Thromboxane A2 or Activated Platelets Slightly Lower Fgf23 Expression in vitro Open Access

Platelets or thrombocytes play a major role in the pathophysiology of cardiovascular and further diseases [1, 2]. When activated in the bloodstream, e.g., by thrombin, they produce thromboxane A2 (TxA2), an eicosanoid, by means of their cyclooxygenase [3, 4]. TxA2 is effective through its receptor, a G protein-coupled receptor mainly expressed by thrombocytes and smooth muscle cells in the vasculature [3, 5]. Stimulation of this receptor elicits platelet aggregation, thereby favoring the formation of a blood clot, and vasoconstriction, both critical events of hemostasis, to stop bleeding [3, 5]. However, enhanced platelet activation and aggregation without a relation to an injured vessel may lead to severe and common medical conditions, i.e., thrombosis [6‒8]. Arterial thrombosis is strongly associated with arteriosclerosis and may lead to frequent cardiovascular diseases such as myocardial infarction, stroke, or peripheral arterial disease (PAD) [9, 10]. Since thrombocyte aggregation is particularly relevant for arterial thrombosis, patients at risk for myocardial infarction, stroke, or PAD are successfully treated with platelet aggregation inhibitors (e.g., acetylsalicylic acid, clopidogrel) [11‒13]. Venous thrombosis may lead to deep vein thrombosis, ultimately resulting in pulmonary embolism, and is mainly due to an activated endothelium as a consequence of inflammation that results in the attachment of blood cells including thrombocytes [14‒16].

Under physiological conditions, fibroblast growth factor 23 (FGF23) is an endocrine regulator of phosphate and vitamin D homeostasis in the kidney derived from bone [17]. In kidney and cardiovascular diseases, elevations of plasma FGF23 concentration occur early and predict outcomes [17‒19]. Also, in myocardial infarction, the plasma FGF23 level goes up [20]. Under pathophysiological conditions, further organs produce FGF23, including polycystic kidneys [21] or heart in pressure-induced cardiac hypertrophy [22]. FGF23 does not only indicate kidney or cardiovascular disease [23], but also drives pathophysiological processes, e.g., induction of left ventricular hypertrophy [24]. FGF23 induces cellular effects through receptor proteins of the FGFR family in the membrane [17, 25]. Whereas the physiological endocrine effects of FGF23 in the kidney are dependent on the transmembrane protein αKlotho as a co-receptor [17], at least some cardiac effects are αKlotho-independent [24, 26]. The pivotal role of αKlotho and FGF23 in controlling phosphate and vitamin D metabolism is illustrated by the fact that either αKlotho- or FGF23-deficient mice age very fast, thereby suffering from many aging-associated diseases typical of human aging [27‒29]. When maintained on low-phosphate or low-vitamin D, the animals are phenotypically quite normal [30].

The regulation of FGF23 production is dependent on other regulators of phosphate and vitamin D metabolism (calcitriol [active vitamin D]) [31, 32], parathyroid hormone [33] and, to a lesser extent, alimentary phosphate [34] and further nutritional (e.g., iron [35], vitamin A [36]), hormonal (e.g., insulin [37], erythropoietin [38]), intracellular (e.g., AMP-dependent kinase [39], mTOR signaling [40]), and extracellular (e.g., inflammation [35], TGFβ [41]) factors. Notably, also prostaglandin E2 (PGE2), another eicosanoid, is a powerful regulator of FGF23 [42]. This and the broad role of both TxA2 and FGF23 for cardiovascular health and disease, tempted us to speculate that TxA2 may also be a regulator of FGF23 synthesis. The present study was conducted to verify this hypothesis. Experiments were performed in two cell lines of bone origin – bone is the main site for FGF23 production under physiological conditions [43] – and FGF23 was determined by qRT-PCR and enzyme-linked immunosorbent assay (ELISA).

