Abstract
Background/Aim: X-linked hypophosphatemia (XLH), the most common form of hereditary rickets, results from loss-of-function mutations in the phosphate-regulating PHEX gene. Elevated fibroblast growth factor 23 (FGF23) contributes to hypophosphatemia in XLH. This study aimed to characterize PHEX variants and serum FGF23 profiles in Taiwanese patients with XLH. Patients and Methods: We retrospectively reviewed the records of 102 patients clinically suspected of having hypophosphatemic rickets from 2006 to 2022. Serum intact Fibroblast growth factor-23 (iFGF23) levels were measured on clinic visit days. PHEX mutations were identified using Sanger sequencing, and negative cases were analyzed using whole-exome sequencing. Results: The majority (92.1%) of patients exhibited elevated FGF23 compared with normal individuals. Among 102 patients, 44 distinct PHEX mutations were identified. Several mutations recurred in multiple unrelated Taiwanese families. We discovered a high frequency of novel PHEX mutations and identified variants associated with extreme FGF23 elevation and tumorigenesis. Conclusion: Our findings revealed the PHEX genotypic variants and FGF23 levels in Taiwanese patients with XLH. These results are crucial given the recent approval of burosumab, a monoclonal FGF23 antibody, for XLH therapy. This study provides key insights into the clinical management of XLH in Taiwan.
- X-linked hypophosphatemia
- hypophosphatemic rickets
- phosphate-regulating endopeptidase gene
- fibroblast growth factor 23
- tumor-induced osteomalacia
X-linked hypophosphatemia (XLH; OMIM #307800), the most prevalent form of hereditary rickets, arises from loss-of-function mutations in the phosphate-regulating endopeptidase (PHEX) gene (1, 2). XLH has an estimated incidence of 1 in 20,000 births and a prevalence of 1.7 to 4.8 per 100,000 individuals (3). XLH was first described in 1937 by Fuller Albright et al., who reported a familial form of vitamin D-resistant rickets (4). The hallmark of XLH is renal phosphate wasting leading to hypophosphatemia due to loss-of-function mutations in the PHEX gene (5, 6). Osteocytic overexpression of FGF23, a phosphaturic hormone, exacerbates renal phosphate wasting, contributing significantly to the pathogenesis of hypophosphatemia in XLH (2). FGF23 excess also causes acquired conditions such as tumor-induced osteomalacia (TIO) (7). Although both PHEX and FGF23 are osteocyte products (8), the mechanisms by which PHEX mutations lead to increased FGF23 levels remain unclear.
The PHEX gene is located on the X chromosome at Xp22.1, consists of 22 exons spanning 220 kb, and encodes a cell surface zinc metalloendopeptidase (9). Over 700 pathogenic PHEX variants have been reported, exhibiting considerable variability in the severity of the XLH phenotype, including short stature, skeletal deformities, dental complications, and nephrocalcinosis (10). Studies in Western populations have identified potential genotype-phenotype correlations and conflicting associations between circulating FGF23 and PHEX mutations (11, 12). However, limited data exist among Asian populations. We have observed certain prevalent PHEX variants in our Taiwanese cohort, including c.1735G>A. Additionally, serum FGF23 levels appear to correlate with tumorigenesis. Therefore, we aimed to characterize the prevalence of XLH and genotype-phenotype correlations in Taiwanese patients with confirmed PHEX mutations.
Patients and Methods
Subjects. Genetic screening was performed on 102 patients with clinically suspected hypophosphatemic rickets, including 27 familial and 75 sporadic cases. The cohort comprised 35 males and 77 females. PHEX mutations were identified in 74 of the 102 screened patients. Medical records, including clinical, laboratory and radiological data, were retrospectively reviewed for these 102 patients, dating back to 2006. Written informed consent was obtained from all patients included in the study. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Chung-Shan Medical University Hospital (IRB number CS2-20098, approved 17 June 2022).
