studies re-determined α-retinol's biopotency compared with retinol and 3, 4-didehydroretinol in a growth assay. Weanling rats (n 40) were fed vitamin A-deficient diet for 8 weeks, divided into 4 treatment groups (n 10/group), and orally dosed with 50 nmol/d retinyl acetate (14.3 µg retinol), α-retinyl acetate (14.3 µg α-retinol), 3, 4-didehydroretinyl acetate (14.2 µg 3, 4-didehydroretinol), or cottonseed oil (negative control). Supplementation continued until control rats showed deficiency signs 5 weeks after supplementation began. In comparison to retinol, body weights and area-under-the-response curves revealed that α-retinol and 3,4-didehydroretinol had 40–50 and 120–130% bioactivity, respectively, compared with retinol. In study 2, rats (n 40) received 70 nmol retinyl acetate and 0, 17.5, 35, or 70 nmol α-retinyl acetate daily for 3 weeks. Although liver retinol differed among groups, α-retinol did not appreciably interfere with retinol storage. In study 3, 3.5 µmol/d α-retinyl acetate was fed to rats (n 15) for 21 d and groups were killed at 1, 2, and 3 week intervals. No hepatic toxicity was observed. In conclusion, α-retinol and didehydroretinol are more biopotent than previously reported with sustained equimolar dosing at 50 nmol/d, which was an amount of retinol known to keep rats in vitamin A balance.

Introduction

Studies conducted in the 1950’s suggested that α-retinol (αR) (Fig. 1) was a biologically inactive form of vitamin A, with approximately 2% biopotency when rat growth and liver storage bioassays of the geometric isomers of vitamin A aldehyde were compared with the US Pharmacopeia Vitamin A Reference Standard⁽¹⁾. Another study reported 2.6% bioactivity⁽²⁾ and Sneider et al.⁽³⁾ confirmed lack of bioactivity but suggested that αR has partial vitamin A function. α-Retinol’s lack of bioactivity was eloquently discussed by Pitt on more than one occasion^(4,5,6). In 1969, it was concluded that αR had only 2.1% of the activity of retinol (ROH) when αR was assayed for growth-promoting activity in rats⁽⁶⁾. α-Retinol was very effective at inducing signs of hypervitaminosis A in both in vitro and in vivo models⁽⁷⁾. On the other hand, Shantz and Brinkman⁽⁸⁾ demonstrated that 3, 4-didehydroretinol (DR; vitamin A₂), which is the predominant form of vitamin A in some freshwater fish, has 40% biological activity compared with ROH (vitamin A₁). Methods to quantify and selectively determine analogue concentrations in tissues were not available when these studies were performed. A more recent study, quantified αR with HPLC after α-carotene dosing to gerbils and found significant liver αR storage⁽⁹⁾. Therefore, it was hypothesized that αR could be used as a chylomicron tag because it appeared to be sequestered in the liver.

Fig. 1.

The chemical structures of β-carotene, α-carotene, retinol, α-retinol, 3, 4-didehydroretinol, and C23-alcohol, which was used as an internal standard.

The difference between the biological activities of αR and DR can be partially explained by the inability of αR compared with the ability of DR to bind to retinol-binding protein (RBP), the specific carrier protein of ROH. RBP is synthesized by hepatic parenchymal cells as a 24-kDa precursor which is then converted to RBP by the cotranslational removal of a 3.5-kDa polypeptide⁽¹⁰⁾. This protein product is called apo-RBP and holo-RBP when complexed with ROH or an analogue⁽¹¹⁾. Moreover, DR is carried by the retinol transporter from the intestine at the same rate as ROH⁽¹²⁾. In 1975, Muhilal and Glover tested the binding affinity of ROH and its analogues to RBP⁽¹³⁾. Unlike ROH, αR failed to bind in vitro with apo-RBP, but DR did bind with a slower saturation time than ROH. As a result, they suggested that the fission of the ring as in γ-retinol or changing the double bond from the 5, 6-position (in β-retinol) to the 4, 5-position as in αR prevents the compound from binding to RBP.

