| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Institute of Animal Physiology, Ludwig Maximilians University (R.G.E., K.W., U.Z.), 80539 Munich, Germany; Institute of Mammalian Genetics, GSF National Research Center for Environment and Health (D.W.S., R.B.), 85764 Neuherberg, Germany; National Center for Scientific Research, UPR 1524, National Institute of Agricultural Research (M.L.), 78 350 Jouy-en-Josas, France; Department of Dermatology D92, Bispebjerg Hospital (R.G.), 2400 Copenhagen NV, Denmark; and Institute of Experimental Genetics, GSF National Research Center for Environment and Health (G.M., J.A.), 85764 Neuherberg, Germany
Address all correspondence and requests for reprints to: Reinhold G. Erben, M.D., D.V.M., Institute of Animal Physiology, University of Munich, Veterinaerstrasse 13, D-80539 Munich, Germany. E-mail: r.erben{at}lrz.uni-muenchen.de.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The role of 1,25-(OH)2D3 in active tubular reabsorption of calcium in the kidney is still controversial. Both in vivo and in vitro studies have demonstrated that 1,25-(OH)2D3 is able to stimulate transcellular calcium transport in the distal nephron (reviewed in Refs. 7 and 8). Transcellular calcium transport in distal tubule cells is a multiple-step process and involves calcium entry at the apical membrane, intracellular calcium diffusion, and active extrusion at the basolateral membrane (reviewed in Ref. 7). The stimulatory action of 1,25-(OH)2D3 on distal renal tubular calcium reabsorption may involve induction of the intracellular calcium-binding proteins calbindin D9 k (9) and calbindin D28 k (10). In the kidney of VDR-ablated mice, the mRNA levels of calbindin D9 k were found to be strongly down-regulated (4, 6, 11), while those of calbindin D28 k were only moderately decreased (11) or unchanged (6). Moreover, the recently cloned and characterized apical epithelial calcium channels (ECaC1 and -2; calcium transport protein1, CaT1, is a synonym for ECaC2) may also play an important role in the 1,25-(OH)2D3-induced increase in renal transcellular calcium transport (12, 13, 14, 15). In this context, it has been shown that VDRE are localized in the promoter region of the human (16) and murine (17) ECaC1 genes, that ECaC1 mRNA and protein levels are decreased in the kidney of vitamin D-deficient rats (18), and that systemic treatment with high dose 1,25-(OH)2D3 increases the renal mRNA levels of ECaC1 and -2 in mice (6). However, data on the physiological importance of these effects in vivo are not available at present.
Besides the role of 1,25-(OH)2D3 in the regulation of gene transcription, a number of cellular and tissue responses to vitamin D metabolites have been described that occur within seconds to minutes (19, 20). These rapid, nongenomic actions of 1,25-(OH)2D3 are believed to be mediated by a putative membrane receptor (20, 21) and include activation of the phospholipase C pathway (22), activation of the adenylate cyclase pathway (23), opening of L-type, voltage-gated Ca2+ channels in the plasma membrane (23, 24, 25), or Ca2+ mobilization from the endoplasmic reticulum (26) in different cell types. However, the physiological significance of nongenomic actions of 1,25-(OH)2D3 is unclear.
To investigate further the functions of 1,25-(OH)2D3 in skeletal and extraskeletal tissues we sought to generate VDR null mutant mice by targeted disruption of the first zinc finger of the VDR. To monitor expression of VDR, a lacZ reporter gene cassette driven by the endogenous VDR promoter was introduced into the targeting vector. Here we show that our gene-targeted mutant mice expressed a VDR with an intact hormone binding domain, but lacking the first zinc finger necessary for DNA binding. These mutant mice show rickets, secondary hyperparathyroidism, growth retardation, and alopecia and are a phenocopy of mice totally lacking the VDR protein (4, 5, 6). After correction of calcium homeostasis by dietary means, homozygous mutants revealed a profound molecular defect in renal calcium reabsorption. Furthermore, genomic and nongenomic responses to vitamin D metabolites were absent in homozygous mutant mice, suggesting that the classical, nuclear VDR is the receptor mediating both genomic and nongenomic actions of 1,25-(OH)2D3.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, A and B). Heterozygous animals showed no defects, and homozygous animals were generated at the expected Mendelian frequency. To our surprise, Northern analysis showed that homozygous gene- targeted mice expressed a VDR transcript of similar length as the wild type in duodenum and kidney (Fig. 1C). Subsequent RT-PCR analysis of RNA from duodenum and kidneys using primers located in exons 1, 2, 3, and 7 revealed the presence of exons 1, 3, and 7, but not of exon 2, in the VDR mRNA in gene-targeted mice (Fig. 1D). When 5' and 3' primers in exons 1 and 7, respectively, were used, the RT-PCR product from homozygous mice was shorter compared with that in wild-type animals (Fig. 1D). The PCR products from homozygous gene-targeted mice were sequenced and found to be authentic VDR with a deletion of 148 bp corresponding to exon 2 (data not shown). The finding that a RT-PCR product could be amplified using primers in exons 1 and 7 indicated that the lacZ and neo cassettes located between exons 1 and 3 of the targeted VDR genomic sequence were obviously spliced out during mRNA processing. The fact that most of intron 2 of the VDR gene was left intact in the targeting construct may facilitate such splice mechanisms. It is not known whether the intron 2 contains regulatory sequences. To rule out tissue-specific splicing of the VDR transcripts in the mutant animals, we performed RT-PCR analysis on RNA from kidney, duodenum, testis, ovary, skin, and primary osteoblast cultures. Using primers located in exons 1, 3, and 7 of the VDR, the same RT-PCR products were present in all tissues examined (Fig. 1E). Thus, we found no evidence for tissue-specific splice mechanisms. The tissue distribution of the mutant VDR mRNA in homozygous mice paralleled that in wild-type mice (Fig. 1E).
