Somatic and Germline Mosaicism for a Mutation of the PHEX Gene Can Lead to Genetic Transmission of X-Linked Hypophosphatemic Rickets That Mimics an Autosomal Dominant Trait

Context: Familial hypophosphatemic rickets is usually transmitted as an X-linked dominant disorder (XLH), although autosomal dominant forms have also been observed. Genetic studies of these disorders have identified mutations in PHEX and FGF23 as the causes of X-linked dominant disorder and autosomal dominant forms, respectively.

Objective: The objective of the study was to describe the molecular genetic findings in a family affected by hypophosphatemic rickets with presumed autosomal dominant inheritance.

Patients: We studied a family in which the father and the elder of his two daughters, but not the second daughter, were affected by hypophosphatemic rickets. The pedigree interpretation of the family suggested that genetic transmission of the disorder occurred as an autosomal dominant trait.

Methods and Results: Direct nucleotide sequencing of FGF23 and PHEX revealed that the elder daughter was heterozygous for an R567X mutation in PHEX, rather than FGF23, suggesting that the genetic transmission occurred as an X-linked dominant trait. Unexpectedly, the father was heterozygous for this mutation. Single-nucleotide primer extension and denaturing HPLC analysis of the father using DNA from single hair roots revealed that he was a somatic mosaic for the mutation. Haplotype analysis confirmed that the father transmitted the genotypes for 18 markers on the X chromosome equally to his two daughters. The fact that the father transmitted the mutation to only one of his two daughters indicated that he was a germline mosaic for the mutation.

Conclusions: Somatic and germline mosaicism for an X-linked dominant mutation in PHEX may mimic autosomal dominant inheritance.

FAMILIAL HYPOPHOSPHATEMIC RICKETS is the most common inherited form of rickets and is characterized by growth retardation, defective bone mineralization, hypophosphatemia secondary to renal phosphate wasting, and an inappropriately low serum concentration of 1,25-dihydroxyvitamin D. Hypophosphatemic rickets is usually transmitted as an X-linked dominant disorder (XLH) (1), although autosomal dominant forms (ADHR) have also been observed (2–4). Genetic studies of these disorders have identified mutations in PHEX and FGF23 as the causes of XLH and ADHR, respectively (5, 6). PHEX is a member of the M13 family of type II cell-surface membrane zinc-dependent proteases, which also includes two endothelin-converting enzymes (ECEs), ECE-1 and -2; neprilysin; and the KELL antigen (7–10). PHEX protein is predominantly expressed in osteoblasts and odontoblasts but not in the kidney (11, 12). On the basis of the tissue distribution of PHEX and the reported actions of neprilysin, ECE-1, and ECE-2 (7, 8, 13, 14), it has been postulated that PHEX plays a role in the activation or inactivation of peptide factors that have roles in osteoblast differentiation and/or mineralization. In addition, PHEX may also participate in the control of renal phosphate handling and vitamin D metabolism (15). However, endogenous PHEX substrates and products have not yet been identified.

In this paper, we report a family affected by hypophosphatemic rickets. In this family, the father and the elder of his two daughters were affected by hypophosphatemic rickets, whereas the second daughter was not. The pedigree interpretation of the family suggested that genetic transmission of the disorder occurred as an autosomal dominant trait. However, molecular analysis disclosed that the transmission occurred as an X-linked dominant trait and that the affected father was a somatic and germline mosaic for the PHEX gene.

Subjects and Methods

Family K (Fig. 1)

Fig. 1.

Pedigree of family K. Black symbols, Patients with hypophosphatemic rickets; arrow, proband.

The proband (III-1) was a 4-yr-old girl of Japanese origin. She was referred at the age of 19 months with lower extremity bowing. Fasting blood and urine samples revealed normal renal function and no acidosis. Serum calcium was 10.1 mg/dl (2.53 mmol/liter), phosphorus was 2.5 mg/dl (0.81 mmol/liter), alkaline phosphatase was 1875 IU/liter (normal range for children < 900 IU/liter), 1,25-dihydroxyvitamin D₃ was 57.4 pg/dl (137.8 pmol/liter) (normal range for children 20–70 pg/dl), and intact PTH was 44 pg/ml (4.6 pmol/liter) (normal range 10–68 pg/ml). Tubular maximum reabsorption of phosphate per glomerular filtration rate was 2.3 mg/dl (0.92 mmol/liter) (normal range for children 4.36–5.53 mg/dl). She had radiographic evidence of severe rickets. She was treated with 1α-hydroxyvitamin D₃ (0.5 μg/day) and sodium phosphate, which led to significant improvement in the radiological signs of rickets after 12 months of therapy.