These discoveries are consistent with a study by Tanumihardjo and Howe⁽⁹⁾ who found that αR was not detected in the serum of gerbils fed α-carotene even though abundant amounts were in the liver. The αR hepatic concentration was very similar to the difference in hepatic ROH concentration between the α-carotene and control groups, consistent with central cleavage of α-carotene in the intestinal brush border. In addition, α-carotene may have a negative effect on β-carotene absorption in rats⁽¹⁴⁾. By feeding a mixture of α-carotene and β-carotene (1:2) to rats, hepatic ROH stores were decreased compared with feeding β-carotene alone⁽¹⁴⁾. This suggests a potential interaction or competition between αR and ROH but the degree and site of interaction have not been clearly defined.

The aims of these studies were, first, to determine the biopotency of αR and DR compared with ROH in a controlled experiment where liver reserves could be accurately quantified with HPLC and ROH was fed at a concentration known to keep rats in vitamin A balance, i.e. 50 nmol/d⁽¹⁵⁾. Second, the influence of αR on liver ROH storage was investigated during graded feeding of αR, while the amounts of ROH were equalised across treatment groups at a higher level than in the first study to allow for some ROH storage. Therefore, the overall influence of physiological graded doses of αR on ROH uptake and storage could be determined. Third, large doses of αR were fed to determine accumulation and disappearance over time and potential liver pathology.

Materials and Methods

Syntheses of α-retinyl and 3, 4-didehydroretinyl acetate

α-Retinyl acetate was synthesized using a previously described method for the synthesis of ¹³C-retinyl acetate⁽¹⁶⁾ except that α-ionone (Sigma Aldrich, St. Louis, MO, USA) substituted for β-ionone as the starting reagent and ¹³C was not added. 3, 4-Didehydroretinyl acetate was synthesized using previously published methods⁽¹⁷⁾ and stored at −70 °C until use. The synthesized acetate esters were purified (>95%) on 8%-water-deactivated neutral alumina using hexanes and diethyl ether. Purity of both compounds was confirmed by thin layer chromatography, UV-VIS spectrophotometry, and an HPLC equipped with photodiode array detection.

Animals and diet

Animal use was approved by the University of Wisconsin (UW)-Madison Animal Care and Use Committee, and all animal procedures adhered to the public health service policy on humane care and use of laboratory animals. The UW College of Agriculture and Life Sciences’ facilities are AAALAC accredited and frequently inspected internally and externally to ensure compliance. Rats (n 98) were housed individually in aspen bedding in a temperature and humidity controlled environment with a 12 h:12 h light:dark cycle. Shaven aspen wood was chosen as bedding because it absorbs moisture, eliminates odor, and has low nutritional value. Corn cobs, which are another option, might have interfered with this bioassay due to kernel contamination. Upon arrival, rats were fed ad libitum with a vitamin A-free purified diet ⁽¹⁸⁾. The vitamin A-deficient diet (TD.04175; Harlan-Teklad, Madison, WI, USA) contained (in g per kg diet): casein (200); dl-methionine (3); sucrose (280); corn starch (215.0436); maltodextrin (150); cellulose (50); soybean oil (55); mineral mix AIN-93G TD.94046 (35); calcium phosphate (3.2); vitamin mix without added A, D, E, and choline TD.83171 (5); vitamin D₃ (0.0044); vitamin E (0.242); choline dihydrogen citrate (3.5); TBHQ (0.01).

