Reactive oxygen species (ROS) generation as an underlying mechanism of inorganic phosphate (P_i)-induced mineralization of osteogenic cells

Abstract

Reactive Oxygen Species (ROS) are a natural byproduct of oxygen metabolism. At physiological levels, ROS regulate multiple cellular processes like proliferation, migration, and differentiation. Increased levels of ROS are associated with pathological conditions, such as inflammation and vascular calcification, where they elicit cytotoxic effects. These contrasting outcomes of ROS have also been reported in osteogenic precursor cells. However, the role of ROS in committed osteogenic cells has not been investigated. Cytotoxic and physiologic effects have also been demonstrated for extracellular phosphate (P_i). Specifically, in committed osteogenic cells P_i stimulates their major function (mineralization), however in osteogenic precursors and endothelial cells P_i cytotoxicity has been reported. Interestingly, P_i cytotoxic effects have been associated with ROS production in the pathological vascular mineralization. In this study, we investigated a molecular mechanistic link between elevated P_i and ROS production in the context of the mineralization function of committed osteogenic cells. Using committed osteogenic cells, 17IIA11 odontoblast-like cell and MLO-A5 osteoblast cell lines, we have unveil that P_i enhances intracellular ROS production. Furthermore, using a combination of mineralization assays and gene expression analyses, we determined that P_i-induced intracellular ROS supports the physiological mineralization process. In contrast, the exogenous ROS, provided in a form of H₂O₂, was detrimental for osteogenic cells. By comparing molecular signaling cascades induced by extracellular ROS and P_i, we identified differences in signaling routes that determine physiologic versus toxic effect of ROS on osteogenic cells. Specifically, while both extracellular and P_i-induced intracellular ROS utilize Erk1/2 signaling mediator, only extracellular ROS induces stress-activated mitogen-activated protein kinases P38 and JNK that are associated with cell death. In summary, our results uncovered a physiological role of ROS in the P_i-induced mineralization through the molecular pathway that is distinct from ROS-induced cytotoxic effects.

Graphical Abstract:

The proposed mechanism underlying differential osteogenic cells responses to P_i-induced intracellular ROS and exogenous ROS: P_i-induced ROS (green) increase mineralization by activating Erk1/2 MAPKs and increasing osteogenic gene expression without inducing osteogenic cell death. Exogenous ROS (red) activates Erk1/2 MAPKs as well stress-associated MAPKs (P38 and JNK1/2), which may drive cytotoxic effect of the extracellular ROS.

Introduction

Reactive Oxygen Species (ROS) production is a physiological consequence of incomplete reduction of molecular oxygen, which results in generation of superoxide anions (O₂⁻), hydrogen peroxide (H₂O₂), highly reactive hydroxyl radicals (OH°) and hydroxide ions (OH⁻) [1]. At physiological levels, ROS regulate multiple cellular processes like proliferation, migration, and differentiation [1]. However, increased levels of ROS may evoke cytotoxic effect and are associated with pathological conditions, such as cancers, neurodegeneration, aging, and vascular calcification [2–5]. ROS are also produced in osteogenic cells. However, in vitro studies using osteogenic cells have shown increased differentiation as well as induction of cell death by ROS [6–9]. Thus, the reported outcomes of ROS activation in osteogenic cells are not only inconsistent, but often contradictory as well. These conflicting results may be due to different experimental models used, as effects of ROS depend on the type, concentration, and duration of a stimulus as well as the type of cells subjected to ROS. For example, H₂O₂ has been reported to inhibit differentiation of osteoblast precursor cells MC3T3-E1 and bone marrow stromal cell line M2–10B4, and enhance pathologic osteogenic differentiation of bovine aortic smooth muscles cells, revealing cell type-specific effects of ROS [4]. In contrast to these results, Naokatu et al. identified ROS as major contributor in MC3T3-E1 osteogenic differentiation [9]. Another study demonstrated that lower concentrations of H₂O₂ induce differentiation of odontoblasts but not MC3T3-E1 cells, while higher concentrations induced cell death in both cell types [10]. The importance of a stimulus and cell type-specificity of a response was shown in a study of deferoxamine (iron chelator). This study has demonstrated that deferoxamine enhances ROS production, which increases odontoblast differentiation but reduces osteoclast differentiation [7, 11]. In summary, there are multiple studies of ROS in osteogenic precursor cells (both osteoblast and odontoblast precursors), in which the role of ROS in differentiation has been assessed, however, the role of ROS in the function of differentiated osteogenic cells has not been studied.

