What Is FGF23? The Bone Hormone That Controls Phosphate

Endocrine physiology

FGF23 is the hormone that settles how much phosphate stays in your blood, and it is made in bone — which was not where anyone expected to find it. It was discovered twice within a year, from two different directions, and both times the clue came from patients whose kidneys were throwing phosphate away for no visible reason.

What is FGF23?

Fibroblast growth factor 23 is a 251-amino-acid protein encoded by the FGF23 gene on chromosome 12p13.32. The first 24 residues are a signal peptide, so the secreted chain runs from residue 25 to 251. Despite the family name, it does almost nothing that growth factors classically do. It is an endocrine hormone: made in one organ, released into blood, acting at a distance. UniProt still lists its older alias, phosphatonin — the name given to the then-hypothetical substance that tumours seemed to be secreting to cause phosphate wasting.

It is produced by osteocytes and osteoblasts — the cells embedded in and lining bone — particularly during active remodelling. That is the conceptual jolt: bone is not just a passive mineral reservoir being told what to do by the parathyroid glands. It sends its own instructions back. If you have read our pieces on parathyroid hormone and the parathyroid glands, FGF23 is the missing third voice in that conversation.

Two papers established it. In November 2000, the ADHR Consortium reported that autosomal dominant hypophosphataemic rickets — a family disease of low blood phosphate, rickets, bone pain and dental abscesses — is caused by missense mutations in a previously undescribed FGF family member. Six months later, Takashi Shimada and colleagues cloned the same gene out of a tumour causing oncogenic osteomalacia, and showed that giving recombinant FGF23 to mice dropped their serum phosphate within twelve hours. One team came from an inherited disease, the other from an acquired one, and they arrived at the same molecule.

One cut decides how much hormone you have

Most of the interesting regulation of FGF23 happens after it is made. The protein carries a recognition motif for proprotein convertases — the same class of enzymes covered in prohormone processing — at residues 176 to 179, reading arginine-histidine-threonine-arginine. Cutting between Arg179 and Ser180 splits the hormone into an N-terminal and a C-terminal fragment and inactivates it.

Two enzymes fight over that bond. GALNT3 attaches an O-linked sugar to Thr178, which physically blocks the convertase; UniProt states plainly that competition between convertase cleavage and this glycan block “determines the level of secreted active FGF23.” Pulling the other way, the kinase FAM20C phosphorylates Ser180 and slows the protective glycosylation down, favouring the cut.

Read the two diseases side by side and the mechanism is impossible to miss. The three ADHR mutations UniProt records — R176Q, R179Q and R179W — all sit inside that four-residue motif and all make the protein resistant to cleavage. The result is too much intact hormone and a lifetime of phosphate wasting. Lose the opposite side of the balance — inactivating mutations in GALNT3, FGF23 itself, or Klotho — and you get hyperphosphataemic familial tumoral calcinosis, in which phosphate is retained and calcified masses form in soft tissue. Same bond, opposite direction, mirror-image disease.

What the C-terminal fragment does afterwards is less settled. In vitro it competes with intact FGF23 for the receptor complex; whether that matters in a living body is, in the words of a 2018 Endocrine Reviews survey, “unclear.”

Why a hormone from bone acts almost only on the kidney

FGF23 binds FGF receptors, and FGF receptors are on nearly everything. So why does a hormone with an almost universal receptor have such a narrow job description? Because it needs a second protein to work.

In 2006, Urakawa and colleagues showed in Nature that Klotho — a protein until then studied mainly for its role in ageing — converts an ordinary FGF receptor into an FGF23 receptor. Klotho binds FGF23; forcing Klotho expression let a kidney cell line respond to FGF23 when it previously could not; and the pairing that reconstitutes the working receptor is FGFR1(IIIc) plus Klotho. The receptor is common. The co-receptor is not.

