comparative embryolethality and teratogenicity of the all-trans isomers of retinoic acid,...

TERATOLOGY !%237-244 (1996)

Comparative Embryolethality and Teratogenicity of the All-trans Isomers of Retinoic Acid, 3,4=Didehydroretinyl Acetate, and Retinyl Acetate in Pregnant Rats PAMELA K. DUITSMAN AND JAMES A. OLSON Department of Biochemistry and Biophysics, Iowa State University, Ames, Iowa 5001 1

ABSTRACT The teratogenic potencies of the all-trans isomers of retinoic acid (RA), 3,4-didehydroretinyl acetate (A2), and retinyl acetate (Al ) were compared. Groups of eight tirned-pregnant Sprague-Dawley rats were administered single equimolar doses (3.5-352 pmol/kg BW) of the retinoids orally in oil on day 8.5 of pregnancy, and dams and fetuses were sacrificed on day 19. The relative teratogenicity and embryolethality of the three tested retinoids were: RA > A2 > A l . The no-effect level of RA and A2 was 3.5 pmol/kg BW and of A1 was 35 pmol/kg BW. Whereas the ad- verse effects of RA and A1 were dose dependent, A2 showed biphasic effects, with a peak of embryolethality at 35 pmol/kg BW. Dams also exhibited weight loss and other toxic manifestations from doses of A2 and RA z 35 Fmol/kg BW. In dosed dams, ( 1 ) Liver concentrations of A1 and A2 increased with the doses of A1 and A2, respectively, (2) RA had little effect on liver A1 except for an increase at the highest toxic dose, and (3) A2 showed a sparing effect on liver A l . RA, although not detected in fetuses from dams treated with Al , was present in significant concentrations (0.5-4.1 nmol/g liver) in fetuses from dams treated with A2. The biphasic change in embryolethality with the dose of A2 correlates with this enhanced concentration of fetal RA. We hypothesize that the actual teratogen in the fetuses of A2-dosed dams is RA. A2 might induce this biphasic effect by inhibiting the catabolism of RA at lower doses and its formation at higher doses. o 1996 WiIey-Liss, Inc.

Physiological concentrations of retinoids are in- volved in regulating cell growth and development. At higher concentrations, retinoids possess teratogenic potential that is species-dependent (Willhite et al., '89). Humans seem to be the most sensitive species on a body weight basis to retinoid-induced terata (Willhite and Book, '90). The structure of the retinoid is impor- tant in determining the embryotoxic and teratogenic potential of the compound (Willhite, '86). Retinoids

that possess an acidic terminus group (e.g., 13-cis-retinoic acid and all-trans retinoic acid) generally have 2- to 3-fold more teratogenic potential in animals than the alcohol forms of vitamin Al; e.g., retinol and its fatty acyl esters. (Willhite, '90).

3,4-Didehydroretinol (A2) differs from retinol (Al) in that it possesses a double bond at the 3,4-position (Fig. 1). Vitamin A2, a naturally occurring analog of Al, is a common component of the food supply (Barua and Nayar, '66). The biological activity of A2 in stimulating the growth of rats is approximately 40% that of A1 (Shantz and Brinkman, '50; Sundaresan and Cama, '61; Howell et al., '67). A2 esters in foods are hydrolyzed to the alcohol (de-

hydroretinol) in the small intestine, absorbed in micel- lar form with other lipids, reesterified to the ester in the gut, transported on chylomicra to the liver, and stored as vitamin A2 ester in the liver (Lederer and Rathman, '38; Goswami and Barua, '86; Tanumihardjo and Olson, '88; Farrer et al., '52; Henbest et al., '55). Vitamin A2, like retinol, is released from the liver as a complex with retinol-binding protein (RBP) (Tanumi- hardjo et al., '87; Wilson and Pitt, '86).

A biological metabolite of vitamin A2, all-truns-3,4- didehydroretinoic acid (ddRA), shows equivalent mor- phogenic properties to its vitamin A counterpart, all- trans retinoic acid (RA) (Thaller and Eichele, '90). Furthermore, ddRA was as active as RA in several models of epithelial differentiation, although their metabolism differed, depending on the cell type used (Torma et al., '94). However, the toxicity and teratogenicity of A2 in vivo has not been studied previously. We report here the relative teratogenicity of A2, Al, and RA in timed-pregnant Sprague-Dawley rats.