|
). The expected difference in molecular mass of about 4 kDa was too small to result in visible changes in protein motility in SDS-PAGE. Immunohistochemistry with a monoclonal anti-VDR antibody directed against an epitope in the region just C terminal of the second zinc finger (amino acids 89–105) revealed expression of the mutant receptor in duodenal epithelial cells of homozygous mice (Fig. 1G). Similar to the wild-type mice, the mutant VDR protein was mainly localized in the nuclei of the mucosa cells. To examine whether the mutant VDR protein would bind 1,25-(OH)2D3 with normal affinity, we performed competitive binding experiments with nuclear extracts from duodenum of wild-type and homozygous mutant mice as well as with human recombinant VDR. The mutant receptor bound radiolabeled 1,25-(OH)2D3 with similar affinity (Kd = 0.80 nM) compared with wild-type (Kd = 0.36 nM) or human recombinant VDR (Kd = 0.62 nM). To investigate whether the mutant VDR could still bind DNA, we performed gel mobility shift experiments with wild-type mouse VDR and the N-terminal truncated form of murine VDR (lacking the first 51 amino acids) present in our mutant mice. Both proteins were cloned from mouse duodenal cDNA and overexpressed in a bacterial system as glutathione-S-transferase (GST) fusion proteins. As shown in Fig. 1H, the mutant VDR did not bind mouse osteopontin VDRE.
These data showed that our gene-targeted mice expressed a VDR protein with a deletion of the first zinc finger, and thus an inability to bind DNA, but with normal ligand binding affinity. Because the usual terminology VDR-/- for homozygous null mutant mice would be misleading for our mice, we use the term VDRwt/
for heterozygous and VDR/ for homozygous gene-targeted mice in the following text. After initial confirmation with Southern blot analysis, genotyping of the mice for all experiments was performed using PCR.
Expression of the lacZ Reporter Gene
To investigate the expression of VDR during development, we studied the expression of the lacZ reporter gene in mouse embryos. The earliest lacZ expression was found in E10.5 embryos (data not shown). In E11.5 embryos, lacZ staining was present in the neural tube, telencephalon, mesencephalon, rhombencephalon, spinal nerves, eyes, heart, urogenital tract, pancreas, liver, and intestine (Fig. 2A). Beginning from stage E12.5 and later, strong lacZ expression was found in bones, cartilage, and intervertebral discs (Fig. 2, B and C). Postnatally, prominent lacZ expression was detected in tissues known to abundantly express VDR, such as duodenum, kidney, and parathyroid glands, confirming the specificity of the reporter gene (Fig. 2D and data not shown). There was a clear gene dosage effect, i.e. homozygous mutants with two alleles of the reporter gene showed higher ß-galactosidase activity than heterozygous mutants with only one gene-targeted allele (Fig. 2D).
|
and VDR/ pups were normal and indistinguishable from wild-type pups. In accordance with our findings, VDR null mutant mice totally lacking the VDR protein are normal at birth (4, 5). Therefore, the absence of a functional VDR during development can be compensated by other as yet unknown factors with redundant functions.
