The authors state that they have no conflicts of interest.
Hereditary vitamin D-resistant rickets (HVDRR) is a rare recessive genetic disorder caused by mutations in the vitamin D receptor (VDR). In this study, we examined the VDR in a young girl with clinical features of HVDRR including rickets, hypophosphatemia, and elevated serum 1,25(OH)2D. The girl also had total alopecia. Two mutations were found in the VDR gene: a nonsense mutation (R30X) in the DNA-binding domain and a unique 3-bp in-frame deletion in exon 6 that deleted the codon for lysine at amino acid 246 (ΔK246). The child and her mother were both heterozygous for the 3-bp deletion, whereas the child and her father were both heterozygous for the R30X mutation. Fibroblasts from the patient were unresponsive to 1,25(OH)2D3 as shown by their failure to induce CYP24A1 gene expression, a marker of 1,25(OH)2D3 responsiveness. [3H]1,25(OH)2D3 binding and immunoblot analysis showed that the patient's cells expressed the VDRΔK246 mutant protein; however, the amount of VDRΔK246 mutant protein was significantly reduced compared with wildtype controls. In transactivation assays, the recreated VDRΔK246 mutant was unresponsive to 1,25(OH)2D3. The ΔK246 mutation abolished heterodimerization of the mutant VDR with RXRα and binding to the coactivators DRIP205 and SRC-1. However, the ΔK246 mutation did not affect the interaction of the mutant VDR with the corepressor Hairless (HR). In summary, we describe a patient with compound heterozygous mutations in the VDR that results in HVDRR with alopecia. The R30X mutation truncates the VDR, whereas the ΔK246 mutation prevents heterodimerization with RXR and disrupts coactivator interactions.
1,25-dihydroxyvitamin D3 [1,25(OH)2D3] is an important regulator of bone and mineral homeostasis.(1) The actions of 1,25(OH)2D3 are mediated by the vitamin D receptor (VDR), a member of the steroid-retinoid-thyroid superfamily of nuclear transcription factors.(2) Patients homozygous for mutations in the VDR develop hereditary vitamin D-resistant rickets (HVDRR), a rare recessive genetic disease, whereas the heterozygotic parents are usually normal.(3) HVDRR patients have low serum calcium levels that leads to rickets and elevated levels of serum 1,25(OH)2D3 that show resistance to the action of the active hormone.
Some HVDRR patients also develop alopecia and skin lesions that are phenotypically similar to the disease atrichia with papular lesions (APL) caused by mutations in the hairless gene (hr).(4,5) The hr gene product HR has been shown to act as a corepressor of the VDR, thyroid hormone receptor (TR), and retinoid-related orphan receptor (ROR).(6–8) HR, RXR, and VDR are thought converge to regulate similar pathways in the hair cycle.(8)
In this report, we describe a patient with compound heterozygous mutations in the VDR gene, a different mutation inherited from each parent, which results in the syndrome of HVDRR with alopecia.
MATERIALS AND METHODS
Informed consent and cultured fibroblasts
Informed consent was obtained from parents under a Stanford University Institutional Review Board-approved protocol. Dermal fibroblasts were cultured from a forearm skin biopsy of the patient and her parents as previously described.(9) Clinical measurements were performed in the hospital clinical laboratory by routine procedures. Measurement of vitamin D metabolites did not distinguish between vitamin D2 and D3. Normal ranges have not been formally established for Czech children, and the values shown in Table 1 reflect normal ranges provided by the suppliers of the kits in routine use in the hospital.
Table Table 1.. Patient's Clinical Course
DNA sequencing, cDNA cloning, and restriction fragment length polymorphism analyses
Exons 2–9 of the VDR gene were amplified from DNA samples by PCR and directly sequenced at the Stanford University protein and nucleic acid core facility. RNA was isolated from cultured fibroblasts using RNAeasy spin columns (Qiagen, Valencia, CA, USA). cDNA was prepared by reverse transcription using superscript III cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) and cloned into TOPO 2.1 vector using TOPO cloning kit (Invitrogen). Individual clones were sequenced. For restriction fragment length polymorphism analysis, exon 2 of the VDR gene was amplified from DNA samples by PCR. The PCR products were digested with DdeI and separated on 2% agarose gels.
