Inflammation regulates FGF23 production

Abstract

Purpose of review

Fibroblast growth factor 23 (FGF23) is a hormone secreted by osteocytes and osteoblasts that regulates phosphorus and vitamin D homeostasis. FGF23 levels increase progressively in chronic kidney disease (CKD), and FGF23 excess might be a causal factor of left ventricular hypertrophy, CKD progression and death. Therefore, understanding the molecular mechanisms that control FGF23 production is a critical to design therapies to lower FGF23 levels. The present review focuses on the role of inflammatory stimuli on FGF23 regulation and summarizes recent studies that support a novel framework linking inflammation to FGF23 regulation.

Recent findings

Inflammation and iron deficiency, which are common occurrences in CKD have emerged as novel FGF23 regulators. Recent findings show that inflammation increases FGF23 production in bone through direct and iron-related indirect mechanisms. In these settings, Hypoxia Inducible Factor (HIF)-1α orchestrates FGF23 transcription in response to inflammation and is primarily responsible for coordinating FGF23 production and cleavage.

Summary

We demonstrate that inflammation increases FGF23 production and may contribute to elevated FGF23 levels in CKD. Osseous HIF-1α may represent a therapeutic target to lower FGF23 levels in CKD patients and minimize the negative consequences associated with FGF23 excess.

Introduction

FGF23 is a bone produced hormone that targets the kidney to regulate renal phosphate (Pi) handling and vitamin D metabolism. Primary FGF23 excess causes hypophosphatemia, aberrant vitamin D metabolism, impaired growth, and rickets/osteomalacia (1). Inversely, FGF23 deficiency results in hyperphosphatemia, excess 1,25(OH)₂D, and soft tissue calcifications (2-4). FGF23 production is regulated by systemic factors that affect mineral balance, mainly Vitamin D, PTH, phosphate and calcium (5-7). Since osteocytes and osteoblasts are the main sources of circulating FGF23(8-11), bone alterations caused by local and systemic factors are susceptible to indirectly impact FGF23 production. Thus, FGF23 regulation might constitute a response to changes in bone and mineral metabolism to adjust vitamin D production, renal phosphate handling, and balance the mineral flux from bone.

Disordered bone and mineral metabolism is a common complication of CKD that begins early and worsens progressively as kidney function declines (12-14). CKD is the most common cause of elevated FGF23 and the clinical setting with the highest circulating levels (15). High FGF23 levels are strongly associated with greater risks of CKD progression, cardiovascular events, and mortality (16-21), and FGF23 excess is a causal factor for left ventricular hypertrophy (22, ••23). These findings underscore the importance of designing therapies to lower FGF23 levels, but our current lack of understanding of the mechanisms regulating FGF23 production limits this approach. Recent findings suggest that inflammatory stimuli are major regulators of FGF23 production and could substantially contribute to the increase in FGF23 levels in CKD (•24, •25), which we further discuss in this review.

FGF23 biology

FGF23 comprises a 24 amino acids hydrophobic signal sequence, an NH2 terminal of 154 amino acids containing the FGF core homology region, and a characteristic 73 amino acids COOH-terminal domain.(26) In the bloodstream, the FGF23 protein circulates in distinct forms: a full-length mature form (25-FGF23-251) and cleaved shorter forms which mainly arise from proteolytic cleavage of full-length FGF23 at the 176RXXR179 site (27-29). O-glycosylation of FGF23 by UDP-N-acetyl-α-d-galactosamine:polypeptide N-acetylgalactosaminyl-transferase 3 (GALNT3) overlaps the 176RXXR179 cleavage site, and this posttranslational modification protects FGF23 from cleavage by furin-like proprotein convertases (26, 29). Indeed, GALNT3 or FGF23 mutations at glycosylation sites leads to low intact FGF23 levels with marked increase of processed COOH-terminal fragments in the circulation and results in hyperphosphatemic familial tumoral calcinosis (HFTC) (30, 31).