Received September 14, 1995; accepted February 5, 1996. Address reprint requests to Dr. Pamela K. Duitsman, Biochemistry and Biophysics, 3260 Molecular Biol. Bldg., Iowa State University, Ames. IA 50011.

0 1996 WILEY-LISS, INC.

238 P.K. DUITSMAN AND J.A. OLSON

3,4-DJDEHYDRORETINOL

RETINOIC ACID

Fig. 1. Formulas of the all-trans forms of retinol (Al), 3,4didehy- droretinol (A2), and retinoic acid (RA).

MATERIALS AND METHODS All-trans-3,4-didehydroretinyl acetate was synthe-

sized from retinoic acid (Barua and Ghosh, '72) in our laboratories. The compound was purified twice on a column of 8% water-deactivated alumina. Its purity, as judged by its absorption at 350 nm during further analysis by HPLC, was 99.9%. All-trans-retinyl acetate and all-trans-retinoic acid were obtained commercially (Sigma Chemical Co., St. Louis, MO, and BASF Corpo- ration, Wyandotte, MI).

Animals Timed-pregnant Sprague-Dawley rats (Holtzman,

Madison, WI) were housed in accordance with Iowa State University and NIH guidelines in virus-free en- vironments on a lightldark cycle of 6 am to 6 pm. The animals were fed a diet of standard rat chow obtained from HarladTeklad, Madison, WI, containing 14.4 nmol retinyl palmitate/g diet.

Study protocol Each dam was administered single, equimolar oral

doses of the retinoids (3.5-352 pmolkg body weight) in approximately 250 pl corn oil, pipetted directly into the mouth of each animal by using a Gilson positive- displacement pipette (Rainin Instruments, Woburn, MA) on day 8.5 of pregnancy. Single oral dosing of rats on day 8.5 of gestation is a standard method used for evaluating the teratogenic effects of retinoids on rats (Willhite et al., '89). Control animals received corn oil only. Gestation was interrupted on day 19.5 by cervical section of etherized dams. Uterine horns were exam- ined, implantation sites counted, and fetuses removed. Maternal serum was collected by cardiac puncture, and maternal livers were excised. Fetuses were weighed, measured, and photographed. Thereafter, fetal livers

were excised, blotted on humid filter paper, and frozen at -20°C.

Retinoid extraction and analysis All operations were performed under yellow light.

Maternal and fetal liver samples (0.5 g) were ground exhaustively with 1.5 g anhydrous sodium sulfate by use of a pestle and mortar (Barua and Olson '89). Meth- ylene dichloride (5 vol) was added along with internal standard (1.4 pg retinyl acetate; 150 pl), and samples were ground further. Methylene dichloride was fil- tered, extracts were pooled and mixed, and the volume was measured. An aliquot (1 ml) of the extract was evaporated under argon, and the residue was redis- solved in 200 pl of methano1:methylene dichloride (1:l vlv) for analysis by HPLC.

Retinoids from the serum of dams were extracted by treating serum (200 pl), with 100% ethanol (400 pl), 10% acetic acid in water (20 pl), and ethyl acetate (400 1.1). Internal standard (1.4 pg retinyl acetate; 250 p1) was added, and the mixture was subjected to further extraction by use of hexane (2 x ) (Barua and Olson, '89). The extracts were washed, pooled, dried under argon, and resuspended in 100 p1 methano1:methylene dichloride (1:l v/v) for injection onto the HPLC column.

Each sample was applied to a 5-pm C18 Waters Re- solve reversed-phase column by use of a WISP autoin- jector (Waters, Milford, MA) by using the method de- scribed by Barua ('90). A Waters 996 photo-diode array (PDA) detector monitored absorbance at 325 nm for retinoic acid, retinol, and retinyl esters and at 350 nm for retinoic acid, 3,4-didehydroretinol and its esters. Two Waters 510 pumps were set at a flow rate of 1.2 ml/min and operated by use of a Waters automated gradient controller. A gradient solvent system consisted of a linear gradient of solvent A (methanol-water, 68:32, v/v) to solvent B (methanol-methylene dichloride, 4 1 , v/v) during a 20-min period. Solvent B was continued for 10 min more, at which time the gradient was changed back to initial conditions over a 5-min period. The column was then equilibrated for 10 min with Solvent A before the next injection. Retention times for retinoic acid, 3,4-didehydroretinol, retinol, 3,4-didehydroretinyl acetate, retinyl acetate, 3,4-didehydroretinyl palmitate, and retinyl palmitate were 12, 18,21,24,25,31, and 32 min, respectively (Fig. 2). The Millennium 2010 software version 1.2, developed by Waters for use with their 996 PDA detector, performed data acquisition, processing, and management of chro- matographic information. All data were acquired and stored in three-dimensional mode, which allowed detailed examination of selected spectra, integration of peak areas, and assessment of the purity of compounds.

