Presence of 25-hydroxyvitamin D3 and 1,25-dihydroxyvitamin D3 24-hydroxylase in vitamin D target cells of rat yolk sac.

In the pregnant rat, the yolk sac, which possesses true placental functions, is a vitamin D target organ. We tested its ability to hydroxylate 25-hydroxy- and 1,25-dihydroxyvitamin D3 (25-OHD3 and 1,25-(OH)2D3). 24,25-Dihydroxy- and 1,24,25-trihydroxyvitamin D3 were produced by rat yolk sac homogenates incubated with tritiated 25-OHD3 and 1,25-(OH)2D3. Rat yolk sac homogenates also formed small amounts of 25,26-dihydroxyvitamin D3. These newly synthesized metabolites were isolated and identified by Sephadex LH-20 chromatography, high performance liquid chromatography, and periodate cleavage. Yolk sac 25-OHD3- and 1,25-(OH)2D3-24-hydroxylases were present in mitochondria and were of a mixed function oxidase nature. They were detected in the yolk sac as early as day 12 in the embryonic period and until the end of gestation. No hydroxylation occurred in maternal liver, amnion, fetal brain, or skin homogenates. Both 24-hydroxylases were detected in pure isolated rat yolk sac endodermal cells. This may be of physiological importance, since they are the 1,25-(OH)2D3 target cells in the yolk sac. Injection of 1,25-(OH)2[3H]D3 into rat yolk sac vitelline veins strongly suggested that the yolk sac vitelline veins strongly suggested that the yolk sac produced 1,24,25-(OH)3D3 in vivo. We conclude that the yolk sac and more precisely its endodermal cells may help to control vitamin D metabolism within the fetoplacental unit.

In the rat, the visceral yolk sac is involved in the physiological mother to fetus transfer of nutrients and persists until the end of gestation (38). It is in addition the embryonic precursor of the intestine. We recently showed that the rat yolk sac is a vitamin D target organ since it contains both 1,25-(OH)2Ds receptors (39) and a vitamin D-dependent calcium binding protein (40,41). However, little is known about the endocrinological properties of this yolk sac. In the present work, we investigated this fetal organ for its ability to metabolize 25-OHD 3 , 24,25-(OH) 2 D3, and 1,25-(OH) 2 D 3 into more polar vitamin D metabolites. Another consideration which prompted this study was the fact that the main vitamin D target organs involved in calcium transfers also produce vitamin D metabolites.

EXPERIMENTAL PROCEDURES
Animals and Materials-Normal pregnant Wistar rats were obtained from Lessieux (France) and fed ad libitum with a normal diet (UAR 103).
In  (42). Five fetuses were injected per mother. After this procedure, animals were allowed to recover for 0.5, 1, or 2 h, when they were again anesthetized and a cesarean section was performed. Blood was collected from the injected fetuses via the axillary vessels and yolk sacs were carefully removed and immediately frozen in liquid nitrogen. Al the samples from one mother were pooled and kept at -20 "C until analysis.
In Vitro Studies: Tissue Preparations-Unless otherwise stated, in vitro experiments were performed on tissues obtained from the normal pregnant rats on days 18 and 19 of gestation. Animals were killed and bled by decapitation.
Homogenates-The uteri were rapidly excised and immediately placed in ice cold 0.15 M NaCl. All subsequent operations were performed at 4 "C. Yolk sacs were dissected free of amnion, fetus, and placenta. The tissue was washed several times with ice cold 0.15 M NaCl and then with 16 mM Tris/acetate (pH 7.4) containing 196 mM saccharose, 24 mM sodium succinate, and 1.8 mM magnesium acetate (Tris buffer). Yolk sacs were cut into small pieces and homogenized in 5 volumes of Tris buffer by 4 passes in a Potter-Elvehjem homogenizer kept in ice.
Mitochondrial Preparations-Yolk sac homogenates were centrifuged at 4 C for 10 mmin at 400 x g. The pellet was discarded and the supernatant was centrifuged at 4 C for 20 min at 5000 x g. The mitochondrial pellet was collected and resuspended in Tris buffer. The enrichment of the mitochondrial preparation was tested by determining the succinic dehydrogenase activity using the p-iodonitrotetrazolium violet method (43).
