Differing responses of osteogenic cell lines to β-glycerophosphate

Ascorbic acid (Asc), dexamethasone (Dex) and β-glycerophosphate (β-Gly) are commonly used to promote osteogenic behaviour by osteoblasts in vitro. According to the literature, several osteosarcoma cells lines appear to respond differently to the latter with regards to proliferation kinetics and osteogenic gene transcription. Unsurprisingly, these differences lead to contrasting data between publications that necessitate preliminary studies to confirm the phenotype of the chosen osteosarcoma cell line in the presence of Asc, Dex and β-Gly. The present study exposed Saos-2 cells to different combinations of Asc, Dex and β-Gly for 14 days and compared the response with immortalised human mesenchymal stromal/stem cells (MSCs). Cell numbers, cytotoxicity, mineralised matrix deposition and cell proliferation were analysed to assess osteoblast-like behaviour in the presence of Asc, Dex and β-Gly. Additionally, gene expression of runt-related transcription factor 2 (RUNX2); osteocalcin (OCN); alkaline phosphatase (ALP); phosphate regulating endopeptidase homolog X-linked (PHEX); marker of proliferation MKI67 and proliferating cell nuclear antigen (PCNA) was performed every two days during the 14-day cultures. It was found that proliferation of Saos-2 cells was significantly decreased by the presence of β-Gly which contrasted with hMSCs where no change was observed. Furthermore, unlike hMSCs, Saos-2 cells demonstrated an upregulated expression of late osteoblastic markers, OCN and PHEX that suggested β-Gly could affect later stages of osteogenic differentiation. In summary, it is important to consider that β-Gly significantly affects key cell processes of Saos-2 when using it as an osteoblast-like cell model.

www.nature.com/scientificreports/ produce data which is comparable. The present study compared the response of the Saos-2 osteosarcoma cell line and a non-carcinogenic cell model with Asc, Dex and β-Gly. Saos-2 is a commonly used osteoblast-like cell line derived from an 11-year-old Caucasian female patient suffering from primary osteosarcoma 14 . Saos-2 presents features characteristic of relatively mature osteoblasts including mineralised matrix synthesis, expression of 1,25-dihydroxyvitamin D3 receptors and elevated production of collagen-1 3 Furthermore, the relative ease of maintenance of Saos-2 in tissue culture and the short doubling time make it an attractive cell model for in vitro research. However, the cancerous nature of the Saos-2 cells causes several biological variations including altered proliferation kinetics and the absence of contact inhibition in vitro 15 . The latter contrasts markedly with the in vivo environment and therefore limits the physiological relevance of data obtained using these cells so a non-tumour derived cell model was used to compare the effects of Asc, Dex and β-Gly with those from Saos-2. An immortalised human bone marrow cell line (hMSC) was selected due to its non-cancerous origin in contrast with Saos-2. The hMSC line was immortalised via serial passaging combined with viral-mediated transfection of SV40 large T antigen 16 . However, despite a shorter doubling time than observed in primary cells, the hMSC line retains contact inhibition properties and the potential for adipogenic, chondrogenic and osteogenic differentiation. As a result, the hMSC line offers a physiologically relevant osteoblast-like model system while still allowing for the convenience of working with a cell line.
Considering the conflicting data discussed, the aim of the present study was to compare the effect of Asc, Dex and β-Gly on the osteogenic behaviour of the Saos-2 cell line and a relatively undifferentiated non-cancer derived hMSC line. This investigation provides a unique set of data demonstrating the relevance of Saos-2 as an osteoblast-like model system in vitro.

