0031-3998/92/3202-0195$03.00/0 PEDIATRIC RESEARCH Vol. 32, No.2, 1992 Copyright © 1992 International Pediatric Research Foundation, Inc. Printed in U.S.A. Transfer and Metabolism of Retinol by the Perfused Human Placenta JOSEPH DANCIS, MORTIMER LEVITZ, JOSEPH KATZ, DAVID WILSON, WILLIAM S. BLANER, ROSEANN PIANTEDOSI, AND DEWITT S. GOODMAN' Department ofPediatrics [J.D., D.Wj and Department ofObstetrics and Gynecology [M.L., J.K.j, New York University Medical Center and the Institute ofHuman Nutrition, Columbia University College ofPhysicians and Surgeons [WS.B., R.P., D.S.G.j, New York, New York 10016 ABSTRACf. The transfer and metabolism of retinol by The importance of vitamin A to the developing mammalian human placenta was investigated using an in vitro perfusion fetus has been amply demonstrated in the experimental animal. system with independent maternal and fetal circulations. Both deficiencies and excesses are teratogenic (1, 2). The perti­ 3H-retinol bound to albumin added to the maternal perfus­ nence of these observations to the human has recently become ate was rapidly taken up and concentrated by the placenta tragically evident. Retinoic acid administered to the pregnant to levels 16.5 ± 5.28 times the maternal perfusate. Ap­ female for the treatment of acne has caused severe congenital proximately 8% of the retinol retained in the placenta was abnormalities in the newborn infant (3). esterified. No metabolites were detected in the perfusates. Despite its obvious importance, little is known about the Perfusion of placenta with retinol bound to retinol-binding placental transfer of retinoids. The question has been approached protein (RBP) reduced the placental concentration to 4.4 in the sheep (4) and monkey (5), but the placental role is difficult ± 1.72 times the maternal concentration and eliminated to deduce from studies in the intact animal and extrapolation of evidence of metabolism. The transfer rate of RBP:3H­ results to the human is uncertain. Measurements have been retinol was less than that of albumin.'r'Ccretinol when made in the human of maternal and cord retinol levels (6) that measured concurrently in three experiments (clearances are interesting but shed little light on placental functions. Torma 0.11 versus 0.75 mL/min, 0.21 versus 1.7 mL/min, and 0.29 and Vahlquist (7) have studied the uptake of vitamin A and RBP versus 0.48 mL/min, respectively). Transfer of the radio­ by fragments of human placenta in vitro and speculated concern­ active retinol was more rapid than 125I_RBP or albumin, ing the transplacental process. The model, however, does not indicating that retinol was transferred independently of the permit differentiation of events at the maternal surface of the proteins. The transfer index of retinol (clearance reti­ placenta from those at the fetal surface. nol:clearance L-glucose) was 0.73 ± 0.085 compared to 2.1 In the present study, we have used the technique of in vitro ± 0.36 for thiamin and 3.4 ± 0.95 for riboflavin, both perfusion of a human placental lobule. By establishing independ­ water-soluble vitamins with active transport systems. The ent "maternal" and "fetal" circulations, it was possible to study retinol transferred to the fetal perfusate is not bound to the transplacental transfer of retinol, the principal circulating RBP, as demonstrated by gel filtration chromatography form of vitamin A. In vitro, retinol circulates bound to a specific and chromatography on a transthyretin affinity column, serum protein, RBP, which in turn complexes with TTR (8, 9). despite the availability of RBP in the cord serum added to We have investigated the placental transfer and metabolism of the perfusate. The endogenous retinol in the cord serum is retinol bound to RBP and compared the results to retinol bound bound to RBP. The conclusions derived are that RBP nonspecifically to serum albumin. The comparison has provided binding reduces the accumulation of retinol in the placenta insights into the physiologic role of RBP and the biochemical and the transfer to the fetus, that retinol is transferred mechanism through which retinol is taken up by the tissues. intact to the fetal circulation where immediate binding is not to RBP, and that redistribution to RBP occurs subse­ MATERIALS AND METHODS quently, possibly in fetal liver. (Pediatr Res 32: 195-199, 1992) Perfusion technique. Term placentas of clinically normal preg­ nancies were obtained after delivery by cesarean section or by vagina. Independent maternal and fetal circulations were estab­ Abbreviations lished to a placental lobule as previously described (10). Flow RBP, retinol-binding protein rates approximated 12-15 mL/min on the maternal side and 4­ TIR, transthyretin 6 mL/min on the fetal side. In the "closed" experiments, the RBP:retinol, retinol bound to retinol-binding protein perfusate was recirculated through a 70-mL reservoir. The