CASE REPORT WITH REVIEW OF LITERATURE
Intragenic duplication of PHEX in a girl with X-linked hypophosphatemia: a case report with review of literature
2025 Volume 72 Issue 4 Pages 413-419
Details
2025 Volume 72 Issue 4 Pages 413-419
Over 70 intragenic copy-number variations (CNVs) of PHEX have been identified in patients with X-linked hypophosphatemia (XLH). However, the underlying mechanism of these CNVs has been poorly investigated. Furthermore, although PHEX undergoes X chromosome inactivation (XCI), the association between XLH in women with heterozygous PHEX variants and skewed XCI remains unknown. In this study, we determined the precise genomic structure and the XCI status of a girl with XLH who showed short stature and bowing of the legs at 2 years old. Laboratory tests revealed low levels of serum phosphate and elevated levels of alkaline phosphatase and fibroblast growth factor 23. Multiplex ligation-dependent probe amplification and targeted long-read sequencing revealed that she carried a 24.6-kb intragenic duplication of PHEX. The duplication was tandemly aligned in a head-to-tail orientation. The duplication breakpoints shared a 2-bp microhomology, indicating that this CNV resulted from a replication-based error. Trio sequencing results showed that the duplication was a de novo CNV that occurred on the paternally-derived allele. DNA methylation analysis demonstrated random XCI. A literature review of 12 previously reported cases of intragenic CNVs of PHEX revealed that the deletions/duplications can be ascribed to replication-based errors. Our findings and those of previous studies indicate that XLH-causative CNVs in PHEX predominantly arise from replication-based errors. Thus, the genomic region surrounding PHEX may be vulnerable to replication-based errors during gametogenesis or early embryogenesis. Our study provides supporting evidence that heterozygous PHEX variants can lead to XLH in women with random XCI.
X-linked hypophosphatemic rickets [OMIM: 307800] is a genetic disorder characterized by short stature and leg deformities [1]. PHEX [MIM: 300550] is the causative gene of X-linked hypophosphatemia (XLH) [1]. Pathogenic variants in PHEX lead to increased blood levels of fibroblast growth factor 23 (FGF23), resulting in renal phosphate wasting and hypophosphatemia [2]. PHEX resides on the X chromosome and consists of 22 exons. In women, PHEX is subjected to X chromosome inactivation (XCI) [3]. Over 870 pathogenic PHEX variants have been reported to date [3-10], including 66 deletions and nine duplications involving one or only a few exons [2]. Thus, PHEX intragenic copy-number variations (CNVs) are important causes of XLH. Generally, CNVs in the human genome are created via four mechanisms, namely, nonallelic homologous recombination (NAHR), non-homologous end joining (NHEJ), replication-based errors, and retrotransposition [11]. These four mechanisms usually entail specific breakpoint structures [11]. Previously, a deletion involving PHEX exons 1–3 was reported to have an NAHR-specific breakpoint structure [12]; however, the underlying mechanisms of other intragenic CNVs of PHEX remain unelucidated. Furthermore, reports on XCI status in female patients with XLH are rare [13-15]. Thus, although random XCI has been reported in some female patients [13], these findings await further validation.
Recent advances in long-read sequencing have enabled the characterization of CNV structure [16]. Targeted long-read sequencing using adaptive sampling is a cost-effective method for enriching the targeted regions 5- to 10-fold [17]. Furthermore, DNA methylation analysis of the androgen receptor (AR) gene in women provides useful information regarding XCI status [18]. In the present study, we investigated the structure of a PHEX intragenic duplication and the XCI status in a girl with XLH. We also reviewed the literature to discuss the mechanism of XLH-causative CNVs within PHEX.