Additionally, phosphorylation of the serine 180 by family with sequence similarity 20, member C protein (FAM20C), inhibits O-glycosylation and promotes cleavage (32). In contrast with GALNT3 deletion, Fam20C knockout (KO) mice have impaired FGF23 cleavage leading to an increase in circulating bioactive FGF23 causing renal phosphate wasting and severe hypophosphatemic rickets(33). Thus, the levels of circulating and biologically active full-length FGF23 are regulated by GALNT3, FAM20C and furin-like proteases, respectively, responsible for the initiation of O-glycosylation, phosphorylation and proteolytic processing of FGF23.

FGF23 production and cleavage

It has become increasingly common to assess FGF23 production and cleavage non-invasively, in vivo, by using two different commercially available assays: the C-terminal FGF23 assay (cFGF23), which captures both intact FGF23 and its C-terminal fragments(34), and the intact FGF23 assay (iFGF23), which exclusively detects the intact hormone(27). Using these 2 measures as a surrogate marker for FGF23 production and cleavage, we (35) and others (36) have shown that in multiple cases of pathological excess of FGF23 the elevation in circulating levels of intact FGF23 is caused by combined elevated FGF23 production and defective FGF23 cleavage.

The major example is found in patients suffering from autosomal dominant hypophosphatemic rickets (ADHR) caused by mutations at the cleavage site of FGF23 (28) that render FGF23 resistant to cleavage. In these patients, the manifestations of the disease can range from mild or no symptoms to late-onset of hypophosphatemic bone disease whereby increased FGF23 production is triggered by physiological states associated with iron deficiency(37), including puberty and pregnancy (38), and cleavage-resistant intact FGF23 accumulates in the bloodstream. As opposed to patients with ADHR, healthy subjects undergoing low iron states(39) display a proportional increase in FGF23 cleavage such that intact FGF23 levels remain unchanged or mildly affected. This strongly suggests that circulating FGF23 levels are tightly regulated by a balance between FGF23 transcription and cleavage under physiological conditions.

Inflammation and bone and mineral metabolism

Inflammation is part of the complex biologic response of any tissue to injury, infection, ischemia or autoimmune diseases. A variety of cytokines, including interleukin-1β (IL-1β), and acute phase proteins are released as a consequence of inflammation, in order to augment or attenuate the inflammatory response. Systemic inflammation alters bone and mineral metabolism at different levels. The mechanisms involved are complex and ultimately impact the bone remodeling cycle. Effects of acute and chronic inflammation on the skeleton are often very different. Acute inflammation generally increases both bone formation and bone resorption(40), while chronic inflammatory diseases of almost any cause are associated with bone loss, due to an uncoupling of bone formation from resorption in favour of excess resorption (41-45).

Multiple studies suggest that systemic inflammation may regulate FGF23 at least indirectly. Indeed, two of the main targets and regulators of FGF23, phosphorus and vitamin D, are strongly associated with systemic inflammation. Hypophosphatemia has long been reported to be associated with sepsis and correlates with sepsis severity(46, 47). Serum Pi is independently associated with inflammatory markers such as interleukin (IL)-6 (48), and dietary Pi loading dose dependently induces inflammation and increases serum tumor necrosis factor (TNF)-α levels in uremic rats(49). Vitamin D acts as an inhibitor of the inflammatory response (50) and decreases the mediators of systemic inflammation, such as interleukin-2 and TNF-α(51). Consequently, an inverse relation has been shown between vitamin D concentrations and C-reactive protein (CRP), a marker of inflammation, in both healthy subjects and patients with rheumatoid arthritis and frailty (52, 53). Finally, decreased vitamin D concentrations have been associated with an increased risk of developing autoimmune diseases, such as multiple sclerosis, rheumatoid arthritis, and type 1 diabetes (54, 55).