[3H]1,25(OH)2D3 binding and Western blotting
Cell extracts from cultured skin fibroblasts from the patient were prepared as previously described.(10) [3H]1,25(OH)2D3 binding assays were performed as previously described.(11) Western blotting was performed as previously described.(12) Protein concentrations were determined by the Bradford method.(13)
CYP24A1 (24-hydroxylase) gene induction
Fibroblasts were treated with 1,25(OH)2D3 for 6 h in medium containing 1% calf serum. CYP24A1 and TBP genes were semiquantified using real-time PCR.(12) Experiments were performed at least twice with duplicate determinations.
Site-directed mutagenesis of the WT VDR cDNA in pSG5 (Stratagene, La Jolla, CA, USA) was performed using the QuickChange II XL site-directed mutagenesis kit (Stratagene).(12)
Transactivation assays were performed as previously described.(10) Transfections were performed in triplicate, and each experiment was repeated at least three times.
Gel mobility shift assay
DNA binding was assessed by gel mobility shift assay (GMSA) as previously described.(10) A consensus VDRE double-stranded oligonucleotide (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was end-labeled using [γ32P]-ATP and polynucleotide kinase (New England Biolabs, Ipswich, MA, USA). Cell extracts from COS-7 cells transfected with WT VDR or the VDRΔK246 mutant expression plasmids were incubated with vehicle (0.1% ethanol) or 500 nM 1,25(OH)2D3 in mammalian protein extraction reagent (M-PER) buffer containing 0.25 μg/ml poly dIdC and 10% glycerol for 20 min. The [32P]labeled VDRE probe (4.8 pmol, ∼50,000 dpm) was added for an additional 20 min. All incubations were performed at ambient temperature. The final concentration of salt in the binding assay was 125 mM KCl. The samples were electrophoresed on 5% polyacrylamide gels in 0.5× Tris-borate buffer at 180 V for 90 min at ambient temperature. The gel was dried and subjected to autoradiography at –80°C.
GST pull-down assays
VDR proteins were synthesized using the TNT Quick-coupled in vitro transcription/translation system (Promega, Madison, WI, USA). GST-RXRα, GST-SRC-1, and GST-DRIP205 were incubated with in vitro synthesized WT and mutant VDR proteins. Bound proteins were subjected to Western blotting and visualized using ECL-plus (GE Healthcare, Piscataway, NJ, USA) as previously described.(12)
For co-immunoprecipitation (co-IP), COS-7 cells were plated in T75 tissue culture plates and transfected with 1 μg/plate each of HR and VDR cDNA expression vectors using 25 μl Polyfect (Qiagen). After 24 h, the medium was replenished and incubation continued for an additional 24 h. Cells were washed with 10 ml PBS and collected by scraping in 5 ml PBS and centrifugation at 2000 rpm for 5 min at 4°C. The cell pellet was resuspended in 0.5 ml RIPA buffer containing a protease inhibitor tablet (1 mini tablet/10 ml RIPA buffer) and incubated at 4°C for 10 min on rotating mixer. Lysates were prepared by centrifugation in a microcentrifuge at maximum speed for 10 min at 4°C. Lysates were precleared by adding 20 μl Protein A/G plus agarose (25% slurry) and incubation at 4°C for 30 min. After centrifugation at 2500 rpm for 30 s at 4°C to remove the agarose, 1 μl of anti-VDR C-20 (2 μg/μl; Santa Cruz Biotechnology) was added and incubated at 4°C for 2 h. Next, 20 μl of Protein A/G plus agarose was added, and incubation continued at 4°C overnight on a rotating mixer. Samples were transferred to spin columns (Harvard Apparatas, Holliston, MA, USA) placed in 2-ml collection tubes. The tubes were centrifuged at 2500 rpm for 30 s at 4°C. The samples were washed three times with 0.5 ml RIPA buffer. Bound proteins were eluted by addition of 50 μl of 2× lithium dodecyl sulfate sample buffer (Invitrogen) containing 10 mM DTT and centrifuged for 1 min at 2500 rpm. Samples were run on 10% NuPAGE gels in MOPS-SDS running buffer (Invitrogen) and transferred to nitrocellulose. Flag-tagged HR was detected with anti-Flag antibodies (Sigma, St Louis, MO, USA) and VDR detected using anti-VDR D6 antibody (Santa Cruz Biotechnology).