The Phenotype of VDR/ Mice Is a Phenocopy of VDR Null Mutants
Heterozygous and homozygous VDR mutant mice were normal until weaning. At 3 wk of age, the body weight of VDR/ mice was still comparable to that of wild-type animals (Fig. 3A). Thereafter, male and female VDR/ mice showed reduced body weight throughout life. The life span of VDR/ mice was greater than 1 yr in most animals. Heterozygous mice did not have an overt phenotype, indicating that the mutant VDR does not act as a dominant negative receptor. Blood ionized calcium (Fig. 3B) was slightly decreased, and serum intact PTH levels (Fig. 3C) were already increased in 3-wk-old VDR/ animals relative to those in wild-type controls. During the rapid growth phase in young mice, severe secondary hyperparathyroidism developed in all homozygous animals with profound hypocalcemia (Fig. 3B), and PTH serum levels greater than 1000 pg/ml were found in all animals (Fig. 3C). Serum 1,25-(OH)2D3 levels were very high in homozygous mutants (mean ± SD, 54 ± 17, 55 ± 23, and 898 ± 593 pg/ml in 10-wk-old wild-type, heterozygous, and VDR/ mice, respectively; P < 0.001, wt/wt vs. VDR/). In agreement with these findings, bones from 10-wk-old VDR/ mice showed histological signs of rickets (Fig. 3D) and severe secondary hyperparathyroidism (Fig. 3E). Surprisingly, PTH serum levels declined with age in VDR/ animals (Fig. 3C), and skeletally mature VDR/ mice did not exhibit overt secondary hyperparathyroidism at the bone level in most cases (Fig. 3F). Interestingly, PTH serum levels tended to rise with age in heterozygous animals (Fig. 3C), but not in wild-type controls, suggesting that the absence of one functioning allele of the VDR throughout life may favor the development of hyperparathyroidism in aged mice. Starting at about 6–8 wk of age, homozygous animals began to develop progressive alopecia. By about 4 months of age, most animals had completely or almost completely lost their hair (Fig. 3G). Sex did not influence the development of alopecia in VDR/ mice.
|
). In good agreement with these findings, serum levels of testosterone and estradiol were normal in VDR/ mice (Fig. 4B). Histological analysis of ovaries and testicles from VDR/ mice revealed active ovaries with follicles at all stages of maturation and corpora lutea in females (Fig. 4C), and normal architecture of seminiferous tubules with active spermatogenesis in males (Fig. 4D).
|
/ mice is a close phenocopy of mutant mice with a total ablation of the VDR protein (4, 5, 6), indicating that the absence of the DNA binding domain abolishes the major physiological functions of the VDR. The reason why the VDR null mutant mice created by Yoshizawa and co-workers (4, 27) show infertility, uterine hypoplasia, and impaired folliculogenesis in females and impaired spermatogenesis in males is unclear, but may be related to a different genetic background. Nevertheless, our findings of morphologically normal gonads together with unchanged blood levels of sex steroids and normal weight of typical target organs for male and female sex hormones strongly support the conclusion that genomic signaling through the VDR is not essential for reproductive functions. This idea is corroborated by the fact that the VDR knockout mice generated by Li et al. (5) are fertile, and that gonadal insufficiencies are absent in human vitamin D-dependent rickets type II (3). Moreover, the VDR knockout mice generated by Yoshizawa and co-workers show normal fertility when maintained on a high calcium diet (28).
Rescue Diet Normalizes Calcium Homeostasis, But Reveals Renal Calcium Leak in VDR/ Mice
It has been shown that a diet enriched with calcium, phosphorus, and lactose (so-called rescue diet) is able to correct calcium homeostasis in VDR knockout mice (29). To explore the possible molecular effects of disruption of the VDR signaling pathway on renal calcium absorption we examined urinary excretion of calcium in 10-wk-old mice on either a normal diet or the rescue diet. In agreement with the study reported by Li et al. (29), we found that the rescue diet totally normalized blood ionized calcium (Fig. 5A), total serum calcium (data not shown), and serum PTH (Fig. 5B) in our VDR/ mice. Interestingly, however, the urinary calcium/creatinine ratio was about 5-fold higher in normocalcemic VDR/ mice on the rescue diet compared with wild-type animals on either the normal or the rescue diet (Fig. 5C). As shown in Fig. 5D, the 24-h urinary excretion of calcium in VDR/ mice on the rescue diet was about 4-fold higher relative to that in wild-type controls on the same diet, but did not differ between wild-type and VDR/ mice on the normal diet. Very similar results were obtained when calcium clearance was calculated (data not shown). In contrast, urinary phosphate excretion was not different between wild-type and homozygous mice on either the normal or the rescue diet (data not shown). These findings clearly indicate that renal tubular calcium reabsorption was impaired in normocalcemic mice with a nonfunctioning VDR.
|
/ mice receiving the normal diet (Fig. 5, E and F). The rescue diet normalized calbindin D28 k mRNA levels, but had no influence on calbindin D9 k in VDR/ mice.
Genomic and Nongenomic Responses to Vitamin D Hormones Are Abrogated in VDR/ Mice
To test whether our mutant mice expressing a VDR with an intact hormone binding domain, but lacking the first zinc finger, would still respond to pharmacological doses of vitamin D analogs, we treated wild-type, heterozygous, and homozygous mice with different vitamin D metabolites. Short-term administration of high doses of active vitamin D metabolites is known to stimulate intestinal calcium absorption and bone resorption and to suppress PTH secretion by the parathyroid glands (30). These effects develop within hours or days and are known to be mainly mediated through altered gene transcription (31). In contrast, it has been shown that vitamin D metabolites stimulate intestinal calcium absorption in perfused chick and rat duodenal loops within minutes (19, 32), and this nongenomic mechanism is thought to be mediated by a putative vitamin D membrane receptor (21). Another distinct membrane receptor for the vitamin D analog 24,25-dihydroxyvitamin D3 [24,25-(OH)2D3] has been functionally described in osteosarcoma cells (33) and chondrocyte matrix vesicles (34). We reasoned that these putative membrane receptors should still be present in our mutant mice if they represent novel gene products different from the classical VDR.
