The aim of the present study was to assess the effects of insulin-like growth factor I (IGF-I) upon the synthesis of 1,25-dihydroxyvitamin D3 [1,25-(OH)2D3] by human placenta trophoblasts in culture. Cytotrophoblastic cells obtained from normal term human placentae were cultured in Dulbecco's modified Eagle's medium with HEPES and glucose (DMEM-HG) during 72 h and further incubated in serum-free DMEM-F12 in the presence of IGF-I prior to the addition of [3H]-25-(OH)D3 used as a precursor. The results showed that 2 h preincubation time with IGF-I was required for maximal production of [3H]-1,25-(OH)2D3. Cultures in the presence of increasing concentrations of IGF-I (0–6.5 nmol/l), added 2 h before incubation with the labelled substrate, resulted in a dose-dependent response increment of [3H]-1,25-(OH)2D3 production with a maximal conversion rate at the dose of 2.6 nmol/l. Higher doses of IGF-I did not result in further stimulatory effects. Co-incubations in the presence of cycloheximide significantly (P < 0.0001) inhibited the IGF-I-mediated effects upon [3H]-1,25-(OH)2D3 production. Identity of putative [3H]-1,25-(OH)2D3 produced by human placenta was confirmed by spectral and receptor binding analysis. These results demonstrate the ability of cultured human syncytiotrophoblast cells to convert 25-(OH)D3 to 1,25-(OH)2D3 and suggest a local protein-dependent regulatory effect of IGF-I upon this biotransformation.
Placenta is considered as an extrarenal source of 1,25-dihydroxyvitamin D3 (1,25-(OH)2D3) (Weisman et al., 1979; Whitsett et al., 1981; Zerwekh and Breslau, 1986; Hollis et al., 1989). This extrarenal source may contribute to the high maternal serum concentrations of this hormone in pregnant women. It may also be important to the fetal pool of the hormone or as a local source for its action on the placenta (Delvin et al., 1985; Zerwekh and Breslau, 1986; Kovacs and Kronenberg, 1997). However, the placenta contribution to vitamin D metabolism has been questioned. Indeed, synthesis of 1,25-(OH)2D3 by trophoblast homogenates requires supraphysiological amounts of 25-hydroxyvitamin D3 [25-(OH)D3] (Whitsett et al., 1981; Zerwekh and Breslau, 1986; Hollis et al., 1989) and 1,25-(OH)2D3 production by placental cells has been inconsistently found (Rubin et al., 1993). Furthermore, it has been previously reported that, whereas synthesis of 1,25-(OH)2D3 in human decidua is an enzymatically-mediated process, trophoblast tissue might use an alternate nonenzymatic mechanism of hydroxylation (Hollis et al., 1989; Glorieux et al., 1995).
In any case, studies in maternal serum suggest that production of 1,25-(OH)2D3 during pregnancy is regulated by factors different from those acting in the non-pregnant state such as calcium, phosphate and parathyroid hormone (PTH) (Verhaegue and Bouillon, 1992; Ardawi et al., 1997; Kovacs and Kronenberg, 1997). Numerous studies in cell cultures, rodents and humans suggest that insulin-like growth factor I (IGF-I) is an important regulatory factor of the activity and/or synthesis of the renal enzymatic complex hydroxylating 25-(OH)D3 into 1,25-(OH)2D3 (Gray, 1987; Halloran and Spencer, 1988; Caverzasio et al., 1990; Nesbitt and Drezner, 1993; Condamine et al., 1994; Menaa et al., 1995; Wong et al., 1997; Bianda et al., 1997, 1998; Wei et al., 1998). These data, in addition to previous studies demonstrating that placenta is a source of IGF-I (Fant et al., 1986; Han et al., 1996), prompted us to investigate the effects of this growth factor upon the ability of cultured human placenta to convert 25-(OH)D3 into 1,25-(OH)2D3.