Materials and methods
Cell culture and osteogenic differentiation. Saos-2 and non-differentiated hMSCs were obtained from laboratory stores and Abm (T0520, Abm, Canada) respectively. Saos-2 cells were maintained in Dulbecco's Modified Eagle's Medium (DMEM)/Ham's Nutrient Mixture F12 (Sigma, UK) supplemented with 1% penicillin/ streptomycin, 1% l-glutamine and 10% foetal bovine serum (FBS). hMSCs were cultured in DMEM (low glucose, pyruvate) (Gibco, UK) with 0.5% penicillin/streptomycin and 10% FBS. Both cell lines were maintained at 37 °C in an atmosphere of 5% CO 2 . Cells were seeded at a density of 8.4 × 10 3 cells cm −2 and expanded for 72 h until confluence which was day 0 of the experiment. Osteogenic supplements were introduced on day 0 using 283 µM Asc, 9.3 mM β-Gly and 10 -8 M Dex. The following combinations of osteogenic supplements were established: control (no osteogenic supplementation); Asc Dex β-Gly; Asc Dex; Asc β-Gly; Dex β-Gly; Asc; Dex and β-Gly. Cells were incubated in variously supplemented culture media for up to 14 days with a change of media every 3 days. Cell growth in terms of numbers was examined every 2 days until day 14 of incubation. First, cell monolayers were detached from the polystyrene surface using 0.05% trypsin in 0.02% EDTA (Sigma, UK) for 10 min at 37 °C, centrifuged and resuspended in 1 ml 10% FBS in DMEM/Ham's F-12. Viable cell counts were performed using trypan blue staining and a Neubauer haemocytometer. Cytotoxicity assay. The lactate dehydrogenase (LDH) cytotoxicity assay was performed to examine cell death in cultures by quantifying the release of LDH. The LDH-Glo™ kit was used (Promega, UK) according to the manufacturer's instructions. Briefly, a 2 µl aliquot of culture media was collected on day 7 of incubation in osteogenic media and diluted in 50 µl LDH storage buffer before 25 µl of this solution was transferred into an opaque walled, transparent bottom 96 well plate and mixed with an equal volume of LDH detection reagent. Luminescence was recorded using a plate reader (Tecan Spark, Switzerland) after 60 min of incubation at room temperature. Two negative controls were used in this experiment: a no cells control-to assess the background luminescence of the culture media and no treatment control-to assess cell death in cultures without the osteogenic supplementation. A positive control with LDH was prepared by permeabilising the no treatment control cells with 10% Triton-X100 for 10 min immediately before collecting the culture media samples.
Alizarin red S staining. To examine mineral deposition in cell cultures Alizarin red S staining (ARS) was performed on day 14 of incubation in osteogenic media as described by Gregory et al. 17 . The cells were fixed with 10% formalin and stained with 40 mM Alizarin red S stain (pH 4.2) for 20 min at room temperature. Stain quantification was performed by extracting the stain with 10% acetic acid for 30 min followed by neutralisation with 0.1% ammonium hydroxide. The optical density of samples was quantified using a plate reader at 450 nm and compared with a standard curve of serial dilutions. Assessment of gene expression using qPCR. Total RNA was isolated every 2 days up to day 14 of osteogenic treatment using the RNeasy Mini kit (Qiagen, UK) according to the manufacturer's instructions. The samples were generated in 3 separate experiments to produce 3 biological replicates. Cell monolayers were dissociated using a lysis buffer followed by dilution with the 70% molecular grade ethanol. Then the suspensions were passed through the RNeasy spin column and centrifuged at 10,000 RPM for 30 s. Any non-specifically bound organic molecules and salts were removed using a wash buffer provided in the kit. Genomic DNA was digested using the On-Column DNAse kit (Sigma, UK) by adding the DNase solution directly to the samples and incubating for 15 min at room temperature. RNA was collected by adding 30 µl of RNAse-free water to the samples and centrifuging at 10,000 RPM for 1 min before storing the RNA samples at − 80 °C until later use. The quality of the resulting RNA was established using spectrophotometry (Genova Nano, Jenway™, UK).
Then, complementary DNA (cDNA) was produced from the isolated RNA via reverse transcription (RT) using the Tetro cDNA synthesis kit (Bioline, UK). The reaction master mix was prepared by combining 1 µl of 10 mM deoxynulceoside triphosphate (dNTP) mix, 4 µl of RT reaction buffer, 1 µl of RNase inhibitor and www.nature.com/scientificreports/ 1 µl of reverse transcriptase (200 U/µl). The reaction contained 8 µl of the master mix, 2 µg of RNA sample and DEPC-treated water to a final reaction volume of 20 µl. The mix was transferred to the thermocycler (Veriti, 96 well thermal cycler, Applied Biosystems, USA) and incubated at 45 °C for 1 h to allow RT followed by 85 °C for 5 min to terminate the reaction. The resulting cDNA samples were stored at − 20 °C before use. Finally, real-time PCR (qPCR) was performed using LightCycler® 480 SYBR Green I Master and 480 LightCy-cler® system (Roche Diagnostics). The expression of the following osteogenic markers was analysed: runt-related transcription factor 2 (RUNX2); osteocalcin (OCN); alkaline phosphatase (ALP) and phosphate regulating endopeptidase homolog X-linked (PHEX). The expression of a marker of proliferation (MKI67) and proliferating cell nuclear antigen (PCNA) were used to provide a comparison of proliferative markers. Sequences of the primers used are provided in supplementary materials. The gene expression levels were normalised to the expression of YWHAZ (tyrosine 3-monooxygenase). Data quantification was performed using the second derivative maximum method. This involved an algorithm identifying crossing points (Cp)-the number of PCR cycles after which the exponential phase of the target sequence amplification began. The starting concentration of target gene samples were found by comparing the Cp values against those in the standard curve of serial dilutions of log primer concentrations.
Determination of cell proliferation. Cell proliferation was examined on day 4 of incubation with full osteogenic culture media (Asc, Dex and β-Gly) to avoid examining confluent cell layers where proliferation may be compromised by contact inhibition. Cultures were incubated in 6 well plates. BrdU (5-Bromo-2′deoxy-uridine) uptake was measured to identify replicating cells according to the manufacturer's instructions (ab287841, Abcam, UK). Briefly, cells were exposed to the BrdU labelling solution for 6 h at 37 °C in 5% CO 2 . Samples were fixed with ethanol (50 mM glycine, 70% ethanol) for 20 min at − 20 °C and washed three times with the washing buffer provided in the kit. Then, samples were incubated with anti-BrdU antibody for 30 min at 37 °C and washed again, followed by incubation with anti-mouse Ig-fluorescein antibody for 30 min at 37 °C. Mounting media containing DAPI was applied for 10 min at room temperature to stain cell nuclei. BrdU and DAPI incorporation were examined using fluorescence microscopy (Eclipse TE300, Nikon, Japan). The following negative controls were established to prevent non-specific staining: (1) no BrdU, (2) no anti-BrdU antibody, (3) no secondary antibody. Images were captures using digital camera (D5100, Nikon, Japan). Brightness and contrast were enhanced equally in all images using Image J without obscuring or eliminating any details. BrdUstained and non-stained nuclei were counted manually.