To test the in vivo response of our mutant mice to pharmacological doses of vitamin D metabolites, mice were adapted to metabolic cages and treated for 3 d with 1 µg/kg 1,25-(OH)2D3 or 5000 µg/kg 24,25-(OH)2D3. Monitoring 24-h urinary calcium excretion is a sensitive way to follow changes in bone resorption and intestinal calcium absorption under these experimental conditions (30). We found that 1,25-(OH)2D3 induced profound hypercalcemia and hypercalciuria in wild-type and heterozygous mice (Fig. 6, A and B). The 5000-fold higher dose of 24,25-(OH)2D3 induced an almost identical calcemic and calciuric response compared with 1 µg/kg 1,25-(OH)2D3 (Fig. 6C and data not shown). However, homozygous animals did not show any increase in blood or urinary calcium in these experiments (Fig. 6, A–C). Similarly, 1,25-(OH)2D3 suppressed serum PTH to undetectable levels in wild-type and heterozygous mice, but not in homozygous animals (Fig. 6D). After a 3-d treatment with high dose 1,25-(OH)2D3, cancellous bone in wild-type and heterozygous mice showed an accumulation of eroded bone surface indicative of increased osteoclastic bone resorption (Fig. 6E), and highly active tartrate-resistant acid phosphatase-positive osteoclasts (Fig. 6F). Again, these changes were absent in bones from VDR/ mice (Fig. 6, E and F). Thus, high doses of different vitamin D metabolites did not stimulate intestinal calcium absorption or bone resorption, did not suppress PTH serum levels, and did not alter renal calcium handling in homozygous mutant mice with nonfunctioning nuclear VDR.
|
). Within 3 d, 1,25-(OH)2D3 and KH1060 (Fig. 6H) induced a profound increase in epidermal thickness in wild-type mice, but the response to both vitamin D analogs was absent in homozygous mice [wt/wt vs. VDR/, mean ± SD, 58.6 ± 17.4 vs. 17.7 ± 7.2 µm (P < 0.05) for 1,25-(OH)2D3; 89.8 ± 11.0 vs. 17.7 ± 2.3 µm (P < 0.05) for KH1060].
Taken together, these data clearly indicated that skin and the organs involved in calcium homeostasis, i.e. intestine, bone, parathyroid glands, and kidney, of homozygous mice expressing the mutant VDR did not respond to high doses of different vitamin D metabolites in vivo. To test whether rapid, nongenomic actions of vitamin D at the cellular level would still be present in our mutant mice, we examined the increase in intracellular calcium concentrations elicited by 1,25-(OH)2D3 in primary, confluent cultures of fura-2-loaded osteoblasts. This system shows an abrupt increase in intracellular calcium concentrations within seconds after the addition of 1,25-(OH)2D3 in a consistent and reproducible fashion (reviewed in Refs. 20 and 36). Interestingly, the rapid response to 1,25-(OH)2D3 was totally absent in osteoblasts from VDR/ mice (Fig. 6I). Also, there appeared to be a gene dosage effect, because osteoblasts from heterozygous animals exhibited a dose-response curve that was shifted to the right by about 1 order of magnitude of 1,25-(OH)2D3 concentrations and a slightly decreased maximal response (Fig. 6I). In contrast, the increase in intracellular calcium in response to 10 nM bovine PTH-(1–84) was identical in osteoblasts of all three genotypes (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In addition to the typical phenotype of the homozygous mice, treatment of VDR/ mice with pharmacological doses of vitamin D metabolites confirmed that the mutant receptor is functionally inactive in vivo. In accordance with these findings, several known point mutations in the DNA binding domain of the human VDR result in hereditary vitamin D-dependent rickets type II (38). Both zinc finger motifs are essential for selective VDR binding to vitamin D response elements, which usually consist of direct hexanucleotide repeats with a three-nucleotide spacing (39, 40). As expected, our mutant VDR lacking the first zinc finger cannot bind to a mouse osteopontin VDRE. In analogy to our results, a naturally occurring mutation in the androgen receptor causing an in-frame deletion of the second zinc finger also results in a functionally inactive protein with nuclear localization (41). Despite normal androgen binding affinity of the mutant receptor, the patients suffer from complete androgen insensitivity. Therefore, the loss of one of the zinc finger motifs of the DNA binding domain of the VDR and of the androgen receptor results in end-organ resistance to the actions of the steroid hormones, but does not interfere with nuclear localization of the receptors.