Materials and methods
Materials and reagents
Dulbecco's modified Eagle's medium (DMEM and DMEM-F12), Hank's balanced salt solution (HBSS), fetal calf serum (FCS), HEPES, and gentamicine were obtained from Gibco (Grand Island, NY, USA). Percoll, 8-bromo adenosine 3′,5′-cyclic monophosphate (8-Br-cAMP), deoxyribonuclease I (DNase I), bovine serum albumin (BSA), glutamine, and cycloheximide were purchased from Sigma Chemical Co (St Louis, MO, USA). All solvents were of high-performance liquid chromatography (HPLC) grade and were obtained from Merck (Darmstadt, Germany). Unlabelled 25-(OH)D3 and 1,25(OH)2D3 were a generous gift from Dr E.-M.Gutknecht and Dr P.Weber (Hoffmann La Roche Ltd, Basel, Switzerland). The 25-hydroxy-[26,27-methyl-3H]-cholecalciferol ([3H]-25-(OH)D3; specific activity: 17 Ci/mmol), 1α,25-dihydroxy-[26,27-methyl-3H]-cholecalciferol ([3H]-1,25-(OH)2D3; specific activity: 130 Ci/mmol) and recombinant human IGF-I were purchased from Amersham (Amersham, Bucks, UK). Human chorionic gonadotrophin (HCG) radioimmunoassay was kindly provided by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (Rockville, MD, USA). All other reagents were of analytical grade.
Cytotrophoblastic cell isolation and culture
The study protocol was approved by the Human Ethical Committee of the Institute. Term placentae were obtained from normal pregnant women after spontaneous delivery. Tissues were brought immediately to the laboratory where several cotyledons were removed and rinsed thoroughly in 0.9% NaCl at room temperature. The isolation and culture of cytotrophoblasts were performed as previously described (Kliman et al., 1986), with minor modifications. Briefly: soft villous tissue (30 g), free of connective tissue and vessels, was collected. Tissue was coarsely minced and digested with 0.125% trypsin, and 0.2 mg/ml DNAase I (1500 kunitz IU/mg) in warmed calcium- and magnesium-free HBSS containing 25 mmol/l HEPES, pH 7.4, at 37°C for 30 min. Cell suspensions were pooled, centrifuged at 1000 g for 30 min, and resuspended in DMEM containing 25 mmol/l HEPES and 25 mmol/l glucose (DMEM-HG). The resultant suspension was placed on 5–70% Percoll (v/v) gradients made up in HBSS. Gradients, which consisted of 5% steps of 3 ml each, were centrifuged at 1200 g at room temperature for 30 min. After centrifugation, the middle band (density 1.048–1.062) containing the cytotrophoblasts was removed, washed once with DMEM-HG, and resuspended in culture medium. Percoll gradient-purified cytotrophoblasts were diluted to a concentration of 1×106 cells/ml with DMEM-HG containing 4 mmol/l glutamine, 50 μg/ml gentamicin, and 20% heat-inactivated FCS, plated in 35 mm Nunclon culture dishes (Nunc, Roskilde, Denmark), and incubated for 72 h in humidified 5% CO2 and 95% air at 37°C.
Morphological and functional aspects of placental cell cultures
Daily, the morphological aspects of cell cultures were examined. HCG in the culture media was measured as previously described (Diaz-Cueto et al., 1994; Queipo et al., 1998) by specific radioimmunoassay using reagents and protocols provided by the NIDDK. Anti-HCG-H80, at a final working dilution of 1:150 000, was used as antiserum. This antiserum exhibits 1.2 and 3.2% crossreactivities with free HCG α- and β-subunits respectively. The sensitivity of the assay was 0.025 ng/tube and the inter- and intra-assay coefficients of variation were <10 and <6% respectively. Total protein content of cell cultures was measured by a previously described method (Bradford, 1976), using BSA as standard.
Metabolism of 25-(OH)D3 by placental cell cultures
On the third day of culture, medium was changed and the cells incubated in 2 ml serum-free DMEM-F12. To assess the ability of placental cell cultures to convert 25-(OH)D3 into 1,25-(OH)2D3, a 3 nmol/l concentration of [3H]-25-(OH)D3 was added in fresh serum-free DMEM-F12 and incubations continued for 120 min. Culture medium was then transferred to glass tubes, and the cells were washed with 0.5 ml of methanol. Protein cell content was determined after addition of 0.5 ml of 1 mol/l NaOH. The [3H]-25-(OH)D3 and its metabolites were extracted from the medium with an additional 3.5 ml of methanol plus 4 ml of chloroform (Bligh and Dyer, 1959). The chloroform phase was dried down under N2, and lipid extracts were redissolved in chromatographic solvent. The samples were co-chromatographed with 100 ng unlabelled 1,25-(OH)2D3 as an elution marker on a Waters HPLC fitted with a photodiode array detector (PDA; model 996; Waters Associates, Milford, MA, USA) using an ultrasphere Si, 5 μm, 4.6×250 mm column (Beckman Instruments, Palo Alto, CA, USA). Vitamin D3 metabolites were separated by two-step straight phase HPLC as previously described (Sempere et al., 1989; Kachkache et al., 1993; Menaa et al., 1995). Fractions (1 min) were collected and an aliquot of each was removed for radioactivity determination. Fractions co-eluting with unlabelled 1,25-(OH)2D3 were pooled and rechromatographed on a second straight phase HPLC using the same column and eluted in methylene chloride:isopropanol (95:5) at a flow rate of 1 ml/min. The conversion rate of [3H]-25-(OH)D3 into putative [3H]-1,25-(OH)2D3 was determined by calculating the percentage of radioactivity co-eluting with unlabelled 1,25-(OH)2D3 after the two successive chromatographies. Results were expressed as fmol/mg protein.
