Iron, inflammation, and fibroblast growth factor 23: Integrating molecular crosstalk, cardiovascular biology, and therapeutic perspectives

Abstract

This narrative review examines how iron availability, inflammation, and erythropoiesis regulate fibroblast growth factor 23 (FGF23), and how these interactions inform cardiovascular biology and therapy. The central framework separates FGF23 production from proteolytic processing: Increased synthesis can coexist with efficient cleavage and little change in intact hormone, whereas insufficient cleavage permits biologically active FGF23 to accumulate. C-terminal assays detect both intact hormone and cleavage products and therefore require different interpretation from intact assays. Experimental studies identify erythroid cells as an inducible FGF23 source and suggest distinct actions of intact FGF23 and C-terminal fragments on erythropoiesis, inflammation, and iron availability. Cellular and animal studies also support cardiac and vascular effects, but inconsistent experimental findings, heterogeneous assays, and predominantly observational human evidence limit causal inference. Iron repletion generally reduces the stimulus for FGF23 production, whereas ferric carboxymaltose can increase intact FGF23 and cause renal phosphate wasting. Global FGF23 neutralization caused hyperphosphatemia, vascular calcification, and increased mortality in experimental chronic kidney disease. Correcting upstream disturbances and testing selective signaling interventions are therefore priorities, but cardiovascular efficacy and safety remain unproven.

Key Words: Fibroblast growth factor 23; Iron deficiency; Inflammation; Hepcidin; Erythropoiesis; Cardiovascular disease; Chronic kidney disease; Hypophosphatemia

Core Tip: Distinguishing fibroblast growth factor 23 production from proteolytic processing explains discordant intact and C-terminal assay results and formulation-specific effects of intravenous iron on phosphate. Experimental cardiac, inflammatory, and erythroid mechanisms require validation in humans. Therapeutic studies should test causality while preserving phosphate homeostasis, because global hormone neutralization caused harm in experimental chronic kidney disease.

INTRODUCTION

Fibroblast growth factor 23 (FGF23) is a bone-derived endocrine hormone produced mainly by osteocytes and osteoblasts. The best-characterized actions of FGF23 occur in the kidney, where signaling through fibroblast growth factor receptors (FGFRs) and the co-receptor α-Klotho promotes phosphaturia and suppresses 1,25-dihydroxyvitamin D (calcitriol) synthesis[1], and in the parathyroid glands, where FGF23 suppresses parathyroid hormone (PTH) secretion and gene expression[2]. Beyond mineral metabolism, FGF23 is increasingly recognized as part of a broader network involving iron homeostasis and inflammation[3,4]. Circulating FGF23 concentrations have also been associated with adverse cardiovascular and metabolic phenotypes, raising the possibility that FGF23 both reflects, and in some settings contributes to, broader systemic disturbances[5].

Circulating FGF23 is regulated at the production and post-translational levels. Phosphate, calcitriol, PTH, and calcium are established regulators[6,7]. Iron deficiency also increases FGF23 transcription in experimental models, but enhanced cleavage can limit the rise in intact hormone[8]. The corresponding human pattern is illustrated by women with iron-deficiency anemia, who may have markedly elevated C-terminal assay results with little change in intact FGF23 (iFGF23)[9].

Inflammation can increase FGF23 production through partially overlapping mechanisms. Acute and chronic inflammatory stimuli directly stimulate FGF23 transcription in bone, although enhanced proteolytic processing may initially limit the rise in circulating iFGF23[10]. Tumor necrosis factor (TNF) is one of the cytokines shown to contribute directly to this response, including in non-renal inflammatory models[11]. At the same time, inflammation alters systemic iron handling through the hepcidin-ferroportin axis. Thus, interleukin-6 (IL-6) induces hepatic hepcidin production, reducing ferroportin-mediated iron export and leading to hypoferremia and functional iron restriction[12].

Erythropoietic activity adds another layer. Acute blood loss raises FGF23 in parallel with erythropoietin (EPO)[13], and recombinant EPO increases C-terminal FGF23 (cFGF23) in mice and humans, with smaller and more context-dependent changes in intact hormone[14]. In the general population, functional iron deficiency and higher EPO concentrations have likewise been associated with higher FGF23 and mortality[15]. FGF23 therefore forms part of the physiological response to anemia and erythropoietic stress rather than exclusively within mineral metabolism.

Experimental evidence suggests a bidirectional relationship. iFGF23 activates FGFR4-dependent calcineurin/nuclear factor of activated T cells (NFAT) signaling in hepatocytes and promotes inflammatory mediator production[3]. In mice with acute inflammation, bone-derived C-terminal peptides interfere with bone morphogenetic protein (BMP) signaling, attenuate hepcidin induction, and preserve iron availability[4]. These fragment effects remain unconfirmed in humans. Table 1 summarizes upstream regulation of FGF23 production and processing.

Table 1 Regulation of fibroblast growth factor 23 production and processing by iron status, inflammation, and erythropoietic signals.

Stimulus	Effect on production	Effect on processing	Biological consequence
Iron deficiency	Stimulates FGF23 transcription through iron- and HIF-sensitive pathways	Increased cleavage usually limits iFGF23 accumulation	Predominant rise in the C-terminal signal with preserved phosphate homeostasis. Experimental and human evidence[8,9]
Iron deficiency with impaired cleavage (e.g., ADHR)	Stimulates FGF23 production	Cleavage-resistant FGF23 accumulates	Increased iFGF23, renal phosphate wasting and hypophosphatemia. Genetic and experimental evidence[8]
Inflammation	Cytokines directly stimulate FGF23 transcription and also restrict iron availability through hepcidin	Processing may increase in parallel, especially early in inflammation	The iFGF23 response depends on the balance between production and cleavage. Experimental evidence[10]
Blood loss and erythropoietic stress	EPO and marrow erythroid cells increase FGF23 production	Cleavage commonly increases	Links the response to anemia with FGF23 regulation. Mouse and human EPO studies[13,14]
Iron repletion	Removes the iron-deficiency stimulus	Cleavage rate is not directly quantified by the fall in cFGF23	Restores iron-dependent regulation. Human intervention evidence[9]
Some intravenous iron formulations	Correct iron deficiency but may transiently alter post-translational regulation	Intact hormone rises after some formulations; the molecular mechanism remains unresolved	Transient iFGF23 elevation, phosphate wasting and hypophosphatemia. Randomized human evidence[16,17]

cFGF23: C-terminal fibroblast growth factor 23 assay result; ADHR: Autosomal dominant hypophosphatemic rickets; EPO: Erythropoietin; FGF23: Fibroblast growth factor 23; HIF: Hypoxia-inducible factor; iFGF23: Intact fibroblast growth factor 23.

Iron treatment illustrates the clinical importance of separating production from processing: Iron repletion generally reduces the production stimulus, but some intravenous formulations increase iFGF23 and cause renal phosphate wasting[16,17].

Experimental studies show that FGF23 can induce cardiomyocyte hypertrophy[18], with FGFR4 mediating a well-characterized pathway independent of membrane-bound α-Klotho[19]. Higher circulating FGF23 is also associated with adverse cardiac remodeling, incident heart failure, and coronary events in population studies[20]. In heart failure and chronic kidney disease (CKD), statistical mediation analyses implicate FGF23 in associations between iron status and outcomes, as discussed in the clinical-settings section[21,22]. These analyses do not establish causal mediation. Vascular tissue expression[23] and experimental myocardial metabolic effects[24] provide further biological context. Recent experimental work further identifies erythroid-derived intact FGF23 as a local inhibitor of erythropoiesis[25], while global FGF23 neutralization in CKD models improves hyperparathyroidism but increases mortality[26]. Figure 1 summarizes the network while distinguishing human evidence, experimental mechanisms, and unresolved therapeutic hypotheses.

