Pharmacological management of X‐linked hypophosphataemia

Abstract

The most common heritable disorder of renal phosphate wasting, X‐linked hypophosphataemia (XLH), was discovered to be caused by inactivating mutations in the phosphate regulating gene with homology to endopeptidases on the X‐chromosome (PHEX) gene in 1995. Although the exact molecular mechanisms by which PHEX mutations cause disturbed phosphate handling in XLH remain unknown, focus for novel therapies has more recently been based upon the finding that the bone‐produced phosphaturic hormone fibroblast growth factor‐23 is elevated in XLH patient plasma. Previous treatment strategies for XLH were based upon phosphate repletion plus active vitamin D analogues, which are difficult to manage, fail to address the primary pathogenesis of the disease, and can have deleterious side effects. A novel therapy for XLH directly targeting fibroblast growth factor‐23 via a humanized monoclonal antibody (burosumab‐twza/CRYSVITA, henceforth referred to just as burosumab) has emerged as an effective, and recently approved, pharmacological treatment for both children and adults. This review will provide an overview of the clinical manifestations of XLH, the molecular pathophysiology, and summarize its current treatment.

Introduction

X‐linked hypophosphataemia (XLH) is a rare metabolic bone disease with significant clinical consequences in both children and adults, due largely to fibroblast growth factor 23 (FGF23) excess and resulting chronic hypophosphataemia. In the past, treatment has been restricted to administering phosphorus and activated vitamin D, with limitations in effectiveness, and significant risk for important adverse events. Over the past 2 decades research into XLH mechanisms has generated novel potential targets for treatment, including the recent regulatory approvals of burosumab‐twza/CRYSVITA (henceforth referred to just as burosumab) for clinical use in patients with XLH. This paper will review the clinical consequences of XLH, the pathophysiology involved and various approaches to its medical management.

Genetic cause

XLH is a rare bone disease affecting an estimated 1:20–25 000 persons 1, 2 (OMIM no. 307800). Deleterious mutations in the phosphate regulating gene with homologies to endopeptidases on the X chromosome (PHEX) 3 cause this X‐linked dominant condition. The female to male ratio is about 2:1. Sporadic cases frequently occur. Other genes having different inheritance patterns cause phenotypically similar disorders. PHEX mutations cause increased plasma FGF23 concentrations, leading to impaired renal reabsorption of phosphorus, hypophosphataemia and impaired activation of vitamin D 2, 4.

Clinical manifestations

XLH typically presents in early childhood with manifestations of rickets, although family history may lead to testing before visible features appear. Unfortunately, clinical laboratories often fail to report age‐appropriate normal ranges, leading to missed diagnoses, since adult normal phosphorus values are much lower than paediatric (especially infant) values 2. Both children and adults with XLH have low phosphorus concentrations for age, and impaired renal tubular phosphorus reabsorption, typically assessed as a low transport maximum for phosphorus adjusted for glomerular filtration rate (TmP/GFR) 2, 4. Serum alkaline phosphatase activity is typically increased for age in children, although this is more variable in adults 2, 4.

At birth, length is usually normal and legs are not usually bowed 5, 6, 7. Bowing and impaired linear growth become more evident with weight bearing, especially between ages 1 and 2 years 2, 5, 6, 8, 9. Childhood features include bowing of the femora and tibiae leading to genu varus or sometimes valgus appearance, widening of the growth plates at the end of long bones, abnormalities of the skull shape, sometimes bone pain and occasionally delays in gross motor milestones (such as difficulties in walking or running) 2, 9, 10. Radiographic growth plate abnormalities are similar to nutritional rickets, along with diaphyseal bowing, often with medial cortical thickening, and torsion of weight‐bearing long bones. Skeletal severity and the response to medical management are highly variable 2, 6, 10. Patients often require corrective surgical procedures to straighten lower extremities. Short stature and gait abnormalities persist into adulthood 2, 6, 8, 10, 11.

Cranial bones are also affected in XLH, most notably with frontal bossing and dolicocephally 12, 13, 14. Flattening of the cranial base leads to decreased depth of the posterior fossa predisposing to Chiari malformations 13, occurring in up to 44% of XLH patients 12. For some patients, craniosynostosis of the sagittal suture requires craniotomy 14.

XLH predisposes to dental abscesses due to a combination of intrinsic effects of Phex deficiency and associated hypophosphataemia leading to under‐mineralized dentin and cementum 15, 16, 17, 18. Severe dental disease (abscesses and periodontitis) affects 61–78% of patients with XLH 19, 20. Retrospective studies suggest that treating XLH with phosphate salts and active vitamin D decreases the occurrence of dental abscesses and periodontitis 19, 20.