Study 1

Rats were chosen as the model for this study because they are a defined model for growth assays and are quickly made vitamin A deficient because they are born with low stores. For the biopotency study, 21-d old weanling male Sprague Dawley rats (n 40) (Charles River; Kingston, NY, USA) were fed vitamin A-deficient diet and weighed daily for the duration of the study. At the beginning of week 9, the rats were divided into four weight-matched treatment groups (n 10/group) and orally dosed with 50 nmol retinyl acetate (14.3 µg ROH), α-retinyl acetate (14.3 µg αR), 3, 4-didehydroretinyl acetate (14.2 µg DR), or plain cottonseed oil vehicle (negative control group). All doses were administered in 100 µl cottonseed oil. This dose level (50 nmol) was chosen because it maintains vitamin A balance in male rats^(15,19,20). The daily dosing regimen was continued until most rats in the control group had abnormal secretions around the eyes and 20% of these were showing severe signs of vitamin A deficiency (i.e. scruffy coat and swollen ankles). All rats were killed 5 weeks after supplemental dosing began and tissues were collected. Aliquots of serum, and weighed samples of liver, kidney, spleen, and lung were analysed for ROH, αR, and DR by HPLC.

Serum analysis

Serum was analysed using a published procedure⁽⁹⁾ with slight modification. Internal standard C-23 alcohol, a synthesised β-apo-carotenol⁽⁹⁾ (Fig. 1), was added to 500 µl serum and an equal volume of ethanol with 0.1% butylated hydroxytoluene was added. The sample was extracted three times with 1 ml hexanes. Supernatant fractions were pooled and dried under argon. The extract was reconstituted in 100 µl methanol:dichloroethane (DCE) (50:50, v/v) and 50 µl was injected into the HPLC. The isocratic HPLC system included a guard column, Waters Symmetry® C18 column (3.5-µm, 4.6 × 75 mm), Waters Resolve™ C18 column (5-µm, 3.9 × 300 mm), Rheodyne injector, Shimadzu SPD-10A UV-VIS detector, Waters Delta 600 pump and controller, and Shimadzu C-R7A Plus data processor. The mobile phase was 87.5:12.5 acetonitrile:water (v/v) with 10 mM ammonium acetate as a modifier at 0.7 ml/min.

Rat tissue analysis

Liver, kidney, lung, and spleen were analysed using published methods^(9,18,21) with minor modification. Tissue (0.5 g) was ground with sodium sulfate (1.5 g) in a mortar. Purified C23-alcohol was added to determine extraction efficiency. The tissue was extracted repeatedly with dichloromethane to 10 ml; 5 ml was dried under argon. The film was redissolved in 0.75 ml ethanol and saponified with 0.4 ml potassium hydroxide:water 50:50 (wt/v) at 45°C for 30 min. The reaction was quenched with 0.5 ml water. The solution was extracted 3 times with 0.5 ml hexane. The hexane layers were pooled, washed with 0.5 ml water, and dried under argon. The film was redissolved in 100 µl methanol:DCE (50:50, v/v) for liver and 200 µl for kidney, lung, and spleen. An aliquot of 50 µl was injected into the HPLC for liver and 25 µl for the other tissues. The Waters HPLC system included the same columns as above, and a Waters 1525 binary HPLC pump, 717 auto sampler, and 996 photodiode array detector. The mobile phase was the same but at a flow rate of 0.8 ml/min. The tissue vitamin A results are reported as “retinol”, which included retinol and retinyl esters because tissues were saponified.

Study 2

For the interaction study, 21-d old weanling male Sprague Dawley rats (n 40) were fed a vitamin A-deficient diet for 2 weeks. Rats were divided into 4 treatment groups (n 10/group). To allow some liver ROH storage, this study used a higher dose of ROH than study 1 (i.e. 70 nmol as retinyl acetate), which resulted in adequate liver reserves over time in prior studies^(18,22). Each group was orally dosed with 70 nmol ROH/d and 0, 17.5, 35, or 70 nmol αR (as α-retinyl acetate) daily. Doses were administered in 100 µl cottonseed oil. All rats were killed 3 weeks after dosing began and tissues were collected. Serum and liver were analysed using the same procedures as in study 1.