Figure 1 Iron-inflammation-erythropoiesis control of fibroblast growth factor 23 production and processing. Iron deficiency, inflammation, and erythropoietic stress increase fibroblast growth factor 23 production, whereas proteolytic processing determines the balance between intact fibroblast growth factor 23 and C-terminal fragments. FGF23: Fibroblast growth factor 23; HIF: Hypoxia-inducible factor; IL-6: Interleukin-6; TNFα: Tumor necrosis factor alpha; EPO: Erythropoietin; FGFR4: Fibroblast growth factor receptor 4; HF: Heart failure.

Compared with earlier broad reviews of FGF23/Klotho pathobiology and biomarker implications[5] and recent reviews focused on osteocytic regulation of FGF23[6], the present review uses the production-processing distinction to integrate assay interpretation with iron and inflammatory signaling, erythroid and C-terminal fragment biology, receptor-specific cardiovascular evidence, formulation-dependent effects of intravenous iron, and therapeutic safety. Its contribution is therefore a cross-domain translational synthesis rather than a claim to have established a new causal pathway. Iron chelation and ferroptosis are considered where they clarify the limits of extrapolating cellular iron biology to circulating FGF23 and human disease.

LITERATURE IDENTIFICATION

This is a narrative review. For the final revision, targeted PubMed searches and consultation of primary reports on publisher websites were updated on 18 September 2026. Searches combined FGF23 or FGF23 with iron chelation, deferoxamine, deferasirox, deferiprone, ferroptosis, cardiac hypertrophy, Mendelian randomization, and relevant recent reviews; the existing reference list was also used to locate primary reports. Reference metadata and publication status, with particular attention to 2025-2026 reports, were rechecked against PubMed on the same date. Foundational studies and peer-reviewed reports available up to that date were considered. Selection emphasized relevance to production and processing, receptor-specific mechanisms, clinical associations, and treatment safety, including contradictory findings. Primary mechanistic studies and clinical trials were prioritized for causal or therapeutic claims. English-language reports and published online-ahead-of-print articles were considered; preprints and conference abstracts were not used to support the newly added mechanistic claims; no systematic multilingual search, formal risk-of-bias assessment, or exhaustive screening was performed. Narrative selection may therefore introduce citation bias.

FGF23 BIOLOGY: PRODUCTION, PROCESSING AND MEASUREMENT

FGF23 belongs to the endocrine FGF subfamily and is synthesized as a secreted precursor protein. Bone is its main physiological source, with expression concentrated in the osteoblast lineage and particularly in mature osteocytes within the lacuno-canalicular network[6,27]. Its production responds to both systemic and local signals. Phosphate, calcitriol, PTH, and calcium can modify FGF23 transcription or secretion, whereas osteocyte-derived proteins such as phosphate regulating endopeptidase X-linked (PHEX) and dentin matrix protein 1 restrain its expression[6,28].

Once released into the circulation, FGF23 acts mainly through FGFRs in the presence of membrane-bound α-Klotho. Because FGF23 has relatively low intrinsic affinity for FGFRs, this co-receptor supports high-affinity canonical signaling. Urakawa et al[29] showed that α-Klotho converts FGFR1c into a specific receptor for FGF23. Structural studies using the shed extracellular domain identified α-Klotho as a non-enzymatic scaffold stabilizing the FGF23-FGFR1c complex[30]. In the kidney, canonical signaling reduces sodium-phosphate cotransporter expression, promotes phosphaturia, suppresses CYP27B1-dependent calcitriol synthesis, and increases CYP24A1-mediated vitamin D catabolism[1,29,30]. Here, membrane-bound α-Klotho denotes the cell-surface co-receptor, and soluble α-Klotho denotes its circulating or experimentally administered extracellular form.

The amount of biologically active FGF23 reaching the circulation also depends on post-translational processing. After removal of its signal peptide, FGF23 can be secreted intact or cleaved at the RXXR sequence, in which R denotes arginine and X any amino acid. Cleavage between Arg179 and Ser180 produces N-terminal and C-terminal fragments. GALNT3-mediated O-glycosylation at Thr178 protects against proteolysis, whereas FAM20C-mediated phosphorylation of Ser180 interferes with glycosylation and favors cleavage[31]. Recognition of Thr178 by GALNT3 is facilitated by prior glycosylation at Thr171[32]. These mechanisms partly uncouple synthesis from the amount of intact hormone released.

Proprotein convertases, proteases that cleave precursor proteins at specific sites, participate in FGF23 processing and change in expression during osteoblast-to-osteocyte differentiation[33]. Furin contributes under some conditions, including iron deficiency, but cleavage persists after EPO or inflammatory stimulation when furin and other candidate convertases are disrupted[34]. Redundancy among convertases or contributions from other proteases are therefore plausible[6,34,35].

The production-processing distinction is relevant to iron deficiency, inflammation, blood loss, and EPO stimulation[9,10,13,14]. Autosomal dominant hypophosphatemic rickets provides a contrasting model: Cleavage-resistant FGF23 accumulates as intact hormone during iron deficiency, causing renal phosphate wasting[8].

Assay choice is therefore central to the interpretation of circulating FGF23. Intact assays use antibodies directed to epitopes on opposite sides of the cleavage site and therefore detect full-length FGF23, whereas C-terminal assays recognize epitopes downstream of that site and detect both intact hormone and C-terminal fragments[7,36,37]. iFGF23 therefore more closely reflects the circulating hormone available for canonical receptor activation, while C-terminal measurements are influenced by both production and proteolytic processing. The distinction is especially relevant when transcription and cleavage change simultaneously, as occurs in iron deficiency and inflammation[9,10]. The complementary information provided by intact and C-terminal measurements is compared in Table 2.

Table 2 Biological information provided by intact and C-terminal fibroblast growth factor 23 measurements.

Measurement	What the assay detects	Biological information	Main interpretive limitation
iFGF23	Full-length FGF23; antibodies recognize epitopes on opposite sides of the cleavage site	Circulating biologically active hormone available for receptor activation	May remain stable despite markedly increased production when cleavage rises in parallel[36,37]
cFGF23	Intact hormone plus C-terminal fragments; antibodies recognize epitopes downstream of the cleavage site	Composite signal influenced by both production and proteolytic processing	Does not measure C-terminal fragments selectively and should not be interpreted as intact hormone[36,37]
Paired iFGF23 and cFGF23	Both measurements obtained in the same clinical or experimental setting	Supports interpretation of discordant assay responses; does not directly quantify production, cleavage, or fragments	Assay methods are not standardized or interchangeable. Research interpretation, not a direct cleavage-rate measure[36,37]
C-terminal fragments	Not selectively quantified by current cFGF23 assays	Potential modulation of BMP-hepcidin signaling and iron availability	Biological activity remains supported mainly by experimental evidence. Mouse and cellular evidence[4]

BMP: Bone morphogenetic protein; cFGF23: C-terminal fibroblast growth factor 23 assay result; FGF23: Fibroblast growth factor 23; iFGF23: Intact fibroblast growth factor 23.

C-terminal fragments should not, however, be regarded simply as inactive products of hormone degradation. Experimental evidence suggests that bone-derived cFGF23 fragments can interfere with BMP signaling, attenuate hepcidin induction, and preserve circulating iron availability during acute inflammation[4]. Whether these effects are quantitatively important in humans remains uncertain, but they further complicate the interpretation of cFGF23 concentrations as a purely analytical surrogate of FGF23 production.