Serum iFGF23 measurement. Serum samples were collected on the day of clinic visits to measure iFGF23 levels. Sixty-three samples were available and stored at −80°C. Serum iFGF23 levels were quantified using an ELISA kit (KAINOS Laboratories, Inc., Tokyo, Japan) according to the manufacturer’s protocol. Data were analyzed according to the manufacturer’s recommendations. The detectable concentration ranged from 3 to 800 pg/ml.
Sanger sequencing of PHEX mutations. Blood samples (3 ml) were collected from each patient in tubes containing EDTA as an anticoagulant. Genomic DNA was extracted from the blood samples using Gentra Puregene Kits (Qiagen, Taipei, Taiwan, ROC) according to the manufacturer’s protocol. The average yield of genomic DNA was 50-100 μg per 3 ml of whole blood. The isolated DNA was then amplified by polymerase chain reaction (PCR) using an established protocol, where each PHEX exon and flanking intronic regions was amplified separately in 20 μl reactions containing 100 ng of genomic DNA template. Primer pairs targeting intronic sequences were designed as described previously (13). PCR products were purified using the Viogene PCR-M Clean Up System (Viogene, Taipei, Taiwan, ROC) and concentrations were measured using a spectrophotometer (Amersham Biosciences, Taipei, Taiwan, ROC). Sanger sequencing was performed on an Applied Biosystems 3730 automated DNA sequencer (Thermo Fisher, Taipei, Taiwan, ROC). Patient sequence data were compared to the PHEX mRNA reference sequence (NM_000444).
Whole-exome sequencing. For whole-exome sequencing, genomic DNA samples from each patient were fragmented to 180-280 bp using the Covaris hydrodynamic shearing system (Covaris, Taipei, Taiwan, ROC), which applies focused acoustic energy to generate uniform DNA fragments. Fragmented DNA (200 ng) was then subjected to end repair to generate blunt ends, A-tailing to ligate 3′ adenosine overhangs, and adapter ligation using the Agilent SureSelect Human All Exon V6 kit (Agilent, Taoyuan, Taiwan, ROC) according to the manufacturer’s protocol. Indexed sequencing libraries were generated by selective enrichment PCR to amplify exonic regions and purified with AMPure XP beads (Beckman Coulter, Taipei, Taiwan, ROC) to remove unwanted primers and small DNA fragments. Library quantification was performed using the Agilent Bioanalyzer 2100 to assess the quality and determine molar concentrations. Paired-end 150 bp sequencing of the prepared libraries was carried out on the Illumina NovaSeq 6000 platform (Genomics BioSci & Tech Co, Taipei, Taiwan, ROC) to obtain whole exome sequences. The resulting sequencing data were analyzed to identify rare genetic variants in exonic regions across the genome.
Bioinformatics analysis. Raw sequencing data were processed using the InheriNext platform (Compass Bioinformatics, Hsinchu, Taiwan, ROC). Adapter sequences were trimmed with fastp and reads were aligned to the hg38 reference genome using DRAGEN 3.7.5 (Illumina Inc.). Variants were annotated using VEP v101 and Jannovar v0.35 with dbNSFP 4.3a. The Human Phenotype Ontology (HPO) terms ranked disease-associated genes using the Emission-Reception Information Content (ERIC) algorithm. Variant pathogenicity weights were calculated based on population frequency, prediction scores and functional consequences.
Statistical analyses. Continuous variables were evaluated for normality using the Anderson-Darling test to determine appropriate statistical methods. Data following a normal distribution are summarized as mean±standard deviation (SD). Non-normally distributed data are expressed as median and interquartile range. All statistical analyses were performed using JMP Pro v16.2.0 software (SAS Institute Inc., Taipei, Taiwan, ROC). Differences with a p-value less than 0.05 were considered statistically significant.