Study 3

For the toxicity study, 21-d old weanling male Sprague Dawley rats (n 18) were vitamin A-depleted for 3 weeks and then given 3.5 µmol α-retinyl acetate/day dissolved in cottonseed oil, or cottonseed oil alone (control) for 21 d. Rats (n 5) were killed at d 1, 8, and 15 after the final dose. Control rats (n 3) were killed at d 1. Serum, liver, and kidneys were collected at each time point. Serum from both the control and αR-treated groups were analysed for biochemical indicators of liver toxicity, which included potassium, urea, albumin, alkaline phosphatase, alanine aminotransferase, gamma GT, cholesterol, and total bilirubin from a routine chemistry panel. Histological sections of livers were analysed using hematoxylin and eosin stain by the UW School of Medicine and Public Health. Serum analyses were performed by UW Veterinary Medical Teaching Hospital Pathology Services.

Statistical analysis

Animal data were analysed using Statistical Analysis System software (SAS Institute Inc., Version 9.1, Cary, NC; 2002–2003). Outcomes of interest (i.e. rat weights, serum and tissue ROH, αR, and DR concentrations) were evaluated using ANOVA. A Test of Effect Slices was used to determine the day where the body weights were different among the groups. Body weight change area-under-the-curves (AUCs) were calculated by trapezoidal approximation. Values presented are means (SD). Significance was assessed at α < 0.05.

Results

Study 1 body weights and weight changes

After dosing began, group differences in body weight were approaching significance (P = 0.054), the day main effect and the interaction of group and day were significant (both P < 0.0001). Using a Test of Effect Slices for group and day, the difference among the groups occurred at day 19 and this remained until the end of the dosing period (Fig. 2). The final body weights did not differ between the ROH and DR groups, but were lower in the αR and control groups. Interestingly, the liver weights were the same for ROH (19.1 (SD 2.1) g) and DR (20.0 (SD 2.5) g) groups but were higher than the αR (14.6 (SD 1.5) g) and control (13.3 (SD 1.6) g) groups, which did not differ from each other. The kidney weights were highest in the ROH group (3.44 (SD 0.28) g), which differed from control (3.07 (SD 0.17) g), and DR (3.06 (SD 0.49) g) groups but did not differ from the αR group (3.32 (SD 0.4) g). Kidney weights did not differ between αR, DR, and control groups.

Fig. 2.

Growth assay in groups of rats (n 10/group) given 50 nmol/d α-retinol (○), 3, 4-didehydroretinol (●, DR), or retinol (∆) compared with a negative control (▲, Oil) group. Using a test of effect slices, body weights began to differ at 19 d after supplementation began (denoted by *). Area-under-the-growth response curves were all different between the groups. Different superscript letters designate a difference (P < 0.05).

Organ retinol concentrations

All rats in study 1 were severely vitamin A deficient (defined as <70 nmol ROH/g liver)⁽¹⁹⁾. Despite adequate serum ROH concentrations in the ROH group (1.37 (SD 0.21) µmol/l), liver reserves were 14-times less than this deficiency cut-off (Table 1) and the kidney had 7-times more ROH than the liver. In fact, kidney ROH concentrations were higher than liver in all treatment groups by 5 to 10 times. Although treatment compounds were fed at the same amount, more ROH was stored in the ROH group than DR or αR in those groups, respectively. Retinol was not detected in the lung and spleen of the control, DR, and αR groups, which supports prior work^(21,23). DR was only detected in the DR group and αR was only detected in the αR group in all tissues analysed.

Table 1.

Concentrations of retinol, 3,4-didehydroretinol and α-retinolin rat liver, kidney, lung, spleen, and serumafter equimolar oral dosing (50 nmol) for 5 weeks^*

Dietary Treatment	Control				Retinol				3, 4-Didehydroretinol					α-Retinol
	ROH		DR	αR	ROH		DR	αR	ROH		DR		αR	ROH		DR	αR
	Mean	SD			Mean	SD			Mean	SD	Mean	SD		Mean	SD		Mean	SD
Liver (nmol/g)	0·01	0·01c	ND2	ND	5·14	2·86a	ND	ND	0·03	0·01b	0·59	0·4	ND	0·01	0·01c	ND	0·38	0·18
Kidney (nmol/g)	0·08	0·13c	ND	ND	37·0	3·78a	ND	ND	014	0·02b	1·22	0·16	ND	0·10	0·01c	ND	0·46	0·19
Lung (nmol/g)	ND		ND	ND	5·58	l·30a	ND	ND	ND		1 43	0·20b	ND	ND		ND	0·37	0·1c
Spleen (nmol/g)	ND		ND	ND	0·33	0·02a	ND	ND	ND		0·32	0·02a	ND	ND		ND	0·22	0·02b
Serum (|a,mol/l)	0·06	0·01c	ND	ND	1·37	0·21a	ND	ND	0·10	003b	0·52	0·14	ND	0·07	0·01c	ND	0·02	0·01