In accordance with previous reports (6, 11), our study has demonstrated that feeding of the rescue diet enriched with calcium, phosphorus, and lactose to VDR/ mice fully corrected blood calcium levels and serum PTH. However, renal tubular calcium reabsorption was distinctly impaired in normocalcemic VDR/ mice. Despite similar ionized, and thus dialyzable, calcium levels in the blood, the daily amount of calcium excreted via the urine was about 4-fold higher in VDR/ compared with wild-type mice on the rescue diet. Therefore, our findings confirm earlier reports that renal tubular calcium reabsorption is impaired in vitamin D deficiency (42) and clearly establish that 1,25-(OH)2D3 indeed has an important physiological role in calcium reabsorption in the kidney in vivo.
In VDR mutants given the normal diet, this renal calcium leak is obviously masked by reduced ionized blood calcium, and therefore reduced filtered calcium load, together with very high levels of circulating PTH stimulating renal tubular calcium reabsorption. Despite the fact that urinary loss of calcium was much more pronounced in VDR/ mice fed the rescue diet relative to those receiving the normal diet, the mRNA levels of renal calbindin D9 k remained at a very low level, and those of calbindin D28 k were normalized by the rescue diet in VDR/ mice. Therefore, the changes in mRNA levels of calbindins do not readily explain the molecular defect in renal tubular calcium reabsorption in normocalcemic mice with a nonfunctioning VDR. Although the newly characterized ECaC have been implicated to play an important role in 1,25-(OH)2D3-stimulated renal tubular calcium reabsorption (12, 13, 43), a recent study conducted in our laboratory using quantitative RT-PCR has shown that despite the presence of a putative VDRE in the promoter region of this gene, the mRNA level of the kidney-specific ECaC1 was down-regulated in the kidneys of homozygous VDR mutant mice given the normal diet, but returned to normal in VDR/ animals fed the rescue diet (17). The mRNA level of the ubiquitously expressed ECaC2 remained unchanged in the kidneys of our VDR/ mice fed either the normal or the rescue diet (17). Recently, it has been reported by Van Cromphaut et al. (6) that renal mRNA levels of ECaC1 and -2 were unchanged in VDR-ablated mice, and that changes in dietary calcium did not influence ECaC expression in VDR null mutants. Taken together, our data and the results provided by the latter study demonstrate that changes in the expression of the apical epithelial calcium channels ECaC1 and -2, at least at the mRNA level, do not appear to be causally linked to the defect in renal calcium reabsorption in normocalcemic mice with nonfunctioning VDR. Based on our present data, the profound down-regulation of the intracellular calcium carrier protein calbindin D9 k in VDR/ mice may provide a possible explanation for the defective renal calcium reabsorption in normocalcemic VDR/ mice. In VDR/ mice fed the normal diet, this defect may be counteracted by secondary hyperparathyroidism increasing renal epithelial calcium transport by a mechanism partially independent of calbindin D9 k. However, although urinary calcium excretion in VDR/ mice given the normal diet did not differ from that in wild-type mice, it appeared to be inappropriately high in light of the approximately 20% lower blood ionized calcium and the more than 100-fold higher PTH serum levels relative to those in wild-type animals. Clearly, further studies are necessary to elucidate the molecular mechanism of this physiologically important renal calcium leak induced by the absence of VDR signaling.
Treatment of our VDR/ mice with pharmacological doses of vitamin D analogs did not elicit any change in calcium homeostasis or keratinocyte proliferation, suggesting that the physiological significance of vitamin D signaling pathways, other than those mediated through the classical, nuclear VDR, is minor, at least in skin and tissues involved in the regulation of calcium homeostasis. Furthermore, deletion of the first zinc finger of the VDR not only abrogated the effects of vitamin D analogs in vivo, but also abolished the rapid nongenomic actions of 1,25-(OH)2D3 in osteoblasts. In the osteoblast model of nongenomic effects, the increase in intracellular calcium concentrations in response to 1,25-(OH)2D3 occurs within seconds and involves opening of L-type channels in the plasma membrane (33) as well as activation of phospholipase C-ß1 linked to G
q/11 and subsequent inositol triphosphate-mediated Ca2+ mobilization from the endoplasmic reticulum (26, 44). Our findings support the conclusion that the classical VDR mediates these rapid actions of 1,25-(OH)2D3 in osteoblasts. Of course, we cannot totally rule out that the lack of nongenomic responses to vitamin D metabolites was caused by a down-regulation of the putative vitamin D membrane receptor in the absence of a functioning classical VDR. However, the total absence of the 1,25-(OH)2D3- induced, rapid effects in osteoblasts of VDR/ mice makes this scenario unlikely. Treatment of osteoblasts from VDR/ mice with PTH-(1–84) induced an identical rise in intracellular calcium compared with that in osteoblasts from wild-type animals, indicating that osteoblasts from VDR/ mutants had the capacity to show a rapid increase in intracellular calcium in response to other hormonal stimuli.