Rat osteoblast-like UMR-106 cells (CRL-1661; ATCC, Manassas, VA, USA) were cultured in Dulbecco’s modified Eagle medium with high glucose (4.5 g/L; Gibco, Life Technologies, Thermo Scientific, Darmstadt, Germany) and 10% fetal bovine serum (FBS; Gibco, Life Technologies) as well as 100 U/mL penicillin and 100 µg/mL streptomycin (Gibco, Life Technologies) at 37°C and 5% CO₂. They were seeded in 6-well plates (200,000 cells per well) (Greiner Bio-One, Frickenhausen, Germany). After 24 h, cells were treated with or without 1–100 ng/mL of pinane thromboxane A2 (pTxA2; Cay19020-500; Biomol, Hamburg, Germany) [44], 1–100 ng/mL of the thromboxane A2 receptor (TxA2 receptor) agonist I-BOP (Cay19600-100; Biomol), and/or 1 µm of the TxA2 receptor antagonist SQ29548 (Cay19025-1; Biomol) for further 24 h in the presence of 10 nm calcitriol (1,25(OH)₂D₃, Tocris, Bio-Techne, Wiesbaden-Nordenstadt, Germany). In another series of experiments, UMR-106 cells were seeded in growth medium supplemented with 10 nm calcitriol, grown for 24 h and then additionally treated with 1–5 µm of thromboxane A2 receptor agonist U46619 (Cay16450-1; Biomol), 1–3 nm serotonin hydrochloride (H9523-25MG; Sigma-Aldrich, Merck, Darmstadt, Germany), 0.3–10 µm adenosine diphosphate (Cay21121-1; Biomol), or 50 nm H-89 dihydrochloride (2910; Tocris, Bio-Techne) for further 24 h. For co-culture experiments, UMR-106 cells were seeded in growth medium supplemented with 10 nm calcitriol, grown for 45 h, and then co-incubated with or without 100,000/µL freshly isolated platelets from human blood with or without additional 1 U/mL human thrombin (T6884-100UN; Sigma-Aldrich, Merck) for further 3 h.

MC3T3-E1 subclone 4 mouse pre-osteoblast cells (CRL-2593; ATCC) were cultured in proliferation medium consisting of α-minimum essential medium (αMEM) with 2 mml-glutamine and nucleosides (Gibco, Life Technologies), 10% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin. For experiments, cells were seeded on rat tail type I collagen-coated 12-well plates (80,000 cells per well) and grown for 24 h in proliferation medium. Then, medium was replaced with αMEM with 2 mml-glutamine and nucleosides, 10% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin, 50 μg/mL ascorbic acid and 4 mm β-glycerophosphate to induce osteogenic differentiation. MC3T3-E1 cells were cultured in differentiation medium for 6 days, with the medium being changed every 2–3 days. Twenty-four hours before cell harvesting, cells were incubated with 10 nm calcitriol. A 6-h treatment with 1–100 ng/mL pinane thromboxane A2, 1–100 ng/mL I-BOP or 1–5 µm U46619 followed. Where appropriate, control cells were incubated with the respective amount of solvent ethanol (Carl Roth, Karlsruhe, Germany), dimethyl sulfoxide (AppliChem, Darmstadt, Germany), methyl acetate (Carl Roth), phosphate-buffered saline, water, or 1× Tyrode’s buffer (pH 7.4).

Venous blood was collected from healthy male and female adults by standard technique into syringes with acid-citrate-dextrose buffer (0.116 m monosodium citrate, 0.07 m citric acid, 0.11 m glucose; all chemicals from Carl Roth) at a pH of 4.6 and a ratio of 1:5 (blood:buffer). The whole blood was centrifuged at 200 g for 20 min in an unrestrained centrifuge. Since the upper fraction contained the platelets, this was carefully removed and transferred to a new vial. Next, 1× Tyrode’s buffer (pH 6.5; 10% 10× Tyrode’s buffer [1.4 m NaCl, 0.12 m NaHCO₃, 0.026 m KCl], 0.1% BSA, 0.1% glucose; all chemicals from Carl Roth) was added. After centrifugation at 900 g for 10 min, the cell pellet was resuspended in 100 µL 1× Tyrode’s buffer (pH 7.4). The platelet count was determined using the hematology device scil Vet abc Plus (scil animal care company, Viernheim, Germany).