Several extracellular stimuli, such as hormones, pro-inflammatory cytokines, glucocorticoids, and inorganic ions, influence ROS levels in osteogenic cells [12–17]. In particular, inorganic phosphate ion (PO₄³⁻, abbreviated here as P_i) is critical for development and homeostasis of all mineralizing tissues by acting as a signaling molecule affecting differentiation and function of cells in skeletal and dental tissues [18, 19]. However, the molecular mechanism underlying P_i-induced mineralization is not well understood. We have previously shown that P_i stimulates the mineralization function of committed osteogenic cells by increasing activity of extracellular signal-related kinases Erk1/2 and expression of genes involved in mineralization, i.e. Dmp1 and Spp1/Opn [20]. The potential role of ROS in P_i signaling has been suggested by a study on mitochondria isolated from rat tissues, where extracellular P_i has been shown to increase mitochondrial H₂O₂ production [21]. Furthermore, studies of endothelial cells and osteoblast precursor cells MC3T3-E1 have demonstrated that P_i-enhanced ROS production leads to cell death and decreased differentiation, respectively [13, 22]. This role of extracellular P_i and ROS production in cell differentiation and death, provide a rationale to analyze outcomes of P_i and ROS in committed osteogenic cells.

In this study, we use an in vitro approach to investigate the mechanistic link between P_i and ROS production specifically in the context of committed osteogenic cells and their primary function of mineralization of extracellular matrix (ECM). We focused on deciphering the molecular pathways that determine physiological versus cytotoxic effects of ROS during the initiation of the mineralization process. We hypothesize that P_i enhances ROS production in osteogenic cells that plays a role in P_i-induced mineralization.

Material and Methods

Cell lines and cell culture condition

We used mouse 17IIA11 odontoblast-like [20, 23–25] and MLO-A5 osteoblast (EKC003, Kerafast) [26] cell lines for this study. Cell lines were maintained as described before [20, 26]. Briefly, 17IIA11 cells were maintained in standard Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco; Thermo Fisher Scientific, Logan, UT) supplemented with 5% FBS (Thermo Fisher Scientific, Logan, UT), 100 units/ml penicillin and 100 μg/ml streptomycin (Cellgro, Manassas, VA) at 37°C and 5% CO₂. MLO-A5 cells were maintained in alpha Modified Eagle Medium (aMEM, Gibco, Thermo Fisher Scientific, Logan, UT) supplemented with 5% FBS, 5% FCS, 100 units/ml penicillin and 100 μg/ml streptomycin and 1x L-Glutamine at 37°C and 5% CO₂ [26]. For analyzing P_i signaling, 1.5 × 10⁶ cells/well were seeded on a 30 mm dish. When cells reached confluence, the growth medium was replaced with low-serum (0.5% FBS) medium for 16–18 h, followed by treatment with 5 mM P_i (in a form of Na-P_i buffer, pH=7.4) for 15 min [20]. For experiments with N-acetyl-cysteine (NAC), NAC was added to the medium, 45 min before P_i application. For analyses of gene expression, confluent cells were stimulated with 5 mM P_i for 24 h with or without NAC.

ROS measurements

ROS were assayed as described before [27]. In brief, 6×10⁵ cells per well were seeded in 8-well Lab-Tek chamber slides (Nalge Nunc, Rochester, NY, USA). After reaching confluence, cells were serum starved for 16–18 h and then stimulated with 5 mM P_i or 1 mM H₂O₂ (positive control), or 20 □l water (negative control) for 15 min followed by staining with 100 nM MitoTracker Red CM-H2XROS (Invitrogen, Eugene, OR, USA) for 30 min at 37°C. Then, medium was replaced with normal growth medium. Images were taken using Nikon Eclipse 90i microscope. For MitoTracker Red, CM-H2XROS 568 nm-filter was used with 1 sec light exposure. For the quantification of signal, grey values were measured using ImageJ software (http://imagej.nih.gov/ij). For each experimental condition, grey values from 100 cells were averaged.

RNA extraction, cDNA synthesis, and quantitative RT-PCR (qRT-PCR)

RNA was extracted using GenElute Mammalian Total RNA Miniprep Kit (Sigma Aldrich, St. Louis, MO, USA). Total RNA (1 μg), after DNase I treatment (Life Technologies, Grand Island, NY, USA), was converted to cDNA with SuperScript III Reverse Transcriptase kit (Life Technologies, Grand Island, NY, USA). Gene expression analyses were performed using AB Biosystems 7500 Fast RealTime PCR System and Fast SYBR Green reaction mix (Roche, Indianapolis, IN, USA). Primer sequences are: Gapdh F: GCAAGAGAGGCCCTATCCCAA R: CTCCCTAGGCCCCTCCTGTTATT; Opn F: ATGAGGCTGCAGTTCTCCTGG R: AAAGCTTCTTCTCCTCTGAGCTGCC; Dmp1 F: TCTTTGCTGTCGCTGGGGGTATCTTG R: AACATCTCTGTGCAAGCGA.