What FGF23 does once it lands

In the proximal tubule of the kidney, phosphate is reabsorbed by two sodium-phosphate cotransporters, NaPi-IIa and NaPi-IIc. FGF23 signalling pulls them off the apical membrane, so phosphate that would have been reclaimed is lost in urine instead. That is the primary effect, and everything else supports it:

• Phosphate goes out. Fewer transporters in the membrane means more phosphaturia.
• Vitamin D activation is shut down. FGF23 suppresses CYP27B1, the enzyme that makes active 1,25-dihydroxyvitamin D, and induces CYP24A1, which degrades it. Less active vitamin D means less phosphate absorbed from food — the same goal by a second route.
• Parathyroid hormone is suppressed — probably. The parathyroid glands express Klotho, and FGF23 has been reported to reduce PTH secretion. This one is genuinely contested: other work found persistent FGF23 signalling driving parathyroid cell proliferation and more PTH.
• Calcium is retained. In the distal tubule, FGF23 increases calcium reabsorption via the TRPV5 channel.

What switches FGF23 on is a shorter and more surprising list. Active vitamin D is one of the strongest stimuli. Dietary phosphate raises it, though whether the trigger is phosphate in the gut or phosphate in the blood is unresolved — acutely infusing phosphate into healthy volunteers does not move FGF23. PTH stimulates its production. And, unexpectedly, iron deficiency and erythropoietin both increase FGF23 transcription, which turns out to matter clinically.

The diseases that named the hormone

Excess FGF23 signalling produces a recognisable syndrome: low blood phosphate, urinary phosphate wasting, inappropriately low active vitamin D, and soft bones — rickets in children, osteomalacia in adults.

• X-linked hypophosphataemia (XLH) — the commonest form, roughly 1 in 20,000 births, caused by mutations in PHEX. More than 300 different mutations have been reported.
• ADHR — the cleavage-resistant FGF23 variants described above.
• Autosomal recessive forms — DMP1 and ENPP1 mutations.
• Tumour-induced osteomalacia — usually a small benign phosphaturic mesenchymal tumour, most often in a limb or the head and neck, quietly secreting FGF23. Finding it can require PET/CT, somatostatin-receptor imaging, or selective venous sampling of FGF23 from ten to thirty sites in the body.

Iron, and a side effect nobody predicted

Certain intravenous iron preparations cause hypophosphataemia, and FGF23 is the reason. The cleanest demonstration is PHOSPHARE-IDA: two identically designed randomised trials, 245 adults with iron-deficiency anaemia, comparing ferric derisomaltose against ferric carboxymaltose.

The mechanism the authors propose is the same cleavage balance again: iron deficiency raises FGF23 transcription, and something about the carbohydrate carrier in ferric carboxymaltose appears to block the cut, so the newly made hormone reaches the blood intact instead of being chopped. They say so carefully — the mechanism “remains unknown” — and the trials were funded by the company that makes the comparator drug. Both facts belong in any honest reading.

The downstream chemistry followed exactly as the physiology predicts: active vitamin D fell further in the carboxymaltose group, its degradation product rose from day 7 onward, ionised calcium dropped and PTH climbed. It is an accidental human experiment in FGF23 physiology.

Burosumab: from gene to drug in eighteen years

Burosumab (CRYSVITA) is a monoclonal antibody against FGF23. The FDA approved it on 17 April 2018 for XLH; the European authorisation followed on 19 February 2018. In the United States it is now indicated for XLH from six months of age and for FGF23-related hypophosphataemia in tumour-induced osteomalacia where the tumour cannot be resected or found.

The pivotal paediatric trial (published in The Lancet in 2019) is unusually informative because the comparator was not placebo but the old standard of care — oral phosphate plus active vitamin D, which these children had already been taking for a mean of four years. Sixty-one children, 64 weeks:

• Total Rickets Severity Score fell from 3.2 to 1.1 on burosumab versus 3.2 to 2.5 on conventional therapy.
• On the Radiographic Global Impression of Change, 21 of 29 burosumab children reached a score of at least +2.0 — the threshold for substantial healing — versus 2 of 32 on conventional therapy.
• Mean serum phosphorus rose from 2.4 to 3.3 mg/dL on burosumab; on conventional therapy it moved from 2.3 to 2.5, and the kidney’s phosphate threshold actually fell.