Total RNA was isolated from UMR-106 and MC3T3-E1 cells and heart tissue from wild-type mice (C57BL/6; Charles River Laboratories, Sulzfeld, Germany) using a phenol-chloroform-based method (TriFast Reagent, VWR, Bruchsal, Germany). For cDNA synthesis with the GoScript Reverse Transcription System and random primers (both from Promega, Mannheim, Germany) on a Biometra TAdvanced thermocycler (Analytik Jena, Jena, Germany), 1.2 µg of total RNA was used.

Quantitative real-time polymerase chain reaction (qRT-PCR) was carried out on a CFX Connect Real-Time PCR Detection System (Bio-Rad, Feldkirchen, Germany) using 2 µL of cDNA. The reaction mix contained: 0.25 µm (rat Fgf23) or 0.5 µm (rat TATA box binding protein [Tbp] and mouse Fgf23, Tbp and Txa2r) primers, 10 µL GoTaq qPCR master mix (Promega), and sterile water. PCR conditions were 95°C for 2 min; 40 cycles of 95°C for 10 s, annealing at primer-specific temperature for 30 s; and 72°C for 25 s.

The following primers (5′ → 3′) and annealing temperatures were used for qRT-PCR: Fgf23, 58°C (rat): TAGAGCCTATTCAGACACTTC and CATCAGGGCACTGTAGATAG; Tbp, 58°C (rat): ACTCCTGCCACACCAGCC and GGTCAAGTTTACAGCCAAGATTCA; Fgf23, 57°C (mouse): TCGAAGGTTCCTTTGTATGGA and AGTGATGCTTCTGCGACAAGT; Tbp, 60°C (mouse): CCAGACCCCACAACTCTTCC and CAGTTGTCCGTGGCTCTCTT; Txa2r, 63°C (mouse): CCTGGTCTCCGACCTTACAC and TAGGGGAGTTCCCTTCGC.

Relative RNA transcript levels were calculated using the 2^−ΔCT method. Tbp was used as internal reference. About 10 µL PCR product as well as a non-template control (NTC) and a minus reverse transcriptase (–RT) control were loaded onto a 2% agarose gel and run for 20 min at 100 V.

Total RNA from UMR-106 and MC3T3-E1 cells was isolated as described above. RNA from rat UMR-106 cells was additionally subjected to DNase digestion using the NucleoSpin RNA Mini kit (Macherey-Nagel, Düren, Germany). For isolation of total RNA from bone (C57BL/6; Charles River Laboratories), femur and tibia were crushed with a mortar and pestle in liquid nitrogen and homogenized in TRI reagent (Thermo Scientific) by means of a tissue lyser. Total RNA was extracted with chloroform and processed with the RNeasy Mini Kit and RNase-free DNase Set (both from Qiagen, Hilden, Germany) according to the manufacturer’s protocol and cDNA was synthesized from 300 ng of total bone RNA. For PCR, 2 µL of cDNA, 0.25 µm (rat Txa2r) or 0.5 µm (mouse Txa2r) primers, 10 μL Green-Mix (Promega), and 6 μL sterile water were used. PCR conditions were 94°C for 3 min; 30 cycles of 94°C for 30 s, annealing at primer-specific temperature for 30 s, 72°C for 30 s, and 72°C for 5 min.

The following primers (5′ → 3′) and annealing temperatures were used for expression analysis: Txa2r, 60°C (rat): CTCCTCCTTCCTCGCTTTGC and GTGAGAAGGGCCGTGTGATG; Txa2r, 63°C (mouse): CCTGGTCTCCGACCTTACAC and TAGGGGAGTTCCCTTCGC. About 10 µL of PCR amplificate, NTC and –RT control were loaded onto a 2% agarose gel and run for 20 min at 100 V.