Results
Serum FGF23 levels in patients with XLH. Our study found serum iFGF23 levels ranging from 30.828 to 1,189.5 pg/ml, with a median of 105.019 pg/ml (interquartile range=72.917-159.659 pg/ml), among patients in families with hypophosphatemic rickets (Table I). The majority (92.1%) of cases had elevated FGF23 levels compared to the normal reference range of 16.1-42.2 pg/ml (11). When stratified by sex, no significant differences in circulating FGF23 concentrations were found between male and female XLH patients in our study cohort. Median FGF23 was 98.662 pg/ml (interquartile range=72.535-134.659 pg/ml) in males and 105.2495 pg/ml (interquartile range=73.0905-183.2085 pg/ml) in females. Extremely high FGF23 levels (>800 pg/ml) were found in three patients, likely related to tumorigenesis.
PHEX gene mutation and iFGF-23 protein concentration in the family with hypophosphatemic rickets.
Identification of PHEX mutations. In our study, we identified 44 distinct mutations in PHEX in the group of 102 patients. These mutations comprised a single silent mutation (2.3%), eight missense mutations (18.2%), 13 nonsense mutations (29.5%), 13 frameshift mutations (29.5%), and nine intron mutations (20.5%), of which seven were located at splice sites (Figure 1). More than half of these mutations (26; 59.1%), were unique and not previously documented in the scientific literature. These mutations were scattered across the PHEX coding region and intronic sequences, but none was observed in the 3′ or 5′ untranslated regions (Figure 1). The most frequently observed mutations were G579R (21.9%) and R747X (7.8%). A comprehensive list of all PHEX mutations identified among X-linked hypophosphatemia patients in this Taiwanese study cohort is provided in Table II. Our analyses revealed a wide spectrum of PHEX mutations among the Taiwanese XLH cases, including missense/nonsense mutations, deletions, insertions, duplications, and splice site mutations (Figure 2).
Distribution of all PHEX mutations identified in this study. A total of 44 different PHEX mutations were detected in 102 patients. Yellow corresponds to silent mutations, green corresponds to missense mutations, red corresponds to nonsense mutations, blue corresponds to frameshift mutations, and purple corresponds to intron insertions.
A comprehensive list of all PHEX mutations identified among X-linked hypophosphatemia patients in this Taiwanese study cohort.
Frequency of PHEX variants in Taiwan. A total of 74 PHEX mutations were detected in 102 patients, including 1 silence mutation (1.4%), 22 missense mutations (29.7%), 21 nonsense mutations (28.4%), 19 frameshift mutations (25.7%), 101 intron mutations (9 at splice-site included) (14.9%).
To corroborate these findings, we analyzed PHEX mRNA extracted from peripheral blood. In a patient harboring the c.1483_1586del mutation, the analysis revealed that this variant induced an exon 14 skipping event, thereby generating a frameshift mutation. Further examination led to the discovery of a sizable deletion (2831bp, c.1483-1823_c.1586+904) in the PHEX gene of the patient.
Elevated serum FGF23 levels in familial cases suggest possible biomarker role in cancer. The serum FGF23 levels were elevated in all 15 familial cases from three families, ranging widely from 46.954 to >800 pg/ml (normal range: 16.1-42.2 pg/ml) (11). Notably, extremely high FGF23 levels (>800 pg/ml) were detected in cases 62 and 63. Case 62 was diagnosed with papillary thyroid microcarcinoma and case 63 with renal cell carcinoma and parathyroid adenoma. Additionally, case 16 had osteochondroma. The role of FGF23 in malignant tumors is best characterized in tumor-induced osteomalacia, whereby FGF23 overproduction by tumors causes renal phosphate wasting and reduced 1,25(OH)2D3, resulting in osteomalacia (7). In cancers affecting the bone, such as multiple myeloma (14) and prostate cancer with bone metastasis (15), FGF23 may directly enhance cancer progression. Elevated FGF23 levels are associated with poor survival in solid tumors with bone metastasis (16). While the pathogenic role of high FGF23 in other cancers is unclear, it may serve as a biomarker of tumor progression.