XLH does not negatively influence lifespan, so one would expect at least three quarters of XLH patients to be adults. As such, although commonly thought of as a paediatric bone disease, adults bear a considerable burden of the consequences of this lifelong condition. The most debilitating features of XLH are experienced by adults: bone pain, pseudofractures, enthesopathy and osteoarthritis 11, 21, 22. Osteomalacic bone pain is common 22 and debilitating. Although complete fractures are not more common in XLH, patients with XLH are prone to pseudofractures 11. Pseudofractures occasionally occur in children, but are seen in up to half of adults in cross sectional studies 21, 22, although the lifetime incidence might be higher. Pseudofractures may progress to complete fractures and may require surgical intervention. Skeletal healing after fracture or surgery may be delayed.

Enthesopathy involves the calcification of tendons and ligaments, typically beginning near the bony attachment sites, and progressing to enthesophyte or osteophyte formation. The pathophysiology of enthesopathy in XLH is uncertain, although it may be due to direct effects of FGF23. Enthesopathy occurs in multiple hypophosphataemic mouse models having FGF23 excess 23, 24. FGF receptor 3 and klotho expression in fibrocartilage of the entheses make direct signalling at least theoretically possible 24.

Enthesopathy causes joint stiffness and limitations in range of motion which may be extreme. When involving the anterior and posterior spinal ligaments, patients lose flexion, extension and rotational movement of the neck and spine. Sometimes patients develop spinal nerve or spinal cord compression requiring surgical intervention, although the incidence of this is uncertain. Enthesopathy is more common in men, and with increasing age and BMI 19, affecting between 30 and 100% of adults 10, 11, 19, 21, 22, 24.

Osteoarthritis is probably a consequence of chronic abnormalities in joint alignment and gait causing degenerative changes in joint cartilage and bony articulation surfaces. Osteoarthritis might be exacerbated by enthesopathy. Joint replacements are common surgeries in XLH patients, although the frequency of such procedures is uncertain 25. Up to 82% of adults with XLH complain of joint pain and stiffness, and about 50% report using pain medications weekly to daily 19, 22.

Muscle weakness, a feature of hypophosphataemia, is present in both the Hyp mouse and patients with XLH 22, 26, 27. However, in XLH patients it is difficult to separate the effects on mobility from skeletal deformation, joint abnormalities, and bone pain from those due to muscle weakness.

Overall, the lifetime burden of illness is high in XLH, complicated often by a lack of understanding of the disease and its management by both patients and clinical providers. Not surprisingly, with the growth abnormalities, skeletal deformities, muscle weakness, and bone and joint pain, quality of life impairments are documented in both children and adults with XLH 28, 29, 30.

Molecular pathophysiology

XLH is fully penetrant with highly variable severity. The Hyp Consortium determined that XLH is caused by inactivating mutations in the PHEX gene 3. PHEX encodes a protein that structurally resembles the M13 family of membrane‐bound metalloproteases. Members of this enzyme class include single transmembrane proteins with large extracellular domains such as neutral endopeptidase and endothelin converting enzymes 1 and 2 (ECE‐1 and ECE‐2) 3, 31. This family is known to cleave small peptides, although the endogenous PHEX protein's substrate has yet to be fully confirmed. Close to 300 inactivating PHEX mutations have been described from XLH patients, including variations that lead to missense, nonsense, frame shift, and exon splicing alterations. PHEX shows the highest expression in bone cells including osteoblasts and osteocytes, and odontoblasts in teeth, as well as lower expression in the parathyroid glands, lung, brain and skeletal muscle 31, 32.

Although the PHEX mutations are predicted to cause loss of function, the mechanisms whereby this leads to changes in bone function remain elusive. The Hyp mouse model of XLH, has a 3′ deletion in PHEX, and has been a very useful tool for the in vivo study of the XLH phenotype 31. Similar to XLH patients, this model has approximately a 10‐fold increase in serum FGF23 (although in patients with XLH, serum FGF23 is typically less severely elevated), and manifests the XLH biochemical syndrome including hypophosphataemia with inappropriately normal 1,25‐dihydroxyvitamin D₃ [1,25(OH)₂D] and normocalcaemia 33. Like patients, the Hyp mice display growth retardation and bone mineralization defects. Interestingly, loss of Phex protein function was associated with a differentiation defect in osteocytes from Hyp mice 34. This mouse line had altered developmental transitions from osteoblasts to osteocytes, with inappropriate expression of cell matrix genes 34. Indeed Hyp bones had over‐expression of type I collagen, as well as altered expression of matrix proteins such as bone sialoprotein 34. Fgf23 mRNA is elevated in Hyp bone 33 and in isolated Hyp osteoblasts/osteocytes 35. These cultures also fail to mineralize fully, consistent with an intrinsic cell defect. However, the cellular pathophysiological mechanisms caused by Phex deficiency that result in altered FGF23 gene expression and sustained FGF23 protein elevation remain elusive.