Western blot analysis

Whole protein extracts from cells were prepared using RIPA lysis buffer (50 mM Tris Cl pH=7.4, 150 mM NaCl, 2 mM EDTA, 1% NP40, 0.1% SDS) supplemented with phosphatase and protease inhibitors (1 mM NaF, 2 mM Na₂VO₄, 2 mM leupeptin, 2 mM pepstatin, 2 mM PMSF, and 10 μM MG132) or with cOmplete™, EDTA-free Protease Inhibitor Cocktail (Sigma). Protein concentration was determined by a Lowry protein assay (Bio-Rad). Protein (10 μg) was subjected to electrophoresis on 4–12% Bis Tris gels (Invitrogen) and transferred to a PVDF membrane (Fisher). The specific proteins were detected by using primary antibodies against pErk1/2 (4370), Erk1/2 (4695), Casp-3 (9662), PARP (9542), pSAPK-JNK (4668), pP38 (4511), from Cell Signaling and tubulin (t5168) from Sigma. The blots were developed with the Pierce™ ECL Western Blotting Substrate (Fisher, 32106). The intensity of the bands was measured by ImageJ (http://imagej.nih.gov/ij) and normalized by tubulin.

Mineralization Assay

The confluent cells on 6-well plate were cultivated in either growth medium or osteogenic medium (50 □g/ml ascorbic acid and 5 mM P_i). Media were changed every other day for 8 days. Mineralization was assessed using alizarin red (for calcium deposits) and von Kossa staining (for phosphate deposits). Cells were fixed in 4% paraformaldehyde and stained either with 40 mM alizarin red-S (Sigma) for 10 min or with 5% silver nitrate according to standard protocols. Stained cells were imaged using Nikon TS100 microscope in a bright field.

Fourier Transform Infrared Spectroscopy (FTIR) analysis

17IIA11 and MLO-A5 cells were cultured in osteogenic or growth media for 9 days. Then cells were washed with saline, collected and lyophilized overnight. Five mg of each lyophilized cell culture sample were ground and mixed with 135 mg of dry potassium bromide (KBr) powder and pressed into a pellet. FTIR spectra from three biological replicates per experimental group were collected at room temperature in the transmission mode with 64 scans per spectrum and the resolution of 4 cm⁻¹, using Bruker Vertex 70 FTIR spectrometer operated by OPUS software (Bruker Optics, Bellerica, MA). The spectra were further processed using Spectrum 10.02 software package (PerkinElmer) in the exactly the same way. Specifically, the automated baseline correction, 5 point running average smoothing and normalization using Amide I peak maximum as 1 were carried out. The crystallinity and mineral to organic ratio analyses were performed using Spectrum software package and Origin Pro 2016 (Origin Lab) according to published procedures [28, 29]. The details of the analyses are described in supplementary materials.

Statistical Analysis

Statistically significant differences were determined using the Student’s t-test except for data where multiple groups were compared with one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparison. A p-value of <0.05 was considered statistically significant. Experiments were repeated at least 3 times. Data are presented as mean ± standard deviation (SD). Statistical analyses were performed using Graph Pad Prism software.

Results

Extracellular P_i stimulates intracellular ROS production in osteogenic cells

To study molecular mechanisms regulating the initiation of the mineralization process we selected 17IIA11 pre-odontoblast and MLO-A5 osteoblast cell lines that have characteristics of committed osteogenic cells, determined by high expression of key osteogenic transcription factors and enzymes required for mineralization. This molecular makeup allows for rapid deposition of mineral in ECM under osteogenic conditions (in the presence of ascorbic acid and elevated P_i; Suppl. Fig. 1A) [23–26]. Additional advantage of using these cell lines is the opportunity to determine the conservation of studied interactions and molecular mechanisms in osteogenic cells of different embryonic origin (17IIA11 - cranial neural crest, MLO-A5 - lateral plate mesoderm). First, we verified that 17IIA11 and MLO-A5 cells respond to P_i by activating Erk1/2 kinases (Suppl. Fig. 1B) and increasing expression of classic P_i-responsive genes Dmp1, Opn (Suppl. Fig. 1C) [18, 30–33]. After validating that 17IIA11 and MLO-A5 cells demonstrate the classic response to P_i, we analyzed P_i-induced ROS production in these cell lines.

Previous studies have shown that P_i enhances ROS production and Oliveira et al. showed that P_i increase mitochondrial ROS production [13, 21, 22]. Therefore, we first compared levels of mitochondrial ROS in 17IIA11 and MLO-A5 cells maintained in standard growth conditions and upon exposure to elevated extracellular P_i (5 mM) for 15 min. We used this concentration and time point on the basis of previously published work from our lab demonstrating that 5 mM P_i activates Erk1/2 within 15 min in 17IIA11 cells [20] and the results of MLO-A5 characterization experiments detecting the same trend in this cellular model of mineralization (Suppl. Fig. 1B). To detect ROS, we used cell permeable reduced MitoTracker Red CM-H2XROS dye that fluoresces upon oxidation. To mimic the oxidative stress, 1 mM H₂O₂ was used as a positive control and, as expected, it showed a significant increase in the intensity of fluorescence. Stimulation of both 17IIA11 and MLO-A5 with 5 mM P_i for 15 min resulted in a significant increase in fluorescence intensity indicating increased oxidation of MitoTracker Red by ROS (Fig. 1). Taken together these results demonstrate that P_i increases mitochondrial ROS production in committed osteogenic cells.