Cytosolic Preparations-Homogenates were centrifuged for 30 min at 135,000 x g in a Beckman L5 75 ultracentrifuge. The final high speed supernatant without the fluffy layer was taken as the cytosol.
Pure Isolated Endodermal Cells-These cells were obtained from intact yolk sacs by the action of EDTA in phosphate-buffered saline at 37 "C, as described in detail elsewhere (44). Cells were collected by low speed centrifugation, washed, and immediately suspended in Tris buffer (about 2 x 106 cells/ml) for incubation with the labeled sterol. The endodermal cells were identified by their appearance under a light microscope and their calcium binding protein content was detected by immunocytochemistry (44). The protein contents of the preparations were determined by a modified version of Lowry's method (45).
Incubation with Labeled Sterols-In a typical experiment, 1 ml of the yolk sac homogenates suspended in 16 mM Tris/acetate (pH 7.4) containing 196 mM saccharose, 24 mM sodium succinate, and 1.8 mM magnesium acetate was transferred into a 25-ml Teflon-stoppered Erlenmeyer flask which was gassed with 02 for 45 s. Ten pmol of the radioactive sterol (0.25 #tCi) dissolved in 10 pl of ethanol were then added, yielding a final concentration of 10 mM. The flask was gassed with pure O2 for a further 45 s and stoppered. Incubation was generally performed for 1 h at 37 C in a water bath with rapid shaking (2 Hz). At the end of incubation, aliquots were taken for duplicate determination of protein concentration and radioactivity. Unlabeled 24,25-(OH) 2 D 3 , 1,25-(OH) 2 D 3 , and 1,24,25-(OH) 3 D 3 were then added in order to minimize losses during extraction and to serve as internal standards during the HPLC separation.
After addition of 2 volumes of methanol and 1 of chloroform, radioactivity was extracted by shaking the flask overnight at 4 C. The mixture was then transferred to a separatory funnel. Addition of 1 volume of CHC13 allowed its separation into 2 phases. The organic phase was collected and the upper layer was washed twice with 1 volume of CHC13. When emulsions appeared, they were easily clarified by centrifugation. The organic phases were then dried by addition of solid Na 2 SO 4 . The residue after rotatory evaporation under vacuum was redissolved in a small amount of ethanol and aliquots were counted for radioactivity. Using this procedure, radioactive recovery was always nearly 100%.
For high performance liquid chromatography analysis, an aliquot of the organic phase was transferred into a conical 0.5-ml vial and evaporated to dryness under nitrogen. The dry residue was redissolved in 200 al of the suitable HPLC solvent. The vial was then centrifuged for 10 min at 2000 x g and the clear supernatant was injected into the HPLC column.
Chromatography. Sephadex (R) LH-20 chromatography of the lipid extract was performed in a column ( 2 D3 by a 75:15 hexane/ isopropyl alcohol solvent system, run at a rate of 3 ml/min. For further identification of purified radioactive metabolites, a Cis Bondapak (Waters) reverse phase column was used. An 82:18 methanol/ water solvent system, run at a rate of 1.2 ml/min, was used to identify 24,25-(OH) 2 D 3 and 25,26-(OH) 2 D 3 , and a 75:25 mixture for 1,24,25-(OH) 3 D 3 . Fractions corresponding to 30 s were collected and counted for radioactivity.
Radioactivity was measured in an Intertechnique SL 40 liquid scintillation counter using Picofluor (Packard, France). When necessary, results were corrected for quenching by means of an external standard.
For quantification of the newly formed metabolites, both Sephadex LH-20 chromatography and HPLC were used and gave the same results. Consequently, HPLC was chosen for routine experiments. Metabolite production was expressed as the ratio of tritium present in the appropriate region versus the radioactivity recovered from the column. In all cases, more than 90% of radioactivity was recovered from the HPLC column.