PHEX mutations in patients from neighboring countries. To better understand XLH pathogenesis, we examined relevant references from Japan (17) and South Korea (18), which are geographically near Taiwan. Integrating the mutations identified in our study with those from these references revealed a total of 94 distinct mutations (Figure 3). Of these, 15 were unique to Japan, 35 unique to Korea, 32 unique to Taiwan, 3 shared between Taiwan and Japan, 4 shared between Taiwan and Korea, and 5 shared among all three countries. No mutations were exclusive to both Japan and Korea (Figure 4).
Distribution of PHEX mutations identified in different countries. Blue corresponds to mutations in Taiwanese patients, green corresponds to mutations in Korean patients, purple corresponds to mutations in Japanese patients, Yellow corresponds to mutations in both Taiwanese and Korean patients, Light blue corresponds to mutations in both Taiwanese and Japanese patients, and red corresponds to mutations in all three countries.
Comparing PHEX mutations in neighboring countries (Japan and Korea). Of all the mutations identified in our study and the references, 15 were unique to Japan, 35 were unique to Korea, 32 were unique to Taiwan, three were shared by Taiwan and Japan, four were shared by Taiwan and Korea, and five were shared by all three countries.
Our analyses revealed shared PHEX mutations between Taiwanese, Japanese, and South Korean XLH cohorts. Notably, the G579R and R747X mutations appeared in multiple unrelated Taiwanese families (#1-3) (Table I) and were also reported in Japanese and Korean cohorts. Four PHEX variants (c.1601C>T, c.1769-1G>A, c.2104C>T, and c.2239C>T) appeared in multiple Japanese families, with c.1601C>T also found in Taiwan and c.2104C>T and c.2239C>T in both Taiwan and South Korea. Additionally, four recurrent PHEX mutations were identified in unrelated South Korean families (c.58C>T, c.931C>T, c.208_212delGTAAA, and c.1177_1178delAT). Of these, c.931C>T which was also found in Taiwan. Several known PHEX mutations, such as c.1601C>T, c.1735G>A, c.1645+1G>A, and c.2239C>T, were shared between these East Asian cohorts. The occurrence of these mutations across Taiwan, Japan, and South Korea indicates their prevalence as disease-causing variants underlying XLH in these countries. Further multi-ethnic analyses are warranted to elucidate pan-Asian PHEX mutation patterns and facilitate genotype-phenotype characterizations for informed prognosis and management for XLH patients.
Negative PHEX-detected cases analysis. Genetic screening was performed in 102 patients clinically suspected of hypophosphatemic rickets. PHEX mutations were detected in 74 patients, while 28 had no identified PHEX mutations. Among these 28 PHEX-negative cases, whole exome sequencing (WES) analysis using the InheriNext (Compass Bioinformatics, Inc.) platform identified pathogenic mutations in 8 cases. However, the underlying pathogenic cause remains unclear in the remaining 20 patients (Table III).
Negative PHEX-detected cases and their pathogenic gene mutations.
Different types of mutations and functional effects. Since key functional domains of the PHEX protein, including the transmembrane domain and zinc-binding motif, are present within residues 1-649 in exons 1-19 (19), mutations in this 5′ region may confer more severe phenotypic effects. To elucidate potential genotype-phenotype correlations, we categorized PHEX variants into two groups - 5′ region mutations spanning exons 1-19 up to residue 649, and 3′ region mutations from residue 650 in exon 20 to the 3′ end. Our analyses identified 64 different PHEX point mutations distributed throughout the coding regions and intronic sequences, but none in the 5′ or 3′ untranslated regions (Figure 1). The majority (78.1%, 50/64) of mutations occurred in the 5′ region encoding key functional domains, while 21.9% (14/64) localized to the 3′ region. These findings reveal a predominance of disease-causing PHEX mutations in the 5′ region encoding functionally important domains, though additional studies are warranted to characterize potential genotype-phenotype correlations. Elucidating such correlations between specific PHEX mutations and phenotypic severity may provide prognostic insights for informed clinical management of XLH patients in the future.