αR, α-retinol; DR, 3, 4-didehydroretinol; ROH, retinol

^*Data represent mean (SD), n 10. Different superscript letters within each row indicate a difference (P < 0·05) between treatment groups. Comparisons are either for retinol concentrations only (i.e. liver, kidney, and serum) or between the test compounds (i.e. lung and spleen).

^†ND, not detected.

Much lower concentrations of DR and αR were detected in the liver, kidney, spleen, and lung of DR and αR groups, respectively, compared with ROH concentrations in ROH group suggesting an increased utilization or impaired uptake of these analogues (Table 1).

Serum concentrations

The rats in the ROH group were able to maintain an adequate serum ROH concentration (1.37 (SD 0.21) µmol/l) when fed 50 nmol/d. Serum ROH was detectable but deficient (≤0.10 µmol/l) in the control, DR, and αR groups. Serum ROH concentration was slightly higher in the DR group than the control and αR groups (P < 0.05), which may indicate a sparing effect considering that the serum DR concentration was substantial (0.52 (SD 0.14) µmol/l). Serum αR concentration was measurable but low in the αR group.

Bioactivity

In the classic growth assay, αR rats maintained about 50% of the growth rate of ROH and rats on DR maintained >100% of that of ROH (Table 2). Bioactivity was calculated two ways. The first was a simple comparison of the ending body weights in reference to the ROH group. In the second calculation, AUC was computed, control values were subtracted from treatment groups, and these values were compared with the ROH group to achieve a percent difference. AUC was different between all of the groups (Table 2). Similar values for bioactivity (i.e. within 8 to 12%) were obtained using both mathematical methods (Table 2).

Table 2.

Determination of the bioactivity of α-retinol and 3,4-didehydroretinol compared with retinol Rats were fed 50 nmol/d of the test compounds for 5 week Body weights were measured daily

Treatment	Final body weight (g)		Difference from control (g)	% of retinol	Area-under-the growth curve (g * d)		Difference from control	% of retinol
	Mean	SD			Mean	SD
Control*	470	46·2c	0	0	867	298d	0	0
α-Retinol	508	34·0b	37·4	49·5	1165	138c	298	37·9
DR	560	41·8a	900	119	1867	188a	1000	127
Retinol	546	362a	75·5	100	1654	273b	787	100

DR, 3, 4-didehydroretinol.

^*Data represent mean (SD), n 10, differences from the control, or % of retinol.

Different superscript letters within each column indicate a difference (P < 0.05) between treatment groups.

Study 2

The final body weights in study 2 (345.1 (SD 20.4) g) did not differ among the groups, nor did liver weight (16.2 (SD 1.6) g). Despite adequate serum ROH concentration (1.21 (SD 0.07) µmol/l), which did not differ among groups, αR was undetectable. On the other hand, the liver ROH concentration did differ among groups (P = 0.042); although when corrected for total liver weight, the difference in total liver ROH was only approaching significance (P = 0.07). Liver ROH concentration was highest in the group that received the lowest amount of αR, which did not differ from the group that received the highest dose. As anticipated, liver αR concentrations responded in a dose-dependent manner (i.e. 262 (SD 34), 449 (SD 23), and 893 (SD 138) nmol/liver for the 17.5, 35, and 70 nmol/d doses, respectively) (P < 0.0001). This represents total dose liver retentions of 71.3, 61.1, and 60.7%, respectively.