FGF23 control of phosphate and active vitamin D

The primary kidney transport protein responsible for phosphate reabsorption in the proximal tubule is the type II sodium‐phosphate cotransporter, or NaPi‐IIa. Transgenic FGF23 mice had dramatic decreases in protein expression of NaPi‐IIa and a related transporter, NaPi‐IIc 36. The Hyp mouse has a 50% reduction in proximal tubule NaPi‐IIa expression 37. In normal individuals, phosphate depletion is a strong stimulus for increasing serum 1,25(OH)₂D 38. However, in XLH patients, hypophosphataemia is accompanied by low or inappropriately normal 1,25(OH)₂D concentrations, due to effects of FGF23. In mice implanted with cells expressing FGF23, or injected with FGF23 itself, mRNA and protein expression levels of the activating enzyme 25‐hydroxyvitamin D₃ 1‐α‐hydroxylase (Cyp27b1) was decreased, while the catabolic 1,25‐(OH)₂D 24‐hydroxylase (Cyp24a1) is increased with high FGF23 39, 40, 41. Thus, the effects of FGF23 on the renal vitamin D metabolic enzymes is responsible for the reductions in serum 1,25(OH)₂D concentrations observed in XLH patients.

Patients with autosomal dominant hypophosphataemic rickets (ADHR) also have elevated FGF23 but some patients display waxing and waning of the disease symptoms 42. Studying a large ADHR kindred, we documented that the ADHR disease state and circulating levels of intact FGF23 correlated strongly with iron deficiency and were reciprocal to serum iron concentrations 43, which was borne out in an ADHR knock in animal model to be due to increased bone Fgf23 mRNA production during experimental iron depletion 44, 45. However, intact FGF23 levels in XLH patients were unrelated to serum iron concentration 46. A clinical trial is underway to place ADHR patients on low‐dose oral iron repletion (ClinicalTrials.gov NCT02233322), thus this regimen may be the optimal treatment for ADHR, but not XLH (see below).

Medical management

Phosphate salts and active vitamin D

For about 4 decades, children and adults with XLH have been medically managed using pharmacological doses of phosphate salts and active vitamin D, attempting to counter the effects of FGF23 excess. Doses reported vary widely in the literature 2. A US‐based group of experts recommended target doses of calcitriol 20–30 ng kg^–1 daily and phosphate 20–40 mg kg^–1 daily, each in divided doses 2. European experts recommended 1–3 μg day^–1 of alfacalcidiol or 0.5–1.5 μg day^–1 of calcitriol, and 40–60 mg kg^–1 daily of phosphate 10. However, no study has systematically evaluated the optimal target dose ranges. Clinicians start at lower doses and titrate to target doses primarily based on gastrointestinal tolerance, with frequent laboratory monitoring to avoid hypercalciuria, hypercalcaemia, hyperphosphataemia or impairments in renal function. This strategy does not effectively normalize the serum phosphorus. In fact, attempts to target normal phosphorus concentrations might be a contributing reason for the high risk of hyperparathyroidism and nephrocalcinosis with these agents 2. Monitoring patients on treatment is complicated by many factors, and most laboratory monitoring addresses safety (serum calcium, phosphorus, creatinine, PTH, urine calcium, urine creatinine).

Treatment with active vitamin D and phosphate decreases alkaline phosphatase and, in bone biopsy studies, improves the osteomalacia of XLH 47, 48. However, the skeletal response varies widely. Some children straighten leg deformities and improve their growth, while others persist in skeletal deformities and short stature. The reason for good responders vs. poor responders is not solely an issue of compliance 49. Treating XLH with vitamin D analogues and phosphate also increases plasma FGF23 concentrations 50, 51, 52, which theoretically could blunt therapeutic effectiveness, although the true consequences remain unknown.

Many children with XLH have a variable but progressive decline in height Z‐scores, often worsening during puberty despite therapy with vitamin D analogues and phosphate 8. Studies suggest that beginning treatment prior to age 1 year improves height outcomes 6.

When leg deformities persist, the optimal timing of corrective surgery remains uncertain. In a retrospective study, 29% of XLH patients had recurrence of deformity after their first corrective surgery 53. Although the differences were not significant, those having earlier corrective procedures were more likely to require additional procedures. This finding may be confounded by more severely affected patients having earlier surgeries.

Unlike nutritional rickets, which is truly cured by adequate vitamin D repletion, the homeostatic defect in phosphorus handling is lifelong in XLH. However, because of the known risks of XLH therapy with active vitamin D and phosphate, general practice has been to stop treatment at the end of growth, and restart if developing clinical issues associated with active osteomalacia 2. This has led to some confusion over whether or not adults benefit from treatment, and misinterpretation of XLH as a childhood disease. Although some adults tolerate stopping treatment with few symptoms, at least for a time, many develop active osteomalacic symptoms, bone pain, muscle weakness and pseudofractures, which demonstrate varying degrees of improvement during treatment with active vitamin D and phosphate 22.