Our investigation uncovered 13 previously unreported frameshift mutations in the FGF23 gene, including 2 insertions and 11 deletions. One patient (case ID 75) displayed an exon 14 deletion and increased intact FGF23 levels. The majority of mutations were small insertions or deletions of less than 14 base pairs, with the exception of the exon 14 deletion. Mutated exons included 3, 5, 7, 11, 13, 16, 17, and 18, with deletions ranging from 1-13 bp (Figure 1). Frameshift mutations can introduce premature stop codons or loss of function in the PHEX protein (Table II).
Discussion
Our study identified 44 distinct PHEX mutations among 102 patients clinically diagnosed with hypophosphatemic rickets. Strikingly, 26 (59.1%) mutations were novel and not previously documented in ClinVar or the literature, highlighting the nature of PHEX mutations (3). Intriguingly, 6 of 35 male patients (17.14%) exhibited double PHEX mutation signals, suggesting possible mosaicism that warrants further investigation.
In summary, our findings reveal a high frequency of novel PHEX mutations in Taiwanese patients with hypophosphatemic rickets, indicate the need to evaluate mosaicism, and provide preliminary evidence linking specific PHEX mutations to tumorigenesis through extreme elevations in FGF23. Further mechanistic studies examining the roles of PHEX and FGF23 in bone malignancies are needed.
Most studies have not detected PHEX mutations in tumor DNA from patients with oncogenic osteomalacia, except one reporting changes at codons 363, 403 and 641 (20). The mutation sites differed from those in our patients. More extensive analyses are required to determine if the sites in our cases 16, 62 and 63 represent novel oncogenic mutations.
FGF23 has recently been revealed as a factor relevant to cancer, typically produced by benign mesenchymal tumors, although its role in neoplastic disease was unclear until recently (21, 22). The role of FGF23 is best characterized in malignant tumors causing tumor-induced osteomalacia, whereby overproduction of FGF23 by tumors leads to renal phosphate wasting and reduced 1,25(OH)2D3, resulting in osteomalacia (23). In cancers affecting the bone, such as multiple myeloma and prostate cancer with bone metastasis, FGF23 signaling may directly enhance cancer progression. Elevated FGF23 correlates with poorer survival of patients with solid tumors and bone metastasis (16). While the pathogenic role of high FGF23 in other cancers remains unclear, it may serve as a biomarker of tumor progression.
Implantation of Chinese hamster ovary cells expressing FGF23 subcutaneously into nude mice can induce tumor formation (24), providing an in vivo model to study oncogenic osteomalacia. Analyses of the PHEX gene for mutations in tumor cell DNA from patients with oncogenic osteomalacia have elucidated disease pathophysiology. However, most studies failed to detect PHEX coding mutations except one reporting changes at codons 363, 403 and 641 in a PHEX cDNA clone (20). The mutation sites differed from those in our cases (IDs 16, 62 and 63). Further extensive analyses are required to determine if the mutations in our cases represent novel oncogenic mutations.
Our study identified 8 distinct missense mutations. G579 appears to be a mutation hotspot in Taiwan, reported in approximately 20 unrelated patients of diverse ethnicities (11, 25, 26). We also detected this mutation in 22.58% of our cohort (including 3 families and sporadic cases). The G579R mutation occurs adjacent to the highly conserved zinc-binding motif (HEFTH) which is a signature feature of the M13 family of zinc metallopeptidases (27). Evidence indicates G579R variants are retained in the endoplasmic reticulum instead of being terminally glycosylated and expressed on the cell surface like wild-type PHEX (25). The G579R mutation in PHEX has been demonstrated to impair protein function through multiple mechanisms, including altered subcellular trafficking, reduced endopeptidase activity, and structural changes (26).
Nonsense and splicing site mutations are believed to generate truncated PHEX proteins, resulting in loss of function. R747X was one of the most common mutations in Taiwan, identified in 8.06% of patients (including 1 familial and 5 cases). R747X is highly conserved across mammalian species, implying this region is important for PHEX function (28). Even the last 3 amino acids appear critical for PHEX activity (28). Among 9 intronic mutations, 4 affected splice donors and 3 splice acceptors. These can significantly alter PHEX secondary structure through exon skipping, intron retention, and creation of new splice sites (29).