Characterization of putative [3H]-1,25-(OH)2D8
For this purpose, placental cells were incubated in the presence of 2.5 μmol/l unlabelled 25-(OH)D3. The putative 1,25-(OH)2D3 produced by these cells was purified as described above with one exception: lipid extracts were co-chromatographed with 1 nCi [3H]-1,25-(OH)2D3 instead of unlabelled hormone used as an elution marker. The amount of putative 1,25-(OH)2D3 was calculated from the absorbance at 265 nm of the single peak co-eluting with [3H]-1,25-(OH)2D3 in the second HPLC. The purified 25-(OH)D3 metabolite was tested for its ability to displace synthetic labelled [3H]-1,25-(OH)2D3 from its specific calf thymus receptor (Reinhardt et al., 1984) using a commercial radioreceptor assay kit (Nichols Institute Diagnostics, San Juan Capistrano, CA, USA).
Regulatory effects of IGF-I upon [3H]-1,25-(OH)2D3 synthesis
In order to determine the effects of IGF-I on placental [3H]-1,25-(OH)2D3 production, cell cultures were preincubated in the presence of IGF-I at different times previous to the addition of 3 nmol/l of tritiated [3H]-25-(OH)D3.
Dose–response curves of IGF-I effects on putative [3H]-1,25-(OH)2D3 production were performed by preincubations with different amounts of IGF-I (0–6.5 nmol/l) during 2 h prior to the addition of [3H]-25-(OH)D3. Furthermore, the effects of IGF-I upon [3H]-1,25-(OH)2D3 production were also assessed in the presence of the protein synthesis inhibitor cycloheximide (30 μmol/l). The IGF-I mitogenic effects were evaluated by measuring protein content in cell cultures.
Data are presented as mean ± SD. All experiments were performed at least three times and each of them consisted of six to nine culture replicates. Statistical significance between groups was established by one-way analysis of variance (ANOVA) using Fisher's protected least-square differences. P < 0.05 was considered to be statistically significant.
Morphological and functional aspects of placental cell cultures
Microscopic examination of cell cultures showed that after 72 h, cytotrophoblast cells consistently aggregated forming well-differentiated syncytiotrophoblast structures. In addition, our placental cell cultures corresponded to functional syncytiotrophoblasts since HCG secretion in the presence of 1.5 nmol/l 8-Br-cAMP increased significantly (P < 0.0001) when compared with cells cultured in the absence of the cyclic nucleotide analogue (Figure 1), as previously reported (Feinman et al., 1986; Queenan et al., 1987; Ulloa-Aguirre et al., 1990; Díaz et al., 1997; Queipo et al., 1998).
Metabolism of 25-(OH)D3 by placental cell cultures
Syncytiotrophoblast cells were able to convert 25-(OH)D3 into more polar metabolites. As shown in Figure 2, one of these metabolites eluting as a single peak co-eluted with 1,25-(OH)2D3 during both the first (Figure 2A) and the second (Figure 2B) HPLC. The amount of this metabolite found after 2 h of incubation with 3 nmol/l [3H]-25-(OH)D3 ranged between 80 and 310 fmoles/mg protein (170 ± 80 fmoles/mg). When cells were incubated in presence of 2.5 μmol/l unlabelled 25-(OH)D3, the amount of 1,25-(OH)2D3 produced by syncytiotrophoblasts averaged 40 ± 15 pmoles/mg protein. Putative 1,25-(OH)2D3 was identified by its spectral and binding properties. Indeed, putative and 1,25-(OH)2D3 presented similar maximum absorbance at 265 nm. Furthermore, the ability of putative 1,25-(OH)2D3 to displace [3H]-1,25-(OH)2D3 from its specific calf thymus receptor was identical to that of 1,25-(OH)2D3 (Figure 3).