Interpretation is also affected by analytical and pre-analytical variability. Commercial assays show substantial method-dependent differences, particularly near the physiological range[36]. They are not standardized or interchangeable, and reference intervals remain method specific[37]. Serum and ethylenediaminetetraacetic acid plasma may yield different iFGF23 concentrations depending on the platform[38], while reference intervals can vary across populations[37,39]. Results should be interpreted alongside kidney function, sample matrix, and clinical context. Paired intact and C-terminal measurements are useful in research when production and cleavage may change simultaneously, but neither their numerical difference nor an unvalidated ratio directly quantifies fragments or cleavage efficiency. Production, processing, and clearance can all influence the circulating signal. Routine paired testing for cardiovascular risk assessment is not an established clinical indication[36,37].

IRON DEFICIENCY AS A REGULATOR OF FGF23

Evidence for iron-dependent regulation of FGF23

The relationship between iron availability and FGF23 has been demonstrated across hereditary, acquired, and experimental settings. Autosomal dominant hypophosphatemic rickets (ADHR) provided an early mechanistic model because disease-causing mutations involving Arg176 or Arg179 render FGF23 relatively resistant to proteolytic cleavage[40,41]. In mutation carriers, lower serum iron is associated with increases in both C-terminal and iFGF23, whereas in healthy individuals low iron is predominantly associated with higher cFGF23[42]. Experimental studies reproduce this difference. Dietary iron deficiency increases bone FGF23 expression in both wild-type and ADHR knock-in mice, but enhanced cleavage in wild-type animals largely prevents accumulation of intact hormone. In contrast, ADHR mice develop increased iFGF23, renal phosphate wasting, and hypophosphatemic osteomalacia[8]. Iron-sensitive regulation is also evident in X-linked hypophosphatemia (XLH), a hereditary phosphate-wasting disorder caused by loss-of-function variants in PHEX and characterized by inappropriately elevated FGF23 levels. In untreated patients with XLH, serum iron is inversely associated with C-terminal but not iFGF23, broadly resembling the pattern observed in healthy individuals[43].

Clinical observations outside hereditary phosphate-wasting disorders support iron-dependent regulation. In women with iron-deficiency anemia, Wolf et al[9] found markedly elevated cFGF23 with relatively preserved intact hormone; correction of iron deficiency rapidly reduced the C-terminal signal. In Gambian children, improvements in iron indices were associated with lower cFGF23[44]. In a secondary analysis of a randomized pregnancy trial, antenatal iron reduced maternal total FGF23 by 62.6% [95% confidence interval (CI): 53.0%-70.3%]. The neonatal total-FGF23 reduction was smaller and statistically uncertain (15.2%; 95%CI: -0.3% to 28.4%; P = 0.06), while neonatal iFGF23 increased[45]. Maternal and neonatal assay responses should therefore not be treated as equivalent.

In a randomized, open-label study of 73 patients with iron-deficiency anemia with or without non-dialysis CKD, ferric citrate hydrate increased ferritin and transferrin saturation and reduced cFGF23, with little change in iFGF23 or serum phosphate[46]. Intravenous formulation-specific effects are discussed below. Colangelo et al[47] studied 26 patients with uncomplicated hereditary hemochromatosis and 19 healthy blood donors before and seven days after phlebotomy or donation. cFGF23 increased in controls, whereas the response was attenuated in patients who remained relatively iron replete. iFGF23 and phosphate were unchanged. Because FGF23 did not correlate significantly with iron markers and blood removal also stimulates erythropoiesis, this pilot study does not isolate an effect of iron availability alone.

Iron replacement studies in ADHR further support iron deficiency as a modifiable trigger. In patients with the FGF23 R176Q variant and iron deficiency, oral iron replacement normalized iFGF23 and serum phosphate in those who were hypophosphatemic at baseline[48]. Observational data similarly link symptomatic episodes of ADHR with periods of increased iron loss or demand, including menarche, menorrhagia, pregnancy, and gastrointestinal bleeding[49]. These observations are particularly informative because impaired FGF23 cleavage in ADHR makes changes in production more readily visible as changes in intact hormone.

Molecular mechanisms

The molecular link between iron availability and FGF23 production involves oxygen-sensing pathways. Prolyl hydroxylase domain (PHD) enzymes require both iron and oxygen to hydroxylate hypoxia-inducible factor (HIF) alpha subunits and promote their von Hippel-Lindau-dependent degradation. Iron deficiency therefore reduces PHD activity and stabilizes HIF signaling even in the absence of severe hypoxia. Experimental studies have implicated HIF-1α in the increase in bone FGF23 transcription induced by iron deficiency, while pharmacological HIF stabilization also increases FGF23 production[10].

Osteocytic PHD2, encoded by EGLN1 in humans and EGLN1 in mice, appears important in this response. Noonan et al[50] found that conditional deletion of EGLN1 (also termed PHD2) in mouse osteocytes increased FGF23 expression. HIF-1α blockade suppressed FGF23 upregulation in PHD2-deficient osteocyte-like cells, which also lost normal iron-mediated suppression of FGF23. These findings identify PHD2 as an experimental link between oxygen/iron sensing and osteocytic FGF23 production.

Iron chelation and phosphate loss

Deferoxamine, deferasirox, and deferiprone are used to treat transfusional iron overload, but their effects on FGF23 should not be assumed to be uniform. Experimental iron chelation increases FGF23 transcription[8]. In osteogenic cell lines, deferoxamine also increased FGF23 through interactions between hypoxia and Notch signaling[51]. Increased transcription alone does not establish an increase in circulating intact hormone or renal phosphate wasting. Clinically, deferasirox-associated Fanconi syndrome provides a separate mechanism of phosphate loss through proximal tubular dysfunction[52]. In a cross-sectional cohort of 213 adults with transfusion-dependent beta-thalassemia, deferasirox use was associated with lower iFGF23 than other chelation regimens[53]. Confounding prevents causal attribution, but these findings do not support a class-wide rise in iFGF23.

Iron deficiency also interacts with erythropoietic signaling. Stabilization of HIF increases renal EPO, and both acute blood loss and exogenous EPO stimulate FGF23 production[13,14]. This pathway is particularly relevant when iron restriction and increased erythropoietic demand coexist and is considered in greater detail below.

Less well understood is how iron deficiency coordinates increased FGF23 transcription with increased proteolytic processing. Changes in GALNT3-dependent glycosylation, FAM20C-mediated phosphorylation, and proprotein convertase activity are potential points of regulation, but the signals that couple iron availability to these post-translational mechanisms remain incompletely defined[31,34].

INFLAMMATION-IRON-FGF23 CROSSTALK

Inflammation-induced FGF23 production

Pro-inflammatory stimuli increase FGF23 expression in osteocyte-like cells and human bone explants. In IDG-SW3 cultures, TNF-α, IL-1β, TNF-like weak inducer of apoptosis, and lipopolysaccharide (LPS) increased FGF23 expression[54]. TNF-α- and IL-1β-induced transcription required nuclear factor κB (NF-κB). These stimuli also reduced PHEX, DMP1, and ENPP1 transcripts and increased GALNT3. Intact protein accumulated prominently only when furin-like activity was inhibited, supporting concurrent regulation of synthesis and cleavage[54].

David et al[10] demonstrated this pattern in vivo. Heat-killed Brucella abortus or IL-1β rapidly increased bone FGF23 transcription and C-terminal FGF23 (cFGF23) while iFGF23 remained comparatively stable. The response coincided with hypoferremia and HIF activation. With more sustained inflammation, iFGF23 also rose, suggesting that the balance between transcriptional drive and proteolytic capacity changes with the intensity and duration of the stimulus.

