FGF23 Induction of O-Linked N-Acetylglucosamine Regulates IL-6 Secretion in Human Bronchial Epithelial Cells

The hexosamine biosynthetic pathway (HBP) generates the substrate for the O-linked β-N-acetylglucosamine (O-GlcNAc) modification of proteins. The HBP also serves as a stress sensor and has been reported to be involved with nuclear factor of activated T-cells (NFAT) activation, which can contribute to multiple cellular processes including cell metabolism, proliferation, and inflammation. In our previously published report, Fibroblast Growth Factor (FGF) 23, an important endocrine pro-inflammatory mediator, was shown to activate the FGFR4/phospholipase Cγ (PLCγ)/nuclear factor of activated T-cells (NFAT) signaling in chronic inflammatory airway diseases such as cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD). Here, we demonstrate that FGF23 increased the O-GlcNAc modification of proteins in HBECs. Furthermore, the increase in O-GlcNAc levels by FGF23 stimulation resulted in the downstream activation of NFAT and secretion of interleukin-6 (IL-6). Conversely, inhibition of FGF23 signaling and/or O-GlcNAc transferase (OGT)/O-GlcNAc reversed these effects. Collectively, these data suggest that FGF23 induced IL-6 upregulation and secretion is, at least, partially mediated via the activation of the HBP and O-GlcNAc levels in HBECs. These findings identify a novel link whereby FGF23 and the augmentation of O-GlcNAc levels regulate airway inflammation through NFAT activation and IL-6 upregulation in HBECs. The crosstalk between these signaling pathways may contribute to the pathogenesis of chronic inflammatory airway diseases such as COPD and CF as well as metabolic syndromes, including diabetes.


INTRODUCTION
Human fibroblast growth factors (FGFs) are classified as intracrine, paracrine, and endocrine FGFs depending on their action process with endocrine FGFs playing key roles in metabolism including bile acid, energy, and phosphate/active vitamin D metabolism (1,2). FGF23 is a 27-kDa protein that has been shown to be strongly associated with the risk of chronic kidney disease progression, systemic inflammation, and mortality (3,4). Our recent data characterized FGF23 signaling as an important mediator in inflammatory airway diseases such as cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD) (5,6). In the COPD lung, FGF23 activated the phospholipase Cγ (PLCγ)/nuclear factor of activated T-cells (NFAT) signaling pathway leading to airway inflammation (5).
NFAT signaling has also been linked to inflammatory cytokine production in hepatocytes, angiogenesis, cardiomyocyte hypertrophy, and many other biological processes (7)(8)(9)(10). On a molecular level, NFAT is regulated by the phosphatase calcineurin, which dephosphorylates NFAT and triggers cytoplasmic to nuclear translocation. Upon nuclear translocation, NFAT interacts with multiple factors to regulate gene expression of molecules involved in the aforementioned disease processes.
Although the activation of the PLCγ/NFAT signaling pathway by FGF23 has been studied, the downstream molecules that are affected have not been fully characterized. Several reports have shown that the activation of the hexosamine biosynthetic pathway (HBP) (11,12), a stress sensor, is linked to NFAT activation and may have a definitive role in inflammation (13,14). It is well-documented that the HBP serves as a precursor for several glycosylation pathways (14)(15)(16). Once activated, the HBP generates the sugar nucleotide UDP-Nacetyl-glucosamine (UDP-GlcNAc), which is a substrate for hyaluronan (HA), N-linked glycosylation, and for the O-linked β-N-acetylglucosamine (O-GlcNAc) modification of proteins (17). The O-GlcNAc modification is a single monosaccharide addition to proteins at unoccupied serine and threonine residues and is similar to protein phosphorylation. Addition or removal of O-GlcNAc is a dynamic process that is regulated by the glycosyltransferase OGT (O-GlcNAc transferase) and the glycosyl hydrolase OGA (β-N-acetylglucosaminidase), each of which serves as a stress sensor and flux mediator (16,18,19) in response to changes in the cellular microenvironment (i.e., stress stimuli) (20,21). An imbalance in the OGT/O-GlcNAc axis has been shown to regulate several cellular functions including cell cycle and proliferation, cardiac hypertrophy, cell metabolism, and inflammation (22,23). We have previously shown that OGT expression/activity was elevated in the pulmonary vascular disease, pulmonary arterial hypertension, regulated pulmonary arterial smooth muscle cell proliferation, and was associated with clinical disease worsening (24). However, its role in chronic inflammatory airway diseases has not been determined. Furthermore, the role the O-GlcNAc modification on PLCγ/NFAT signaling in airway inflammation, specifically upon FGF23 activation, has not been established. Here, we demonstrate that FGF23 stimulates the O-GlcNAc stress response in human bronchial epithelial cells, which is essential for NFAT transcriptional regulation of the inflammatory cytokine, IL-6, and may be involved in chronic inflammatory airway diseases such as COPD and CF.