In agreement with our results, it has been shown that antibodies to the classical estrogen receptor modulate rapid PRL release from pituitary tumor cells (45). Furthermore, the rapid, nonnuclear effects of estrogen on endothelial nitric oxide synthesis were reported to be mediated through the binding of classical estrogen receptor in a ligand-dependent manner to the p85 regulatory subunit of phosphatidylinositol-3-OH kinase (46). In addition, a recent study has indicated that the progesterone receptor directly interacts with SH3 domains of cytoplasmic signaling molecules in a hormone-dependent fashion to elicit rapid nongenomic activation of c-Src tyrosine kinases and MAPK (47). Thus, there is accumulating evidence that the nongenomic effects of steroid hormones are mediated through the classical receptors or subpopulations thereof. This idea is supported by the fact that cloning and characterization of any novel bona fide membrane steroid receptors have never been accomplished.
Our results favor a model in which the classical VDR interacts with as yet unknown proteins to elicit rapid, nongenomic effects in osteoblasts. Contrary to this idea, it has been shown that the ligand specificity for membrane effects of vitamin D analogs in osteoblasts are distinct from those for gene trans-activation- mediated effects (20, 36). However, it may be possible that specific protein cofactors alter the ligand specificity of the classical VDR in this nongenomic signaling pathway. At present, we do not know whether the absence of the first zinc finger, the lack of the phosphorylation site at Ser51, or conformational changes in the mutant VDR interfere with binding of the mutant receptor to VDR-interacting proteins or intracellular trafficking of the protein. More extensive experimentation is required to define the cytoplasmic or membrane proteins possibly interacting with the classical VDR to mediate the nongenomic effects of vitamin D metabolites.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-DASHII phage library (a gift from J. Rossant, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Canada). The four overlapping clones covered 30 kb of VDR genomic DNA. We constructed a targeting vector containing 8.7 kb of homology to the region of the VDR locus, a lacZ cassette as a reporter gene, a neo phosphotransferase (neo) cassette for positive selection, and a thymidine kinase cassette for negative selection. The lacZ and neo cassettes replaced exon 2 and part of intron 2. The lacZ cassette was introduced into the vector directly downstream of the start codon of exon 2, so that lacZ gene transcription used the endogenous start codon and was driven by the endogenous VDR promoter. The neo cassette was introduced into intron 2 in the opposite translation direction. Most of intron 2 remained intact. The SalI-linearized vector was electroporated into R1 and E14 mouse embryonic stem (ES) cells, and the ES clones were selected in the presence of the neomycin analog G418 and gancyclovir as previously described (48). DNA prepared from resistant clones was digested with HindIII and ClaI for analysis of the 5' region and with HindIII for analysis of the 3' region and was screened by Southern blot using 0.8-kb SalI-XhoI and 1.3-kb SmaI-EcoRI fragments as 5' and 3' probes, respectively. Eight correctly targeted clones from R1 ES cells and one positive clone from E14 ES cells were introduced into CD1 morulae by aggregation and implanted into pseudopregnant CD1 females. Chimeric male mice were obtained from four ES clones and were mated with C57BL/6 females. Subsequently, heterozygous F1 mice were backcrossed to the C57BL/6 genetic background for four or five generations. Homozygous mutant mice were produced by intercrossing heterozygous animals. All experiments were performed on wild-type, heterozygous, and homozygous offspring of heterozygous matings.
Animal Maintenance
All mice were kept at 24 C with a 12-h light, 12-h dark cycle and were allowed free access to diet and tap water. The normal diet (Altromin, Lage, Germany) contained 0.9% calcium, 0.7% phosphorous, 0% lactose, and 600 IU vitamin D/kg. The rescue diet (Altromin), containing 2% calcium, 1.25% phosphorous, 20% lactose, and 600 IU vitamin D/kg, was fed starting from 16 d of age. To examine alterations in renal excretion of minerals, 10-wk-old wild-type, heterozygous, and homozygous mice fed either the normal or the rescue diet (n = 12–13 each) were kept in metabolic cages for a 15-h period overnight for urine collection. During that period the mice had free access to tap water, but were deprived of food. For measurement of ionized blood calcium, 100 µl blood were drawn from the retroorbital sinus into heparinized capillaries under ether anesthesia. Immediately thereafter, blood was drawn from the abdominal vena cava under anesthesia with ketamine/xylazine for serum collection. All animal procedures were approved by the local ethical committee and the government authorities.
Immunohistochemistry, lacZ Staining, and Histology
Immunohistochemical detection of VDR expression in duodenum was performed on methacarn-fixed, paraffin-embedded tissues. Sections were deparaffinized, microwaved for 10 min in 10 mM citrate-buffer, pH 6.0, and, after blocking with 20% rabbit serum, incubated overnight at 4 C with a monoclonal rat anti-VDR antibody (clone 9A7, Affinity BioReagents, Inc., Golden, CO) diluted 1:200. Bound antibody was detected with biotinylated rabbit antirat IgG and peroxidase-conjugated avidin-biotin complex (Vector Laboratories, Inc., Burlingame, CA), using Vector VIP as enzyme substrate. To visualize lacZ reporter gene expression, X-gal staining of mouse embryos at pH 8.0 and subsequent clearing with benzyl benzoate/benzyl alcohol were carried out as previously described (49). Bone specimens were processed as previously described (50).