The proband’s father (II-1) had a history of hypophosphatemic rickets/osteomalacia. Treatment with 1α-hydroxyvitamin D₃ and sodium phosphate started at 2 yr of age but was interrupted from 15 to 25 yr of age because of noncompliance with the medication. At 20 yr of age, he had an osteotomy due to severe deformities of the lower limbs. At the time of this study, he was 35 yr of age and had a serum calcium level of 9.6 mg/dl (2.40 mmol/liter), phosphorus level of 1.7 mg/dl (0.55 mmol/liter), and alkaline phosphatase level of 215 IU/liter with 1α-hydroxyvitamin D₃ and sodium phosphate treatment. His body height was 148.0 cm (mean height of Japanese males 170.7 cm). Karyotype analysis of blood lymphocytes displayed a 46, XY karyotype. The proband’s younger sister (III-2) was 6 months of age at the time of this study. At 1 month of age, her fasting blood and urine samples revealed a serum calcium level of 11.0 mg/dl (2.75 mmol/liter), phosphorus level of 7.3 mg/dl (2.36 mmol/liter), alkaline phosphatase level of 971 IU/liter, and a percent tubular reabsorption of phosphate of 99%. A repeat evaluation after 2 wk demonstrated similar laboratory findings. Therefore, she was presumed to be unaffected. The proband’s mother (II-2) was a normal healthy female aged 35 yr at the time of this study. The paternal grandparents (I-1 and I-2) showed normal adult body height with no history of leg deformities or bone pain.

The parents of the proband gave informed consent for this study. The study was approved by the Institutional Review Board of Kobe Children’s Hospital.

Mutation analysis

Leukocyte DNA was extracted using standard methods. Single hair root DNA was obtained from the proband’s father (II-1) using a DNA extraction kit (QIAamp DNA Micro, QIAGEN, Hilden, Germany). The coding regions and related exon-intron boundaries of FGF23 and PHEX were investigated by automated direct nucleotide sequencing after amplification of the target sequences by PCR. Oligonucleotide PCR primers for each gene were designed using Primer3 software (http://frodo.wi.mit.edu/primer3/primer3_code.html) (16) and are listed in Table 1.

TABLE 1.