The patient is a young girl born in 2002 in the Czech Republic. She initially presented with early-onset severe rickets and secondary hyperparathyroidism at the age of 2 yr. The patient had total alopecia and bowed legs (Fig. 1). The mother, born in Indonesia, was a healthy 29-yr-old woman with no family history of bone disease or genetic disease. The father, from the Czech Republic, was 42 yr old and healthy and also had no family history of bone disease or genetic disease. The family denies consanguinity. The child was born after a normal pregnancy (birth weight, 3500 g; length, 50 cm), without complications during the perinatal period except for total alopecia. She received prophylactic vitamin D supplementation with vitamin D from age 14 days to 1 yr (except for a 4-mo gap) and then with fish oil until age 2. The child had no unusual health problems. The parents sought medical attention for bowing of the child's legs at 2 yr of age. Physical examination at that time showed partial growth retardation (body weight, 10.5 kg; length, 86 cm), swelling of the wrists and knees, bowed legs, and alopecia (Fig. 1). Radiographs of wrists and legs showed classical changes of rickets. Laboratory data are shown in Table 1 and reveal slightly reduced serum calcium levels, hypophosphatemia, and elevated PTH and alkaline phosphatase. Her total protein (72.3 g/liter; normal range, 60–80 g/liter) and albumin (49.8 g/liter; normal range, 38–54 g/liter) were normal. 25-hydroxyvitamin D was low, and 1,25(OH)2D was elevated to 196 pg/ml or three times the upper limit of the normal range. The child was treated with 3000 mg of calcium carbonate orally per day and 3 μg/kg of calcitriol. The family refused intravenous therapy because of frequent travels between the Czech Republic and Indonesia. Over time, the waddling gait and features of rickets clinically and on X-ray partially improved (Fig. 1). Her ionized calcium rose to the lower limit of normal. The phosphate, alkaline phosphatase, PTH, and 1,25(OH)2D levels all improved slightly but did not return to normal (Table 1). There was no change in the alopecia. The child had no unusual illnesses. Based on the findings of a child with rickets and alopecia and laboratory data showing elevated 1,25(OH)2D levels and secondary hyperparathyroidism, the diagnosis of HVDRR was considered, and blood specimens from the child and parents and skin biopsy specimen from the child were sent to our laboratory at Stanford for analysis.
Genotyping of parents and child
We first sequenced the patient's VDR gene from DNA isolated from peripheral blood and identified two different heterozygous mutations. A nonsense mutation was found in exon 2 that changed the codon for arginine to a stop codon at amino acid 30 (R30X). The R30X mutation created a new DdeI restriction site in exon 2. The heterozygosity of the R30X mutation in the patient's DNA and her father's DNA sample was confirmed using DdeI restriction endonuclease digestion (Fig. 2A).
A second mutation was found in exon 6 of the VDR gene. Sequence analyses of exon 6 from the patient's genomic DNA sample resulted in a mixed sequence (Fig. 2B). To identify the mutation in exon 6, we cloned and sequenced the patient's VDR cDNA from RNA isolated from her cultured skin fibroblasts. A 3-bp in-frame deletion that resulted in the deletion of lysine 246 (ΔK246) was identified in one cDNA clone (Fig. 2C). A second cDNA clone contained the wildtype sequence of exon 6, showing that the patient was heterozygous for the 3-bp deletion (Fig. 2C). Sequence analyses of the mother's VDR gene showed that she was heterozygous for the 3-bp deletion as evidenced by the mixed sequence in exon 6 (Fig. 2B). The patient and both parents were also genotyped for the FokI polymorphism located in the translation initiation start site. The patient and her father were heterozygous (F/f) for the FokI site, whereas the mother was homozygous (F/F; data not shown). These data indicated that the child inherited the F allele with the 3-bp deletion from the mother and the f allele with the R30X mutation from the father.