Inflammation regulates FGF23 production and cleavage

Associations between inflammatory markers and FGF23 have been reported in many inflammatory diseases (•56-62). Recently, we and others have shown that inflammatory cytokines are direct regulators of FGF23 production in cardiac fibroblasts (63) and osteoblasts (•24, •25). Using two separate models of acute inflammation resulting from the administration of either heat-killed Brucella abortus (BA) or IL-1β in vivo, we observed approximately a ten-fold increase in Fgf23 mRNA expression and serum cFGF23 while iFGF23 levels remained unchanged. The discrepancy between cFGF23 increase and the unaltered levels of intact hormone was due to concomitant increases in FGF23 production and cleavage. Indeed, co-administration of IL-1β with a furin/furin-like protease inhibitor, which blocks FGF23 cleavage, significantly increased circulating iFGF23 compared to IL-1β treatment alone. Also, furin inhibition did significantly increase intracellular and secreted iFGF23 in IL-1β -treated cells versus controls only receiving IL-1β. Collectively, these data suggest that inflammation is a powerful stimulator of FGF23 production and that FGF23 cleavage by furin/furin-like proteases plays a key role in maintaining adequate iFGF23 levels in response to a coinciding increase in circulating cFGF23.

Although the current data show that Fgf23 expression and cFGF23 levels peak in the acute setting, Fgf23 expression remains markedly elevated during chronic inflammation and, to different degrees, both cFGF23 and iFGF23 are increased during chronic inflammation. For unclear reasons, FGF23 processing is perturbed during chronic inflammation so that more biologically active FGF23 enters circulation. We speculate that prolonged states of amplified Fgf23 expression overwhelm the cleavage capabilities within osteocytes but, at this stage, additional studies are needed to investigate the difference between chronic and acute inflammation on FGF23 production.

Direct and indirect effects of inflammation on FGF23 production

Systemic inflammation may affect FGF23 production indirectly. Indeed, inflammation-induced alterations of known regulators of FGF23, such as systemic calcium, phosphate and vitamin-D metabolism, might contribute to FGF23 elevation during inflammation. Paracrine stimuli secreted by osteoclasts, release of matrix bound components and/or local increase in phosphate and calcium due to inflammation-induced excessive bone resorption may also stimulate FGF23 (57, •64). In this context, excess calcium released by bone resorption may stimulate Fgf23(5) through calcineurin/NFAT pathway in osteoblasts(65). In support of this, inhibition of bone resorption reduces FGF23 levels in patients with osteogenesis imperfecta.⁽⁶⁶⁾

While these mechanisms are important, the major indirect effect of inflammation on FGF23 production is exerted through the regulation of iron metabolism. In vivo, the onset of inflammation is closely associated with the decline in circulating iron (•24, 67). As opposed to “true” iron deficiency, which reflects a state of low total body iron stores, “functional” iron deficiency is a consequence of chronic inflammation in which iron sequestration in the reticulo-endothelial system decreases the amount of circulating iron despite adequate total body iron stores (68, 69). In mice or humans, hypoferremia increases FGF23 transcription but also cleavage, resulting in high circulating cFGF23 and normal or mildly elevated iFGF23 (•24, 36, 37, 39). In contrast, administration of iron to iron deficient patients decreases FGF23 (39). The effects of inflammation-induced functional iron deficiency on FGF23 are strikingly similar to those of true iron deficiency. Administration of exogenous murine hepcidin to WT mice, a peptide produced by the liver in response to inflammation, which increases iron sequestration and decreases gastrointestinal iron absorption, increased cFGF23 six hours post-injection (•24). However, these effects can only partially explain the FGF23 elevation in vivo and they do not account for the direct effects of inflammatory cytokines on FGF23 production in osteoblasts. Therefore, inflammation affects FGF23 production and cleavage directly through cytokine mediated mechanisms, and indirectly by limiting the amount of circulating iron.