Fig. 1. P_i increases intracellular ROS production in osteogenic cells.

Summary graphs of quantification of MitoTracker Red CM-H2XRos fluorescence intensities per cell volume for 17IIA11 (A) and MLO-A5 (B) cells revealing increase in ROS-induced fluorescence intensity in the cells stimulated with 5 mM P_i or 1 mM H₂O₂ for 15 min in comparison with cells receiving water (ctrl). Values are mean ± S.D. of three independent experiments; *p<0.05, **p<0.01.

Depletion of intracellular ROS reduces P_i-stimulated mineralization

In cell culture system, mineralization is typically induced by supplementing growth medium with ascorbic acid and phosphate (osteogenic medium). Commonly, β-glycerophosphate (βGP) is used as a source of phosphate, which requires alkaline phosphatase production by the cells in order to utilize phosphate provided in this form [25, 34]. Considering previous studies in limb bud mesenchymal cell cultures showing differences in formed mineral depending on the source of phosphate (organic μGP and inorganic Na-P_i), we first tested if using osteogenic medium with μGP versus P_i results in any difference in the quality or quantity of mineral deposited by 17IIA11 cells. FTIR analyses demonstrated that 17IIA11 cells deposit hydroxyapatite when grown in either osteogenic medium as evident from overlapping peaks of pure HA crystal and mineral deposits from 17IIA11 cells (Suppl. Fig. 2A and B). Most importantly, there is no difference in the amount and crystallinity of deposited mineral between the groups exposed to these two sources of phosphate (Suppl. Fig. 2C and D).

After validating our experimental approach, we addressed the biological significance of P_i-induced ROS in mineralization. For this, we depleted ROS from the system using NAC that quenches ROS and is commonly used to identify and test ROS inducers [35]. We compared the extent of mineralization in cells grown in osteogenic medium and in cells treated with NAC prior to exposure to osteogenic medium. As expected, robust mineralization, assessed by quantity of calcium (alizarin red staining, Fig. 2A and B) and phosphate deposits (von Kossa staining, Fig. 2C and D), was detected in both analyzed mineralizing cell models grown in the osteogenic medium for 9 days. However, addition of 10 mM NAC significantly reduced P_i-induced mineralization as demonstrated by decreased alizarin red and von Kossa staining (Fig. 2A–D). We further confirmed by FTIR (Fig. 2E and F) that NAC significantly reduced the ratio of mineral to organic components in both 17IIA11 and MLO-A5 cell cultures (Fig. 2G and H). Due to the low mineral content it was impossible to measure crystallinity in these samples. Collectively, these results revealed that P_i-induced ROS contributes to mineralization as depleting ROS significantly reduces P_i-induced mineral deposition by osteogenic cells.

Fig. 2. Depletion of ROS by NAC reduces P_i-induced mineralization of osteogenic cells.

Representative images of alizarin red (A, B) and von Kossa staining (C, D) of 17IIA11 (left side) and MLO-A5 cells (right side) maintained in growth medium (Ctrl) or osteogenic media (OM) (50 ug/ml ascorbic acid and 5 mM Pi) with or without 10 mM NAC for 9 days. Representative FTIR spectra (E, F) and quantification of mineralization (G, H) for 17IIA11 and MLO-A5 cells. Values are mean ± S.D. of three independent experiments; *p<0.05. Scale bar = 100 μm.

P_i-induced ROS is not detrimental for osteogenic cells

In an in vitro model of vascular pathology caused by hyperphosphatemia, high extracellular P_i has been shown to cause endothelial cell death through ROS-dependent mechanisms [22]. Considering this and other studies showing that ROS as well as P_i may have cytotoxic effects [36, 37], our next step was to determine whether the uncovered involvement of ROS in P_i-induced mineralization is a physiological response of osteogenic cells or if it reflects cytotoxicity of elevated P_i. First, we compared levels of cleaved forms of caspase-3 and PARP (molecular markers of apoptotic cell death) in 17IIA11 and MLO-A5 cells incubated in osteogenic medium for 24 h and in cells maintained in growth medium. Along with this, we employed H₂O₂ as an exogenous source of ROS to determine its effect on 17IIA11 and MLO-A5 cells. Protein analysis using Western blot revealed no significant difference in cleaved PARP or caspase-3 between cells in osteogenic medium and standard growth medium (Fig. 3A), while H₂O₂ treated cells have increased cleaved PARP and caspase-3 levels indicative of apoptosis. We further evaluated 17IIA11 cell viability upon P_i treatment by MTT assay. Our results demonstrate that 24 h P_i exposure significantly increased metabolic activity/proliferation of 17IIA11 cells. In contrast, 1 mM H₂O₂ treatment for the same duration significantly reduced cellular survival when compared to cells treated with 5 mM P_i and untreated cells (Suppl. Fig. 3A). These results demonstrate that P_i does not induce death of 17IIA11 and MLO-A5 cells. Hence, unlike the extracellular ROS (provided in the form of H₂O₂), P_i-induced endogenous ROS is not detrimental for osteogenic cells.