Regulatory effects of IGF-I upon 1,25-(OH)2D3 placental production
The simultaneous addition of IGF-I and [3H]-25-(OH)D3 to day 3 cultures did not result in a significant increase in the conversion of substrate to the active vitamin D metabolite (Figure 4). As shown in this figure, time-response experiments in cultures preincubated with IGF-I (3.3 nmol/l) at various times (0, 2, 8, 16 h) prior to the addition of [3H]-25-(OH)D3, demonstrated that a minimum of 2–8 h preincubation period was required to significantly increase (P < 0.0001) the conversion of substrate to [3H]-1,25-(OH)2D3. However, a significant decrease (P < 0.0001) in [3H]-1,25-(OH)2D3 synthesis was observed with longer preincubation periods. This stimulatory effect of IGF-I was not associated with an increase in protein synthesis linked to the mitogenic action of this growth factor, since cell protein contents were similar in IGF-I-treated (81 ± 5 μg/well) and untreated cells (83 ± 5 μg/well).
Effects of increasing concentrations of IGF-I added to cultures 2 h prior to incubation with [3H]-25-(OH)D3 were also evaluated. As shown in Figure 5, a dose–response effect of IGF-I upon [3H]-1,25-(OH)2D3 production was obtained with a maximal conversion rate at a concentration of 2.6 nmol/l (P < 0.0001 versus control). Higher doses of IGF-I did not further increase the stimulatory effect, but synthesis of [3H]-1,25-(OH)2D3 remained significantly above controls (P < 0.005).
In addition, the stimulatory effects of IGF-I upon [3H]-1,25-(OH)2D3 production were also evaluated in the presence of cycloheximide. As shown in Figure 6 (Panel A), the presence of 30 μmol/l of cycloheximide significantly inhibited (P < 0.0001) the production rate of [3H]-1,25-(OH)2D3 in 2 h-preincubated cultures stimulated with IGF-I (2.6 nmol/l). A similar effect was observed on 8-Br-cAMP-stimulated HCG secretion (Figure 6B).
The results presented in this communication indicate that cultured human syncytiotrophoblasts are able to produce 1,25-(OH)2D3 when incubated in the presence of physiological concentrations of 25-(OH)D3. The culture system used converted 25-(OH)D3 into a metabolite with chromatographic behaviour, UV spectral patterns and binding abilities to thymus cytosol receptors identical to 1,25-(OH)2D3. Moreover, these data provide further evidence showing that not only decidual cells (Delvin et al., 1985; Kachkache et al., 1993) but also syncytiotrophoblasts may contribute to the maternal or fetal pool of 1,25-(OH)2D concentrations during human pregnancy. These results are in agreement with other studies showing the ability of the trophoblast tissue or cells to produce 1,25-(OH)2D in vitro (Hollis et al., 1989; Rubin et al., 1993). In the present study, consistent production of 1,25-(OH)2D3 was detected using physiological amounts of 25-(OH)D3, whereas in earlier similar studies (Hollis et al., 1989; Rubin et al., 1993) the results were inconsistent or required supraphysiological doses of the precursor. In addition, previous observations suggest that 1,25-(OH)2D3 production involves a different hydroxylation system than the 25-(OH)D-1α-hydroxylase (1αOHse) (Hollis et al., 1989). Although it is not possible to ascertain whether the presently observed production of 1,25-(OH)2D3 involved the mitochondrial 1αOHse system, the conditions under which 1,25-(OH)2D3 production was observed [cell incubations with low doses of 25-(OH)D3 during a short period of 2 h], make it likely. In addition, part of the discordance with previous studies may derive from differences in the placental tissues or cells tested, including trophoblast tissue (Hollis et al., 1989; Rubin et al., 1993) or cultured syncytiotrophoblast cells (Rubin et al., 1993; and present work). It is also important to mention that differences in the culture medium and/or the concentration of FCS used may also have influenced the differentiation state of the cells and thus the ability to produce 1,25-(OH)2D3. Indeed, DMEM containing 3.75% FCS was used by Rubin et al. (1993) while DMEM supplemented with glucose and 20% FCS was used in the present work. Finally, the incubation of cells with DMEM or Roswell Park Memorial Institute (RPMI) 1640 medium may also have influenced the activity of the 1αOHse, as has been previously shown in cultures of decidual cells (Kachkache et al., 1993).