Statistics All data were analyzed using two-way analysis of

variance (ANOVA) with compound, dose, and compound-dose interaction as the effects in the model.

< 0.15

0.10

0.05

0.00 I ~ " ~ I ~ ~ " I ~ ~ ' ~ I ~ J ~ ~ I ~ ~ ~ ' I ~ ' ' '

5.00 10.00 15.00 20.00 25.00 30.00 3! Minutes

Fig. 2. Chromatogram obtained with reversed-phase gradient HPLC of a standard mixture of retinoids.

Comparisons between each treatment group and the control group were made using the test of least signif- icance difference (LSD). ANOVA, standard errors, means, standard deviations, LSD, and P-values were determined by use of general linear model (GLM) pro- cedures of the Statistical Analysis System, version 6.07 (Cary, NC). GLM employs the method of least-squares to fit general linear models among the statistical methods.

RESULTS Statistics

The F-values for the treatment differences among all dependent variables are reported in Table 1. F-values were derived using two-way ANOVA with compound, dose, and compound-dose interactions as the effects in the model. Differences between compounds, differences between doses, and differences due to compound-dose interactions were highly significant for nearly all dependent variables. Thus, the null hypothesis was re- jected. Comparisons were then made between treatment groups and controls by use of the LSD test. The mean for each treatment group, standard errors, and P-values based on the LSD test are reported in Tables 2-4.

Toxic effects on dams Dams dosed with 352 pmol/kg of RA or A2 showed

overt retinoid toxicity, exhibited by loss of hair and hair discoloration. At doses 235 pmoukg A2 and 2113 pmol/kg RA, dams lost weight but did not show overt toxicity. Thus, weight loss seemed to be a more sensitive indicator of toxicity in the dams for both retinoids than hair loss or decoloration.

Dams dosed with 3.5 pmolkg BW of either RA or A1 were significantly younger and smaller than other animals used in this study. Throughout the experiment, we perceived that these animals showed a higher level of stress than older animals. Thus, the anomalous data

TERATOGENICITY OF RETINOIDS 239

acquired for the progeny from these two groups, as in- dicated later in specific cases, may well be a reflection of the very young dam's response to pregnancy rather than to the compounds dosed.

Fetal resorption As summarized in Table 2 and Figure 3, control an-

imals experienced no resorption. This finding is unusual inasmuch as average background embryolethality in rodents has been reported to be approximately 4% (Manson, '86). In similar earlier studies in our lab- oratory, we found 1.8% resorption in control Sprague- Dawley rats (Gunning et al., '93). RA at 352 pmolkg caused 100% resorption, which decreased in a dose-dependent manner to 3.6% at 3.5 pmolkg. All doses of A1 gave similar resorption rates of approximately 4%, although others have reported dose-dependent embryolethality with A1 in Wistar rats (Langman and Welch, '66). Interestingly, the vitamin A2 dose-response curve for resorption was nonlinear, with most resorption occurring at the 35-pmoVkg dose, but with very little (1% and 2.6% respectively) at the lowest and highest doses tested.

Terata RA was the most potent teratogen of the congeners

studied, increasing terata significantly (p 5 0.0001) a t 113 pmol/kg (Table 2, Fig. 3). Of the terata present, 75% consisted of exencephaly, 17% abdominal protrusion, 8% microcephaly, and 8% bulging crown. No terata were evident for doses of RA 535 pmol/kg. A2 at 352 pmolkg induced 5% terata, compared with 3% for Al. Furthermore, the type and extent of terata for the A2 group were more profound than were those caused by Al. Fetuses exhibiting terata at this dose of A2 had bulging crowns (60%) and exencephaly (40%). One stillborn fetus of the A2-treated dams exhibited gross terata, with complete craniofacial malformation, exencephaly, and protrusion of the abdomen. Terata appar- ent at the same dose of A1 included eye abnormalities (25%) and bulging crowns (75%). Both A2 and A1 also induced terata at the 113 Fmolkg dose; A2 primarily produced malformations of craniofacial features, and A1 generated slight cranial bulges. No terata were evident for doses 535 pmol/kg of A1 or A2 (Table 2, Fig. 3).