IL-8 and IL-6 ELISA and mRNA Assessment
IL-8 and IL-6 enzyme-linked immunosorbent assays (ELISA) from Invitrogen (Thermo Scientific) were used according to the manufacturer's protocol. HBECs were stimulated for 24 h with FGF23, TG, or OSMI-1, and 100 µl of the medium (undiluted) was used for measurement.
RNA was extracted using the GeneJET RNA purification kit (Thermo Scientific, Grand Island, NY, USA). For gene expression analysis, qPCR was performed by using Taqman probes (Life technologies/Applied Biosystems, Carlsbad, CA, USA) with the following: Hs00174103_m1 for IL-8, Hs00174131_m1 for IL-6, and Hs02758991_g1 for GAPDH.
Gene Silencing of OGT and NFAT Isoforms Using siRNA 16HBE cells were used for NFAT and OGT siRNA-mediated knockdown (KD) experiments as previously described (5,25). Briefly, 5 × 10 4 cells were seeded in coated 24-well plates and transfected for 6 h in 0.5 mL OptiMEM with 5 nmol of either AllStar negative control or OGT siRNAs (Thermo Scientific, Grand Island, NY, USA) or NFATC2 (NFAT2c) or NFATC3 (NFAT3c) siRNAs (Qiagen; Hilden, Germany) using 1.5 µL/well of Qiagen HiPerFect transfection reagent. Following the transfection, medium was replaced with complete medium and the cells were subjected to an additional 48 h incubation to allow for NFAT or OGT knockdown. After a 24 h treatment with FGF23, wells were washed with 1.0 mL cold phosphate buffered saline (pH 7.4) and RNA was extracted using the GeneJET RNA purification kit (Thermo Scientific, Grand Island, NY, USA). For some experiments, HBECs were transfected with siRNA against OGT or NFAT2c/3c and 100 ul of conditioned medium was collected for ELISA.

NFAT Luciferase Reporter Assay
As described previously (5), 16HBE cells plated at 1.5 × 10 4 per well in a coated 96-well plate were transfected with 100 ng of DNA mixture containing the constitutively-active Renillaluciferase construct, as a transfection control, together with a Fire Fly Luciferase reporter construct, and a NFAT reporter construct serving provided with the NFAT Cignal Reporter Assay Kit (Qiagen; Hilden, Germany). Transfection was performed overnight under serum-free conditions in OptiMEM using Lipofectamine 2000 transfection reagent (Thermo Scientific, Grand Island, NY, USA). Cells were collected after an additional 24 h treatment and a Luciferase assay was performed using the Dual-Luciferase Reporter Assay System (Promega; Madison, WI, USA) as directed by the manufacturer (Promega; Madison, WI, USA). Relative light units (RLUs) were measured utilizing a SpectraMax i3x plate reader equipped with dual injectors (Molecular Devices; Sunnyvale, CA, USA).

Statistics
Data were analyzed with Prism5 (GraphPad Software, Inc., La Jolla, CA) and shown as mean ± SEM using Student's t-test and analysis of variance or Kruskal-Wallis H-test with one-way ANOVA with appropriate post tests for at least three independent experiments. Significance was accepted at p < 0.05.

Modulation of O-GlcNAc Is Regulated Through the PLCγ Signaling Pathway Upon Stimulation With FGF23
To determine whether FGF23 activates PLCγ (FGFR4-mediated), ERK (FGFR1-mediated), or both FGF23 signaling pathways in HBECs, we immunoblotted for total and phosphorylated PLCγ and ERK proteins Phosphorylation of PLCγ and total PLC was significantly increased without change in ERK phosphorylation following To determine whether the changes in O-GlcNAc following FGF23 are regulated through the PLCγ signaling pathway, we blocked PLCγ activation using a PLCγ inhibitor (U-73122), which has been shown to block FGF23 signaling through FGFR4 (27,28). As shown in Figure 1H, O-GlcNAc was dosedependently reduced following PLCγ inhibitor administration ( Figure 1I: NT: 0.92 ± 0.090; 0.1 µM PLC inhibitor: 0.76 ± 0.07; 1.0 µM PLC inhibitor: 0.54 ± 0.02; p = 0.0095). In addition, OGT and OGA protein levels were decreased after PLCγ blockade (Figures 1H,J