Study 3

No toxicity was evident in αR dosed rats compared with control rats after 21 d. No stellate cell hypertrophy was detected (Fig. 4) or increases in serum markers of liver toxicity (Table 3). Hepatic and renal αR quickly decreased between 1 and 8 d but did not change at 15 d (Table 4).

Fig. 4.

Hematoxylin and eosin staining of liver sections. No hepatotoxicity was noted with 21 d of α-retinyl acetate dosing (3.5 µmol/d). Rats were killed 1, 8, and 15 d after the last dose.

Table 3.

Serum chemistry profile from rats given 3.5 µmolα-retinyl acetate/day (n 5) compared with rats given cottonseed oil alone (control, n 3) for 21 days.

Test	Control*		α-Retinol treated
	Mean	SD	Mean	SD
Potassium (mmol/1)	6·1	1·42	5·32	0·46
Urea (mg/dl)	16·0	1·73	15·8	1·30
Albumin (g/dl)	3·47	0·38	3·08	0·23
Alkaline phosphatase (U/l)	224	88·3	248	34·4
Alanine aminotransferase (U/l)	44·3	9·45	38·0	4·06
Gamma GT (U/l)	<5		<5
Cholesterol (mg/dl)	78·0	25·5	81·0	20·6
Total bilirubin (mg/dl)	0·83	0·85	0·64	0·30

^*Data represent means(SD). No differences were noted between groups. Serum was collected 1 day after the final dose.

Table 4.

Total retinol and α-retinol in rat liver, kidney, and serum after given 3.5 µmolα-retinyl acetate/day for 21 days. All groups were fed a vitamin A-deficient diet for 3 weeks. Rats were killed on d 1, 8, and 15 after the final dose. Control rats were given cottonseed oil alone^*

Time	Control				Day 1				Day 8				Day 15
	ROH		αR		ROH		αR		ROH		αR		ROH		αR
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Liver (nmol/g)	1·47	1·11c	2·1	3·64c	24·5	7·57a	2300	227a	18·8	6·45b	1750	208b	20·4	4·08ab	1770	150b
Kidney (nmol/g)	6·88	4·32a	0·31	0·20b	7·00	4·69a	3·96	2·08a	2·31	0·51b	1·99	0·37b	2·85	0·50ab	1·83	0·39b
Serum, alcohol form (µmol/l)	0·81	0·39a	0·01	0·02c	0·67	0·05a	0·05	0·01b	0·58	0·14a	0·07	0·02a	0·71	0·11a	0·08	0·01a
Serum, ester form (µmol/l)	ND2		ND		ND		0·08	0·01b	ND		0·10	0.01ab	ND		0·11	0·02a

αR, α-retinol; ROH, retinol.

^*Data represent mean(SD), n 5 for rats for d 1, 8, and 15 timepoints, and n 3 for the control group.

Superscript letters within each row indicate a difference (P < 0.05) in concentration of retinol or α-retinol (αR) between timepoints.

^†ND, not detected.

Discussion

In the 1950’s, estimates of bioactivity in rats for αR and DR were 2% and 40% that of ROH, respectively. Bioactivities of 40–50% and 120–130% were demonstrated for αR and DR, respectively, in the same animal model using an equimolar feeding approach. Large doses of αR were given in prior studies; therefore, sustained chylomicron delivery from repeated dosing to replenish the tissues with small amounts did not occur. Thus, most of the dose was shunted to the liver and not recirculated to meet tissue needs because of inability to bind to RBP. The data presented here on αR are consistent with the finding by Clamon et al.⁽²⁴⁾ that α-retinyl acetate sustains growth of hamsters when given intraperitoneally. In prior studies with DR, the preparation was isolated from crude fish liver oil and a distilled fraction in oil fed to the rats⁽²⁵⁾. Therefore, it is difficult to discern exactly how much 3, 4-didehydroretinyl ester was fed during that 28-d study, because only the amount of oil fed was measured and reported. The current preparations were chemically synthesised, carefully purified, and an accurately measured amount equalised to retinyl acetate known to maintain balance in rats was fed each day.