The girl was born as the first child of non-consanguineous Japanese parents. Both parents were 27 years of age and phenotypically unremarkable. The heights of the father and mother were 175 cm and 146 cm, respectively. The patient was born with normal body length (50.0 cm, +1.4 SD) and weight (2.678 kg, +0.4 SD). At the age of 2 years, she visited our hospital because of short stature and bowing of the legs (Fig. S1). Her height and weight were 78.1 cm (–2.1 SD) and 12.0 kg (+0.9 SD), respectively. She was otherwise healthy, with no other phenotypic abnormalities. Laboratory tests showed low levels of serum phosphate and tubular maximum phosphate reabsorption per glomerular filtration rate (TmP/GFR) and elevated levels of serum alkaline phosphatase (ALP) and fibroblast growth factor 23 (Table 1). She was diagnosed with hypophosphatemic rickets and treated with alfacalcidol and monobasic sodium phosphate monohydrate, dibasic sodium phosphate anhydrous. Following treatment initiation, her growth rate, skeletal findings, and laboratory data improved. On the latest visit at 6 years and 8 months of age, her height and weight were normal (109.7 cm [–1.2 SD] and 20.9 kg [–0.1 SD]), and she exhibited mild bowing of the legs and a normal level of serum ALP (Table 1).
| At diagnosis (2 y) | At the last visit (6 y 8 m) | |||
|---|---|---|---|---|
| Serum calcium (mg/dL) | 9.1 | (8.8–10.5) | 9.4 | (8.5–10.2) |
| Serum phosphate (mg/dL) | 2.7 | (3.8–6.0) | 2.6 | (3.9–5.8) |
| Serum alkaline phosphatase (U/L) (JSCC)* | 2,464 | (410–1,150) | 1,210 | (460–1,250) |
| Serum creatinine (mg/dL) | 0.27 | (0.17–0.45) | 0.40 | (0.25–0.48) |
| Serum 25-OHD (pg/mL) | 15 | (>20) | No data | — |
| Serum intact PTH (pg/mL) | 122 | (10–65) | 37 | (10–65) |
| Intact FGF23 (pg/mL) | 90 | (<30) | No data | — |
| TmP/GFR (mg/dL) | 2.3 | (4.9–5.7) | 1.8 | (4.9–5.7) |
The reference ranges are shown in parentheses.
Abnormal values are boldfaced.
* IFCC = JSCC × 0.35
Abbreviations: 25-OHD, 25-hydroxyvitamin D; FGF23, fibroblast growth factor 23; PTH, parathyroid hormone; TmP/GFR, tubular maximum reabsorption of phosphate
This study was approved by the Institutional Review Board Committee at the National Center for Child Health and Development and performed after obtaining written informed consent. We performed molecular analyses using genomic DNA extracted from peripheral blood samples collected from the patient and her parents. Sanger sequencing of the patient identified no rare variant in the exons or intron-exon boundaries of PHEX. Multiplex ligation-dependent probe amplification (MLPA, catalog number, P223; MRC Holland, Amsterdam, Netherlands) detected a heterozygous duplication involving PHEX exons 17–20 (Fig. 1A). This variant was absent from the Genome Aggregation Database (https://gnomad.broadinstitute.org/), the Integrative Japanese Genome Variation Database (https://jmorp.megabank.tohoku.ac.jp/), and Novel PHEX locus-specific database (https://www.rarediseasegenes.com).
Next, we performed long-read sequencing (Oxford Nanopore Technologies, Oxford, UK) for the patient. A ~0.4 Mb genomic region surrounding PHEX (chrX:21,932,325–22,351,310; GRCh38/hg38) was subjected to adaptive sampling according to the manufacturer’s instructions [17]. The results revealed a heterozygous 24,644 bp duplication encompassing PHEX exons 17–20 (chrX:22,216,638–22,241,282) (Fig. 1B). The results of long-read sequencing were confirmed by Sanger sequencing (Fig. 1C). The amplified regions were tandemly aligned in a head-to-tail manner (Fig. 1C). The breakpoints in introns 16 and 20 were located within Alu and L1 repeats, respectively (Table 2). The two breakpoints shared a 2-bp microhomology, but no long sequence similarity. No nucleotide insertion was observed at the fusion junction. Sanger sequencing of the parental samples detected no duplication in PHEX. Trio genotyping of a single nucleotide polymorphism within the duplication (rs5904514) suggested that the duplication of the patient resided on the paternally inherited X chromosome (Fig. 2A). We performed PCR targeting PHEX to determine whether the father had somatic mosaicism for the duplication. We prepared PCR reaction mixes containing DNA of the patient’s father or a control male at a final concentration of 5 ng/μL. As positive controls, we prepared solutions containing the genomic DNA of the patient at final concentrations of 5 ng/μL or 25 pg/μL, and an artificially mixed sample containing 25 pg/μL DNA of the patient and 5 ng/μL DNA of the control male. These five samples underwent PCR-amplification using two primer sets that amplified either the normal sequence of PHEX intron 16 or the fusion junction between introns 16 and 20. Following PCR, we observed amplicons of the normal PHEX sequence in all samples. However, amplicons of the fusion junction were only observed in the samples containing the patient’s DNA (Fig. 2B), indicating the absence of mosaicism in the father.