Major qualitative differences were observed upon examination of skeletal tissues (unpublished observa- tions). Both RA and A2-dosed groups differed significantly from controls in length of mandible, rib, radius, ulna, femur, and tibia bones. Al-dosed animals differed only slightly from controls in length of mandible. A detailed analysis of skeletal and histological examina- tions of tissues will be published separately.

Pup length and weight Growth of the fetus was significantly reduced by

doses of RA 211 pmol/kg (Table 2). At an RA dose of


TABLE 1. The variance ratio (F) and probability (P) of values for all dependent variables tested in the pregnancy outcome of rats administered all-trans retinoids on day 8.5 of gestation

Effect of compound' Effect of dose2 Effect of compound.dose interaction Dependent variable F p3 F P3 F P3 Resorption 45.72 0.0001 9.80 0.0001 20.17 Terata 2.72 0.0469 3.53 0.0089 4.08

Fetal length 2.49 0.0628 2.23 0.0694 4.03 Dam weight 2.61 0.0540 13.18 0.0001 10.56 Dam liver/BW Ratio (%) 3.61 0.0150 1.13 0.3436 3.67

Dam liver total A2 (pmol/g liver) 192.81 0.0001 45.48 0.0001 45.15 Dam liver RA (pmol/g liver) 1.04 0.3838 1.04 0.3970 1.04 Fetal liver total A1 (nmoVg liver) 22.02 0.0001 28.78 0.0001 3.51 Fetal liver total A2 (nmol/g liver) 457.01 0.0001 186.97 0.0001 159.82

'Effect of compounds (vehicle only, RA, A2, Al) as a source of variation for the dependent variables. 2Effect of dose (0, 3.5, 11, 35, 113, or 352 pmolkg BW) as a source of variation for the dependent variables. 3Probability of a value of F larger than that observed if the null hypothesis is true.

Fetal weight 16.31 0.0001 0.63 0.6430 4.79

Dam liver total A1 (pmoUg liver) 5.22 0.0039 4.05 0.0075 5.35

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.4306 0.0047 0.0001

TABLE 2. Pregnancy outcome for rats administered all-trans retinoids on day 8.5 of gestation

Mean pup Mean pup Dose No. of Implantation % % weight (g) length (cm)

Treatment (umol/kg BW) litters sites Stillbirths resorotion terata (mean f SEM) (mean &SEM)

A1

A2

None control 0 RA 3.5

11 35

113 352

11 35

113 352

3.5

3.5 11 35 113

10 I 6 7 5 7 5 6 8

10 9 7 9 6 8

126 84 81 98 79 88 50 75 96

103 91 96

108 53 88

0 0 0 0 1

0 0 0 0 0 0 0 0 1

-

0 0 2.61 t 0.09 3.15 f 0.05 3.6 0 2.45 t 0.11 3.29 f 0.06 7.4 0 2.18 rfr 0.12l 3.09 5 0.07 6.1 0 2.34 t 03.11' 3.34 f 0.06l

52.03 15.03 1.66 k 0.133 2.87 2 0.072

4.0 0 2.34 rfr 0.13 3.24 f 0.07 3.0 0 2.43 2 0.12 3.35 f 0.07 4.2 0 2.30 2 0.10 3.27 f 0.06 5.8 1 2.53 t 0.09 3.22 ? 0.05 3.3 3.3 2.45 t 0.09 3.13 f 0.05 1.0 0 2.17 k 0.1l2 3.09 5 0.06 6.5 0 2.21 2 0.09' 3.16 & 0.05

22.62 0 2.55 t 0.12 3.22 f 0.06 11.4 2.3 2.60 2 0.11 3.27 2 0.06

100.03 - - -

~~- 352 9 78 1 2.6 5.1 2.47 rfr 0.24 3.15 2 0.05

'Significantly different from control (P 10.05). 2Significantly different from control (P 50.01). 3Significantly different from control (P ~0.0001).

113 pmolkg, the mean pup weight and height were only 64% and 91%, respectively, of those of control pups. At all concentrations A1 and A2 had no consistent effect on pup length or weight. The mean weights but not the lengths of pups from dams treated with 3.5 or 11 pmolkg BW A2 were less than those of the control group and of all other Al- and A2-treated groups.