Quadro et al.⁽²⁶⁾ have shown that RBP-knockout mice survive well and show little evidence of vitamin A deficiency when fed a vitamin A-adequate diet. Presumably their extrahepatic tissues obtain enough vitamin A from retinyl esters in lipoproteins to support growth and function. This explains why αR can support growth in rats despite its inability to bind to RBP. α-Retinol can be transported to tissues via chylomicra and their remnants. It can be utilized as a tag to traffic ROH, but can be distinguished from ROH that arrives to tissues through RBP. This concept was recently applied to a lactating sow-nursing piglet dyad where αR was followed from mother to offspring⁽²⁷⁾. Retinol trafficking could be studied during pregnancy with αR and without the use of RBP-deficient animals⁽²⁸⁾.

The growth effects of vitamin A are mediated by retinoic acid via retinoic acid nuclear receptor (RAR)⁽²⁹⁾, which may indicate that the growth effect of αR was functioning by the same mechanism. The apparent conclusion is that the analogues are oxidized to carboxylic acid analogues and that they bind to retinoid nuclear receptor proteins. This phenomenon might have implications for pharmaceutical or cosmetic applications. Due to the limited knowledge of the role of α-retinoic acid in growth promotion, a future study to look for the interaction of α-retinoic acid on RAR or perhaps retinoid X receptor⁽²⁹⁾ is certainly important. 3, 4-Didehydroretinoic acid is a known ligand of RARα, β, and γ with the same affinity as retinoic acid⁽³⁰⁾. 3, 4-Didehydroretinoic acid supports avian growth and development and binding assays for RARβ were nearly identical between chick and mouse receptors⁽³¹⁾. Furthermore, at high doses of 3, 4-didehydroretinyl acetate and retinyl acetate given to dams, both caused terata in fetuses but not to the extent of retinoic acid alone⁽³²⁾. These data support similar bioactivities. The finding of a higher growth AUC for DR should be investigated with 3, 4-didehydroretinoic acid compared with retinoic acid in a similarly designed rat study.

Furthermore, the difference between this and previous work may be partly due to changes in animal feeding and care since the 1950’s. Currently, animal care facilities are often sterile or near-sterile environments behind barriers. Veterinary care and personal protective equipment are common place. Considering the severe degree of vitamin A deficiency in the growth assay study, even a mild infection among the rats would have certainly caused death in all groups. The normal serum ROH concentrations in the ROH-treated rats, which were severely vitamin A deficient based on liver reserves, indicate that the rats were healthy and not suffering from infection, which elicits the acute phase response reducing serum ROH concentrations⁽³³⁾. Moreover, diet formulations have improved over the past few decades and are more than adequate to meet nutritional requirements. Thus, co-nutrient depletion is not a confounder when a diet is designed to be single nutrient deficient.

One of the notable improvements in this study over those done in the 1950’s, is that the treatment compounds were fed at equimolar amounts at a level known to keep rats (weight ~420 g) in vitamin A balance⁽¹⁵⁾. Although this level of ROH resulted in very little ROH storage, it maintained serum ROH concentrations. A disadvantage of this study is that lower doses of DR and αR were not included. Future studies should investigate lower doses of each of these compounds in maintenance of rat growth. At the end of this study, rats were continuing to grow on the various analogues and therefore, longer term studies with lower doses may help to define the value of growth assays versus hepatic storage. Although storage rates of the compounds were not equal and not included in the estimation of bioactivity presented here, the severe ROH deficiency of the animals, even in the ROH-supplemented group, supports higher bioactivity than previously published for both DR and αR. Furthermore, extreme differences in liver retention of αR were noted among studies 1, 2, and 3. In study 1 where ROH was not made available, only 0.3% of the cumulative αR doses were recovered in the liver. This is in stark contrast with study 2 where 61–71% of the cumulative doses were recovered, further supporting αR bioactivity. The huge decrease in αR concentration between d 1 and 8 in study 3 supports utilization considering no ROH was administered during that study. Liver ROH concentration in study 2, suggests that αR does not interfere with ROH in intestinal absorption and hepatic metabolism when fed at physiological levels. However, a sparing effect of ROH may have occurred at the low dose of αR (17.5 nmol/d), which could also be used for growth as shown in study 1.