Analyses of the ΔK246 allele
Because the ΔK246 mutation was located in the VDR ligand binding domain (LBD) we examined whether the mutation affected the binding affinity and expression of the patient's VDR using [3H]1,25(OH)2D3 ligand-binding assays and immunoblotting. As shown in Fig. 3A, the patient's VDR and both parents' VDRs exhibited saturable [3H]1,25(OH)2D3 binding. Scatchard analyses of the binding data showed that the patient's cells expressed a high-affinity VDR with a Kd of 3 × 10−10 M that was similar to the parent's VDRs (Kd = 3 × 10−10 M for the mother and Kd = 2 × 10−10 M for the father; Fig. 3A). The total amount of VDR expressed by the patient's fibroblast (Nmax = 10 fmol/mg protein), however, was approximately one third of the amount expressed by the parent's fibroblasts (Nmax = 28–33 fmol/mg protein). Immunoblot analyses confirmed that the patient expressed the VDRΔK246 mutant protein but at low levels compared with VDR protein levels expressed by the parent's fibroblasts and normal control fibroblasts (Fig. 3B, compare lane 1 with lanes 2–4). Both the fibroblasts from the mother and father expressed levels of VDR similar to the normal control fibroblasts (Fig. 3B, compare lanes 2 and 3 with lane 4).
Functional analyses of fibroblasts from parents and child
To determine whether the patient's cells exhibited resistance to 1,25(OH)2D3 treatment, we treated the patient's fibroblasts with graded concentrations of 1,25(OH)2D3 for 6 h and examined CYP24A1 induction using real-time RT-PCR. As shown in Fig. 4, the fibroblasts from the patient's mother and father both showed a dose-dependent increase in CYP24A1 gene expression in response to 1,25(OH)2D3 treatment. In contrast, the patient's fibroblasts exhibited no induction of CYP24A1 gene expression when treated with up to 1 μM 1,25(OH)2D3, clearly showing that the patient's cells were resistant to 1,25(OH)2D3.
Functional analyses of recreated ΔK246 mutant VDR
We next recreated the ΔK246 mutation in the VDR cDNA and examined the effects of the mutation on VDR function. To assess whether the mutation affected VDR transactivation, we transfected COS-7 cells with the WT VDR or VDRΔK246 mutant expression vectors and a CYP24A1 promoter reporter construct. As shown in Fig. 5A, in the cells transfected with the WT VDR, 1,25(OH)2D3 treatment stimulated transactivation activity. In contrast, cells transfected with the VDRΔK246 mutant were unresponsive to 1,25(OH)2D3 treatment. Immunoblotting showed that the VDRΔK246 mutant was expressed at levels similar to the WT VDR (Fig. 5B). These results showed that deleting amino acid K246 abolishes VDR transactivation.
Analyses of defect in VDR function caused by the ΔK246 mutation
Because VDR transactivation involves heterodimerization with RXR and DNA binding, we examined the ability of the VDRΔK246 mutant to form a complex on a VDRE using gel shift assays. As shown in Fig. 6, a band shift was generated by the WT VDR in the presence of 1,25(OH)2D3. On the other hand, the VDRΔK246 mutant failed to generate a band shift in the presence of 500 nM 1,25(OH)2D3, showing that the mutation affected DNA binding. We determined whether the inability of the mutant VDR to form a complex on a VDRE was caused by a defect in RXR heterodimerization using GST pull-down assays. As shown in Fig. 7A, the WT VDR was bound to GST-RXR in the absence of 1,25(OH)2D3 and addition of 1,25(OH)2D3 further increased the binding in a dose-dependent manner. In contrast, the VDRΔK246 mutant failed to bind to GST-RXR in the absence or presence of up to 1 μM 1,25(OH)2D3. These results showed that deletion of amino acid K246 disrupted VDR-RXRα heterodimerization.