Extra-skeletal production may contribute in acute inflammation. After LPS, Bansal et al[55] detected increased tissue FGF23 expression and cFGF23 without an early parallel increase in iFGF23; during an acute-on-chronic inflammatory challenge, both molecular forms increased. The spleen contributed substantially in that model. Its importance in human inflammatory disease is less certain, and osteocytes remain the dominant physiological source.

Cytokine-dependent regulation and inflammatory mediators

Several cytokine pathways converge on FGF23 regulation. IL-1β and TNF can activate osteocytic FGF23 transcription through NF-κB[54], while IL-1β-induced hypoferremia and HIF signaling provide an additional indirect stimulus[10]. TNF is notable because the effect is not confined to kidney failure: In experimental CKD and non-renal inflammation, TNF increased FGF23, whereas TNF neutralization reduced it and suppressed inflammation-associated ectopic renal FGF23 expression[11].

IL-6 acts through a separate pathway. In adenine-induced CKD, IL-6 deficiency markedly attenuated the rise in bone FGF23 and circulating FGF23[56]. In cell and organ culture, IL-6/soluble IL-6 receptor trans-signaling increased FGF23 transcription through signal transducer and activator of transcription 3 phosphorylation and a conserved promoter element[56]. NF-κB, HIF, and signal transducer and activator of transcription 3 signaling can therefore operate in parallel rather than as a single linear inflammatory pathway.

Lipocalin 2 (LCN2), also known in humans as neutrophil gelatinase-associated lipocalin, links innate immunity, iron handling, kidney injury, and bone. In Col4a3-deficient mice, genetic deletion of LCN2 markedly lowered FGF23, prevented left ventricular hypertrophy, and improved survival without correcting kidney dysfunction or anemia[57]. The deletion blunted the FGF23 response to inflammation but not to iron deficiency or phosphate loading, whereas exogenous LCN2 increased FGF23 through cyclic adenosine monophosphate-dependent signaling in bone cells[57]. Whether LCN2 has a comparable role in human FGF23 regulation remains unresolved.

Hepcidin and functional iron deficiency

The hepcidin-ferroportin system is the main bridge between inflammation and functional iron deficiency. IL-6-induced hepcidin limits iron export from enterocytes, macrophages, and hepatocytes, lowering circulating iron despite preserved or increased tissue stores[12]. In chronic inflammation, this creates a second route to FGF23: Cytokines can stimulate FGF23 directly, while hepcidin-driven iron restriction reduces PHD activity and reinforces HIF-dependent FGF23 transcription[10].

FGF23 may also feed back on this pathway. In an LPS model, Agoro et al[58] used the C-terminal tail of FGF23 to antagonize endogenous full-length FGF23 signaling. The intervention lowered hepatic and circulating hepcidin, reduced iron sequestration in liver and spleen, and attenuated the fall in serum iron and transferrin saturation[58]. In that setting, signaling by iFGF23 contributed to inflammatory hypoferremia.

Cleavage products may have the opposite effect. Courbon et al[4] found that osteocytes generated most of the excess cFGF23 fragments during acute inflammation; reducing osteocyte-derived fragments led to greater hepcidin induction and more severe hypoferremia. The peptides bound BMP2 and BMP9 and attenuated BMP-dependent hepcidin signaling. Thus, iFGF23 signaling and C-terminal fragments may influence inflammatory iron handling in different directions. This attractive model still rests mainly on experimental data and has not been established quantitatively in humans.

Reverse signaling: Effects of FGF23 on inflammation, iron metabolism and innate immunity

FGF23 can also act upstream of inflammation. Singh et al[3] showed that hepatocytes, despite lacking α-Klotho, respond to FGF23 through FGFR4 and phospholipase Cγ/calcineurin/NFAT signaling. This pathway increased expression and secretion of inflammatory mediators, including C-reactive protein, whereas genetic or pharmacological interruption of FGFR4 reduced hepatic inflammatory responses in experimental CKD[3].

Its immune effects are not uniformly pro-inflammatory. Rossaint et al[59] found that FGF23 impaired neutrophil recruitment in CKD by activating FGFR2/protein kinase A signaling, inhibiting Rap1, and preventing chemokine-induced beta2-integrin activation. Neutralizing FGF23 restored leukocyte recruitment and host defense in mice with kidney failure. In patients with CKD, Wallquist et al[60] likewise reported associations between FGF23, inflammatory cytokines, and leukocyte adhesion markers; FGF23 exposure reduced CD11b expression on human granulocytes in vitro.

Human observational data are compatible with an FGF23-inflammation link but cannot establish its direction. Among 3879 Chronic Renal Insufficiency Cohort (CRIC) participants, higher FGF23 was independently associated with IL-6, C-reactive protein, TNF-α, and fibrinogen[61]. In patients with CKD stages 3-4 undergoing femoral endarterectomy for peripheral arterial disease, Donate-Correa et al[62] found independent associations of circulating FGF23 with TNF-α and TNF messenger RNA in peripheral blood cells, together with greater TNF-α immunoreactivity in femoral arterial tissue at higher FGF23 concentrations.

In 3655 CRIC participants followed for a median of 6.5 years, higher baseline cFGF23 was associated with hospitalization for major infection; the adjusted hazard ratio for the highest vs lowest quartile was 1.51 (95%CI: 1.23-1.85)[63]. Additional adjustment for inflammatory and mineral-metabolism biomarkers did not remove the association, but residual confounding by disease severity remains possible.

Not every inflammatory state is accompanied by higher iFGF23. A study of 79 children included 26 with bacterial infections, 26 with viral infections, and 27 healthy controls[64]. Hepcidin increased and functional iron deficiency was common in the infection groups, whereas iFGF23 was lower and inversely related to inflammatory markers. Because cFGF23 was not measured, accelerated cleavage is only a possible explanation and was not demonstrated. The study illustrates the need to specify the molecular form assayed.

Inflammation may also change tissue responsiveness to FGF23. Using single-cell transcriptomics and complementary chromatin and cell-based approaches, Agoro et al[65] mapped Klotho-dependent FGF23 responses across renal tubular populations and identified crosstalk between FGF23-mitogen-activated protein kinase and TNF receptor/NF-κB signaling. TNF/NF-κB activation blocked renal FGF23 bioactivity in vitro and in vivo, showing that high circulating FGF23 does not necessarily imply preserved signaling at its target tissue.

Ferroptosis and the limits of mechanistic inference

Ferroptosis is iron-dependent regulated cell death driven by lipid peroxidation[66]. Glutathione peroxidase 4 protects against ferroptotic injury, whereas ferritinophagy, the selective autophagic degradation of ferritin, can increase available intracellular iron and promote ferroptosis[67,68]. These processes concern intracellular iron and membrane damage; systemic iron restriction and circulating FGF23 are not interchangeable measures of ferroptotic susceptibility.

A possible connection with the FGF23-α-Klotho axis has been described in rheumatoid arthritis-associated osteoporosis. Ji et al[69] found lower α-Klotho in patients and studied mechanisms in mouse MC3T3-E1 preosteoblasts. In those cells, α-Klotho overexpression enhanced FGF23-FGFR1 interaction, reduced NF-κB signaling, and increased glutathione peroxidase 4 in association with protection against ferroptosis. The human component was observational and the mechanistic experiments were cellular. These results do not establish an FGF23-driven ferroptosis pathway in human CKD or cardiovascular disease, nor justify transferring protective effects of α-Klotho to FGF23 itself.