| Cases | Sporadic /Familial | Sex (Number of cases) | Affected exons | Position (GRCh38/hg38) | Breakpoint features | Reference | ||
|---|---|---|---|---|---|---|---|---|
| Repetitive sequence | Microhomology | Nucleotide insertion | ||||||
| Duplication | ||||||||
| 1 | Familial | M (2), F (1) | Exons 13–15 | chrX:22,147,071–22,198,783 | None | 5 bp | None | 22 |
| 2 | Sporadic | F (1) | Exons 16–20 | chrX:22,216,638–22,241,282 | Alu (5′), L1 (3′) | 2 bp | None | Our case |
| Deletion | ||||||||
| 3 | No data | No data | Exon 1 | chrX:22,032,977–22,033,657 | None | 3 bp | None | 3 |
| 4 | Familial | M (1), F (1) | Exons 1–3 | chrX:22,010,936–22,063,079 | L1 (5′) | 7 bp | None | 24 |
| 5 | Sporadic | M (1) | Exons 1–3 | chrX:22,016,715–22,056,805 | L1 (5′) | 4 bp | None | 12 |
| 6 | Sporadic* | F (1) | Exons 1–11 | chrX:22,017,560–22,130,115 | None | Blunt end | None | 26 |
| 7 | No data | No data | Exon 5 | chrX:22,077,649–22,077,752 | None | Blunt end | None | 23 |
| 8 | No data | No data | Exons 6–10 | chrX:22,077,702–22,114,460 | None | Blunt end | None | 25 |
| 9 | No data | No data | Exons 9–12 | chrX:22,099,085–22,159,034 | None | 4 bp | None | 3 |
| 10 | Sporadic | M (1) | Exon 14 | chrX:22,177,469–22,178,897 | None | 4 bp | None | 27 |
| 11 | Sporadic | F (1) | Exons 16–17 | chrX:22,212,919–22,219,061 | None | 5 bp | None | 28 |
| 12 | No data | No data | Exon 22 | chrX:22,247,875–22,248,186 | None | Blunt end | None | 3 |
| 13 | No data | No data | Exon 22 | chrX:22,247,875–22,248,189 | None | Blunt end | None | 6 |
* This case had a mosaic karyotype: mos 45,X[32]/46,X[18], arr[hg19] 2q37.1q37.3 (232,808,400–242,783,384) x1 arr[hg19] X –p22.11 (22,035,678–22,148,232)x1
Abbreviations: F, Female; M, Male
We attempted to examine the effects of the PHEX variant at the mRNA level. Thus, we analyzed the expression level of PHEX in a peripheral blood sample collected from a healthy individual and used GAPDH as a positive control. RT-PCR yielded products of GAPDH amplicons only, and no PHEX amplicons were present (data not shown). These findings suggested the lack of PHEX expression in blood, which was consistent with the data in the Genotype-Tissue Expression database (https://gtexportal.org/home/). Thus, it was impossible to investigate PHEX mRNA using the blood sample collected from the patient.
Then, we analyzed the XCI status of the patient by AR methylation analysis. To determine the XCI ratio, we performed Hpa II-mediated DNA methylation analysis for the polymorphic trinucleotide locus in the AR gene [18]. In brief, the genomic region flanking this locus was PCR-amplified and subjected to microsatellite analysis before and after Hpa II digestion. Hpa II cleaves unmethylated target sequences. We calculated the ratio of the unmethylated alleles to all alleles from the area under the curve of each PCR product. The ratio of unmethylated alleles in the shorter to longer alleles in the patient was 49:51 (Fig. 2C), indicating random (unskewed) XCI [13].