Multiple mechanisms control FGF23 production during inflammation

One of the mechanisms at play during iron deficiency, is the normoxic stabilization of Hypoxia Inducible Factor (HIF)-1α at the cellular level. Induction of HIF-1α in osteoblasts/osteocytes, in response to iron-deficiency or hypoxia, increases FGF23 production (36, 70). In addition, HIF-1α activation frequently occurs during inflammatory diseases (71) and many studies show that HIF-1α functions in an adaptive manner by increasing ischemia tolerance and controlling excessive inflammation(72). In MC3T3-E1 osteoblast-like cells and bone marrow stromal cells, treatment with deferoxamine (DFO), an iron chelator, or IL-1β significantly increased Fgf23 mRNA expression (•24, 36). While both mechanisms involve HIF-1α stabilization, DFO had no effect on Hif-1α mRNA expression, but potently induced its nuclear abundance. Treatment with IL-1β increased Hif-1α mRNA expression and nuclear translocation in osteoblasts line suggesting that inflammation directly stimulates Fgf23 mRNA expression through a HIF-1α -dependent mechanism. In vivo, Hif-1α mRNA expression was acutely elevated by IL-1β but returned to baseline after four days of daily IL-1β injections. While stabilization of HIF-1α occurs when reduced cellular iron inhibit prolyl hydroxylase (73), mainly during iron deficiency and chronic inflammation, in acute inflammation there is an additional increase in Hif-1α transcription, resulting in higher total HIF-1α activity(74). In line with these findings, co-treatment of IL-1β injected mice with a HIF-1α inhibitor (2-methoxyestradiol (M2E2)), partially inhibited IL-1β -induced increase in Fgf23 promoter activity. Co-treatment with M2E2 also significantly reduced the abundance of nuclear HIF-1α, cytoplasmic FGF23, and secreted FGF23 versus treatment with IL-1β alone. In mice, M2E2 and an additional HIF-1α inhibitor (BAY-87-2243) attenuated IL-1β -induced increase in Fgf23 mRNA expression, demonstrating a key role of HIF-1α in FGF23 production.

Additional mechanisms, triggered by inflammation to increase FGF23 production are also involved. One study found that activation of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), the prototypical proinflammatory pathway, potently increased FGF23 transcription in vitro (•25), and that NF-κB inhibition also dramatically reduced FGF23 response to IL-1β and TNFα. However, a 50 to 100-fold increase in FGF23 transcription remained unaccounted for even after NF-κB inhibition. In line with our findings, this suggests the existence of several mechanisms controlling FGF23 production in response to inflammation, but these mechanisms might be closely interrelated. Indeed, NF-κB is a critical transcriptional activator of HIF-1α (75). In turn hypoxia and HIF-1α stabilization lead to NF-κB activation (76) and amplify the NF-κB pathway activation by increasing the expression and signaling of Toll-like receptors (TLR)s(77). Taken together these data suggest that normoxic HIF-1α activation is a central mechanism to inflammation-dependent FGF23 production (Figure1).

Figure 1. FGF23 regulation during inflammation.

A. Acute Inflammation and consequent functional iron deficiency results in markedly increased HIF-1α expression and stabilization HIF-1α. Transcriptionally active HIF-1α/HIF-1β heterodimers translocate to the nucleus and bind hypoxia responsive elements (HRE) to simultaneously upregulate Fgf23 and Furin expression. Inflammation also stimulates FGF23 production through alternate mechanisms. Furin overexpression counters the FGF23 elevation so that the majority of the secreted protein is cleaved. B. Chronic inflammation and functional iron deficiency induces HIF-1α stabilization, but not HIF-1α expression resulting in lower Furin/Furin-like protease activity compared to acute inflammation while the inflammatory stimulus driving Fgf23 expression persists. Consequently, the cleavage is unable to match the rate of FGF23 production. In contrast to acute inflammation, more biologically active intact FGF23 enters circulation despite a lesser amount of total protein. C. In CKD, chronic inflammation, true and functional iron deficiency further stimulate FGF23 production. Additionally, unknown mechanisms reduce FGF23 cleavage in this setting. The net effect is a substantial elevation in circulating intact FGF23. We postulate that either Furin inhibition, increased O-glycosylation by Galnt3, or decreased phosphorylation by FAM20C reduce FGF23 cleavage in CKD.