ERYTHROPOIESIS AND THE FGF23 AXIS

EPO as a regulator of FGF23

The erythropoietic response is an important non-mineral stimulus for FGF23. Acute blood loss raises EPO and cFGF23 within hours, with little early change in iFGF23, and the accompanying increase in FGF23 expression is concentrated in bone marrow rather than cortical bone[13]. Recombinant EPO produces a similar pattern in mice and humans[14], consistent with increased synthesis accompanied by substantial proteolytic processing.

Erythroid progenitors contribute directly. In experimental acute kidney injury, Toro et al[70] showed that these cells express FGF23 and that EPO receptor activation increases marrow FGF23; EPO receptor blockade prevented the marrow response and partly reduced circulating FGF23. Hanudel et al[71] extended the observation across models of endogenous and exogenous EPO excess. In mice, EPO increased total FGF23 disproportionately to iFGF23, and in four human CKD cohorts, serum EPO or exogenous EPO dose was associated with total FGF23. Among 680 kidney transplant recipients, EPO was associated with total but not iFGF23.

The response is not obligatory. Repeated low or micro-doses of recombinant EPO for two weeks did not change cFGF23 in a randomized trial of 24 healthy men[72]. Dose, duration, iron availability, kidney function, and baseline erythropoietic drive probably determine whether an EPO-FGF23 signal becomes measurable. The context-dependent effects of iron deficiency, inflammation, blood loss, and EPO on FGF23 production and processing are summarized in Table 1.

Erythroid cells as an extra-skeletal source of FGF23

Human hemolytic anemia provides evidence of an association with erythropoietic drive, but does not itself identify the cellular source of circulating FGF23. In 90 patients with hereditary hemolytic disorders, EPO was associated with total FGF23 after adjustment for iron load, inflammation, and kidney function, but not with iFGF23[73]. EPO also correlated with erythroferrone (ERFE), whereas ERFE did not correlate with FGF23. These observations do not support ERFE as the principal mediator of the measured EPO-FGF23 association.

FGF23 as a regulator of erythropoiesis

FGF23 can, in turn, influence erythroid development. Coe et al[74] found expansion of hematopoietic stem cells and increased erythropoiesis in FGF23-null mice, whereas FGF23 administration suppressed erythroid production in wild-type animals. In a 5/6 nephrectomy model, inhibition of FGF23 signaling increased erythroid commitment, reduced apoptosis of erythroid precursors, improved red-cell indices, and attenuated iron deficiency and inflammation[75].

These effects do not appear to be explained solely by phosphate. Park et al[76] used dietary and genetic models in which phosphate and FGF23 varied independently and found impaired erythropoiesis and lower iron availability when FGF23 was high even with normal serum phosphate; hypophosphatemic Hyp mice showed a similar erythroid phenotype.

Courbon et al[25] provided direct experimental evidence for a local marrow mechanism. In iron-deficient mice, lineage-specific FGF23 deletion distinguished osteocytic contributions to circulating intact hormone from erythroid contributions to cleaved peptides. Erythroid-specific deletion corrected anemia, whereas erythroid FGF23 overexpression induced anemia in iron-replete mice. Erythroid Furin deletion increased local iFGF23 and worsened iron-deficiency anemia. In cultured erythroid progenitors, iFGF23 impaired differentiation and mitochondrial function, and FGFR1 inhibition abolished these effects. Erythroid FGF23 deletion also prevented anemia in experimental progressive CKD. These findings support paracrine inhibition of erythropoiesis in mice, but do not establish that circulating iFGF23 measures marrow signaling in humans.

CLINICAL SETTINGS OF THE IRON-INFLAMMATION-FGF23 AXIS

Iron-deficiency anemia

Iron-deficiency anemia commonly produces high cFGF23 with relatively preserved iFGF23 and serum phosphate[9,42]. The production-processing framework explains this pattern and its response to iron therapy[44-46,48]. Cleavage-resistant FGF23 in ADHR instead permits intact hormone to accumulate and drive phosphate wasting[8,42].

Chronic inflammatory disorders

Chronic inflammatory diseases combine cytokine activation, hepcidin-mediated iron restriction, and variable absolute iron deficiency, but direct clinical evidence for FGF23 remains limited. In 61 patients with rheumatoid arthritis, circulating FGF23 correlated with erythrocyte sedimentation rate, C-reactive protein, and Disease Activity Score in 28 joints disease activity[77]. In inflammatory bowel disease, where intravenous iron is frequently used, a prospective multicenter study of 21 patients with iron-deficiency anemia found that high-dose iron isomaltoside 1000 (now ferric derisomaltose) corrected anemia without a marked increase in iFGF23 or clinically important hypophosphatemia[78]. These small studies support biological plausibility but are not sufficient to establish FGF23 as a reproducible marker of inflammatory activity across immune-mediated diseases.

Heart failure

In BIOlogy Study to TAilored Treatment in Chronic Heart Failure, cFGF23 was measured in 2399 patients with new-onset or worsening heart failure[79]. Higher concentrations were associated with congestion, lower estimated glomerular filtration rate, atrial fibrillation, and less successful up-titration of renin-angiotensin system blockade. The adjusted hazard ratio for death or heart-failure hospitalization was 1.15 (95%CI: 1.08-1.22) per log increase in cFGF23. Van der Wal et al[21] reported statistical mediation of the association between iron deficiency and mortality by FGF23 in worsening heart failure. Shared determinants of iron restriction, congestion, kidney dysfunction, and FGF23 limit causal interpretation.

CKD

CKD brings together phosphate retention, inflammation, iron restriction, anemia, and endogenous or therapeutic EPO exposure. Experimental data suggest that cleavage may become insufficient relative to production[10]; declining clearance and altered mineral metabolism further complicate interpretation in patients. In 3747 CRIC participants, the iron-deficiency phenotype was associated with mortality (adjusted hazard ratio, 1.28; 95%CI: 1.04-1.58) and heart failure (1.34; 1.05-1.72), and cFGF23 most strongly attenuated these associations among the candidate mediators[22]. The functional iron-deficiency category was not significantly associated with the evaluated outcomes. Thus, risks differed across study-defined iron phenotypes, which should not be pooled into a single iron-deficiency construct.

General population

Community cohorts extend the association beyond overt cardiorenal disease. In 6544 Prevention of Renal and Vascular End-stage Disease participants, lower ferritin and higher EPO were associated with higher total FGF23 and mortality[15]. In Multi-Ethnic Study of Atherosclerosis, 6547 participants initially free of cardiovascular disease were followed for 7.5 years; each 20 pg/mL higher iFGF23 was associated with incident heart failure (adjusted hazard ratio, 1.19; 95%CI: 1.03-1.37) and coronary disease (1.14; 1.01-1.28), but not stroke[20]. A later Prevention of Renal and Vascular End-stage Disease analysis of 6830 participants without heart failure associated cFGF23 with incident heart failure, particularly heart failure with reduced ejection fraction, after multivariable adjustment[80]. The association with preserved-ejection-fraction heart failure was no longer significant after adjustment. These cohorts differ in assay and outcome phenotype. Table 3 summarizes the main clinical settings.

Table 3 Clinical contexts linking iron status and fibroblast growth factor 23 biology.