Maternal liver Liver weightlbody weight ratios of dams were not

consistently affected by any dose of any retinoid, except for a reduction at the largest RA dose (Table 3). Liver concentrations of A1 and A2 expectedly increased with the doses of A1 and A2, respectively. RA had little effect on liver A1 except at the greatest toxic dose, which induced an increase of Al. A2 showed a sparing effect on A1 at larger doses (Table 3). RA was detected

only in the livers of dams treated with 35 Wmolkg of A2. The A1 value at the RA dose of 3.5 pmolkg BW was exceptionally high. As mentioned previously, this group of dams was much younger and smaller than dams of other groups; therefore, no importance is at- tached to this anomalous observation.

Fetal liver Mean concentrations of total A1 in fetal livers tended

to decrease with increases in dosing, as summarized in Table 4. A2 was detected in fetal livers from animals dosed with 11 to 352 p.mol/kg A2, with individual values ranging from 0.3 to 3.9 nmol/g liver. RA was detected in fetal livers from dams treated with the 3.5 and 113 pmolkg of RA and 11-352 pmolkg of A2. RA values in individual fetal livers in the RA and A2- treated groups ranged from 0 to 2.2 nmollg liver and


TABLE 3. Liver to body weight ratios, and means f standard deviations of retinoids recovered from liver tissues of dams on day 19.5 of gestation (n = 3)

Dose Livedbody Total A1 Total A2 all-trans RA Treatment (pmolkg BW) weight ratios (%) (pmol/g liver) (pmollg liver) (pmoVg liver) None (control) 0 3.97 f 0.07 0.68 f 0.04 ND ND RA 3.5 4.11 f 0.10 2.18 f 0.5g2 ND ND

11 4.19 f 0.10 0.42 f 0.06 ND ND 35 4.26 f 0.Og2 0.30 f 0.09 ND ND

113 4.27 -+ 0.11' 0.42 f 0.10 ND ND 352 3.62 f 0.Og2 0.79 k 0.15 ND ND

A1 3.5 4.37 f 0.11~ 0.71 f 0.49 ND ND 11 3.95 f 0.10 0.57 f 0.04 ND ND 35 4.06 f 0.08 1.52 k 0.40 0.001 t 0.001 ND

113 4.19 * 0.07 1.67 f 0.15l 0.001 f 0.001 ND 352 4.12 k 0.08 2.37 k 0.134 0.003 f 0.001 ND

A2 3.5 4.02 f 0.09 0.61 f 0.02 0.01 f 0.003 ND 11 4.16 f 0.08 0.35 f 0.01 0.03 f 0.01 ND 35 3.77 k 0.10 1.83 k 0.97' 0.31 2 0.12~ 0.002 f 0.0032

113 4.10 t 0.09 1.01 f 0.94 0.64 f 0.134 ND 352 4.15 f 0.08 1.06 f 0.50 1.01 t 0.11~ ND

ND, not detected. 'Significantly different from control (P 50.05). 2Significantly different from control (P 50.01). 3Significantly different from control (P ~0.001) . 4Significantly different from control (P ~0.0001).

TABLE 4. Mean concentrations (k standard deviations) of retinoids recovered from liver tissues of fetuses on day 19.5 of gestation (n = 9)

Dose Total A1 Total A2 all-trans RA Treatment (pmolkg BW) (nmol/g liver) (nmol/g liver) (nmol/g liver) None (control) 0 8.7 * 1.1 ND ND RA 3.5 16.1 f 1.63 ND 0.23 ? 0.33

11 9.0 f 1.0 ND ND 35 9.8 f 0.9 ND ND

113 15.7 * 3.22 ND 0.73 f 1.03' 352 - - -

A1 3.5 21.4 * 0.74 ND ND 11 15.1 f 1.32 ND ND 35 10.7 k 1.1' ND ND

113 13.6 f 0.7 ND ND 352 11.4 * 3.0 ND ND

A2 3.5 14.6 t 2.62 ND ND 11 9.8 t 1.3 0.40 k 0.032 1.12 f 0.443 35 8.8 f 1.3 0.49 f 0.1l2 0.60*

113 11.2 t 1.3 3.36 f 0.324 4.07 k O.03Ei4 352 7.3 * 1.1 3.52 * 0.3g4 0.54 f 0.005

ND, not detected. 'Significantly different from control (P 50.05). 2Significantly different from control (P 10.01). 3Significantly different from control (P ~0.001). 4Significantly different from control (P ~0.0001). *Only one animal was available for this data point. RA was found, but the single value cannot be considered as representative of a group.