Northern and Western Blot Analyses, and Competitive Ligand Binding
Total RNA was extracted from duodenum and kidney using the RNeasy Midi Kit (QIAGEN, Valencia, CA), and was separated on formaldehyde-containing agarose gels. Nuclear extracts from duodenal mucosa were prepared as previously described (51) and were assessed by Western blot analysis (52) using a polyclonal antibody directed against the C- terminal part of the human VDR (H-81, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at 1 µg/ml for 16 h at 4 C. The peroxidase-labeled secondary antibody was used at a dilution of 1:1000. For competition studies, nuclear extracts were isolated from the duodenal mucosa of 10-wk-old wild-type and homozygous mice (51). Subsequently, competitive binding experiments with the duodenal nuclear extracts and recombinant human VDR (Oxford Biomedical Research, Oxford, MI) were performed according to the protocol described by Koszewski et al. (53) using 3H-labeled 1,25-(OH)2D3 (Amersham Pharmacia Biotech, Freiburg, Germany) at a fixed concentration of 0.12 nM and concentrations of cold 1,25-(OH)2D3 ranging from 0.001–1000 nM. Binding curves were analyzed using PRISM 3.02 software (GraphPad Software, Inc., San Diego, CA).
RT-PCR
For tissue distribution studies of the wild-type and mutant VDR with RT-PCR, sex-matched littermates were killed, and kidney, duodenum, skin, ovary, testis, and larynx including thyroid and parathyroid glands were isolated. Primary cultures of osteoblasts from calvariae of newborn mice were prepared as previously described (54). Total RNA was extracted from tissues using the RNeasy Mini Kit (QIAGEN) and with TRIzol (Life Technologies, Inc., Grand Island, NY) from cultured cells. Two micrograms of total RNA were reverse transcribed using the RevertAid First Strand cDNA Synthesis Kit (MBI Fermentas, St. Leon-Rot, Germany). PCR involved annealing at 58 C for VDR exon analysis or at 59 C for genotyping for 30 cycles. The following primers were used: VDR exon 1 forward, 5'-CAGAGTTCTTTTGGTTGGACAG-3'; VDR exon 2 forward, 5'-CCACGGGCTTCCACTTCAA-3'; VDR exon 3 forward, 5'-CAAGGACAACCGGCGACACTG-3'; VDR exon 7 reverse, 5'-GACTTAAGCAGGACAATCTGGTG-3'; VDR wild-type and mutant allele (genotyping) forward, 5'-GCCTGCTCTTCTTACAGGGATG-3'; VDR wild-type reverse, 5'-GGACTCACCTGAAGAAACCCTTGC-3'; and VDR mutant allele reverse, 5'-GGCCTCAGGAAGATCGCACTCC-3'.
Construction of Bacterial Expression Vectors and Protein Purification of Wild-Type and Mutant VDR
DNA fragments coding for the full-length VDR (422 amino acids) and the N-terminal-truncated form of murine VDR (lacking the first 51 amino acids, VDR ) were amplified from mouse duodenal cDNA by PCR using primers introducing BamHI and KpnI restriction sites (underlined). For VDR, the forward primer 5'-TTTGGATCCATGGAGGCAATG CAGCCAGC-3' and the reverse primer 5'-TTTGGTACCTCAGGAGATCTCATTGCCAAACACCTC-3' were used. For VDR , we used the forward primer 5'-TTTGGATCCATGAAGCGCAAGGCCCTGTTC-3' and the same reverse primer as shown above. The PCR products were directionally cloned into modified pGex 2T vector (52) for the bacterial expression of GST fusion proteins. The recombinant plasmids were transformed into Escherichia coli BL21 DE3 codon plus RIL (Stratagene, Amsterdam, The Netherlands). After harvesting, the bacteria were lysed by four freeze-thaw cycles in the presence of lysozyme. GST fusion proteins in the supernatant of the lysate were bound to GT-Sepharose (Amersham Pharmacia Biotech) and eluted by Tris-HCl/glutathione after several washing steps as previously described (52). The eluates were directly used for the gel mobility shift assays.
Gel Mobility Shift Assay
Ten picomoles of double-stranded mouse osteopontin response element (5'-ACAAGGTTCACGAGGTTCACGTCT-3') were end-labeled with 20 pmol [
-33P]ATP by T4 kinase (MBI Fermentas). Ten femtomoles of labeled osteopontin response element were incubated with 100 ng VDR proteins or without protein in binding buffer [20 mM Tris HCl (pH 7.9), 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 0.05% Nonidet P-40, and 10% glycerol] with 50 µg/ml poly(deoxyinosine-deoxycytidine) for 30 min on ice. Samples were then resolved by electrophoresis on a 7% nondenaturing acrylamide gel in 0.5x Tris-boric acid-EDTA buffer at 26 V/cm at 4 C. The gels were dried, exposed 1 h to a phosphor screen, and scanned using a phosphor imager XLA-3000 (Fuji Photo Film Co., Ltd., Tokyo, Japan).