To determine C-terminal and intact FGF23 concentration in the cell culture supernatant, UMR-106 cells were seeded in the presence of 10 nm calcitriol, grown for 24 h and then treated with pTxA2, I-BOP or U46619 for further 24 h. Cell culture supernatant was collected and stored at −70°C. After thawing, it was concentrated using Vivaspin 2 centrifugal concentrators (Sartorius, Göttingen, Germany). The FGF23 protein concentration was determined using a mouse/rat C-terminal and a mouse/rat intact FGF23 ELISA kit (both from Quidel, San Clemente, CA, USA).

Data are given as arithmetic means ± standard error of the mean. N represents the number of independent experiments. Shapiro-Wilk test was used to test for normal distribution. Two groups were analyzed with two-tailed paired t test. More than two groups were analyzed with one-way analysis of variance followed by Dunnett’s multiple comparison test or Šidák’s multiple comparison test or were analyzed with two-way analysis of variance. Differences were considered significant if p < 0.05. Statistics were analyzed using GraphPad Prism version 10.3.1 (GraphPad Software; Boston, MA, USA).

This study sought to verify the hypothesis that TxA2 regulates FGF23 production. UMR-106 osteoblast-like cells and MC3T3-E1 cells, cultured for 7 days in differentiation medium, were used to verify it. As a first step, we studied expression of TxA2 receptor in these cells by means of PCR analysis. As depicted in Figure 1a, TxA2 receptor was expressed in UMR-106 cells. Also in mouse bone, TxA2 receptor expression was confirmed (Fig. 1b). The same holds true for mouse MC3T3-E1 cells and mouse heart (Fig. 1c). However, qRT-PCR revealed that TxA2 receptor expression was higher in mouse heart compared to MC3T3-E1 cells (Fig. 1d).

Next, UMR-106 cells were simultaneously exposed to calcitriol and increasing concentrations of pTxA2 for 24 h, and Fgf23 transcripts were quantified as a proxy for FGF23 production. According to Figure 2a, 100 ng/mL pTxA2 slightly but significantly reduced Fgf23 gene expression in UMR-106 cells. Also in MC3T3-E1 cells pretreated with calcitriol for 24 h and then incubated with and without pTxA2 for further 6 h, pTxA2 significantly reduced Fgf23 mRNA abundance (Fig. 2b). Adenosine diphosphate and serotonin, two other mediators of platelet function and activation, did not significantly affect Fgf23 gene transcription in UMR-106 cells (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000545696).

We additionally tested two pharmacological TxA2 receptor agonists, I-BOP and U46619. None of the pharmacological manipulators of TxA2 significantly affected viability of UMR-106 cells (online suppl. Fig. 2). As illustrated in Figure 3a, ≥100 ng/mL I-BOP significantly suppressed Fgf23 gene expression within 24 h in UMR-106 cells treated simultaneously with calcitriol. The same effect was observed in MC3T3-E1 cells pretreated with calcitriol for 24 h and further incubated with or without I-BOP for 6 h (Fig. 3b). The I-BOP effect on Fgf23 was not significantly affected by H-89 (online suppl. Fig. 3), an inhibitor of protein kinase A (PKA), suggesting that the cAMP-PKA-axis, relevant for at least some of the effects of TxA2 [45], is not required by TxA2 signaling to regulate Fgf23 transcription. Finally, UMR-106 cells were first treated with calcitriol for 24 h and then further incubated in the presence or absence of varying concentrations of TxA2 agonist U46619 for additional 24 h. According to Figure 4a, ≥5 µm U46619 significantly lowered Fgf23 gene expression. Similarly, U46619 reduced Fgf23 transcripts in MC3T3-E1 cells (Fig. 4b) pretreated with calcitriol for 24 h and additionally incubated with and without U46619 for 6 h.

Next, we explored whether the effect of TxA2 signaling on Fgf23 transcripts translates into altered FGF23 protein abundance. We treated UMR-106 cells with or without pTxA2, I-BOP, or U46619 and determined the protein concentration of both, C-terminal FGF23 and intact FGF23 in the cell culture supernatant. As a result, treatment with pTxA2 (Fig. 5a) tended to lower and treatment with I-BOP (Fig. 5b) as well as U46619 (Fig. 5c) significantly reduced C-terminal FGF23 protein concentration in the cell culture supernatant of UMR-106 cells. Intact FGF23 protein concentration was very low under all conditions tested, and we were unable to detect statistically significant effects of either treatment (online suppl. Fig. 4).