0-4.1 nmol/g liver, respectively. RA was not detected in fetal livers from dams treated with any concentration of Al . Furthermore, no A2 was found in fetal livers from dams dosed with any concentration of A l .

DISCUSSION The maternal toxicity evident in the present study

(decrease in weight gain, hair loss, and hair discoloration) is consistent with classic toxic manifestations of


100 H A2

- W A1 M RA M Control -

1'

50 -

K 40 0

- .- c P - 8

30 -

2 Dose pmol/kg BW

Dose pnol/kg BW

Fig. 3. Percent resorption and terata for all litters of rats adrnin- istered all-trans retinoids on day 8.5 of gestation.

retinoid administration. Corroborative of previous studies, RA produced well-defined patterns of terata and embryolethality in a dose-dependent manner (Gun- ning et al., '93; Nau et al. '94). A1 showed teratogenic

A1 - RA - inactivemetabolites

\ Inhibits at Inhibits at

highdoses \ / moderatedoses

A2

Fig. 4. Hypothesized effects of A2 on the formation and synthesis of RA in vivo.

Although A2-induced terata also followed a linear dose-response curve, A2-induced embryolethality did not. Embryolethality for A2 at 35 pmolkg BW was nearly nine times that observed at 352 pmolkg BW. To validate this finding, the A2 dose at 35 Fmol/kg was repeated, in a different group of eight pregnant dams. Similar results were obtained.

The recovery of retinoids from dam livers points both to similarities and differences in the uptake and storage of A1 and A2 (Table 3). Liver storage of A1 in Al-dosed rats and of A2 in A2-dosed animals increased in a dose-response manner. The net storage of A1 in the dam's liver a t larger doses (235 pmol/kg BW), however, was roughly twice that of A2 at equivalent doses. This observation accords with the reported relative ef- fectiveness of (40%) vitamin A2 in stimulating the growth of vitamin A depleted rats (Shantz and Brink- man, '50). On the other hand, when smaller equimolar oral doses (7 pmolkg) of A1 and A2 were given simul- taneously to vitamin A-deficient rats, the ratio of AllA2 in the liver at 5-7 days was approximately 0.8 (Thaitawee, '84). Thus, the bioavailability of A1 and A2 seem to be comparable. Retinoic acid, although somewhat more hydrophilic, is also absorbed well from the GI tract. Thus, it seems unlikely that the differences in observed teratogenicity are primarily caused by differences in bioavailability. Nonetheless, to abet our studies with maternal oral dosing, and t o evaluate the possible importance of differences in bioavailability, the relative pharmacokinetic behavior (peak plas- ma concentrations and areas-under-the-curve) of the three compounds in dams should be conducted.

Compared with control animals, A2 was maximally effective in sparing A1 in livers of animals dosed at 35 pmol/kg BW A2 and probably showed sparing effects at higher doses as well. Thus, A2 seemingly reduced the utilization rate of liver Al. An additional curious finding was that RA was found only in the livers of dams dosed with A2 at 35 pmoVkg BW, but not at other doses of A2. That RA was not found in tissues of RA dosed animals is not surprising, because tissues were analyzed 11 days after RA was administered. RA is known to be metabolized quickly, and its inactive metabolites are excreted in the urine and feces within 7 days (Rob- erts and DeLuca, '67). Furthermore, RA was not found

effeks only at the highest dose tested. in the livers of Al-dosed dams. The ability of A2 at 35


TABLE 5. Comparison of the concentration of vitamins A1 and A2 in pregnant rats and their fetuses on day 19.5 of gestation

Rat Process Vitamin A1 Vitamin A2 Dam Storage in liver after dosing Dose-dependent Dose-dependent Dam Efficiency of liver storage Good Fair to good Dam Liver RA after dosing None Some at 35 p,moVkg BW Fetus Uptake by fetal liver Inversely related to dose Dose-dependent Fetus RA in fetal liver after dosing Not present Present

pmol/kg BW to increase liver concentrations of A1 and RA correlates with the increased embryolethality seen at this dose of A2.