Interactions of the ΔK246 mutant VDR with coactivators were also analyzed using GST pull-down assays. As shown in Fig. 7B, the WT VDR was bound to GST-SRC-1 and GST-DRIP205 when treated with 1 μM 1,25(OH)2D3. In contrast, the VDRΔK246 mutant failed to bind to GST-SRC-1 or GST-DRIP205 when treated with 1 μM 1,25(OH)2D3 (Fig. 7B). These results clearly showed that deletion of amino acid K246 abolished interactions with these coactivators.
Patients with HVDRR often have alopecia that is a phenocopy of APL that is caused by mutations in the hr gene.(4,5) The VDR has been shown to interact with HR, leading to the hypothesis that VDR and HR converge on similar pathways involved in hair growth.(8) Because the patient reported here had alopecia, we were interested in determining whether the ΔK246 mutation affected VDR interactions with HR. Using co-IP assays, we showed that FLAG-tagged HR was co-immunoprecipitated with the WT VDR as well as with the VDRΔK246 mutant (Fig. 7C). These results showed that deletion of amino acid K246 did not affect VDR-HR interactions in intact cells.
The child described here exhibited the classical clinical pattern of HVDRR. She initially presented with early-onset rickets, secondary hyperparathyroidism, hypocalcemia, hypophosphatemia, and elevated serum 1,25(OH)2D levels. The patient also had alopecia. The patient's parents were from different parts of the world and were unrelated and asymptomatic. We identified two mutations in the VDR gene: a nonsense mutation in exon 2 that created a premature stop at amino acid 30 (R30X) and a novel in-frame 3-bp deletion in exon 6 that deleted a lysine at amino acid 246 (ΔK246) in the VDR LBD. The father was heterozygous for the R30X mutation, whereas the mother was heterozygous for the 3-bp deletion. The R30X mutation was previously identified in two unrelated HVDRR patients of French-Canadian and Brazilian origins.(14,15) In addition to this case, two other cases of HVDRR were shown to be caused by compound heterozygous mutations in the VDR.(4,16) Both patients also had alopecia. One patient had two missense mutations in the VDR LBD that resulted in L263R and R391S amino acid substitutions.(16) The second patient had a single base deletion in exon 4 that led to a frameshift and downstream termination codon in exon 4 and a missense mutation in the VDR LBD that resulted in an E329K amino acid substitution.(4)
The clinical case had some slightly unusual features. First, despite marked rickets and alopecia, the serum calcium was only minimally decreased. Total calcium was actually within the normal range as were her albumin and total protein, whereas ionized calcium was low. She did have hypophosphatemia, secondary hyperparathyroidism, and elevated alkaline phosphatase. Of interest was the partial clinical response to oral calcium and calcitriol therapy. Although she had very reduced levels of a nonfunctional VDR, there was a clear clinical improvement to oral therapy with calcium and calcitriol based on clinical and X-ray findings as well as slight improvement in serum chemistries. The best formulation is that the oral calcium must have improved her bone mineralization based on a non-vitamin D-dependent increase in intestinal calcium absorption. We predict that she would have had a greater clinical improvement if treated with intravenous calcium administration.