Knockdown of OGT Abrogates the O-GlcNAc Modification of Proteins and the Effects of FGF23 Signaling
Previous reports have shown that there is an extracellular OGT (eOGT), which resides in the ER and transfers the GlcNAc moiety to epidermal growth factor-like domains (29). Interestingly, it has been shown to have high expression in the lung (30). Therefore, we wanted to determine whether the increase O-GlcNAc levels, stimulated by FGF23, is transferred by OGT and not the ERresident eOGT in HBECs. To confirm that OGT is the sole enzyme responsible for O-GlcNAc transfer in these cells and affected by FGF23 signaling, we used siRNA targeted knockdown of OGT. As shown in Figure 2, FGF23 induced O-GlcNAc and

FGF23 Activates NFAT Through O-GlcNAc Augmentation in HBECs
Previously published reports have shown an association between O-GlcNAc signaling and NFAT regulation in cardiomyocyte hypertrophy (13,31) and in lymphocyte activation (32). In addition, we have recently shown that bronchial epithelial cells express NFAT2c and 3c as main isoforms and are activated by FGF23, which results in airway inflammation (5). To determine the role of FGF23 and O-GlcNAc in NFAT activation, we performed NFAT2c/3c activation assays using a luciferaseconjugated NFAT reporter gene in HBECs. As shown in Figure 4A, FGF23 significantly increased NFAT activation (Ctrl: 4.68 ± 0.32; FGF23: 6.68 ± 0.70, p = 0.035), p = 0.012), which is similar to our previous report (5) To determine whether knockdown of NFAT effects O-GlcNAc modification of proteins, we silenced NFAT2c/3c using siRNA.   Figure 4B, knockdown of NFAT2c/3c did not affect O-GlcNAc levels, or OGT/OGA protein expression. Altogether, these data combined suggests that NFAT activation is regulated through the FGF23 increase in O-GlcNAc levels, which lies upstream of NFAT in HBECs.