Fig. 3. P_i-induced ROS is not detrimental for osteogenic cells.

Representative Western blots showing cleaved PARP and Caspase-3 in 17IIA11 and MLO-A5 cells after 24 h (A) and 9 days (B) in either growth medium (Ctrl) or OM (50 ug/ml ascorbic acid and P_i 5 mM) with and without 10 mM NAC. α-tubulin was used as a control for protein loading. (n = 3 independent experiments).

Next, to evaluate whether reduced mineralization observed in osteogenic cells treated with ROS scavenger NAC (Fig. 2A–D) is due to deleterious effects of NAC on osteogenic cells, we also analyzed levels of cleaved PARP and caspase-3 in cells incubated in P_i-containing osteogenic medium supplemented with 10 mM NAC. During the short-term 24 h incubation, we observed no difference in cleaved PARP and caspase-3 levels between the cells incubated with either growth medium, P_i-containing osteogenic medium or P_i-containing osteogenic medium supplemented with 10 mM NAC (Fig. 3A). Microscopic analyses of 17IIA11 and MLO-A5 cells and nuclei morphology (Suppl. Fig. 3B) further corroborated our conclusion that neither P_i nor NAC treatment exert toxic effects on cells. We also assessed potential increase of cell death of 17IIA11 and MLO-A5 cells exposed to elevated P_i with or without NAC for 9 days, when mineralization reaches the plateau. Western blot analyses of cleaved PARP and caspase-3 detected no increase in apoptosis markers during long-term exposure of cells to these external stimuli (Fig. 3B). These results suggest that the cell death is not a predominant osteogenic cells response to elevated extracellular P_i neither to NAC at those experimental conditions.

In summary, these results demonstrate that NAC-mediated inhibition of P_i-induced mineralization is due to depletion of ROS rather than osteogenic cell death.

Extracellular ROS and P_i increase Erk1/2 activity in osteogenic cells

Although little is known about P_i signaling, it has been shown that in osteogenic cells, P_i activates Erk1/2 kinase and expression of osteogenic markers osteopontin (Opn) and Dentin Matrix Protein 1 (Dmp1) [20, 25, 32, 38]. Therefore, to identify a molecular link between P_i and ROS production in the mineralization process, we first compared H₂O_2-stimulated Erk1/2 activation with P_i-stimulated Erk1/2 activation. As expected, protein analyses by Western blot detected a significant increase in Erk1/2 phosphorylation in cells exposed to 5 mM P_i (Fig. 4A and C). Importantly, these analyses also revealed a dose-dependent increase in Erk1/2 phosphorylation after 15 min of H₂O₂ treatment in both 17IIA11 and MLO-A5 cells (Fig. 4A and C). Next, we analyzed whether extracellular ROS, in a form of H₂O₂, affects expression of osteogenic genes Dmp1 and Opn, which are targets of P_i signaling. Considering that 1mM H₂O₂ has a toxic effect on cells exposed to such high concentration of this form of ROS for 24h (Fig. 3), we used lower concentrations of H₂O₂ to test the hypothesis that ROS at lower levels stimulate osteogenic response but at higher concentration are detrimental. qRT-PCR analysis of RNA from 17IIA11 and MLO-A5 cells treated with H₂O₂ for 18 h revealed that 0.03 and 0.06 mM H₂O₂ is sufficient to significantly increase Dmp1 and Opn expression in 17IIA11 cells. In MLO-A5 cells a significant increase in Dmp1 and Opn expression was detected at 0.12 and 0.25 mM H₂O₂ (Fig. 4B and D). Together, these results show that extracellular ROS elicit a similar molecular response as P_i does in osteogenic cells, which suggests an existence of a cross-talk between their molecular pathways.

Fig. 4. Exogenous ROS (H₂O₂) increase Erk1/2 activity and osteogenic gene expression in osteogenic cells.

Representative Western blots and graphs (showing densitometry-based quantification of the Western blot results by ImageJ) of activated Erk1/2 in 17IIA11 cells 15 min after stimulation with H₂O₂ and 5 mM P_i (A). α-tubulin was used as the protein loading control. qRT-PCR showing Dmp1 and Opn expression in 17IIA11 cells 18 h after stimulation with H₂O₂ (B). Representative Western blots and graphs (showing densitometry-based quantification of the Western blot results by ImageJ) of activated Erk1/2 in MLO-A5 cells 15 min after stimulation with H₂O₂ and 5 mM P_i (C). α-tubulin was used as the protein loading control. qRT-PCR showing Dmp1 and Opn expression in MLO-A5 cells 18 h after stimulation with H₂O₂ (D). Values are mean ± S.D. of three independent experiments; *p≤0.05, **p≤0.01, ***P≤0.001.