Another interesting feature of the case is that the ΔK246 mutation described here had no affect on the affinity of the mutant VDR for 1,25(OH)2D3. This is not surprising because K246 is not a contact site for 1,25(OH)2D3. However, the double mutation resulted in a significant (∼67%) decrease in VDR protein levels compared with the heterozygotic parents and wildtype VDR levels. Because the second mutation is a premature stop that truncates the VDR, all of the expressed VDR is the ΔK246 mutant. Both parent's fibroblasts expressed VDR levels similar to VDR levels expressed in normal control fibroblasts. Thus, the father, who also carries the R30X mutation, and the mother, who carries the ΔK246 mutation, can compensate, and their total complement of VDR is similar to wildtype cells. The child with only the ΔK246 allele expressed has substantially decreased levels of VDR. A possible explanation for these findings is that the VDR gene is regulated by 1,25(OH)2D3 and contains VDREs in its promoter region.(17) Therefore, the levels of VDR in the parents may be upregulated by 1,25(OH)2D3 acting through the normal VDR allele, whereas the child's VDR level remain low because of the absence of a functional VDR.
The VDR LBD is composed of 12–13 α-helices forming a hydrophobic core that the ligand occupies.(18) K246 is located in helix H3 in the LBD (Fig. 2C). It has been suggested K246 is involved in forming a salt bridge with E420 in helix H12.(19) However, the crystallographic studies of the VDR showed that the salt bridge was formed by K264 in helix H4 and E420.(18) Mutation of K264 to K264A or K264E also disrupts transactivation presumably by disrupting coactivator interactions, although this has not been directly tested.(20,21) The salt bridge along with both hydrophobic contacts and polar interactions with other residues in the LBD are thought to stabilize the positioning of helix H12 during occupancy by the ligand. The repositioning of helix H12 is a critical event that occurs as a consequence of ligand binding and is essential to create the correct surface interface for coactivator binding required for transactivation. Whether K246 is involved in forming the salt bridge with E420 remains unresolved; however, it is clear from our data that K246 is important for providing a surface for interaction with coactivators.
Investigators have shown previously that mutation of K246 to alanine (K246A) or glutamic acid (K246E) had no affect on heterodimerization with RXR and DNA binding but the conversion of amino acid 246 disrupted interactions with the coactivators SRC-1 and GRIP-1.(22–24) Our data, on the other hand, showed that the K246 deletion as opposed to the conversion also prevented heterodimerization with RXR and DNA binding in addition to abolishing interactions with the coactivators SRC-1 and DRIP205 (Fig. 6B). Because the K246A mutation did not alter RXR heterodimerization, these results suggest that deletion of K246 changed the spacing in the VDR LBD that disrupted the contact with RXR.
Unliganded actions of the VDR are thought to be involved in the regulation of the hair cycle. RXRα and HR are also thought to be essential partners in transmitting the unliganded actions of the VDR during the hair cycle. This hypothesis is based on several findings. First, a patient with HVDRR caused by an E420K mutation in helix H12 in the VDR AF2 domain that abolished coactivator interactions and transactivation but not ligand binding, DNA binding, or RXR heterodimerization did not have alopecia.(21) Second, targeted expression of a transactivation defective VDR or a ligand binding defective VDR to keratinocytes in VDR knockout mice, which have alopecia, restored hair growth.(25) Third, targeted deletion of RXR in skin keratinocytes in mice also caused alopecia,(26) and previous cases of HVDRR with alopecia have been linked to defects in RXR heterodimerization.(9,27) The finding that the patient in this study also had alopecia indicates that the ΔK246 mutation disrupts both the liganded and unliganded functions of the VDR. Although we could show that the ΔK246 mutant interacted with HR, the inability of the mutant VDR to heterodimerize with RXRα would disrupt the HR-VDR-RXRα complex. These findings suggest that the development of alopecia in the patient was caused by the failure of the VDRΔK246 mutant to heterodimerize with RXR. However, additional factors besides RXR may also be needed for the unliganded actions of the VDR.
In conclusion, we identified a compound mutation (R30X/ΔK246) as the molecular basis for HVDRR in a patient with HVDRR and alopecia. The R30X mutation severely truncates the VDR, whereas the ΔK246 mutation abolishes RXR heterodimerization, coactivator interactions, and gene transactivation.
The authors thank Dr Milan Uskokovic for the 1,25(OH)2D3. This work was supported by NIH Grant DK42482 (to DF).