Periodate Reaction-The purified metabolites were submitted to periodate cleavage. Fifty pl of a 5% aqueous sodium meta-periodate (or 50 I of water as a control) was added to the metabolite dissolved in 50 / 1 of methanol. The reaction was allowed to proceed overnight at room temperature and under nitrogen. The reaction mixture was dried under vacuum and twice extracted with 200 1 l of CHC1 3 which was then evaporated under vacuum. Results are expressed as the per cent of radioactivity lost by the NaIO 4 -treated sample compared to the control. As another test of the periodate action, aliquots of NaIO 4treated and of untreated samples were injected into an HPLC column eluted with the appropriate straight phase solvent system. Radiochromatograms were obtained as described in the HPLC section.
Sucrose Density Gradient Analysis-The binding of the putative 24,25-(OH)2[H]D 3 to the rat serum vitamin D binding protein was tested by sucrose density gradient analysis as previously described (30).

Metabolism of 25-OH[ 3 H]D 3 by Yolk Sac Homogenates-
After incubation of yolk sac homogenates with 10 nM 25-OH[ 3 H]D 3 , the radioactive lipid extract was analyzed by HPLC and Sephadex LH-20 chromatography, and the radioactivity profile obtained showed that radioactive products more polar than 25-OHD 3 were formed during incubation. After HPLC separation, significant radioactivity peaks were detected in the 24,25-(OH) 2 D 3 and 25,26-(OH) 2 D 3 regions, but not in the 1,25-(OH) 2 D 3 region (Fig. 1). These results were confirmed by Sephadex LH-20 chromatography (results not shown). The radioactivity peaks were not detected by HPLC or Sephadex LH-20 chromatography in the control experiments in which 25-OH[ 3 H]D 3 was incubated with Tris buffer alone or when the homogenates were boiled for 15 min prior to incubation. These experiments indicate that the radioactive peaks observed were formed by enzymes and not by degradation. The use of several chromatographic steps, including Sephadex LH-20 chromatography and HPLC with straight and reverse phase systems, enabled newly formed radioactive metabolites to be characterized. The lipid extract was first ap-.plied to a precalibrated Sephadex LH-20 column. The fractions corresponding to the 24,25-(OH) 2 D 3 and 25,26-(OH) 2 D 3 legions were collected and analyzed by HPLC in the straight phase system. As shown in Fig. 2, A and B respectively. In addition, the remaining radioactivity was not found to be associated with the corresponding untreated metabolite standard after HPLC in the straight phase system. This is consistent with the presence of hydroxyl groupings in positions 24,25 and 25,26 of the molecules, taking into account that the radioactivity of the starting 25-OH[ 3 H]D 3 was located on methyls 26 and 27. To further confirm the 24,25-(OH) 2 D3 structure, the ability of the putative 24,25-(OH) 2 [ 3 H]D 3 to be bound by serum vitamin D binding protein was tested. This purified radioactive metabolite was incubated for 1.5 h at 4 °C with 1:20 diluted plasma from vitamin D-deficient rats, and the mixture was submitted to sucrose density gradient centrifugation analysis. The radioactivity was found to be associated with a protein sedimenting at 4 S. The 4 S radioactive peak was not detected when incubations were performed in the presence of a 500-fold excess of unlabeled 24,25-(OH) 2 Fig. 3. Compared to the controls, two new radioactive metabolites were detected. The first peak, X, which eluted in a region between 1,25-(OH) 2 D 3 and 1,24,25-(OH) 3 D 3 , did not correspond to any known vitamin D metabolites. The second radioactive peak was located in the 1,24,25-(OH)D 3 region. This metabolite was purified by HPLC in the straight phase system with an 85:15 hexane/isopropyl alcohol solvent mixture. Fig. 4 shows that the radioactivity of the HPLC-purified putative 1,24,25-(OH) 2 [ 3 H]D 3 fraction co-eluted as a single peak with standard 1,24,25-(OH) 3 D 3 during HPLC in both the straight and reverse phase systems.