Biological Chemistry
Blood ionized calcium was measured with an AVL 9140 electrolyte analyzer (Roche, Mannheim, Germany). Serum and urinary calcium levels were measured with flame photometry. Phosphorus and creatinine concentrations were determined using a Hitachi 766 autoanalyzer (Roche Molecular Biochemicals, Mannheim, Germany). PTH concentrations were assessed using a two-sided ELISA specific for intact mouse and rat PTH (Immutopics, San Clemente, CA). Total serum testosterone (Beckman Coulter, Inc., Krefeld, Germany), and estradiol (Diagostics Systems Laboratories, Inc., Sinsheim, Germany) were determined by RIA after serum extraction with diethyl ether. Serum concentrations of 1,25-(OH)2D3 were measured using a radioreceptor assay (Immundiagnostik, Bensheim, Germany).
Systemic and Topical Treatment of Skin with Vitamin D Metabolites
For systemic treatment with vitamin D metabolites, 3- to 4-month-old mice (n = 6/genotype and treatment) were adapted to metabolic cages and were sc treated with vehicle (5% ethanol in 0.15 M NaCl), 1 µg/kg 1,25-(OH)2D3, or 5000 µg/kg 24,25-(OH)2D3 (Solvay Pharmaceuticals, Inc., Weesp, The Netherlands) for 3 consecutive days. Blood was drawn at baseline and at the end of the experiment, i.e. 24 h after the last treatment. The femurs were harvested for histological analysis at the end of the trial. Urine was collected in 24-h samples during the experiment. For topical treatment with vitamin D metabolites, the back skin of 3-month-old mice (n = 4/genotype and treatment) was shaved and was topically treated with placebo, 1,25-(OH)2D3 at 100 pmol/cm2, or the vitamin D analog KH1060 at 10 pmol/cm2 once a day for 3 consecutive days (35). After 72 h, skin was harvested, and epidermal thickness was measured in quintuplicate in interfollicular areas on hematoxylin- and eosin-stained microscopic sections as previously described (55).
Rapid, Nongenomic Effects of 1,25-(OH)2D3 in Osteoblasts
Osteoblasts were isolated from parietal bones of newborn mice by sequential enzymatic digestion (54) and were seeded and grown on coverslips for 4 d in phenol red-free MEM supplemented with 10% heat-inactivated fetal calf serum. Subsequently, cells were incubated for 72 h in phenol red-free medium containing 1% heat-inactivated fetal calf serum and transferred to serum-free medium 24 h before use. After washing, cells were loaded with 1 µM fura-2/AM for 30 min in Hank’s HEPES buffer at pH 7.4, and the increases in intracellular calcium concentration in response to different concentrations of 1,25-(OH)2D3 were measured with a spectrofluorometer, as previously described (44), 10 sec after the addition of the drug. Each measurement on fura-2-loaded cells was followed by a parallel experiment under the same conditions with non-fura-2-loaded cells.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Present address: MBT Munich Biotechnology GmbH, Fraunhoferstrasse 10, 82152 Martinsried, Germany. 
2 Present address: GBF Gesellschaft für Biotechnologische Forschung, Mascheroder Weg 1, 38124 Braunschweig, Germany.
Abbreviations: /, Homozygous; 1,25-(OH)2D3, 1,25- dihydroxyvitamin D3; 24,25-(OH)2D3, 24,25-dihydroxyvitamin D3; ECaC, epithelial calcium channel; ES, embryonic stem; GST, glutathione-S-transferase; KH1060, 20-epi-22- oxa-24a,26a,27a-trihomo-1,25-(OH)2D3; VDR, vitamin D receptor; VDRE, vitamin D-responsive element; wt/, heterozygous; wt/wt, wild-type.
Received for publication November 15, 2001. Accepted for publication March 18, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
,25-dihydroxyvitamin D3. Trends Endocrinol Metab 9:419–427[Medline]
,25(OH)2D3: past, present, and future. J Bone Miner Res 13:1360–1369[CrossRef][Medline]
,25-(OH)2 vitamin D3 and testosterone potentiation of calcium channels in osteosarcoma cells. J Biol Chem 271:33335–33343,25-(OH)2D3 and 24R,25-(OH)2D3. J Cell Physiol 167:380–393[CrossRef][Medline]
,25-dihydroxyvitamin D3 and receptor-active 20-epi analogues of 1,25-dihydroxyvitamin D3. Biochem Pharmacol 49:621–624[CrossRef][Medline]