In fetal liver tissues, A1 concentrations were inversely related to dose, suggestive of placental regula- tion of its transfer to the fetus. Indeed, whereas RA dosing had no consistent effect on liver Al, increasing doses of A1 and A2 were associated with lower total A1 values. In contrast, A2 appears in fetal livers of A2- dosed animals in a dose-dependent manner. Interest- ingly, RA is found in the fetal livers of all A2-dosed animals, with the exception of the smallest A2 dose, but in none of the A1 dosed animals. Even when RA was administered, fetal liver RA at 19.5 days was less than in A2-dosed animals. RA accumulation from an A2 dose was maximal a t two midrange levels of A2 (1 1 and 113 pmolkg BW). Because of the high incidence of embryolethality at the 35 pmolkg BW dose of A2, only one fetal liver was available for analysis of RA concentration. Although this liver did contain RA (0.6 nmoY g), the fetus was taken from a litter free from terata and resorption. Hence, this one sample cannot be considered representative of the entire group.

A comparison of the uptake and storage of A1 vs. A2 in the pregnant rodents in our study is summarized in Table 5. The detection of RA in the livers of A2 dosed dams and fetuses correlates well with the biphasic dose-response curve of A2 relative to embryolethality. Thus, the increased teratogenicity and embryolethality of A2 relative to that of A1 may be due to the presence of RA in tissues of A2 dosed animals. One possible ex- planation for the observed biphasic response is that at high concentrations, A2 may inhibit conversion of inactive precursors to RA, whereas a t slightly lower concentrations, A2 may inhibit catabolism of RA to inactive products (Fig. 4). Thus, as the concentration of A2 in the fetus increases in a dosedependent manner, the steady-state concentration of RA will first rise and then fall. This hypothesis can be tested by determining the rate of formation of RA from radioactive retinol and the rate of catabolism of radioactive RA in fetal prep- arations in vitro as a function of A2 concentrations in the media.

Other explanations for this biphasic response to A2 exist. Retinoid nuclear receptor proteins (RARs and RXRs) are specifically activated by retinoids in a manner unique to each ligand. Consequently, distinct pathways may be initiated by different retinoids, leading to

specific developmental, toxic, and/or teratogenic effects. 3,4-Didehydroretinoic acid (ddRA), although not found in rat tissues from this study, has been shown to be as potent as RA at invoking digit duplications in the chick wing bud (Thaller and Eichele, '901, and the nuclear receptor protein binding profiles for ddRA and for RA in vitro are very similar (Torma et al., '94). None- theless, it is conceivable that differences in the metabolism of Al, A2, and their corresponding acids may well lead to differential activation of retinoid nuclear receptors and thereby account for different toxic and teratogenic effects. As a case in point, the activity of glycerol-3-phosphate dehydrogenase in Ob17 cells shows a biphasic response to RA concentrations in the medium (Safanova et al., '94). Comparative pharmacokinetic studies of A1 and A2 in the pregnant animal and their relative effects on specific reactions in RA metabolism may clarify the unusual outcome data reported here.

ACKNOWLEDGMENTS This work was supported in part by ISU-CDFIN-

USDA-CSRS 94-34115-0269, journal Paper No. 5-16548 of the Iowa Agriculture and Home Economics Experiment Station, Ames, Iowa; Project No. 3335.

LITERATURE CITED Barua, A.B. (1990) Analysis of water-soluble compounds: glucu-

ronides. Methods Enzymol., 189:136-145. Barua, R.K., and P.G. Nayar (1966) Naturally occurring anhydrovi-

tamin A2: Transformation into retinene2. Biochem. J., 101:302- 307.

Barua, A.B., and M.C. Ghosh (1972) Preparation and properties of 4-0x0-retinoic acid and its methyl ester. Tetrahedron Lett., 18:

Barua, A.B., and J.A. Olson (1989) Chemical synthesis of all-tmns- [ll-3Hlretinoyl p-glucuronide and its metabolism in rats in vivo. Biochem. J., 263:403-409.

Farrer, K.R., J.C. Hamlet, H.B. Henbest, and E.R.H. Jones (1952) Studies in the Polyene Series. Part XLIII. The structure and synthesis of vitamin A2 and related compounds. J. Chem. SOC., 2657- 2668.

Goswami, B.C., and A.B. Barua (1986) Metabolism of dehydroretinyl ester in white leghorn chicks. Br. J. Nutr., 55~379-385.

Gunning, D.B., A.B. Barua, and J.A. Olson (1993) Comparative ter- taogenicity and metabolism of all-trans retinoic acid, all-trans retinoyl p-glucose, and all-trans retinoyl p-glucuronide in pregnant Sprague-Dawley rats. Teratology, 47.29-36.