per­ albumin:retinol, retinol bound to albumin fusions were performed under reduced lighting, as were all pro­ TCA, trichloroacetic acid cedures with retinol, because of its light sensitivity. Perfusate. The basic perfusate consisted of Earle's buffered salt solution equilibrated against Oz-COz, 95-5%, pH 7.4, with 150 mg/dl. of n-glucose and mixed amino acids in physiologic con­ centration. In the experiments with albumin-bound retinol, 10 g/L BSA was added to the perfusate with 3H-retinol -20000 cpm/ml, and retinol 1.7 JLmol/L. Ten g/L albumin in the Received January 2, 1992; accepted March 20, 1992. perfusate solubi1izes the retinol and provides sufficient binding Correspondence and reprint requests: Joseph Dancis, M.D., New York Univer­ to prevent absorption to the perfusion tubing and apparatus sity Medical Center, 550 First Ave., TH 528, New York, NY 10016. Supported by grants from the Langer Foundation and the NIH (DK05968, similar to previous experience with vitamin D (11). Binding was DK43097). accomplished by incubating 3H-retinol with BSA for 1 h. For the 'Deceased. RBP experiments, 3H-retinol was bound to RBP by incubation 195 196 DANCIS E T AL. at 3rC .for 1 h or at 5°C overnight, while rotating gently. The RESULTS RBP:retmol was added to perfusate containing 20% maternal serum. Transfer and Metabolism ofAlbumin-Bound Retinol. 3H-reti­ nol was incubated with serum albumin at 3rC for 1 h which Preparation of samples for analysis. Perfusate samples were add ed directly to scintillant to determine radioactivity. If there as demonstrated by TCA p;ecipita­ was an insufficient concentration ofcounts for chromatographic tion. The albumin-bound retinol was added to the maternal perfusate, which contained 10 g/L serum albumin and recircu­ an alysis, samples were first concentrated with an Amicon-BI5 concentrator (Amicon Corp., Beverly, MA) for gel filtration lated for duration ofthe perfusion. The fetal pe;fusate, which analysis or by extraction with hexane for reverse phase chro ma­ also IO .g/L serum albumin, was kept open (two tography. Th.roughout the analytical procedures, retinol was expenments) or recirculated (four experiments). protected against oxidation by the addition ofbutylated hydrox­ Placental uptake. Uptake of 3H-retinol from the maternal ytoluene, 6.1 mM (1 mg/ml.), or a-tocopherol, 100 JLg. perfusate was very rapid during the first 20 min and continued At the end of the perfusion, 2-4 g of villous tissue was dissected at a slower rate throughout the perfusion. At the end of the from the perfused placenta and homogenized in 10 volumes of perfusion , the concentrations of 3H-retinol in the placenta and methanol:hexane (1:9). The extraction was repeated and both m the maternal perfusate were measured. The concentration in extracts combined and evaporated under vacuum. The residue the placenta was 16.5 ± 5.28 times that in the perfusate (Table was dissolved in 3 mL hexane containing 6.1 mM butylated 1). hydroxytoluene, transferred to a small tube, and taken to dryness . In five experiments, the albumin-bound 3H-retinol was added The residue was dissolved in 0.5-1 .0 mL of meth anol and passed to the fetal perfusate that was recirculated. The ratio of the through a 0.22-JLm filter, and 100 to 500 JLL were analyzed by concentrations of 3H-retinol in the placenta to the fetal perfusate HPLC. was 16.0 ± 4.79, suggesting that similar mechanisms were in­ Chromatographic analysis. Reverse phase HPLC was used to volved in uptake and concentration of retinol from the maternal identify retinoic acid, retinol, and the retinyl esters. Separation and fetal perfusate. was on a Waters RPC-I8 column (Waters Associates Milford Metabolism. The maternal and fetal perfusates were analyzed MA) with a UV detector at 325 nm . The mobile was 80% by reverse phase HPLC. No radioacti ve metabolites were de­ acetonitrile in 1% ammonium acetate with a flow rate of 1 mLI tected. In four experiments, the placenta was extracted at the min. Retinol and retinoic acid, 100 JLg, were added to the sample completion of the perfusion and also analyzed. Over 90% of the be,fore injection to serve as standards, the former eluting at 7 radioactivity coeluted with retinol ; 8, 4, 7, and 11% eluted with retinyl esters. Retinoic acid was not detected. nun and the latter at 4.5 min . For the retinyl esters, 100% 3H-retinol methanol was the mobile phase, retinyl palmitate was the stand­ Placental transfer. Transplacental clearances of were ard, and elution time was 13.4 min (12). calculated (Table 2). 14C-inulin was included in these experiments High performance gel filtration was performed to separate to provide a simultaneous ofdiffusion ofa large (5-kD), according to molecular size (13). Samples were injected water-soluble molecule. Retinol was transferred IO-fold more into a Waters TSK 2000 SW column equipped with a UV or rapidly than inulin (10.2 ± 2.58).