Extracellular ROS but not P_i activate stress activated protein kinases (SAPK)

After determining that either H₂O₂ or P_i increase Erk1/2 (MAPK) activity and Dmp1 and Opn expression (Fig. 4), but only high (1mM) concentration of H₂O₂ for 24 h induces cell death (Fig. 3A and Suppl. Fig. 3A), our next step was to identify the molecular mechanism underlying the observed difference in the cell survival outcome. Previous studies have shown that activation of stress-activated protein kinases JNK and P38 (MAPKs) by ROS induce cell death in various cellular models [36, 39–42]. Therefore, we first analyzed activation of P38 and JNK upon H₂O₂ application. Western blot analyses revealed that 1 mM H₂O₂ exerts statistically significant increases in both P38 and JNK kinases activity within 15 min of application in 17IIA11 and MLO-A5 cells (Fig. 5A). Next, we compared the effect of H₂O₂ and P_i on P38 and JNK activity. Our results reveal that stimulation of cells with 5 mM P_i had no effect on stress-associated kinases activity, as opposed to 1 mM H₂O₂, which significantly increases P38 and JNK activity (Fig. 5B). As a positive control of activated cellular response to P_i, we used pErk1/2 detection by Western blot, which confirmed activation of P_i signaling in this experiment. These results reveal a difference in osteogenic cell responses to extracellular ROS and P_i, which may explain different outcomes of osteogenic cells exposure to these stimuli, i.e. increased mineralization in response to P_i and increased cell death in response to high extracellular ROS.

Fig. 5. Exogenous ROS but not Pi activate stress-activated MAPKs.

Representative Western blots and graphs (showing densitometry-based quantification of the Western blot results by ImageJ) showing increased levels of activated P38 and JNK1/2 in 17IIA11 cells stimulated with indicated concentration of H₂O₂ for 15 min (A). Western blots and quantification graphs showing increased levels of activated P38 and JNK in 17IIA11 cells 15 min after stimulation with 1 mM H₂O₂ but not by 5 mM P_i (B). Values are mean ± S.D. of three independent experiments; *p≤0.05, **p≤0.01.

Depletion of ROS impairs P_i-induced Erk1/2 activity and osteogenic gene expression

To understand the role of P_i-induced ROS in regulation of early molecular events supporting osteogenic cell function, we analyzed the consequences of ROS depletion on P_i-induced Erk1/2 activity and gene expression. For that, ROS scavenger NAC was used 45 min before the stimulation of 17IIA11 and MLO-A5 cells with 5 mM P_i. Quantitative Western blot analyses revealed that 10 mM NAC completely abolished activation of Erk1/2 in response to P_i (Fig. 6A and B). The dampen cellular response to P_i upon ROS depletion was further demonstrated at the gene expression level, where qRT-PCR data revealed that activation of Opn expression by P_i is significantly reduced by NAC (Fig. 6C, F). These results demonstrate that ROS plays an important role in activation of cellular response to P_i.

Fig. 6. Depletion of ROS impairs activation of Erk1/2 and osteogenic gene expression in response to P_i.

Representative Western blots and quantification graphs of Erk1/2 activity in 17IIA11 (A, B) and MLO-A5 cells (D, E) after 15 min of stimulation with P_i. NAC 10 mM was applied 45 min before P_i. qRT-PCR analysis of P_i-induced expression of Opn in 17IIA11 (C) and MLO-A5 (F) after 24 h of stimulation with P_i ±10 mM NAC. Values are mean ± S.D. of three independent experiments; *p≤0.05, **p≤0.01, ***p≤0.001, ****p≤0.0001 (ANOVA).

Discussion

In the majority of cell types, elevated extracellular P_i, similarly to high extracellular ROS, causes cell stress that may lead to cell death [4, 10, 13, 22]. However, in osteogenic cells, P_i stimulates their physiologic function, namely formation of mineralized ECM, raising a question about the involvement of P_i-induced ROS in mineralization. In this study, we have used committed osteogenic cells to investigate the role of ROS in determining the outcomes of the exposure of osteogenic cells to elevated P_i. We have uncovered that P_i enhances intracellular ROS production in osteogenic cells. Furthermore, our results revealed that, unlike extracellular ROS, which is detrimental for osteogenic cells, P_i-induced intracellular ROS plays a critical role in the physiological mineralization process. By comparing molecular signaling induced by extracellular ROS and P_i, we identified differences in signaling routes that determine physiologic versus toxic effects of ROS on the osteogenic cells. Specifically, while both extracellular and P_i- induced intracellular ROS utilize Erk1/2 signaling mediator, only extracellular ROS induces stress-activated MAPKs P38 and JNK that are associated with cell death. In summary, our results demonstrated a novel physiological role of ROS in the P_i-induced mineralization through the molecular pathway that is distinct from ROS-induced cytotoxic effects.