Henbest, H.B., E.R.H. Jones, and T.C. Owen (1955) Studies in the Polyene Series. Part LI. Conversion of vitamin A1 into vitamin A2. J. Chem. SOC., 2765-2819.

Howell, J. McC., J.N. Thompson, and G.A.J. Pitt (1967) Reproduction

1823-1825.

244 P.K. DUITSMAN AND J.A. OLSON and vision in rats maintained on a retinol-free diet containing 3-de- hydroretinol (vitamin A2). Br. J. Nutr., 21:373-376.

Langman, J., and G.W. Welch (1966) Effect of vitamin A on development of the central nervous system. J. Comp. Neurol., 128:l-16.

Lederer, E., and F.H. Rathman (1938) A physico-chemical and bio- chemical study of vitamin A2. Biochem. J., 32:1252-1261.

Manson, J.M. (1986) Teratogens. In: Casarett and Doull's Toxicology: The Basic Science of Poisons, 3rd Ed. C.D. Klaassen, M.O. Amdur, and J. Doull, eds. Macmillan, New York, pp. 195-220.

Nau, H., I. Chahoud, L. Dencker, Ed. Lammer, and W.J. Scott (1994) Teratogenicity of vitamin A and retinoids. In: Vitamin A in Health and Disease, Rune Blomhoff, ed. Marcel Dekker, New York, pp. 615-663.

Roberts, A.B., and H.F. DeLuca (1967) Pathways of retinol and retinoic acid metabolism in the rat. Biochem. J., 102:600-605.

Safonova, I., C. Darimont, E. Amri, P. Grimaldi, G. Ailhaud, U. Re- ichert, and B. Shroot (1994) Retinoids are positive effectors of adi- pose cell differentiation. Mol. Cell. Endocrinol., 104:201-211.

Shantz, E.M., and J.H. Brinkman (1950) Biological activity of pure vitamin A2. J. Biol. Chem., 18.3~467-471.

Sundaresan, P.R., and H.R. Cama (1961) Biological activity of crys- talline vitamin A2 aldehyde. Br. J. Nutr., 15:225-230.

Tanumihardjo, S.A., A.B. Barua, and J.A. Olson (1987) Use of 3,4- didehydroretinol to assess vitamin A status in rats. Int. J. Vitam. Nutr. Res., 57:127-132.

Tanumihardjo, S.A., and J.A. Olson (1988) A modified relative dose- response assay employing 3,4-didehydroretinol (vitamin A2) in rats. J. Nutr., 118:598-603.

Thaitawee, P. (1984) Determination of vitamin A status by using vitamin A2. M.S. thesis, Faculty of Graduate Studies, Mahidol Uni- versity, Bangkok, Thailand, pp. 80.

Thaller, C., and G. Eichele (1990) Isolation of 3,4didehydroretinoic acid, a novel morphogenetic signal in the chick wing bud. Nature

T6rma H., D. Asselineau, E. Andersson, B. Martin, P. Reiniche, P. Chambon, B. Shroot, M. Darmon, and A. Vahlquist (1994) Biologic activities of retinoic acid and 3,4-didehydroretinoic acid in human k e r a t i n o m are similar and correlate with receptor affinities and transactivation properties. J. Invest. Dermatol., 102:49-54.

Willhite, CC. (1990) Molecular correlates in retinoid pharmacology and toxicology. In: Chemistry and Biology of Synthetic Retinoids. M.I. Dawson and W.H. Okamura, eds. CRC Press, Boca Raton, FL, pp. 539-574.

Willhite, C.C. (1986) Structure-activity relationships of retinoids in developmental toxicology. 11. Influence of the polyene chain of the vitamin A molecule. Toxicol. Appl. Pharmacol., 83563-575.

Willhite, C.C., and S.A. Book (1990) Utility of disposition data in evaluating retinoid developmental toxicity. Methods Enzymol.,

Willhite, C.C., P.J. Wier, and D.L. Berry (1989) Dose response and structure-activity considerations in retinoid-induced dysmorpho- genesis. Crit. Rev. Toxicol., 20:113-135.

Wilson, T.C.M., and G.A. F'itt (1986) 3,4Didehydroretinol (vitamin A2) has vitamin A activity in the rat without conversion to retinol. Biochem. Soc. Trans., 14:960-951.

345:815-819.

190:406-448.

comparative embryolethality and teratogenicity of the all-trans isomers of retinoic acid,...

Documents