DISCUSSION
In this report, we show that FGF23 can increase O-GlcNAc levels as well as OGT and OGA protein expression in HBECs. This effect seems to be a downstream target via the FGFR4/PLCγ signaling pathway (FGFR1/ERK signaling is not affected) since PLC blockade resulted in decreased O-GlcNAc following FGF23 adminstration ( Figure 1H). Furthermore, both FGF23 and O-GlcNAc lead to increased secretion of IL-6, but not IL-8 (Figure 3). Conversely, reduction of O-GlcNAc levels (by OSMI-1) reduced IL-6 secretion. Upon assessing effects of the NFAT activation in HBECs (Figure 4), we found that FGF23 activated NFAT, which is consistent with our previous report (5). In line with this, O-GlcNAc modulation resulted in either an increase in NFAT activation (by blocking O-GlcNAc removal) or a decrease in NFAT upon inhibiting O-GlcNAc transfer (OSMI-1) or FGF23 signaling (through blockade of FGFR4). Interestingly, knockdown of NFAT 2c or 3c did not affect the O-GlcNAc levels or OGA/OGT expression ( Figure 4B). However, NFAT2c silencing did affect both IL-6 expression and secretion, while NFAT3c knockdown only affected IL-6 secretion ( Figure 5).
In our results, there were discrepancies in the mRNA expression of IL-6 compared to IL-8 with the different treatments (Figures 3, 5). In addition, there was higher IL-6 protein secretion compared to IL-8, which did not correlate with the respective mRNA levels. This was also similar with the IL-6 results for NFAT3c in Figure 5. These discrepancies in the mRNA expression and protein levels of cytokines have been documented in other studies (33). Therefore, caution should be used when interpreting mRNA expression as a proxy to protein levels. In addition, we cannot rule out the fact that the increased levels of IL-6 (or no change in levels of IL-8) in our study may be due to altered cytokine uptake/turnover following treatments. Nevertheless, these data combined suggest that FGF23 stimulation of O-GlcNAc levels is upstream of NFAT signaling in HBECs and regulates the secretion of the proinflammatory cytokine IL-6 ( Figure 6).
FGF23 has been characterized as a hormonal regulator of circulating phosphate and vitamin D levels as well as a prognostic risk factor for cardiovascular mortality in patients with chronic kidney disease (28,(34)(35)(36). The role of FGF23 role as an inflammatory facilitator has also been recently studied (5,37,38). Interestingly, it may be involved in several metabolic processes, including glucose and fat metabolism. For example, FGF23 was shown to contribute to insulin sensitivity in obese adolescents (39) and was altered in vitamin D deficient patients following an oral glucose load (40). Based on these findings, administration of FGF23 may alter metabolic pathways that are involved in glucose dysregulation and/or inflammation. The O-GlcNAc modification has long been studied and defined as cellular nutrient/stress sensor and the connection between the FGF23 and O-GlcNAc levels is plausible based on these previous reports and our results.
The O-GlcNAc modification regulates nuclear and cytosolic protein function and cellular signaling. Previous reports have shown a pro-and anti-inflammatory role for O-GlcNAc (23,41), and this biphasic effect is dependent on different cell types and diseases. Previous reports have also shown that glucosamine activation of the HBP attenuates NFκB activation in chondrocytes (42) or IL-1β mediation chondrocyte activation (43). In addition, inhibition of the NFκB pathway by O-GlcNAc has been shown in acute vascular injury models (44). On the other hand, the O-GlcNAc modification has been shown to activate NFκb under increase glucose concentrations in vascular smooth muscle cells in diabetes and obesity, suggesting a proinflammatory phenotype (45). This pro-inflammatory phenotype has been shown in other reports where HBP flux augments the oxidative stress pathways and the expression of other pro-inflammatory markers vascular cell adhesion molecule-1 (VCAM-1), IL6, IL-1β, TNFα, and NFκB (46). To our knowledge, we are the first to show the link between FGF23 and O-GlcNAc regulating IL-6 expression (Figure 3).
Several reports have shown that "outside-in" signaling can be modulated by O-GlcNAc. In T cell activation, antigen peptide binding to the T cell receptor have been shown to increase the O-GlcNAc modification of proteins that were associated with inflammatory cytokine production and cellular proliferation (47). In bone morphogenic protein (BMP) signaling on osteoblast differentiation, hyperglycemic conditions or activators of the HBP were shown to alter the O-GlcNAc levels and affect osteogenic differentiation (48) suggesting a role for glucose dysregulation on normal BMP function and signaling. In another report, FGF signaling was also shown to be altered by loss of function of a gene that encodes an enzyme in the HBP, which resulted in defective O-GlcNAc modification of the FGFR (49,50). This was shown to impair FGF mediated migration of mesodermal and tracheal cells during fly development. In addition, Miura, T. and colleagues demonstrated that O-GlcNAc modification of PKCζ blocks the signaling effects of FGF4, which resulted in the maintenance of an ESC undifferentiated state (51). A role for the O-GlcNAc modification was also shown for PLC inactivation and subsequent reduction of IP3 and Ca2+ mobilization in myoblast even in response to bradykinin (52). Similar to these findings, we show that FGF23 signaling, through PLCγ, increases the O-GlcNAc modification of proteins that may be involved in inflammatory cytokine production in HBECs (Figures 1-3). Interestingly, we also observed an increase in total PLCγ protein expression (Figure 1). The increase in total PLCγ expression has been shown in other inflammatory associated complications, including cancer and skin conditions (53) that may be associated with altered FGF23 levels (54)(55)(56) and is consistent with our findings. Collectively, we show that FGF23 signaling, through PLCγ, increases the O-GlcNAc modification of proteins that may be involved in inflammatory cytokine production in HBECs (Figures 1-3).
As stated above, NFAT signaling has also been linked to inflammatory cytokine production in hepatocytes, angiogenesis, cardiomyocyte hypertrophy, and many other biological processes (27,28,37,57). In addition, T-and B-lymphocytes activation, which has been documented to be regulated by NFAT (32), can be regulated by the O-GlcNAc modification, which may be required for its nuclear translocation. A similar phenomenon was shown in cardiomyocyte hypertrophy where the activation of NFAT has been linked to increased O-GlcNAc modification (13,58). Our previous report suggested a role for FGF23 activation of NFAT in the airway (5). However, no experiments have been done to determine the role of O-GlcNAc on NFAT activation through FGF23. Based on our results, combined with our previously published data, we put forth a model (Figure 6), whereby FGF23 regulates NFAT activation through the modulation of O-GlcNAc (Figures 4, 5), and stimulates IL-6 expression and secretion in HBECs (Figures 1-3).
Our findings in this report are the first to describe a role for FGF23 in the augmentation of O-GlcNAc levels. In addition, the role for FGF23 in the activation of NFAT through O-GlcNAc stimulation increases our knowledge of the molecules that may be involved in the process. The impact of O-GlcNAc by way of FGF23 will open new avenues for research in lung diseases associated with chronic airway inflammation such as COPD, cystic fibrosis, and asthma as well as metabolic disorders including diabetes and heart failure.

AUTHOR CONTRIBUTIONS
SK and JB conceived, designed, and analyzed all experiments and wrote the manuscript. EH, SH, SB, JG, RD, RZ, and HW performed experiments that were conceptualized and guided by JB and SK.