Bone is a mineralized connective tissue that is continuously remodeled through the specialized actions of bone-forming osteoblasts and bone-resorbing osteoclasts. Many systemic and local factors contribute to bone homeostasis by regulating bone cells function, survival, and death. ROS have been implicated in bone homeostasis by regulating fundamental processes including proliferation, differentiation, and cell death [17, 43–45]. Stimuli like glucocorticoids, Ca²⁺, PTH, cytokines have been shown to produce ROS in osteogenic cells in various experimental settings with different functional outcomes [14–18]. The role of ROS in attenuating osteogenic differentiation has been demonstrated in rat bone marrow mesenchymal stem cells (BMSCs) and OB-6 osteoblastic cell line [46, 47]. However, the cellular effects of ROS may differ depending on cell differentiation stage, due to different molecular makeup of progenitor versus mature cells. Therefore it is important to distinguish functions of ROS in progenitor cells from those in mature cells as some defects of mineralized tissues, like osteomalacia [48–50], are due to disturbed function of differentiated osteoblasts. However, the role of ROS in the physiology and function of mature osteogenic cells has not been evaluated yet. The major function of mature osteogenic cells is secretion and mineralization of ECM. In this study, we showed that extracellular P_i, which is a major stimulus initiating and supporting the mineralization process, increases levels of intracellular ROS (Fig. 1). Furthermore, by depleting ROS in P_i-stimulated osteogenic cells, we determined that P_i-generated ROS are a functional protagonist in the mineralization process (Fig. 2).

ROS have long been considered as a potential detrimental byproduct of the aerobic metabolism that has a negative effect on cell functions and survival (oxidative stress). It is now recognized that ROS generated at sub-toxic levels by different intracellular systems act as signaling molecules and mediate important physiological functions (redox biology). This dual constructive/destructive effect has also been suggested for P_i-induced ROS. For example, in an in vitro model of osteogenic differentiation (MC3T3-E1 cell line), P_i-induced ROS suppressed differentiation by inhibiting expression of osteogenic markers: alkaline phosphatase, osteocalcin, and Runx2, suggesting the negative effect of P_i-induced ROS in physiological osteogenic differentiation [13]. However, P_i-induced ROS has been also shown to enhance transdifferentiation of vascular smooth muscle cells into osteoblastic cells, which is one of the suggested pathomechanisms underlying medial vascular calcification [51]. In addition, P_i-generated ROS has been shown to induce apoptosis in endothelial cells, adding to the diversity of effects of P_i-induced ROS in various cell types [22]. To the best of our knowledge, our study is the first one that addressed the role of ROS in the function of differentiated osteogenic cells. Our results revealed that P_i-induced intracellular ROS support physiologic mineralization, as demonstrated by FTIR analyses (Fig. 2), without inducing cell death (Fig. 3). These results suggest that P_i, as a stimulus of physiological mineralization, generates levels of ROS that are in physiological range and support the physiological function of mature osteogenic cells.

The effects evoked by ROS differ not only depending on the cell type, but also on the site of ROS production (intracellular versus extracellular). This means that the same cell type may respond to the extracellular ROS differently than to the ROS produced intracellularly in response to extracellular stimuli. However, this has not been well-studied in osteogenic cells [17, 46]. In this study we have compared the effect of extracellular ROS (in the form of H₂O₂) and P_i-induced intracellular ROS in osteogenic cells. We found that both sources of ROS activate Erk1/2 kinase (Fig. 4 and 6). Extracellular ROS also increased the expression of classic P_i-responsive genes Dmp1 and Opn, although the magnitude of this activation was far lower than activation of these genes by P_i (Fig. 4 and 6C, F, and published data [25, 52]). However, unlike P_i, extracellular ROS is also detrimental for osteogenic cells as demonstrated by activation of apoptotic markers (cleaved PARP and caspase-3) and decreased metabolic activity of cells (Fig. 3, Suppl. Fig. 3). These results reveal that osteogenic cells respond differently to ROS depending whether ROS are provided as an extracellular stimulus or generated inside cells in response to P_i.

Contrasting outcomes of extracellular and P_i-induced intracellular ROS in osteogenic cells prompted us to investigate a potential overlap and differences in molecular pathways activated by these two types of extracellular stimuli. Although P_i signaling cascade is largely unknown, it has been established that P_i activates Erk1/2 in osteogenic cells [18, 20, 53, 54]. While Erk1/2 kinases can be activated by various stimuli, hence are considered “non-specific”, our previous studies demonstrated a specific role of P_i-activated Erk1/2 in release of matrix vesicles (MVs), which support the initiation of mineralization [20]. In this study, we found that in mature osteogenic cells both P_i and extracellular ROS activate Erk1/2 (Fig. 4). Importantly, we demonstrated that activation of Erk1/2 in response to P_i is diminished by ROS scavenger NAC (Fig. 6), placing ROS in the P_i signaling cascade in mature osteogenic cells.