Important side effects limit this therapy. Doses of phosphate may be limited by gastrointestinal symptoms, nephrocalcinosis, ectopic calcification and hyperparathyroidism, while active vitamin D may also contribute to nephrocalcinosis or other ectopic calcification. Nephrocalcinosis is reported in 50–80% of XLH patients receiving active vitamin D and phosphate, and appears to relate to episodes of hypercalciuria 54, 55, 56, 57. Most nephrocalcinosis is mild, but occasionally patients develop chronic kidney disease (CKD), the prevalence of which is uncertain, but progression to end‐stage renal disease is rare. Hypertension has been reported in 27% of XLH patients and may relate in part to nephrocalcinosis and CKD 58.

Hyperparathyroidism in XLH is complex. Nearly half of treatment naïve children with XLH have elevated serum PTH 59, which often decreases after starting active vitamin D. Serum PTH correlates with FGF23 concentrations in untreated patients 52. However, phosphate doses also lead to secondary hyperparathyroidism, probably through transient decreases in serum calcium following every dose. Some patients with secondary hyperparathyroidism progress to tertiary hyperparathyroidism with multigland hyperplasia and require surgical intervention to control hypercalcaemia 60, 61. Prior to the development of parathyroid autonomy, adequate doses of active vitamin D and lowering doses of phosphate may combat the rise of PTH.

Novel targets

Given the limitations of effect and the significant risks of treating XLH with active vitamin D and phosphate salts, the need for improved treatment options has led to development of other strategies. We will focus on those targets that have some clinical/translational data in humans.

Calcimimetics

PTH stimulates FGF23 expression in osteocytes 62, which could exacerbate hypophosphataemia in XLH. Since hyperparathyroidism also complicates XLH treatment, cinacalcet, a calcimimetic, has been used to manage hyperparathyroidism in cases of XLH 63, 64. Short term studies of cinacalcet in children with XLH demonstrated suppression of the PTH surge that follows doses of phosphate 65, and improved TmP/GFR and serum phosphorus. In a case report, one child developed a severe aversion to, and refusal of, his oral phosphate doses. Although he did not have hyperparathyroidism, treatment with cinacalcet, calcitriol and calcium led to improvements in serum phosphorus and rickets, although serum FGF23 still increased 66. Further studies are needed to determine what role cinacalcet should play, although it may be useful in tertiary hyperparathyroidism.

Calcitonin

Calcitonin stimulates 1α‐hydroxylase in the proximal renal tubule in Hyp mice 67, and transiently increases 1,25(OH)₂D in subjects with XLH after a single injection 68. Osteocytes also express the calcitonin receptor 69, and as such, calcitonin could influence FGF23 secretion. Calcitonin decreased serum FGF23 in a patient with tumour‐induced osteomalacia (TIO) 70. A single calcitonin injection decreased serum FGF23 and increased serum phosphorus transiently in patients with XLH, although not in healthy controls 71. To make the calcitonin story more confusing, case reports indicate calcitonin increases urinary phosphorus excretion in hyperphosphatemic tumoural calcinosis (a condition of FGF23 deficiency rather than excess) 72, which would be opposite of the goal of XLH management. A 3‐month‐long blinded randomized controlled clinical trial of monotherapy with nasal calcitonin 400 units daily for XLH failed to demonstrate improvements in serum phosphorus, TmP/GFR or FGF23 73. Given the current lack of data indicating benefit, calcitonin should not be used for XLH.

Fibroblast growth factor receptor antagonism in TIO

TIO (OMIM 605380) is another syndrome of excess fibroblast growth factor receptor (FGF23) 39, 74. These tumours are classified under the collective term of phosphaturic mesenchymal tumour, mixed connective tissue variant (PMTMCT) or phosphaturic mesenchymal tumour 75. Complete tumour resection is the most straight forward cure for TIO. For tumour localization, imaging techniques using radiolabelled octreotide 76 as well as magnetic resonance imaging 77, computed tomography 78, whole body sestamibi scanning 79, and Ga68‐DOTA‐octreotide positron emission tomography/computed tomography imaging 80 may be useful. Selective venous sampling for FGF23 levels has been attempted to locate tumours as well 77, 81. Unfortunately, many PMTMCTs are small in size and remain difficult to localize, necessitating medical management.

In a group of 15 TIO tumours analysed through next‐generation sequencing, a fibronectin (FN1)‐FGFR1 fusion gene was detected in 60% of the tested samples 82. The fusion protein is predicted to express portions of the three extracellular FGF‐binding (Ig‐like) domains 82, therefore ligand‐activated receptor signalling could perhaps occur. It is not established however, whether the fusion gene is causative of the TIO tumour or is a consequence of the tumourigenesis. Activating FGFR1 mutations are associated with osteoglophonic dysplasia, a disease of dwarfism as well as craniosynostosis 83. Some osteoglophonic dysplasia patients also have significantly elevated FGF23 and hypophosphataemia. Thus, inhibitors of FGF binding or FGFR activity could be useful pharmacological treatments for TIO. Such inhibitors may block FGFR auto‐ and ligand‐ and Klotho‐dependent FGFR dimerization, and may reduce signalling through FGFRs along with directly inhibiting FGF23 bioactivity. However, the ubiquitous nature of FGFR1 could lead to off‐target effects. Considering the difficulty with targeting FGFR inhibitors to a specific tissue, another plausible approach for treating TIO patients until the tumour is found may be monotherapy with anti‐FGF23 antibody (see below) as is currently being tested in TIO and XLH (ClinicalTrials.gov NCT02304367).