Our whole exome sequencing analyses identified several mutations in genes involved in collagen formation, a key process underlying bone matrix development. Specifically, mutations were found in COL1A1 and COL1A2, which encode the α1(I) and α2(I) chains of type I collagen, the predominant structural collagen comprising the organic matrix of bone (30). Over 85% of osteogenesis imperfecta patients have mutations in collagen genes family (31). COL2A1 encodes the alpha-1 chains of type II collagen, synthesized by chondrocytes (32). COL2A1 mutations cause type II collagenopathies including hypochondrogenesis and spondyloepiphyseal dysplasia congenita (33).
Other mutated genes are involved in ion transport. CLCN5 encodes the chloride channel ClC-5 (34). CLCN5 mutations cause X-linked renal tubular disorders including nephrolithiasis, hypophosphatemic rickets and low molecular weight proteinuria (35). TRPV4 encodes a mechanosensitive calcium channel contributing to osteocytic calcium influx (36). In mice, TRPV4 deficiency reduces bone’s mechanical sensitivity, decreasing bone formation and increasing bone resorption in response to unloading (37).
Finally, LRP5 is a well-established regulator of bone mineral density (BMD), with mutations causing disorders of both low and high BMD (38). Patients with osteoporosis-pseudoglioma syndrome, characterized by low bone mass, have loss-of-function LRP5 mutations (39). Many studies confirm the importance of VDR in the bone. VDR in osteoblasts negatively regulates bone mass, as impairments in bone formation and mineralization occur in VDR knockout mice (40).
Conclusion
In summary, this study delineates the clinical and genetic basis of XLH through retrospective review of medical records for 102 patients since 2006 to 2022. Among 28 PHEX-negative cases, whole exome sequencing revealed pathogenic mutations in 8 cases. Given the recent approval of burosumab, a monoclonal anti-FGF23 antibody, for XLH therapy, understanding the genotype and FGF23 profiles of Taiwanese XLH patients is crucial. This study provides key insights that will enable future expansions of the cohort to systematically characterize the clinical and genetic lineage of XLH in Taiwan and investigate potential correlations between circulating FGF23 and clinical/genetic features.
Acknowledgements
The Authors thank all people for supporting this study, including those delivering the samples of patients with XLH to our laboratory and contributing patient information.
Footnotes
Authors’ Contributions
Conceptualization: Pen-Hua Su and Ching-Yuang Lin; Methodology: Ju-Shan Yu; Software and validation: Yu-Shen Tsai; Writing—original draft preparation: Yu-Zhen Wu; Writing—review and editing: Pen-Hua Su; Visualization: Jo-Ching Chen; Supervision: Ching-Yuang Lin. Contributed data: Pen-Hua Sua, Fu-Sung Lo, Ju-Li Lin, Mei-Chyn Chao, Chia-chi Hsu, Yu-Yuan Ke, Pao-Chin Chiu, Jo-Ching Chen, Ying-Hua Huang, Shuan-Pei Lin, Yen-Yin Chou, Wei-Hsin Ting, Shuo-Yu Wang, Chiao-Fan Chiu, Yen-Chun Huang, Hui-Pin Hsiao, Chao-Hsu Lin, Chung-Hsing Wang, Da-Tian Bau, Ching-Yuang Lin. All Authors read and approved the final manuscript.
Funding
This research was funded by Chung-Shan Medical University Hospital, grant number CSH-2021-C-033 and by the China Medical University Hospital, grant numbers DMR-110-063, DMR-111-066, DMR-112-049, and DMR-112-161.
Conflicts of Interest
All Authors declare no conflicts of interest associated with this study.
- Received September 21, 2023.
- Revision received October 14, 2023.
- Accepted October 16, 2023.
- Copyright © 2024, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.