Hif1a controls FGF23 production and cleavage

Beyond its role on FGF23 production, activation of HIF-1α might also be responsible for the regulation of FGF23 cleavage. Indeed, HIF-1α induces furin expression and increases the bioavailability of mature furin substrates in hypoxic conditions (78). In IL-1β -injected mice, HIF-1α inhibition results in increased circulating iFGF23. Conversely, increasing nuclear HIF-1α abundance by prolyl-hydroxylase inhibitors increases Fgf23 mRNA expression and circulating cFGF23. Finally, co-administration of furin inhibitors that partially inhibit FGF23 cleavage, with prolyl-hydroxylase inhibitors, that stabilize HIF-1α, further increase iFGF23. Other findings show that pro-inflammatory stimuli acutely increased Galnt3 mRNA which should protect FGF23 from cleavage in IDG-SW3 cells, while surprisingly detecting mainly FGF23 fragments (•25). These observations suggest that HIF-1α coordinates cleavage during periods of FGF23 overproduction by increasing furin/furin-like proteases expression thus limiting the secretion of biologically active FGF23.

Inflammation: a potent stimulus for FGF23 in CKD

Chronic inflammation promotes renal, cardiac and vascular injury and is a major risk factor for systemic bone loss leading to fractures and substantial morbidity and mortality (79-81). Associations between kidney function and inflammation have been reported by multiple investigators (82-84), and inflammation has been shown to predict the long-term risk of developing chronic kidney disease (CKD) (85). FGF23 is associated with increased inflammatory burden in CKD (60) and we recently identified that FGF23-responsive genes are associated with renal damage and chronic inflammation (86). The Col4a3^KO mice, analog of human Alport’s Syndrome and a validated model of progressive CKD, exhibit a time-dependent loss of kidney function in association with progressive elevation of cFGF23 and iFGF23 levels(14). Injection of IL-1β in Col4a3^KO mice and WT mice resulted in a similar increase in cFGF23 and osseous Fgf23 mRNA expression, and a similar decline in serum iron. Unexpectedly, serum iFGF23 was further increased in CKD mice when challenged acutely with IL-1β, which suggests that FGF23 cleavage is impaired in CKD (•24). Given the prominent rise in iFGF23 in CKD compared to the very modest response in WT mice, it would seem that renal disease impairs FGF23 cleavage. Although it is unlikely that FGF23 cleavage occurs in the kidney (87), perhaps injured kidneys secrete a hormone that acts on bone to reduce intracellular FGF23 cleavage. Alternatively, a modification in furin, fam20c or galnt3 expression levels in osteoblasts/osteocytes might constitute an initial adaptative mechanism to CKD settings. Regardless, these observations strongly suggest that systemic inflammation might be a potent stimulus for iFGF23 increase in CKD.

Conclusion

Inflammation and iron deficiency anemia are two hallmarks of CKD. Both inflammation and iron deficiency are potent inducers of FGF23 production and these pathways are of particular importance in a CKD setting. In a rather important twist, FGF23 also exhibits pro-inflammatory and immuno-modulatory effects, affecting macrophages and neutrophils (•88, ••89) suggesting the existence of a “feed-forward” loop in CKD (86), whereby the effects of inflammation and FGF23 amplify each other, leading to negative outcomes. Whether HIF-1α, which regulates FGF23 production and cleavage, is also a target of excess FGF23, remains to be further investigated. Future exciting studies will determine if a dysfunctional osseous HIF-1α could be a therapeutic target in CKD.

Key Points.

• Herein we summarize recent studies that show compelling evidence of a close relationship between FGF23 production and processing and inflammation.

• We show that inflammation is a major stimulus for FGF23 production and metabolism and potentially a main component of FGF23 excess in chronic kidney disease.

• We explore different inflammation-driven mechanisms that stimulate FGF23 transcription and cleavage and highlight the importance of Hypoxia Inducible Factor 1 alpha as a central pathway controlling both FGF23 transcription and cleavage.