FGF23‐blocking antibodies

Investigators at Kirin Pharma developed murine anti‐FGF23 monoclonal antibodies that bind separate isotopes on either the C‐terminal or N‐terminal portions of FGF23 84, providing information on the structural interactions of FGF23 with its receptor/co‐receptor. Both antibodies interfered with FGFR signalling as indicated by declines in Egr‐1 reporter activity (a molecular marker for MAPK activation). The N‐terminal antibody interfered with binding of FGF23 to the FGFR, while the C‐terminal antibody interfered with binding to the coreceptor klotho. Injecting either antibody into mice produced generally similar effects: transient increases in 1,25(OH)₂D beginning within 3–5 hours, and in TmP/GFR and serum phosphorus after 16 h 84. Combining the antibodies was synergistic.

Hyp mice receiving subcutaneous injection of these antibodies in combination demonstrated a similar biochemical response and time course after a single dose 85. When juvenile, 4‐week‐old Hyp mice were treated with weekly injections for 5 weeks, renal sodium phosphate cotransporter expression increased on immunohistochemistry, and serum phosphorus improved in a dose‐dependent manner along with increases in Cyp27b1 mRNA expression and serum 1,25(OH)₂D. As a result, mice had bone histomorphometric improvements, including decrease in osteoid thickness and volume toward normal, as well as increased mineral apposition rate, bone formation rate, and bone volume/total volume (BV/TV). This translated into improvements in the rachitic growth plate morphology, the growth and shape of their tails, femora and tibiae, and in overall size of the mice 26, 85. Treating adult Hyp mice with anti‐FGF23 antibodies also improved osteomalacia on histology, and increased muscle grip strength toward normal without changing muscle weight 26.

Based on successes in the Hyp mouse, a fully human IgG1 monoclonal antibody to bind FGF23 was developed for human trials (KRN23, now called burosumab or CRYSVITA). Table 1 summarizes key findings from the published clinical trials. In the initial dose finding study, single doses of intravenous or subcutaneous burosumab (or placebo) were administered to adults with XLH after stopping phosphate or active vitamin D 86. Burosumab increased TmP/GFR, serum phosphorus and 1,25(OH)₂D in a dose dependent manner. The half‐life was longer for subcutaneous dosing (13–19 days) than with intravenous dosing (8–12 days). Baseline FGF23 concentration did not predict AUC changes for any pharmacodynamic parameter.

Table 1.

Key findings from the primary burosumab clinical trial publications in X‐linked hypophosphataemia (XLH) to datea

Study [ref] /ClinicalTials.gov number	Carpenter et al. JCI 2011 86 NCT00830674	Imel et al. JCEM 2015 87, Ruppe et al. Bone Reports 2016 28 NCT01340482, NCT01571596	Carpenter et al. NEJM 2018 29 NCT02163577	Insogna et al. JBMR 2018 21 NCT02526160
Population	Adults with XLH	Adults with XLH	Children with XLH, ages 5–12 years at enrollment	Adults with XLH and Brief Pain Inventory worst pain score ≥ 4
Treatment arms	Single dose   Intravenous burosumab 0.003–0.3 mg kg–1 (n = 17) vs. placebo (n = 5) or   Subcutaneous burosumab 0.1–1 mg kg–1 (n = 12) vs. placebo (n = 4)	Multidose for 16 months:   4‐month dose escalation phase Every 4 weeks burosumab 0.05–0.6 mg kg–1 (n = 28),   followed by   12‐month extension phase Every 4 weeks burosumab 0.1–1 mg kg–1 (n = 22) b	Multidose for 64 weeks, with dose titration period c Every 2 weeks burosumab (mean dose at week 40 was 0.98 mg kg–1) (n = 26) or Every 4 weeks burosumab (mean dose at week 40 was 1.5 mg kg–1; n = 26)	Multidose for 24 weeks:   Every 4 weesk burosumab 1 mg kg–1 (n = 68)   or   Every 4 weeks placebo (n = 66)
Primary outcome	Safety and tolerability	Proportion of subjects achieving maximum fasting serum Pi in normal range	Change in RSS from baseline to week 40 and 64	Percent of subjects achieving mean serum Pi in the normal range across the midpoint of dosing intervals
Biochemical findings	Increased serum Pi, serum 1,25(OH)2D3, and TmP/GFR	Increased serum Pi, serum 1,25(OH)2D3, and TmP/GFR	Increased serum Pi, serum 1,25(OH)2D3, and TmP/GFR; decreased serum alkaline phosphatase.	Increased serum Pi, serum 1,25(OH)2D3, and TmP/GFR
Radiographic outcomes			Improved rickets by RSS and RGI‐C. Greater changes seen with every 2‐week dosing, and in those with higher baseline RSS	Improved healing of active fractures/pseudofractures from baseline: burosumab (43.1% of fractures healed) compared to placebo (7.7%)
Other outcomes		At baseline: Impairment of physical function   Over 4 months of treatment: Improved role limitations due to physical health (SF‐36v2), physical functioning and stiffness (WOMAC)	Increased standing height z‐score;   Among those with baseline impairments in 6MWT, the distance walked improved   On PODCI, among those with impairments, burosumab improved: Sports and physical functioning domain, Pain and comfort domain, Global functioning	WOMAC: Decreased stiffness scores