In the culture system and under the experimental conditions used throughout this study, a clear and significant stimulatory effect of IGF-I on 1,25-(OH)2D3 production was observed. These data agree with previous observations that synthesis of 1,25-(OH)2D3 is under the control of a number of modulators. It is known that in addition to calcium and phosphate, and hormones such as PTH and calcitonin, including 1,25-(OH)2D itself, other factors may influence the activity and possibly the expression of renal 1αOHse (Garabédian et al., 1972; DeLuca, 1978; Fraser, 1980; Kawashima and Kurokawa, 1986; Breslau, 1988; Reichel et al., 1989). IGF-I is likely to be one of these additional physiological regulators, as it increases in vivo the serum concentrations of 1,25-(OH)2D and stimulates its renal production in vitro (Gray, 1987; Halloran and Spencer, 1988; Caverzasio et al., 1990; Nesbitt and Drezner, 1993; Condamine et al., 1994; Menaa et al., 1995; Wong et al., 1997; Bianda et al., 1997, 1998; Wei et al., 1998). These observations are of physiological importance since parallel increases in serum IGF-I and 1,25-(OH)2D concentrations have been observed during pregnancy (Petraglia et al., 1996; Kovaks and Kronenberg, 1997), suggesting that the increase in serum maternal 1,25-(OH)2D results from an IGF-I-dependent stimulation of renal 1αOHse. However, based on the present findings, IGF-I may also stimulate the production of 1,25-(OH)2D in placenta, thus representing an additional source of maternal and/or fetal pools of this metabolite.
In the absence of precise kinetic data, it is not possible in this study to know whether the IGF-I-dependent increase in placental 1,25-(OH)2D3 production results from a stimulation of the 1αOHse activity or from a decreased catabolism of the produced 1,25-(OH)2D3. Yet kinetic studies with renal cells suggest that stimulation of the 1αOHse is a major component of the IGF-I effect on vitamin D metabolism (Nesbitt and Drezner, 1993; Menaa et al., 1995). In so far as the mechanism of the IGF-I effect on trophoblast cells remains to be elucidated, results obtained with renal cells (Caverzasio and Bonjour, 1989, 1992; Caverzasio et al., 1990; Quigley and Baum, 1991; Menaa et al., 1995) indicate that this pathway may involve changes in phosphate transport or the stimulation of calcium uptake by placental cells. Whatever this mechanism, results of cell incubations with cycloheximide suggest that the IGF-I effect on 1,25-(OH)2D3 production by trophoblast cells involves de-novo protein synthesis. Recently IGF-II, rather than IGF-I, was shown to be preferentially expressed in the placenta trophoblast (Han et al., 1996), thus suggesting that both local and/or systemic IGFs acting via the same or the specific receptor are involved in 1,25-(OH)2D3 production by the placenta. However, whether IGF-II is the relevant peptide deserves to be further investigated.
In summary, this study demonstrates a clear ability of cultured human syncytiotrophoblast cells to synthesize 1,25-(OH)2D3 with a marked stimulation of this production by physiological doses of IGF-I. The physiological relevance of these findings has not been evaluated, but interactions between IGF-I and 1,25-(OH)2D may play an important role in a tissue which expresses both the vitamin D and IGF-I receptors (Marshall et al., 1974; Tanamura et al., 1995). If the placenta appears not to be considered as an important contributor to the 1,25-(OH)2D concentrations in maternal blood, IGF-I may locally regulate the production of 1,25-(OH)2D by trophoblast cells and hence control some of the ion transport mechanisms, as in the case of renal cells (Caverzasio and Bonjour, 1989, 1992; Caverzasio et al., 1990; Quigley and Baum, 1991; Menaa et al., 1995), through the placental barrier or other vitamin D-dependent functions on this tissue. It may also contribute to the increase in the fetal pool of 1,25-(OH)2D during fetal growth.
The authors are indebted to the National Institute of Diabetes and Digestive and Kidney Diseases for HCG radioimmunoassay reagents, and to F.Hoffmann–La Roche LTD for 25(OH)D3 and 1,25-(OH)2D3 standards. This work was supported in part by grants (3508-M9310, M9607, and 26238-M) from the National Council of Science and Technology (México) and the Special Programme of Research, Development and Research Training in Human Reproduction of the World Health Organization (Geneva, Switzerland). The authors acknowledge with thanks to Hospital General M. Gea Gonzalez, México D.F., for placentae donation.