PCR primers for the FGF23 and PHEX genes

	Forward primer	Reverse primer
FGF23
Exon 1	CCGACAGGAGTGTCAGGTTT	GGTTGGATTAGCCCTCCAGT
Exon 2	ATCAATCCAGGGAGGTTTCA	GGAAACAGGTCACCAGGGTA
Exon 3A	GCTCAACGCCCTAAGAACTG	GGCTTCAGCACGTTCAGG
Exon 3B	CTTCAACACCCCCATACCAC	CCCAGAGAAGCAGCAAATTC
PHEX
Exon 1	TTTCCTGACGGCAGTTTCTT	ACCTATGAACGCAGGCAAAC
Exon 2	TGGGTTTTGGAATACCGTGT	GCTCCACTGTTTCACACCAA
Exon 3	AAGGCTTGGAAACTGGTTGA	AGTCATGCTTCAAATCCCAAA
Exon 4	GACTTCCAACTTGGCACCAT	TCCAGTCTTTCACAATCATTCC
Exon 5	CCACCCCACCTCTTTTACCT	GCACCCCAAAAGGCTAATCT
Exon 6	AATATGGCTGGGATGCAGAC	TCCTGCATTGGGAATATGGT
Exon 7	TCTGCTCTTCCATGTCTCTCAA	CAATGGGCAATGACACAAAA
Exon 8	ATGCAGATGTTTTGGCACAT	GGCATCCCAATACACAGACA
Exon 9	GGATGGCAATGATCAGGAGT	ACCGGGATTTTCCCTATGAC
Exon 10	GGAGCTTTGCCAACTGTTTC	GGCACACACACACACACAAA
Exon 11	GGGTTAGGGTGTGCAGTGTT	GACAATACCCACAGGCCACT
Exon 12	CAGAGCATGGAGTCAAGCTG	GCATGAACATCCATTAAACCA
Exon 13	ATTTTTGCCCTTCACAGTGG	GAAAGGCACAAGGCCAGTAA
Exon 14	CATGGCTTTGTGACTTCTGG	AGAGACTCCGCTTCTCACCA
Exon 15	GTCCAACATCCCCATTGTTC	CAACCTTCCTTCACCAGCAT
Exon 16	GAGGAGTGCCTTTCAGATGG	TTCCATGGCTTCTTTCTGCT
Exon 17	GCAGTTTATCTTGGCTTTCCA	TTATTGCAAGCCATCACAGC
Exon 18	TTTTGAAGGCTTGTCGAGGT	TTCAGCAGGTATGGGGTAGG
Exon 19	TTGATGCCTCTTGCTGAATG	GGTCAATGGGGAGACACACT
Exon 20	GGTGTACCTGCCTCACTGGT	GGGAGCAAACTCAAGTCCTG
Exon 21	TTGGAGCAGTTAAAACAGCAGA	ATGGAAATCACACGTCCACA
Exon 22	GGGCTTTAGTTGTCTCCCTGT	TCTCCAGGCCTAAAGCAATG

Mosaicism analysis using leukocyte and hair root DNA

Because the sequence peaks of PHEX obtained from leukocyte DNA from the proband’s father (II-1) indicated that the father, who was expected to be hemizygous, was actually heterozygous for mutant and normal alleles, we performed single-nucleotide primer extension and denaturing HPLC (DHPLC) on DNA extracted from leukocytes and hair roots. After primer extension using Thermo Sequenase (Amersham Bioscience, Piscataway, NJ), the wild-type and mutant alleles were discriminated and quantified by DHPLC (17, 18).

Before the primer extension reaction, 5 μl of a PCR product containing the mutation were treated with 2 μl of ExoSAP-IT (Amersham Bioscience) to remove unincorporated primers and deoxynucleotide triphosphates. Primer extension reactions were carried out in a final volume of 20 μl containing the purified PCR product (50–60 ng), 50 μm of a mixture of 2′,3′-dideoxycytidine-5′-triphosphate (ddCTP) and 2′,3′-dideoxythymidine-5′-triphosphate (ddTTP), 15 pmol of a primer (5′-CCTTTCTTTTGGGGAACAGAATATCCT-3′) located upstream of the mutation site, and 0.5 U Thermo Sequenase in the buffer provided by the manufacturer. The reaction was performed in a thermal cycler with an initial denaturation step of 1 min at 96 C followed by 50 cycles of 96 C for 10 sec, 43 C for 15 sec, and 60 C for 1 min. At the end of the thermal cycling, the reaction was incubated at 96 C for 30 sec and then immediately placed on ice. Separation of the extended primer was performed by DHPLC using an HPLC machine (Transgenomic Wave System; Transgenomic, Omaha, NE) as previously described (17).

Haplotype analysis using polymorphic markers on the X chromosome

Genomic DNA from leukocytes obtained from the proband (III-1), her sister (III-2), her parents (II-1 and II-2), and her paternal grandparents (I-1 and I-2) were analyzed using polymorphic markers located throughout the full length of the X chromosome. PCR amplification was performed with an ABI PRISM linkage mapping set (version 2.5 MD,10 panel 28; Applied Biosystems, Foster City, CA). The PCR products were run on an ABI PRISM 310 automated sequencer and the allele sizes were determined by the GENESCAN software (Applied Biosystems).