^aIn each of these trials, subjects had a washout period from phosphate salts and active vitamin D analogues, and burosumab (or placebo) was given as monotherapy. Dose titration was based on serum phosphorus concentration, targeting normal serum phosphorus ranges

^bDuring the extension phase over 80% of subjects received doses of 0.6–1 mg kg^–1 87

^cChildren in this trial started initially at very low doses, which were titrated based on serum phosphorus concentration, and sequential entry cohorts began at higher doses 29

Pi, phosphorus; PODCI, Pediatric Outcomes Data Collection Instrument; RGI‐C, Radiographic Global Impression of Change scale for rachitic features (positive scores indicate improvement); RSS, Thacher Rickets Severity Score; Sf‐36v2, Medical Outcomes Study Short Form Health Survey version 2; TmP/GFR, transport maximum phosphorus adjusted for glomerular filtration rate; WOMAC, Western Ontario and McMaster Osteoarthritis Index; 1,25(OH)₂D₃ 1,25‐dihydroxyvitamin D₃; 6MWT, 6‐min walking test

Twenty‐eight adults with XLH stopped all XLH medications and enrolled in a Phase 2 dose‐escalating multidose trial of burosumab administered subcutaneously every 4 weeks, with doses ranging from 0.05 to 0.6 mg kg^–1 and 22 subjects continued in a 12‐month extension with doses ranging from 0.1–1 mg kg^–1 87. During the multidose trial, burosumab exhibited first‐order absorption and elimination kinetics, similar to other monoclonal antibody therapies, with an elimination half‐life of 17.8 days 88.

The peak value of 1,25(OH)₂D and serum phosphorus occurred about 3–7 days after injection, while peak TmP/GFR occurred about 7 days after injection 87. After each peak, the TmP/GFR, serum phosphorus and 1,25(OH)₂D decreased toward a trough level 4 weeks after injection that remained generally above the baseline values. Peak and trough values increased during the dose escalation phase, then remained generally stable through the 12 month extension, except for the peak of 1,25(OH)₂D, which decreased in magnitude over later doses 87. Most subjects at least transiently achieved normal serum phosphorus concentrations.

Despite the transient increase in 1,25(OH)₂D during each cycle, there was no systematic change in serum or urine calcium or in PTH. Kidney function also remained stable. Ten subjects had baseline nephrocalcinosis or nephrolithiasis, and no subject had worsening nephrocalcinosis 87.

Although this first multidose trial did not assess clinical bone outcomes, the bone biomarkers P1NP, osteocalcin and bone alkaline phosphatase increased during the study period. Quality of life instruments (SF‐36v2 and WOMAC) indicated baseline impairments in bodily pain, physical function and role limitations, along with stiffness. After 4 months there were improvements in the patients' perception of their physical functioning and stiffness 28.

In a recently published, open‐label, phase 2 trial, 52 children aged 5–12 years with XLH in Europe and the USA were randomized to receive burosumab subcutaneously every 2 weeks or every 4 weeks for 64 weeks 29. Patients were mostly prepubertal, although Tanner stage 2 subjects were allowed. All patients stopped phosphate and vitamin D analogues 2 weeks prior to screening. Since this was the first trial conducted in children, the initial cohort started at very low burosumab doses and titrated upwards targeting serum phosphorus levels between 3.2 and 6.1 mg dl^–1 (1.0–2.0 mmol l^–1). Rickets was assessed using a rickets severity score (RSS) that rated radiographic appearance at the wrist and knee on a scale from 0–10 (10 being the worst). Differences between baseline and follow‐up radiographs were also assessed using a seven‐point ordinal Radiographic Global Impression of Change (RGI‐C) scale (ranging from –3 severe worsening, to 0 no change, to +3 complete healing) 89.

One limitation of this paediatric study was its lack of a control group. However, all but one subject had been receiving standard treatment with active vitamin D analogues and phosphate salts up to the time of enrolment (for a mean of 6.9 years prior treatment). Thus, the baseline rickets severity at enrolment (ranging from RSS 0–4.5, mean 1.8 ± 1.1) represents residual rickets despite prior treatment. Any improvements seen in trial occurred in the setting of switching from prior standard therapy to burosumab.