Clinical setting	Typical FGF23 pattern	Potential relevance
Iron-deficiency anemia	High cFGF23 with relatively preserved iFGF23	Efficient cleavage may protect phosphate homeostasis despite increased production. Human observational and interventional evidence[9,42]
ADHR with iron deficiency	Increased iFGF23 because the mutant hormone resists cleavage	Hypophosphatemia and osteomalacia; oral iron may correct the biochemical phenotype. Small open-label human trial[48]
X-linked hypophosphatemia	Iron is associated mainly with variation in cFGF23	Iron-sensitive regulation persists in a primary FGF23-excess disorder. Human observational evidence[43]
Chronic inflammatory disorders	Variable FGF23 pattern influenced by cytokines, hepcidin and iron restriction	Clinical evidence remains limited and disease-specific. Small human studies; uncertain generalizability[77,78]
Heart failure	Iron deficiency and elevated FGF23 frequently coexist	Associations with congestion, remodeling, hospitalization and mortality; causality remains uncertain. Human observational evidence; causality unproven[21,79]
Chronic kidney disease	Elevated intact and C-terminal measurements; production, processing, clearance, and mineral metabolism contribute	Cardiovascular, heart-failure and mortality associations. Human observational evidence with experimental support; causal contribution uncertain[10,22]
General population	Higher FGF23 associated with incident heart failure and coronary events	May integrate iron status, erythropoietic drive, kidney function and inflammation. Human observational evidence; no established testing indication[15,20,80]

ADHR: Autosomal dominant hypophosphatemic rickets; cFGF23: C-terminal fibroblast growth factor 23 assay result; FGF23: Fibroblast growth factor 23; iFGF23: Intact FGF23.

FGF23 AND CARDIOVASCULAR DAMAGE

Cardiac hypertrophy and FGFR4-dependent signaling

The most developed experimental evidence for a direct cardiovascular action concerns cardiomyocyte growth. Faul et al[18] demonstrated hypertrophy in isolated rat cardiomyocytes and mice. Grabner et al[19] identified a pathway involving FGFR4, phospholipase Cγ, calcineurin, and NFAT that operates without membrane-bound α-Klotho; genetic or pharmacological inhibition attenuated hypertrophy in their models. Subsequent work showed regression of established hypertrophy with selective FGFR4 blockade[81]. Han et al[82], using a separate cardiac-specific deletion model, also found that myocardial FGFR4 was required for recombinant FGF23-induced hypertrophy and that soluble α-Klotho attenuated the response. Several supportive studies share investigators and related models, so they should not all be counted as independent replications.

Replication has not been uniform. Moench et al[83] observed elevated intact and total FGF23 without cardiac hypertrophy in rat 5/6-nephrectomy and mouse high-phosphate models. Differences in strain, age, disease severity, exposure, and assessment may contribute, but do not resolve the discrepancy. Exposure also matters: The Grabner experiments used 25 ng/mL FGF23 in neonatal rat cardiomyocytes and 100 ng/mL in neonatal mouse cardiomyocytes[19], far above healthy circulating concentrations, which are assay dependent[39]. Relevance to advanced CKD requires comparison with intact hormone measured on a comparable platform, as well as exposure duration and tissue availability; C-terminal assay values cannot be converted into these concentrations. This evidence supports a context-dependent experimental mechanism rather than a universal response to FGF23 elevation.

Metabolic, fibrotic and electrophysiological remodeling

Myocardial effects can precede visible hypertrophy. In experimental CKD, Fuchs et al[24] detected mitochondrial and metabolic abnormalities before overt left ventricular hypertrophy. FGF23-FGFR4 signaling shifted cardiomyocyte metabolism toward glycolysis, impaired mitochondrial function, and produced bioenergetic stress; cardiac-specific FGFR4 loss prevented these changes. Metabolic remodeling may therefore be an early component of the myocardial response to sustained FGF23 excess.

FGF23 also affects excitation-contraction coupling. In adult ventricular cardiomyocytes, it reduced L-type calcium current, systolic calcium transients, sarcoplasmic reticulum calcium load, and contractility while increasing spontaneous diastolic calcium release through calcium/calmodulin-dependent protein kinase II-dependent ryanodine receptor 2 hyperphosphorylation[84]. In intact mouse hearts, FGF23 produced ventricular ectopy, episodes of ventricular tachycardia, and heart rate-corrected QT prolongation, all attenuated by FGFR4 blockade[85]. Cardiac FGF23 deletion has also been associated with lower susceptibility to atrial fibrillation, whereas FGF23 overexpression increased calcium currents, calcium transients, ryanodine receptor 2 expression, and triggered activity[86]. The evidence remains mainly experimental, but it provides a mechanistic basis for electrical instability at high FGF23 exposure. The direction of calcium-current changes differs between adult ventricular and neonatal atrial models, emphasizing that arrhythmogenic effects depend on chamber, maturation, and exposure rather than a single uniform calcium response.

Experimental fibrosis findings also depend on cellular context. In mouse infarction and ischemia-reperfusion models, FGF23 promoted fibrosis alongside β-catenin and transforming growth factor beta signaling[87]. Lee et al[88] found increased proliferation and migration of human atrial fibroblasts at 25 ng/mL FGF23, but not at 1 ng/mL or 5 ng/mL, through FGFR1-phospholipase C/inositol 1,4,5-trisphosphate-dependent calcium signaling; collagen and α-smooth muscle actin expression were not increased. Shen et al[89] reported FGFR4-mitogen-activated protein kinase/extracellular signal-regulated kinase-dependent activation in transforming growth factor beta-stimulated human primary cardiac fibroblasts and worsened fibrosis in pressure-overloaded mice. Proliferation, myofibroblast activation, and tissue collagen deposition are distinct endpoints. Receptor expression, cell origin, and pre-existing profibrotic stimulation may help explain the different findings.

Vascular dysfunction, stiffness and atherosclerosis

Endothelial responses vary with the model and α-Klotho context. Silswal et al[90] found impaired endothelium-dependent relaxation in mouse aortic rings at 9000 pg/mL FGF23, with increased superoxide and reduced nitric oxide availability. In human coronary endothelial cells expressing membrane-bound α-Klotho, Richter et al[91] found FGFR1-protein kinase B-endothelial nitric oxide synthase activation and increased antioxidant defenses at 10 ng/mL FGF23; inhibition of α-Klotho shifted the balance toward oxidative stress. This cultured-cell result should not be generalized to all vascular beds or taken as proof that soluble and membrane-bound α-Klotho are interchangeable.

A direct role in vascular calcification is particularly uncertain. In 1501 CRIC participants, FGF23 was not independently associated with coronary or thoracic aortic calcification, and it did not induce calcification in vascular smooth muscle cells, in contrast to phosphate[92]. Other vascular phenotypes may be more responsive. Vergara et al[93] found that high FGF23 promoted a shift from a contractile to a synthetic smooth-muscle phenotype through FGFR1-extracellular signal-regulated kinase signaling and downregulation of miR-221/222, with greater arterial stiffness in experimental models and concordant associations in patients with CKD stages 2-3. FGF23 expression is also increased in calcified and atherosclerotic human vascular tissue[23], although tissue expression by itself does not establish a causal role.

In ApoE-deficient mice, Lindberg et al[94] found that sustained FGF23 overexpression increased plaque burden in males but not females, with inflammatory cell recruitment. FGF23 rose more than 400-fold, and phosphate, PTH, and renal function also changed, limiting attribution of the vascular phenotype to a direct hormone effect. In patients after myocardial infarction, Vázquez-Sánchez et al[95] associated higher FGF23 with worse cardiac function and mortality; tissue analyses also showed increased FGF23 expression after ischemic injury. Vetrano et al[96] reported associations with mortality, heart failure, and atherosclerotic events across CKD studies, with more consistent mortality and heart-failure associations for cFGF23 than iFGF23. This assay dependence cautions against equating a prognostic C-terminal signal with toxicity from intact hormone.