When the purified putative 1,24,25-(OH) 2 [ 3 H]D 3 was submitted to the periodate reaction, no significant radioactive loss was observed. However, when the periodate-treated fraction was analyzed by HPLC, radioactivity was no longer present in the 1,24,25-(OH) 3 D 3 region. These results are consistent with the presence of a vicinal glycol grouping probably in position 24,25 of the molecule, since the tritium atoms of the starting 1,25-(OH) 2 [ 3 H]D3 were located in positions C23 and C24.
All these results strongly suggest that the new metabolite was 1,24,25-(OH) 2 D 3 .  and 1,25-(OH) 2 D 3 24-hydroxylases, respectively. When homogenates were flushed with carbon monoxide prior to incubation, the formation of both 24-hydroxylated metabolites was greatly lowered (up to 70% inhibition). Such decreases were never observed in control experiments in which nitrogen was used instead of carbon monoxide.

Inhibition of the 25-OHD 3 and 1,25-(OH) 2 D 3 24-Hydroxyl
In other respects, in the absence of succinate in the incubation medium, there was no formation of 24-hydroxylated metabolites. All these experiments confirmed the cytochrome P-450 mixed function oxidase nature of the yolk sac 24-hydroxylase system.

Mitochondrial Location of the Enzymatic Activities-Yolk
sac mitochondrial preparations were incubated in Tris buffer with 10 nM 25-OH[ 3 H]D 3 or 1,25-(OH) 2 [ 3 H]D 3 . The HPLC radioactive profiles obtained for the two lipid extracts were identical with those shown for homogenates in Figs. 1 and 3. Conversely, when yolk sac cytosols were incubated under the same conditions, the radioactivity was recovered after HPLC as the untransformed starting metabolite. These results show that the enzymatic activities were mainly of mitochondrial origin. Tissue Distribution Studies-Tissue distribution studies indicated that 25-OHD 3 and 1,25-(OH) 2 D 3 were metabolized by rat yolk sac homogenates but not by amnion, the other component of fetal membranes, or by fetal brain, fetal skin, or maternal liver.

Metabolism of 25-OH[ 3 H]DD and 1,25-(OH) 2 [3H]D 3 by Isolated Yolk Sac Endodermal
In Vivo 24-Hydroxylation of 1,25-(OH) 2 D 3 by Yolk Sac-To test whether 24-hydroxylation occurred in vivo, we chose to follow the metabolism of 1,25-(OH) 2 D 3 , since in vitro experiments had indicated that 24-hydroxylation of the latter appeared to be greater than that of 25-OHD 3   In vivo 24-hydroxylation of 1,25-(OH) 2 D 3 by yolk sac 1,25-(OH)2h 3 H]D 3 was injected into the vitelline vein of the yolk sac of normal pregnant rats. One-half, 1, or 2 h later, the animals were killed. Lipids were extracted from fetal plasma and yolk sac and counted for radioactivity. The total radioactivity present in the lipid extracts of yolk sac or fetal plasma is expressed as counts/min/g of yolk sac or per ml of fetal plasma. 1,24,25-(OH) 3 D 3 is quantified after HPLC analysis of the lipid extracts and expressed as the percentage of the radioactivity present in the 1,24,25-(OH) 3

DISCUSSION
This study clearly demonstrates that rat yolk sac homogenates metabolize 25-OHD 3 to more polar metabolites, i.e. 24,25-(OH) 2 D 3 and to a lesser extent 25,26-(OH) 2 D 3 . This organ also hydroxylates in vitro 1,25-(OH) 2 D 3 in position 24, leading to 1,24,25-(OH) 3 D 3 . The identification of these three metabolites is reasonably certain, since it is based on the results of several biochemical techniques, including Sephadex LH-20 chromatography, HPLC in straight and reverse phase systems, and periodate sensitivities.
The 25-OHD 3 and 1,25-(OH) 2 D 3 24-hydroxylase activities of the yolk sac were found to be associated with mitochondria. They were inhibited by metyrapone, SKF 525 A, and carbon monoxide and thus are of a cytochrome P-450 mixed function oxidase nature. These properties are similar to those reported for the 24-hydroxylase in adult kidney (7,8). Under our experimental conditions, 1,25-(OH) 2 D 3 seemed to be a better substrate than 25-OHD 3 for the rat yolk sac 24-hydroxylase enzymes. Such preferential transformation of 1,25-(OH) 2 D 3 has been reported for rat intestinal 24-hydroxylase (9).