Results

Mutation analysis

Direct nucleotide sequencing of FGF23 in the leukocyte DNA from the proband did not reveal any putative disease-causing mutations. Subsequent sequencing of PHEX in the leukocyte DNA from the proband identified a heterozygous C-to-T transition at nucleotide 1699 in exon 16, which resulted in an arginine (CGA)-to-stop codon (TGA) substitution at codon 567 (R567X). Unexpectedly, the proband’s father (II-1) was heterozygous for the mutant and wild-type alleles, whereas the proband’s sister (III-2) did not carry the mutation. The proband’s mother (II-2) and paternal grandparents (I-1 and I-2) did not carry the substitution. Because the mutation eliminates an MnlI restriction enzyme recognition site and creates a StyI site, the heterozygous states of the proband and her father were confirmed by restriction enzyme analysis (Fig. 2).

Fig. 2.

Restriction enzyme analysis. PCR products of exon 16 of PHEX were digested with MnlI or StyI and separated on a 2% agarose gel. A, The resultant fragment lengths are shown schematically. B, The restriction enzyme analysis confirmed heterozygous states for the mutant and normal alleles in the proband and her father.

Mosaicism analysis

Single-nucleotide primer extension and DHPLC analysis of leukocyte DNA from the proband’s father (II-1) revealed that 40% of the tested DNA had the wild-type genotype, whereas 60% had the mutant form. In the proband (III-1), her sister (III-2), and her grandfather (I-1), the proportions of leukocyte DNA carrying the mutant sequence of PHEX were 56, 0, and 0%, respectively (Fig. 3). The same analysis performed on DNA from eight single hair roots from the proband’s father showed four different ratios of the mutant allele: 6% (two samples), 38% (one sample), 64–73% (four samples), and 94% (one sample) (Fig. 4). These ratios are largely consistent with the hypothesis that each hair root originates from three different progenitor cells. Theoretically single hair roots from a man with a mosaic mutation located on the X chromosome show four different ratios of the mutant allele: 0% (no mutant progenitor cells and three normal progenitor cells), 33% (one mutant progenitor cell and two normal progenitor cells), 67% (two mutant progenitor cells and one normal progenitor cell), and 100% (three mutant progenitor cells and no normal progenitor cells) (Fig. 4) (18–20).

Fig. 3.

Single nucleotide primer extension and DHPLC analysis using leukocyte DNA and the corresponding direct sequence electrophoregrams. The proportions of leukocyte DNA carrying the mutant sequence of PHEX in the proband’s grandfather (I-1), the proband’s father (II-1), the proband (III-1), and the proband’s sister (III-2) were 0, 60, 56, and 0%, respectively. Asterisks, Unextended primers; C, wild-type alleles; T, mutant alleles.

Fig. 4.

A, Representative profiles of single-nucleotide primer extension and DHPLC analysis using DNA from eight single hair roots from the proband’s father (II-1). Four different ratios of the mutant allele were observed: 6, 38, 64, and 94%. Asterisks, Unextended primers; C, wild-type alleles; T, mutant alleles. B, Theoretical models for single hair roots from a man with a mosaic mutation are schematically shown. Single hair roots show four different ratios of the mutant allele: 0% (no mutant progenitor cells and three normal progenitor cells), 33% (one mutant progenitor cell and two normal progenitor cells), 67% (two mutant progenitor cells and one normal progenitor cell), and 100% (three mutant progenitor cells and no normal progenitor cells). Open circles, Normal progenitor cells; closed circles, mutant progenitor cells.

Haplotype analysis

The genotypes for 18 markers on the X chromosome of the proband’s father (II-2) were equally inherited by the proband (III-1) and her sister (III-2). The father inherited the genotypes for these markers on the X chromosome from the proband’s grandmother (I-2) (Fig. 5).

Fig. 5.

Haplotype analysis. The genotypes for 18 markers on the X chromosome of the proband’s father (II-2) were equally inherited by the proband (III-1) and her sister (III-2). There are several possible chromosomal arrangements of the mother (II-2) and paternal grandmother (I-2). At present, however, the correct arrangement is not known because additional information derived from other family members is not available to us.