Critical interpretation and receptor context

Calcification, stiffness, endothelial dysfunction, and atherosclerosis are different vascular outcomes. The negative calcification experiments[92] therefore do not exclude effects on smooth-muscle phenotype[93], while the latter do not establish that FGF23 causes mineral deposition. Variable receptor expression, α-Klotho availability, experimental dose, and cell differentiation state offer plausible explanations that require direct testing. In patients, kidney dysfunction, phosphate retention, PTH and vitamin D disturbances, anemia, inflammation, and hemodynamic load can affect both FGF23 and cardiovascular risk. Statistical adjustment and mediation analysis cannot eliminate residual confounding or reverse causation.

Human genetic findings also qualify the causal argument. Donovan et al[97] found no association between genetically predicted FGF23 and major cardiovascular outcomes or cardiac structural abnormalities. Akwo et al[98] found no association with overall heart failure in the Heart Failure Molecular Epidemiology for Therapeutic Targets consortium, but reported signals for preserved-ejection-fraction heart failure in BioVU, particularly with genetically lower estimated glomerular filtration rate. These results differ by instrument and population and do not establish a consistent causal effect. Genetic estimates of lifelong differences may also fail to capture nonlinear effects of extreme hormone excess or local paracrine signaling in advanced CKD.

Receptor identity should be interpreted with tissue and co-receptor context. FGFR1c with membrane-bound α-Klotho mediates canonical renal signaling[29,30], whereas experimental cardiomyocyte and hepatocyte effects involve FGFR4 without that membrane co-receptor[3,19,82]. FGFR1-dependent erythroid and vascular responses and FGFR2-dependent neutrophil effects arise in distinct experimental systems[25,59,88,91,93]. Alpha-Klotho independence should not be assumed when it was not directly tested. Table 4 summarizes these distinctions and explains why receptor-selective inhibition cannot automatically be equated with organ-specific benefit.

Table 4 Tissue context and receptor specificity of fibroblast growth factor 23 signaling.

Receptor	Tissue	Alpha-Klotho context	Evidence and interpretation
FGFR1c	Renal tubular epithelium	Membrane-bound α-Klotho in canonical signaling	Structural, cellular, and animal evidence supports phosphate and vitamin D regulation[1,29,30]
FGFR4	Cardiomyocytes	Membrane α-Klotho not required in the cited models; soluble α-Klotho can modulate responses	Rodent/cell studies support hypertrophy and metabolic effects, with negative replication in other models; human benefit unproven[19,24,82,83]
FGFR4	Hepatocytes	Membrane α-Klotho-independent in the cited models	Cell and rodent evidence for inflammatory mediator production[3]
FGFR1	Erythroid progenitors	Requirement not established by the cited study	Mouse lineage deletion and progenitor culture support local inhibition of erythropoiesis[25]
FGFR1	Atrial fibroblasts and vascular smooth muscle	Do not infer independence without direct testing	Cell and animal evidence for proliferation or phenotype switching; these endpoints differ from calcification[88,93]
FGFR1	Coronary endothelial cells	Membrane α-Klotho detected; inhibition altered the response	Human cells in culture, not a clinical intervention: Context-dependent nitric oxide and redox effects[91]
FGFR2	Neutrophils	No α-Klotho requirement demonstrated here	Mouse and ex vivo human evidence for impaired integrin activation and recruitment[59]

These are selected experimentally implicated pathways, not an exhaustive receptor map. Human cells studied in culture are classified as cellular evidence. Absence of a demonstrated requirement is not proof of co-receptor independence. FGFR: Fibroblast growth factor receptor.

THERAPEUTIC MODULATION OF THE IRON-INFLAMMATION-FGF23 AXIS

Iron replacement and the intravenous iron paradox

Correcting iron deficiency usually removes a stimulus for FGF23 production, as shown in hereditary and acquired iron-deficient states[44-46,48]. Intravenous iron is different because the effect depends on the formulation. In the randomized FIRM physiological substudy, ferric carboxymaltose (FCM) caused a rapid rise in iFGF23 followed by phosphaturia, lower serum phosphate and calcitriol, and secondary increases in PTH; ferumoxytol did not produce the same mineral response[16]. Two randomized trials comparing FCM with ferric derisomaltose likewise found substantially more hypophosphatemia after FCM[17].

PHOSPHARE-IDA showed that severe or persistent hypophosphatemia occurred only in FCM-treated participants and was accompanied by larger increases in iFGF23, PTH, and alkaline phosphatase and greater suppression of 1,25-dihydroxyvitamin D[99]. PHOSPHARE-IBD reproduced the formulation effect in inflammatory bowel disease: Hypophosphatemia occurred in 51.0% of patients receiving FCM and 8.3% receiving ferric derisomaltose despite comparable correction of anemia[100]. ExplorIRON-CKD found similarly divergent effects on iFGF23, phosphate, and mineral metabolism in non-dialysis CKD[101]. The molecular basis of the FCM-specific rise in intact hormone remains unresolved; clinically, however, it is clear that hypophosphatemia is not an inevitable consequence of intravenous iron itself.

Phosphate loss during iron chelation should be distinguished from this intravenous iron effect. Deferasirox-associated proximal tubular dysfunction can cause hypophosphatemia[52], whereas the FCM trials demonstrate an intact-FGF23-mediated endocrine response[16,17]. Establishing the mechanism requires evidence on iFGF23, phosphate handling, and other features of tubular dysfunction; hypophosphatemia alone is insufficient.

Modulation through HIF, erythropoiesis and inflammation

HIF-prolyl hydroxylase inhibitors illustrate the competing influences within this axis. HIF stabilization can increase EPO and FGF23 transcription, whereas improved erythropoiesis and iron mobilization may reduce other drivers of FGF23 excess. In CKD mice, molidustat corrected anemia and reduced circulating iFGF23 by more than 60%[102]. Hepatic hepcidin fell and iron utilization improved, but renal indices also changed, limiting attribution to one pathway. Vadadustat increased total FGF23 disproportionately to iFGF23 in mice without CKD but lowered both forms in CKD as anemia and renal parameters improved[103]. In 28 hemodialysis patients switched from darbepoetin to roxadustat, hepcidin fell and changes in iFGF23 and cFGF23 varied with phosphate and baseline iron status[104]. These agents should not be regarded as FGF23-lowering drugs; their net effect depends on the clinical context.

Anti-inflammatory approaches provide proof of principle rather than established FGF23-directed therapy. TNF neutralization lowers FGF23 in experimental CKD and non-renal inflammation[11], and LCN2 deletion reduces FGF23 and left ventricular hypertrophy in experimental CKD[57]. The hepcidin-ferroportin axis is also an attractive target for anemia of inflammation, but there is no clinical evidence that benefit from hepcidin-directed therapy is mediated through FGF23. The C-terminal peptide used experimentally to reduce inflammatory hypoferremia[58] likewise remains a mechanistic tool.

Direct vs pathway-selective targeting of FGF23

Direct inhibition is effective in primary FGF23-mediated phosphate wasting: In the AXLES-1 trial, burosumab improved phosphate homeostasis and fracture healing in adults with XLH[105]. The safety implications in CKD differ. Shalhoub et al[26] treated CKD rats fed a high-phosphate diet with an FGF23-neutralizing antibody. Although secondary hyperparathyroidism and bone measures improved, serum phosphate and aortic calcification increased, together with mortality. This is a central translational constraint: Blocking an adaptive phosphaturic response can worsen mineral balance even when selected biochemical endpoints improve. These animal findings do not quantify human risk, but they preclude assuming that benefits in XLH extend to CKD.