It was important to test whether the 25-OHD 3 and 1,25-(OH) 2 D 3 24-hydroxylases are active in vivo. Injection of tritiated 1,25-(OH) 2 D 3 into the vitelline vein of the yolk sac demonstrated that 1,25-(OH) 2 D 3 24-hydroxylase may function in vivo. Yolk sac may thus be considered as one of the sources of 1,24,25-(OH) 3 D 3 that we detected in the fetal plasma. The presence of this trihydroxylated vitamin D metabolite has not so far been demonstrated in the fetus. In the adult, 1,24,25-(OH) 3 D 3 is often considered as an inactivated form of 1,25-(OH) 2 D 3 (1). However, one cannot exclude that it would be active per se in the fetal unit (46). With respect to 24,25-(OH) 2 D 3 , several works recently drew attention to this metabolite during fetal life (24,25,31). 24,25-(OH) 2 D3 is known to be specifically accumulated in the fetal skeleton and may be important in early bone formation (26). In addition, the plasma levels of this metabolite exhibit a mother to fetus (23,30,31). However, the origin of 24,25-(OH) 2 D 3 in the rat fetus remains unknown. The present study suggests that yolk sac is a potential source of 24,25-(OH) 2 D 3 for fetus. Thus, our findings concerning vitamin D metabolism by the yolk sac, together with those describing its ability to transform progesterone (47), extend the function of the yolk sac in steroid hormone metabolism.
We found that yolk sac homogenates formed 24,25-(OH) 2 D 3 and 1,24,25-(OH) 3 D 3 throughout the second half of gestation. During this period, changes in the in vitro metabolism of 25-OHD 3 and 1,25-(OH) 2 D 3 were observed. This may be of physiological significance, although it should be remembered that several uncontrollable parameters, such as endogenous substrate or inhibitor concentrations, may interfere in the enzymatic determinations. It is worth noting that as early as day 12 of gestation, 25-OHD 3 and 1,25-(OH) 2 D 3 24 hydroxylases are present in the yolk sac and also in the embryo. This is the first evidence for vitamin D metabolism during embryonic life in the rat.
The presence of 25-OHD 3 and 1,25-(OH) 2 D 3 24-hydroxylases in the yolk sac makes this vitamin D target organ (39)(40)(41) comparable to the main target organs like intestine or bone, which are also extrarenal sites of 24-hydroxylation (9)(10)(11). In these organs, the identification of the cells containing the 24hydroxylases is obviously of interest. The relatively simple structure of the yolk sac allowed us to elucidate this point. Anatomically, the rat yolk sac is composed of three layers: the mesothelium, the mesenchyme containing the vitelline vessels, and the endoderm (48). With respect to the rat yolk sac transfer function, the endodermal cells, facing the uterine epithelium, are the most important element of the organ (49). In addition, calcium binding protein was shown by immunocytochemistry to be exclusively located in the endodermal cells, suggesting that they are the 1,25-(OH) 2 D 3 target cells of the yolk sac (44). One of us recently succeeded in isolating pure rat yolk sac endodermal cells (44). The present study demonstrates that they contain both 25-OHD 3 and 1,25-(OH) 2 D 3 24-hydroxylases. These findings may have some physiological importance, since the 24-hydroxylase systems we located in these noncultured well defined isolated vitamin D target cells might indeed be a means for a local regulation of vitamin D metabolism and action. This idea is supported by recent reports indicating that an established pig kidney cell line possesses both 1,25-(OH) 2 D 3 receptors and 25-OHD 3 24hydroxylase activity stimulated by 1,25-(OH) 2 D 3 . 2 3 The model of cultured pure endodermal cells now being developed in our laboratory will provide a new convenient tool for studying the relationships between the 1,25-(OH) 2 D 3 hormonal mechanism of action and the 24-hydroxylase systems.