Discussion

Familial hypophosphatemic rickets is the most common inherited form of rickets and is usually transmitted as XLH, although ADHR has also been observed (1, 2–4). Genetic studies of these hypophosphatemic disorders identified gene mutations in PHEX and FGF23 as the causes of XLH and ADHR, respectively (5, 6). Under a model with X-linked inheritance, an affected man is expected to transmit the mutation to all of his daughters. Therefore, the pedigree interpretation of family K suggested that genetic transmission of the disorder occurred as an autosomal dominant trait. However, molecular analysis demonstrated that the proband (III-1) was heterozygous for a R567X mutation in PHEX, rather than FGF23, suggesting that the genetic transmission occurred as an X-linked dominant trait. Unexpectedly, the proband’s father (II-1) was heterozygous for this mutation. These findings suggested that the R567X mutation was present in a mosaic fashion in the father’s leukocytes.

We used single-nucleotide primer extension and DHPLC analysis to confirm and measure the degree of mosaicism. Single-nucleotide primer extension and DHPLC analysis of leukocyte DNA (mesodermal derivatives) from the proband’s father revealed that 60% of the total leukocyte DNA had the mutant genotype. The same analysis performed on DNA from eight single hair roots from the father demonstrated four different proportions of the mutant allele. According to the theory that every human hair root originates from three progenitor cells, this finding indicated that the mutation was not present in all of the father’s ectodermal cells (18–20). In other words, single hair root DNA analysis revealed that the proband’s father was a somatic mosaic for this mutation of PHEX. Haplotype analysis revealed that the affected father (II-1) transmitted the genotypes for 18 markers on the X chromosome equally to his daughters (III-1 and III-2), but the second daughter did not inherit the mutation of PHEX located on the X chromosome. The fact that the affected father transmitted the mutation to only one of his two daughters indicated that he was a germline mosaic for the mutation. Therefore, the mutation may have occurred during early embryogenesis before the commitment to germ cells.

At an early stage of development, dramatic methylation changes have been observed. During preimplantation development, gamete-specific patterns of methylation are erased by genome-wide demethylation, tending toward a ground state of absence of methylation in the inner cell mass of the blastocyst. Except for a small number of methylated CpG sites in imprinted genes, the vast majority of methylated CpG sites are unmethylated by the stage of cavitation (16-cell stage) (21, 22). The point mutation (CGA-to-TGA) identified in family K occurred at a CpG site, and this site may have been subject to demethylation during the early steps of embryo development. An aberrant process of demethylation could lead to a mosaic mutation at this CpG site during early embryogenesis.

One interesting issue is why the Phex/PHEX gene mutation produces a dominant phenotype. Hyp mice, the mice homolog of human XLH, and XLH patients are truly dominant conditions with similar degrees of bone and renal defects in hemizygotes, heterozygotes, and homozygotes for the disease allele (23–27). A number of studies have suggested that the dominant pattern of Hyp/XLH inheritance is due to haploinsufficiency, in which the normal Phex/PHEX protein produced by the wild-type allele in heterozygotes does not reach the threshold level necessary for normal Phex/PHEX function (25, 28, 29). In the proband’s father (II-1), 40% of the leukocyte DNA had the wild-type genotype, whereas 60% had the mutant form. Although the genotype in cartilage, bone, and teeth, in which PHEX is predominantly expressed (30, 31), could not be analyzed in the proband’s father, the normal PHEX protein produced by the wild-type allele in these tissues may not reach the threshold level necessary for normal PHEX function.

In summary, we report here a family affected by hypophosphatemic rickets. The pedigree interpretation of this family suggested that genetic transmission of the disorder occurred as an autosomal dominant trait. However, molecular analysis revealed that the transmission occurred as an X-linked dominant trait and that the affected father was a somatic and germline mosaic for a mutation of PHEX. To date, somatic and germline mosaics have been reported for only a few X-linked and autosomal dominant monogenic disorders, and this is the first report of mosaicism for a mutation of PHEX. Somatic and germline mosaicism for an X-linked dominant mutation in the PHEX gene may mimic autosomal dominant inheritance. Mosaicism has important implications for molecular diagnosis interpretation, clinical evaluation, and genetic counseling.

Acknowledgments

Received August 5, 2005. Accepted November 11, 2005.

Abbreviations:

 
• ADHR,

  Autosomal dominant XLH form;

 
• DHPLC,

  denaturing HPLC;

 
• ECE,

  endothelin-converting enzyme;

 
• XLH,

  X-linked dominant disorder.