The label’s warnings follow directly from the biology. Oral phosphate and active vitamin D analogues must be stopped a week before starting, and using them together is contraindicated, because blocking FGF23 while still supplementing phosphate risks hyperphosphataemia and nephrocalcinosis. It is also contraindicated in severe renal impairment — where, as the next section explains, FGF23 is already abnormal.

Kidney disease, and the causality question

In chronic kidney disease, FGF23 is the earliest thing to change. In 3,879 patients from the Chronic Renal Insufficiency Cohort, raised FGF23 was more common than either high phosphate or secondary hyperparathyroidism at every level of kidney function, and it started rising at a higher eGFR than PTH did. It behaves like an early warning system: the kidney is losing capacity, so bone shouts louder to keep phosphate moving.

High FGF23 also tracks with bad outcomes. In a 2008 NEJM study of patients starting haemodialysis, those in the highest quartile had roughly a 5.7-fold odds of death in the first year compared with the lowest. And in 2011, Faul and colleagues showed FGF23 causing pathological hypertrophy of isolated rat heart cells and left ventricular hypertrophy in mice — notably without Klotho, meaning a different, off-target mechanism.

Why two FGF23 tests can disagree

There are two assays. The intact assay measures only full-length hormone. The C-terminal assay measures full-length hormone plus its C-terminal fragments. When cleavage is working normally, these two can move in opposite directions.

A 2011 study made the point elegantly. In ADHR patients, low serum iron correlated with both intact and C-terminal FGF23. In healthy controls, low iron raised C-terminal FGF23 but left intact FGF23 flat — because in people with an intact cleavage site, production rises and cutting rises with it, so the active hormone never accumulates. That is homeostasis working, visible only because the two assays disagree. Note also that reference intervals and units differ between them, so numbers from different studies are not interchangeable.

For diagnosis, an intact-FGF23 assay is used to distinguish FGF23-driven hypophosphataemia from other causes, with a proposed cut-off around 30 pg/mL against a normal range of roughly 10 to 50 pg/mL. The ranges overlap, and the interpretation is contextual: a “normal” FGF23 in someone who is frankly hypophosphataemic is not normal at all — it is inappropriately high for the circumstances.

Frequently asked questions

Is FGF23 the same as FGF21?

No. They are relatives in the endocrine FGF subfamily and both need a Klotho-family co-receptor, but they do different jobs and pair with different partners. FGF23 uses alpha-Klotho and governs phosphate; FGF21 uses beta-Klotho and acts on metabolism. Sharing a design principle is not sharing a function.

Why does bone control phosphate at all?

Because bone is where most of the body’s phosphate is stored, and osteocytes are well placed to notice when mineral supply and demand are out of step. Having the storage organ signal directly to the excreting organ closes the loop faster than routing everything through the parathyroid glands.

Should I get my FGF23 measured?

FGF23 testing is a specialist investigation, mainly used when someone has unexplained low serum phosphate and a clinician needs to know whether the cause is FGF23-driven. It is not a general-purpose bone or metabolic screening test, and there is no established role for it in routine health checks. That is a conversation for a physician, not a lab menu.

Can an iron infusion really lower phosphate?

Some intravenous iron formulations can, and ferric carboxymaltose is the best documented. In PHOSPHARE-IDA about three-quarters of patients given it dropped below 2.0 mg/dL, versus about 8% given ferric derisomaltose. It is a recognised, formulation-specific effect that clinicians monitor for — not a reason to avoid treating iron deficiency.

Informational only — not medical advice · 21+. Nothing here is dosing or diagnostic guidance. Disorders of phosphate metabolism are managed by specialists; any decision about testing or treatment belongs with a qualified healthcare professional.