Erk1/2 is one of three members of mitogen-activated protein kinases (MAPK). The other two, c-Jun N-terminal kinase (JNK) and the P38 kinase, also called stress activated kinases, can induce pathological cell death [39, 40, 42, 55]. Studies in MC3T3-E1 osteogenic precursor cells have demonstrated that P_i does not alter P38 and JNK activity [56]. Our data from 17IIA11 and MLO-A5 cells demonstrate that this is also true in mature osteogenic cells (Fig. 5). In contrast to P_i, stress-activated kinases JNK and P38 are induced by H₂O₂ (Fig. 5). These results reveal that mature osteogenic cells do not perceive elevated extracellular P_i as a stress factor, but as a physiological cue that supports their major function. Thus, these results increase our understanding of both the P_i signaling and ROS signaling pathways in mature osteogenic cells.

In this study, we implicated ROS in P_i-initiated molecular events leading to mineralization, by demonstrating that exogenous P_i stimulates ROS production in cells committed to formation of mineralized ECM. However, it is important to note that cellular ROS homeostasis is balanced by ROS production and detoxification of ROS by the antioxidant system. This system includes enzymatic ROS scavengers such as, Superoxide dismutases (SOD), Glutathione Peroxidase (GPxs), Thioredoxin (Trx), as well as nonenzymatic/exogenous antioxidants (for example: ascorbic acid, vitamin E and carotenoids) [57, 58]. In the in vitro cellular models of mineralization, ascorbic acid is used together with P_i to stimulate osteogenic differentiation and formation of mineralized ECM. Therefore, it is likely that the role of ascorbic acid in osteogenic differentiation and mineralization is not limited to stimulation of collagen production [59–61], but extends to maintaining ROS homeostasis. In this study our focus was placed on early signaling events initiated by P_i. Our follow up study will address the regulatory role of the antioxidant system in P_i-induced mineralization to investigate whether this arm of intracellular ROS regulation can be targeted in mineralization-associated pathologies.

A clear understanding of the role of ROS in various signaling pathways regulating cell differentiation and function, as well as in pathological processes will allow us to develop therapeutic strategies enhancing positive effects of ROS and minimizing the negative ones. Recent clinical trials targeting redox status of cancer cells demonstrate a significant benefit of redox-based therapies [58, 62, 63]. We have shown that intracellular ROS is involved in P_i-induced mineralization function of committed osteogenic cells, which provides a novel physiological function of ROS. Based on our data, we anticipate that by modulating P_i-induced ROS production it is possible to modulate the mineralization process. Such approaches could be used to stimulate mineralized tissue repair or to prevent vascular calcification in hyperphosphatemic conditions.

Supplementary Material

MMC1

Supplementary Fig. 1. Verification of cellular models of Pi-induced mineralization. Alizarin red staining showing calcium deposits (mineralization) in 17IIA11 and MLO-A5 cells grown in osteogenic medium (50 ug/ml ascorbic acid and 5 mM P_i) for 6 and 8 days (A). Representative Western blots revealing dose-dependent increase of pErk1/2 after 15 min of 17IIA11 and MLO-A5 cells stimulation with 2, 5 and 10 mM P_i. α-tubulin was used as a protein loading control (B). qRT-PCR analysis revealing increased expression of Dmp1 and Opn (fold change) in 17IIA11 and MLO-A5 cells after 24 h of stimulation with 5 mM P_i. (C). Values are mean ± S.D. of three independent experiments; *p<0.05, **p<0.01.

Supplementary Fig. 2. Replacing βGP with inorganic P_i in osteogenic medium did not affect mineral quality and quantity. Representative FTIR spectra of 17IIA11-produced ECM: (A) full range spectrum encompassing inorganic and organic material analyses (0–4000 cm-1) and (B) higher resolution of the spectrum in the range for mineral analysis (400–800 cm-1) obtained by FTIR. The values obtained from FTIR analyses with respect to average Cap/Amide I ratio (C) and crystallinity (amorphous CaP to crystal ratios) (D). Values are mean ± S.D. of three independent experiments; *p<0.05.

Supplementary Fig. 3. P_i-induced ROS is not detrimental for osteogenic cells. (A) MTT assay results for 17IIA11 cells cultured in the growth medium containing either 5 mM P_i or 1 mM H₂O₂ for 24 h. Values are mean ± S.D. of three technical repeats; ***p ≤ 0.001). (B) Microscopic images of DAPI fluorescent staining and bright field images of 17IIA11 and MLO-A5 cells. Cells were maintained in growth medium (Ctrl), OM or OM+NAC medium. Scale bar = 100 μm.

Highlights.

• Inorganic phosphate (P_i) generates reactive oxygen species (ROS) in committed osteogenic cells.

• Both exogenous ROS and P_i-induced intracellular ROS utilize Erk1/2 signaling mediator, however only exogenous ROS induces stress activated kinases P38 and JNK.

• In contrast to exogenous ROS, P_i-induced endogenous ROS is not detrimental to osteogenic cells.

• ROS generated in response to P_i contributes to physiological function of osteogenic cells.