A notable limitation of this study is that vision testing was not performed. Nonetheless, predictions from the literature indicate that the control and αR groups were likely night blind and 3, 4-didehydroretinal likely replaced retinal in the rod cells at the high serum DR concentration. Indeed, rats maintained on DR exhibited normal retina integrity⁽³⁴⁾. Furthermore, when fish oil concentrate was fed to humans, night vision acuity shifted in favor of red light likely due to 3, 4-didehydroretinal (λ-max of ~400 nm) replacing retinal (λ-max of ~370 nm) in the retina⁽³⁵⁾.

Significant serum concentrations of ROH and DR were detected in rats dosed with retinyl acetate and 3, 4-didehydroretinyl acetate, respectively, which is an indication of increased recycling considering low liver storage and comparative higher concentrations in the kidney. On the other hand, only small amounts of αR were detected in the serum of α-retinyl acetate-dosed rats supporting previous reports that αR cannot bind to RBP^(9,13). Serum DR concentrations in the rats dosed with DR were 2.6-times less than serum ROH concentrations in the rats dosed with ROH in study 1, but growth was maintained likely due to sustained formation of 3, 4-didehydroretinoic acid. Serum ROH concentrations drive utilization of ROH^(20,36). It appears that the rats given ROH maintained a high utilization rate even though their liver ROH reserves were exhausted. Furthermore, identical serum ROH concentrations in study 1 rats given 50 nmol/d compared with study 2 rats given 70 nmol/d and no αR (data not shown) may indicate increased recycling by the kidney given that liver concentrations in study 2 were 9.3-times higher than those in study 1.

The finding that αR did not appreciably affect ROH storage suggests that αR has little influence on ROH metabolism at physiological doses in ratios consistent with typical dietary intake, unlike the finding that α-carotene has a negative effect on β-carotene absorption in rats, which was measured by liver ROH storage⁽¹⁴⁾. If α-carotene was being fed in study 2, the biopotency of αR might differ from that demonstrated in this study. α-Carotene may affect β-carotene uptake earlier in the digestion process either at the site of absorption, competition at the carotenoid transporter on the apical surface of enterocytes⁽³⁷⁾, chylomicron assembly, or competition for the 15, 15-carotenoid monooxygenase cleavage enzyme.

These data on bioactivity of αR could change the calculated vitamin A value of some foods, especially orange carrots, which contain substantial α-carotene. Currently, αR is not considered to have vitamin A activity, yet it clearly supports growth in rats and is not toxic at moderate intakes. During vitamin A deficiency, αR, presumably through the formation of α-retinoic acid as needed, most likely supports similar functions as retinoic acid. Thus, growth assays with α-retinoic acid and retinoic acid should be performed to determine the difference in bioactivities. The results from the toxicity study suggested that at very high hepatic αR concentrations, no hepatotoxicity occurred but αR was quickly cleared. Nonetheless, at αR concentrations >1.7 µmol/g liver, significant α-retinyl esters were circulating in the blood (>10% of total retinol) most likely on lipoproteins. Future studies should determine at what liver ROH concentration this begins to occur because liver retinyl esters >10% of total are considered a biomarker of excessive vitamin A status in humans⁽³⁸⁾.

Fig. 3.

Liver retinol concentration for each group of rats (n 10/group) supplemented with different ratios of α-retinol to retinol. Each group was orally dosed with 70 nmol retinol/d (as retinyl acetate) and either 0, 17.5, 35, or 70 nmol α-retinol (as α-retinyl acetate). Different superscript letters designate a difference (P < 0.05).

Abbreviations

αR

α-retinol

DCE

dichloroethane

DR

3, 4-didehydroretinol

RAR

retinoic acid nuclear receptor

ROH

retinol

RBP

serum retinol-binding protein

UW

University of Wisconsin