Selective inhibition of extrarenal signaling is a testable strategy, but receptor selectivity does not by itself establish tissue specificity or clinical safety. FGFR4 blockade or deletion protects against structural and metabolic remodeling in several experimental models[19,24,81,82], while other models have yielded negative results[83]. Cardiovascular benefit in humans has not been established. Hu et al[106] found that a cFGF23 fragment improved heart and kidney measures in three murine models with high endogenous FGF23. This pharmacological antagonism should be distinguished from endogenous fragment binding to BMP2/9[4]; neither establishes selective preservation of renal FGF23 signaling or safety in human CKD.

In a secondary EVOLVE analysis, cinacalcet produced at least a 30% FGF23 reduction in 68% of participants vs 28% with placebo[107]. Among cinacalcet-treated participants, this reduction was associated with fewer cardiovascular deaths, sudden deaths, and heart-failure events. EVOLVE did not meet its primary endpoint in the unadjusted intention-to-treat analysis. Biomarker responders were not randomized to achieve FGF23 reduction, so the secondary association cannot establish benefit from lowering FGF23 itself. Table 5 compares the principal interventional studies and their interpretive limits.

Table 5 Clinical trials informing iron treatment and fibroblast growth factor 23 modulation.

Ref.	Design and population	Intervention	Findings and limits
FIRM[16]	Randomized double-blind trial; 1997 adults with iron-deficiency anemia; 185 in the physiological substudy; follow-up to week 5	Ferric carboxymaltose 750 mg vs ferumoxytol 510 mg, administered intravenously at baseline and week 1	Hypophosphatemia < 2.0 mg/dL: 50.8% vs 0.9% over 5 weeks. Intact FGF23 rose after ferric carboxymaltose. Biochemical and mechanistic outcomes; no cardiovascular efficacy inference
PHOSPHARE-IDA[17,99]	Two randomized open-label trials; 245 adults; reduced kidney function excluded; follow-up to day 35	Ferric derisomaltose 1000 mg once on day 0 vs ferric carboxymaltose 750 mg on days 0 and 7	Hypophosphatemia < 2.0 mg/dL through day 35: 7.9% vs 75.0% and 8.1% vs 73.7%. Severe or persistent hypophosphatemia occurred only with ferric carboxymaltose in the secondary analysis. Different total iron doses and short follow-up limit extrapolation
PHOSPHARE-IBD[100]	Randomized double-blind trial; 97 adults with inflammatory bowel disease and iron-deficiency anemia; follow-up to day 70; primary endpoint through day 35	Ferric derisomaltose vs ferric carboxymaltose: 1000 mg on day 0, followed by 500 or 1000 mg on day 35 according to calculated iron requirement	Hypophosphatemia < 2.0 mg/dL from baseline to day 35: 8.3% vs 51.0%; adjusted risk difference -42.8% (95%CI: -57.1% to -24.6%). Both corrected anemia. Longer-term clinical consequences remain uncertain
ExplorIRON-CKD[101]	Single-center randomized double-blind exploratory trial; 26 patients with CKD stages 3a-5; follow-up to 3 months	Ferric derisomaltose vs ferric carboxymaltose: 1000 mg at baseline, followed by 500-1000 mg at 1 month as required for iron repletion	Greater intact FGF23 rise and phosphate and calcitriol reduction with ferric carboxymaltose. Small mechanistic study, not powered for clinical events
AXLES-1[105]	Randomized double-blind placebo-controlled phase 3 trial; 134 adults with XLH; 24 weeks	Burosumab 1 mg/kg subcutaneously every 4 weeks vs placebo	Mean midpoint serum phosphate above the lower limit of normal: 94.1% vs 7.6%; fracture healing improved at week 24. Demonstrates efficacy in primary FGF23 excess, not CKD or cardiovascular disease
EVOLVE Biomarker analysis[107]	Secondary analysis of a randomized placebo-controlled trial in hemodialysis with secondary hyperparathyroidism; 2602 participants with paired baseline/week-20 samples	Cinacalcet 30-180 mg/day vs placebo, added to conventional treatment; dose titrated according to PTH and calcium	At least 30% FGF23 reduction by week 20: 68% vs 28%. Within the cinacalcet group, biomarker reduction was associated with fewer events; the secondary responder analysis does not establish causality

Trial acronyms identify the cited studies. Percentages are reported in the intervention order shown. Trials differ in eligibility, dosing, sampling, and endpoints and should not be compared as a pooled dataset. CI: Confidence interval; CKD: Chronic kidney disease; FGF23: Fibroblast growth factor 23; XLH: X-linked hypophosphatemia; PTH: Parathyroid hormone.

FUTURE RESEARCH DIRECTIONS

The first priority is cardiovascular causality. Studies should test whether interrupting a defined FGF23 pathway improves cardiac function beyond changes in phosphate balance, blood pressure, kidney function, iron status, and inflammation. Early human studies require evidence of target engagement and prespecified monitoring of phosphate, calcium, PTH, calcitriol, kidney function, and adverse events before cardiovascular outcome trials are justified. The safety signal from global FGF23 neutralization makes preservation of renal endocrine function a prerequisite[26].

The second priority is measurement. Prospective cohorts and mechanistic trials should report the assay, molecular form, sample matrix, handling conditions, and reference interval. Paired measurements are particularly informative during iron, EPO, or HIF-prolyl hydroxylase inhibitor interventions, but cannot directly quantify the production or cleavage rate without additional validation. Assay harmonization and fragment-specific methods are needed to determine whether apparent clinical differences reflect intact hormone, fragments, or analytical effects[36,37].

A third priority is validation of local mechanisms in humans. Erythroid FGF23-FGFR1 signaling should be examined in human marrow and relevant anemia phenotypes[25], while endogenous C-terminal fragment effects on BMP-hepcidin signaling require direct confirmation[4]. Ferroptosis studies should manipulate FGF23, its receptors, and α-Klotho separately, measure lipid peroxidation and cell death, and test rescue by ferroptosis inhibitors rather than infer ferroptosis from isolated marker changes.

Mechanistic replication should compare clinically relevant iFGF23 exposure, duration, age, sex, and kidney disease models. Trials of selective signaling interventions should establish that any benefit persists independently of changes in mineral metabolism. Variation in access to FGF23 assays and intravenous iron formulations also limits generalizability; clinical utility and feasibility must be demonstrated in the populations in which testing or treatment is proposed.

Iron-treatment studies also need to move beyond short-term serum phosphate. Head-to-head comparisons of intravenous formulations, longer follow-up after repeated dosing, and inclusion of bone and cardiovascular end points could determine whether transient disturbances of FGF23 processing have consequences beyond biochemical hypophosphatemia. Similar mechanistic phenotyping in anti-inflammatory and anemia trials may identify patients in whom treatment corrects the upstream stimulus and those in whom iFGF23 rises despite clinical improvement.

CONCLUSION

FGF23 production and proteolytic processing link iron availability, inflammation, and erythropoiesis, and explain why intact hormone and C-terminal assay results can diverge. Human intervention studies establish formulation-specific effects of intravenous iron on iFGF23 and phosphate, whereas the proposed erythroid, inflammatory, and cardiovascular actions remain supported largely by experimental or observational evidence. FGF23 may be both an adaptive hormone and a marker of systemic illness; a direct contribution to human cardiovascular disease remains unresolved. Global neutralization increased hyperphosphatemia, vascular calcification, and mortality in experimental CKD, making preservation of phosphate homeostasis essential. Future work should prioritize cardiovascular causality, assay standardization, and the safety of selective interventions. FGFR4 inhibition and C-terminal fragment-based approaches are investigational hypotheses, not established cardiovascular therapies. Interpretation of paired assays remains method- and context-dependent.