At week 40 the mean burosumab dose had been titrated to 0.98 mg kg^–1 every 2 weeks or 1.5 mg kg^–1 every 4 weeks 29. Children receiving every 4 weeks dosing had pharmacodynamic profiles of serum phosphorus, 1,25(OH)₂D and TmP/GFR similar to that seen in adults 29. However, the every 2 weeks dosing group had more sustained improvements in these parameters, with less pronounced biochemical trough effects, and serum phosphorus values more consistently in the normal range. Alkaline phosphatase is commonly elevated in XLH patients (mean 459 ± 105 units l^–1 at baseline of this trial), which decreased by 20% at week 64. Both groups demonstrated improvements in rickets by week 40 that were sustained at week 64. Rickets improvements were numerically greater in those receiving every 2 weeks injections and also in those having worse baseline rickets severity. Total RSS score decreased from mean 1.9 to 0.8 after every 2 weeks dosing, and from 1.7 to 1.1 after every 4 weeks dosing. The mean total RGI‐C was +1.57 at the end of study, indicating healing of rickets. Substantial healing of rickets (RGI‐C ≥ +2) occurred in approximately half the patients in each group, but in 94% of those with baseline RSS ≥1.5 that received burosumab every 2 weeks. In addition, small improvements in standing height Z‐score were observed (+0.15 ± 0.04). Additional functional improvements were noted. Impairments in 6‐min walking distance were present at baseline in 46% of patients, and among these the 6‐min walking distance improved by 10% of the predicted normal range for age. Similarly, baseline scores indicated impairments for sports and physical functioning, and in pain and comfort domains in 54% of the children using the paediatric outcomes data collection instrument, which were seen to improve during treatment.

In a Phase 2 trial, 13 children with XLH, age 1–4 years, were enrolled to receive open‐label burosumab 0.8 mg kg^–1 every 2 weeks. This trial indicated similar improvements in serum phosphorus, alkaline phosphatase and rickets severity to that seen in the 5–12‐year‐old children 90, 91. Some details of this trial have been presented at scientific meetings and are indicated in the Food and Drug Administration (FDA) drug label 90, 91, but have not yet been fully published.

The pivotal Phase 3 adult trial randomized 134 adults with XLH to receive burosumab at doses of 1 mg kg^–1 (maximum 90 mg) every 4 weeks vs. a placebo for 24 weeks (without an active comparator treatment), after which all subjects received burosumab for the subsequent 24 weeks 21. Subjects were required to have a brief pain inventory worst pain ≥4 to enrol. Nearly all subjects (99%) had radiographic evidence of enthesopathy, half had nephrocalcinosis, while 47% in the burosumab group and 57% in the placebo group had pseudofractures/fractures with several subjects having multiple pseudofractures. The primary endpoint (the percent of patients achieving mean serum phosphorus concentrations in the normal range across the midpoint of dosing intervals) was achieved in 94% of subjects after burosumab vs. 7.6% after placebo (Figure 1A). Trough serum phosphorus remained in the normal range in 67.6% of burosumab treated subjects.

Figure 1.

Data from a 24‐week randomized, placebo‐controlled trial of burosumab 1 mg kg^–1 vs. placebo in adults with X‐linked hypophosphataemia are shown. (A) Mean serum phosphorus is within the normal range at the midpoint of dosing interval during treatment with burosumab, compared with placebo treated subjects, whose phosphorus remained low. (B) Active fractures or pseudofractures were more likely to heal during 24 weeks of burosumab compared with placebo. Adapted and reproduced from Figures 1A and 3A of Insogna et al. J Bone Miner Res 2018;33:1383–1393; (Reference [21]); Wiley Publishers, https://doi.org/10.1002/jbmr.3475. Creative Commons Attribution License (CC BY 4.0, https://creativecommons.org/licenses/by/4.0)

The burosumab group had improvements at week 24 in the brief pain inventory worst pain score, and WOMAC physical function and stiffness subscales compared to placebo, but only stiffness remained significant after adjustment for multiple comparisons. Most importantly 43.1% of baseline fractures/pseudofractures were healed after 24 weeks in the burosumab group vs. 7.7% in the placebo group (Figure 1B).

Regarding safety, the most common drug‐related adverse events have been local injection site reactions that were often characterized as welts, and usually lasted a few to several hours, resolving without treatment. Injection site reactions occurred in 57.7% of children 29 and 11.8% of adults 21, although, interestingly, injection site reactions also occurred in 12.1% of placebo treated adults. Additional adverse events occurring in ≥25% of children included headache, vomiting, pyrexia, extremity pain and decrease in vitamin D. Additional adverse events occurring in ≥10% of adults included back pain, tooth abscess, nasopharyngitis, headaches, nausea and dizziness. However, most listed adverse events were not considered drug related, while some are consistent with symptoms common to XLH patients. Restless legs syndrome was noted in the initial multidose trial (5/28 or 17.9%) 87. Restless legs syndrome occurred more often in burosumab‐treated subjects in the placebo controlled adult study (8/68 or 11.8%) but was also present to some degree in the placebo arm (5/66 or 7.6%) 21. No serious adverse events were attributed to burosumab itself.