In order to confirm the contribution of TxA2 receptor to I-BOP-mediated suppression of Fgf23 transcripts, we next employed TxA2 receptor antagonist SQ29548. As shown in Figure 6, I-BOP again down-regulated Fgf23 gene expression, an effect reversed by SQ29548 in UMR-106 cells.

The natural source of TxA2 in the body is activated thrombocytes in the bloodstream. We isolated human thrombocytes from healthy volunteers and co-incubated them for 3 h with UMR-106 cells pretreated with calcitriol for 45 h, resulting in a total cell culture duration of 48 h. The mere co-incubation of the isolated platelets with UMR-106 cells did not significantly affect their Fgf23 transcript levels within 3 h (Fig. 7). Thrombin in plasma is a physiological activator of thrombocytes. Thrombin alone did not significantly affect Fgf23 gene expression of UMR-106 cells when added to the cell culture medium for 3 h (Fig. 7). However, the addition of both, thrombin and freshly isolated platelets, significantly reduced the Fgf23 mRNA abundance of UMR-106 cells within 3 h (Fig. 7).

This study found that TxA2 negatively regulates Fgf23 gene transcription and C-terminal FGF23 protein production through TxA2 receptor. Our expression analysis revealed that UMR-106 osteoblast-like cells and MC3T3-E1 differentiated for 7 days express TxA2 receptor, the major physiological ligand of which is TxA2 [46]. Bone, the main source of FGF23 under physiological conditions [17], has already been demonstrated to be a target of TxA2 [47], and we confirmed bone expression of the TxA2 receptor. We now demonstrated that pTxA2 down-regulates Fgf23 mRNA abundance in a dose-dependent manner when administered directly. Albeit small in size, the effect was statistically significant.

In order to better characterize the significance of TxA2 receptor for the regulation of Fgf23 gene expression, we performed additional experiments with pharmacological manipulators of TxA2 receptor. Two distinct and stable TxA2 receptor agonists, I-BOP and U46619 even more potently suppressed Fgf23 gene expression in both, UMR-106 and MC3T3-E1 cells. Moreover, TxA2 receptor agonists slightly reduced the C-terminal FGF23 concentration in the supernatant of UMR-106 cells, however, to a varying degree and not statistically significant for some experiments. We did not find a significant effect of either treatment on intact FGF23 protein concentration. When interpreting this experiment, it is mandatory to keep in mind that the intact FGF23 concentration was very low in all experiments and below the range of the sensitivity of the ELISA. Therefore, it is not possible to draw safe conclusion for the change in intact FGF23 in response to TxA2 signaling. Nevertheless, it is fair to say that TxA2 agonism down-regulates Fgf23 gene expression, an effect likely associated with lower FGF23 protein production. Further experiments with SQ29548, a TxA2 receptor antagonist, confirmed that the I-BOP effect on Fgf23 is indeed dependent on the TxA2 receptor. Thus, it appears safe to conclude that TxA2 receptor is a negative regulator of FGF23 production.

In vivo, activated platelets are the main source of TxA2 [3, 5]. Since our cell culture experiments suggested that platelet activation may suppress FGF23 formation in bone, we performed further experiments with freshly isolated platelets from healthy donors. Their co-incubation with UMR-106 cells in the absence of platelet activators did not significantly affect Fgf23 transcripts. This indicated that unstimulated platelets do not affect FGF23. However, if UMR-106 cells were co-cultured with thrombocytes activated with thrombin that induces TxA2 production [48, 49], Fgf23 transcripts were significantly reduced. Notably, thrombin alone did not affect FGF23 significantly. Therefore, similar to direct agonism of TxA2 receptor, activated platelets suppressed Fgf23 gene expression in UMR-106 cells.