We presented a girl with XLH who carried a hitherto unreported intragenic duplication within PHEX. Long-read sequencing revealed that the 24,644-bp duplication contained PHEX exons 17–20 and was tandemly aligned in a head-to-tail manner (Graphic Abstract). This study provides further evidence for the usefulness of the combination of MLPA and long-read sequencing for characterizing genomic CNVs [11]. The duplication of our patient resulted in an insertion of 370 nucleotides into PHEX mRNA. There are three possible mechanisms whereby this duplication results in the loss of function of the PHEX protein. First, the mutant mRNA harbors a terminal codon at the 704th position, and therefore, is predicted to encode a truncated protein (p.Val691Ilefs*14). Since it was demonstrated that a mutant PHEX protein lacking only 36 amino acids at the C-terminus (p.Gln714*) led to XLH [3, 19, 20], the mutant protein in our patient is unlikely to retain residual activity. Second, the p.Val691Ilefs*14 variant satisfies the condition of nonsense-mediated mRNA decay [21]. Thus, the mutant mRNA may undergo rapid degradation. Third, the duplication may cause abnormally spliced PHEX mRNA. We were unable to examine these possibilities because PHEX expression was absent in the blood.
In our patient, the two duplication breakpoints shared a 2-bp microhomology. The duplication lacked long sequence similarities between the two breakpoints indicative of NAHR, short nucleotide insertions at the fusion junction characteristic of NHEJ, or transposable elements associated with retrotransposition. Thus, this duplication most likely arose from a replication-based error. In this context, 12 XLH-causative CNVs within PHEX have been characterized (Table 2) [3, 6, 12, 22-28]. Similar to the duplication in the present case, the characteristic features of NAHR, NHEJ, or retrotransposition were absent in these CNVs. Furthermore, breakpoint microhomologies of 2–7 bp were observed in seven of the 12 cases (one duplication and six deletions), indicating that these CNVs were also the result of replication-based errors [11]. The remaining five deletions had blunt-end breakpoints, which can be ascribed to replication-based errors or NHEJ [11]. Thus, the genomic region surrounding the PHEX exons may represent a hotspot for replication-based errors. Additionally, the trio genotyping revealed that the duplication occurred on the paternally inherited allele, and PCR analysis of the patient’s father revealed no evidence of mosaicism. These findings suggested that the duplication in the present case was a de novo CNV on the paternally inherited X chromosome and was assumed to have occurred during gametogenesis or soon after fertilization. The duplication was independent of advanced paternal age, which is a known risk factor for de novo rearrangements [29]. Further studies are needed to clarify the factors associated with PHEX intragenic CNVs.
Our patient showed random XCI in the blood and exhibited typical clinical features of XLH. These results are consistent with Orstavik et al., who reported that the XCI patterns of XLH patients were similar to those of unaffected females [13]. These findings indicate that PHEX haploinsufficiency is sufficient to cause XLH in women with random XCI. However, we cannot exclude the possibility of skewed XCI exclusively in the skeletal tissue of our patient. Owen et al. and Kayser et al. reported that selective inactivation of the mutant allele resulted in a normal phenotype in a woman with a heterozygous PHEX variant [14, 15]. Thus, XCI likely exerts some effects on the clinical consequences of PHEX variants in women.
To summarize, the results of our study, in conjunction with those of previous studies [3, 12, 22, 24, 27, 28], indicate that XLH-causative CNVs in PHEX predominantly arise from replication-based errors. Moreover, this study provides supporting evidence that heterozygous PHEX variants can lead to XLH in women with random XCI. Further studies are necessary to clarify the factors associated with the genomic predisposition of the PHEX-flanking region to submicroscopic CNVs.
Maki Fukami is a member of Endocrine Journal’s Editorial Board.
This study was supported by grants from The Japan Endocrine Society, the National Center for Child Health and Development, and the Takeda Science Foundation.
Data will be made available on request.
alkaline phosphatase
ARandrogen receptor
CNVscopy-number variations
GAPDHGlyceraldehyde-3-phosphate dehydrogenase
MLPAMultiple Ligation-dependent Probe Amplification
NAHRnonallelic homologous recombination
NHEJnon-homologous end joining
SNVsingle nucleotide variants
TmP/GFRtubular maximum phosphate reabsorption per glomerular filtration rate
XCIX chromosome inactivation
XLHX-linked hypophosphatemia.