There has been no systematic evidence of change in PTH, serum calcium or urine calcium excretion in any trials. Nephrocalcinosis was present at baseline in 35% of children in the Phase 2 trial 29 and 54% of adults in the Phase 3 trial 21. So far, there has been no clear indication of either worsening or improvement of nephrocalcinosis overall with burosumab treatment. In fact, similar numbers of patients had increases in nephrocalcinosis score in the burosumab and placebo groups, while similar numbers also had decreases in nephrocalcinosis scores in each group 21. No significant changes in echocardiograms were reported.

An advantage of burosumab is that it achieves normalization of serum phosphorus through improvements in TmP/GFR, in contrast to standard therapy with active vitamin D and phosphate salts. One hopes that this will translate into lower long‐term risk of nephrocalcinosis and hyperparathyroidism, provided that episodes of hyperphosphataemia are avoided. In the phase 3 placebo controlled adult trial, five patients (7.4%) had dose reductions according to protocol, due to serum phosphorus above the target range, four of which were classified as hyperphosphataemic 21. These subjects maintained in the trial on smaller doses. Thus, some patients may respond well to smaller doses to manage XLH.

Conceptually, anti‐drug antibodies could bind and interfere with the action of burosumab. The initial multidose trial in adults, and the Phase 2 trial in children did not find anti‐drug antibody development 29, 87. However, one published trial (and the drug labels) notes that anti‐drug antibodies have been detected in small numbers of subjects at baseline before burosumab administration 21, 91. Anti‐drug antibodies persisted in some subjects, but further development of antibodies during trial participation, was not seen. To date, there is no evidence for neutralizing anti‐drug antibodies attenuating effect of burosumab in human subjects.

Patients with moderate to severe CKD were excluded from the burosumab trials. In a rat model of severe CKD, treatment with a different FGF23 antibody hastened death, probably due to the frank hyperphosphataemia that developed 92. Thus, the use of burosumab in severe kidney disease is not recommended, and hyperphosphataemia should be avoided. The safety of burosumab (or for that matter of phosphate and active vitamin D) in pregnancy has not been established.

Burosumab was approved as monotherapy in 2017 by the European Medicines Agency (EMA; conditionally) and in 2018 by the FDA (fully). The EMA conditionally approved use of burosumab in children aged 1 year and older with radiographic evidence of bone disease, through adolescence during skeletal growth. The EMA approved starting dose is 0.4 mg kg^–1 subcutaneously every 2 weeks in children with 0.8 mg kg^–1 as typical maintenance dosing, and a maximum dose of 90 mg 93. The FDA approved starting dose is 0.8 mg kg^–1 every 2 weeks in children aged 1 year and older, and 1 mg kg^–1 every 4 weeks in adults with a maximum dose of 90 mg 91. Doses are titrated on the basis of peak and trough serum phosphorus concentrations.

Several questions remain. A true active comparator study in growing children aged 1–12 years is underway (ClinicalTrials.gov NCT02915705). This trial will identify the effect of 64 weeks of burosumab vs. phosphate and active vitamin D on multiple clinical parameters with a primary outcome of differences in healing of radiographic rickets. The effect on adult height in growing children is unknown. However, this will take years to evaluate, given the mean ages of children enrolled in the clinical trials. Since enthesopathy and other joint related complications develop and progress slowly, many years of treatment with burosumab will be necessary to interpret whether or not these outcomes are altered. Similarly, years of treatment are likely to be necessary to determine whether the chronic risk of nephrocalcinosis changes.

Conclusion

XLH is a complex musculoskeletal condition, requiring careful medical and surgical management. Recent approval of burosumab represents a substantial advancement in management of XLH, directed at its pathophysiology of FGF23 excess. Further studies are needed to fully characterize long‐term risks and benefits of this drug, especially regarding disease features that take years to develop.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 94, and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 95, 96.

Competing Interests

E.A.I. receives research funding that contributes to this work from the NIH/NIAMS 1P30AR072581. K.E.W. would like to acknowledge support by NIH grants DK063934, DK095784, and AR059278. E.A.I. has received research funding and fees for consulting with Kyowa Hakko Kirin and Ultragenyx, Pharmaceuticals. K.E.W. receives royalties from Kyowa Hakko Kirin Co. Ltd for licensing FGF23.

Imel E. A., and White K. E. (2019) Pharmacological management of X‐linked hypophosphataemia, Br J Clin Pharmacol, 85, 1188–1198. doi: 10.1111/bcp.13763.