Our major finding, i.e., regulation of FGF23 formation by TxA2 receptor, may have broader clinical relevance in view of the fact that enhanced platelet activation is the culprit in common cardiovascular diseases [2, 50], including stroke [51, 52], myocardial infarction [53], or PAD [50, 54]. These and further cardiovascular diseases, in particular if they are associated with enhanced inflammation, have already been demonstrated to be associated with elevated FGF23 plasma levels [17, 19]. That platelet activation lowers FGF23 production, may therefore, at least at first glance, come as a surprise. However, it could be interpreted as an effort to counteract too strong elevations of the plasma FGF23 concentration associated with cardiovascular diseases [17‒19] as these could derange phosphate and vitamin D metabolism or amplify pathophysiological effects of FGF23. Moreover, it must be kept in mind that in myocardial infarction, a condition associated with activated platelets, cardiac FGF23 production has been observed, a finding that may explain why this condition leads to higher FGF23 levels despite activated platelets suppressing FGF23 production in bone [55]. Moreover, common platelet aggregation inhibitor therapy with aspirin targeting TxA2 formation can be expected to impact FGF23 levels, a hypothesis that needs to be addressed in clinical studies. It is a limitation of this study that FGF23 expression was ramped up by calcitriol in both cell models and not by proinflammatory cytokines as those may induce Fgf23 expression in a different manner [56]. A recent study uncovered another eicosanoid, PGE₂, as a powerful stimulator of FGF23 production [42]. Chronic kidney disease (CKD) is a particularly prominent condition associated with very high levels of both, PGE₂ and FGF23. However, also TxA2 production is ramped up in CKD [57] and would, according to our study, be expected to lower FGF23. However, it must be kept in mind that the FGF23-lowering effect of TxA2 signaling was low in our study. In addition, it appears to be likely that the FGF23-lowering effect of TxA2 is clearly outweighed by several powerful FGF23 inducers effective in CKD including PGE2 [42], TNFα [58], IL-1 [59], IL-6 [60], EPO signaling [61], or parathyroid hormone [62].

Myocardial infarction induces an inflammation-dependent impairment of the function of bone vasculature [63]. It is tempting to speculate that activated platelets contribute to this impairment and may regulate bone FGF23 production. Activated platelets are also involved in bone fracture healing [44] and may therefore impact FGF23 production upon bone fracture.

The lack of in vivo experiments in our study is a clear limitation. Without doubt, further in vivo studies are needed to more precisely define the role of TxA2 and platelet activation for FGF23. Moreover, studies addressing the impact of platelet aggregation inhibition therapy on FGF23 are warranted.

Our study demonstrated that TxA2 signaling is a negative regulator of FGF23 production in UMR-106 and MC3T3-E1 cells, as is platelet activation. These results may have clinical relevance in view of the broad use of platelet aggregation inhibition therapy.

The authors would like to thank Dr. Patrick Münzer and Katharina Hammerschmidt for their experimental support and Katharina Kohm for graphical assistance.

This study protocol was reviewed and approved by the Ethics Committee of the State Physician Board of Baden-Württemberg, Approval No. F-2022-054. Written informed consent was obtained from all study participants. The data were collected in accordance with the Declaration of Helsinki. The animal study was reviewed and approved by the state government of Baden-Württemberg, Approval No. RPS35-9185-99/388, and were conducted in accordance with the federal law for the welfare of animals.

Michael Föller received speaker fees from Kyowa Kirin without relevance for this study. All other authors declare no conflict of interest.

This study was supported by Deutsche Forschungsgemeinschaft (DFG) (Grant No.: Fo695-2/2).

Michael Föller designed the research. Elena Kohm, Steffen Rausch, Julia Vogt, and Martina Feger performed the experiments and analyzed the data. Elena Kohm and Steffen Rausch made statistical analysis. Elena Kohm, Steffen Rausch, and Michael Föller wrote the manuscript.

Elena Kohm and Steffen Rausch contributed equally and thus share first-authorship.

All data generated or analyzed during this study are included in this article and its online supplementary material files. Further inquiries can be directed to the corresponding author.