Statistical analysis.
All experiments were established with three technical replicates (i.e., 3 wells) and undertaken independently 3 times. The data was presented as the median of 3 biological replicates as well as the minimum, maximum, lower quartile, and upper quartile values. Differences in parametric data were compared using Brown-Forthyse and Welch ANOVA tests and P < 0.05 was considered statistically significant. Statistical analysis and figures were generated using GraphPad Prism, Version 9.5.0 (GraphPad Software, San Diego, California US).

Results
Cell growth. The effect of osteogenic supplements on cell growth was studied by quantifying viable cell numbers for 14 days. Saos-2 cultures exposed to β-Gly demonstrated a significantly lower number of viable cells compared with Dex and/or Asc supplemented cultures after day 4 of incubation ( Fig. 1) and at all subsequent time points examined. In contrast, the number of viable hMSCs was not statistically significantly affected by any of the osteogenic supplements in comparison with controls ( Fig. 2). Cytotoxicity assay. To evaluate the cause of the reduced Saos-2 cell numbers in β-Gly-positive cultures an LDH assay was performed. No significant change in LDH release was observed between the osteogenically treated cells and controls (Fig. 3) in either Saos-2 (A) or hMCSs (B).

Proliferation assay. Saos-2 cells incubated in the presence of β-Gly showed significantly fewer (3.1%)
BrdU-stained nuclei than cells not exposed to β-Gly (Figs. 4, 6) with 35.6% of all nuclei being stained with BrdU. Hence, the proliferative activity of Saos-2 was decreased in the presence of β-Gly. No statistically significant difference was identified in the number of BrdU-stained nuclei in hMSCs with 23.7% of BrdU stained nuclei in β-Gly exposed cells and 25% of BrdU-stained nuclei in negative controls (Figs. 5, 6). Therefore, cell proliferation was not affected by the presence of β-Gly in hMSCs relative to control cultures. Negative controls with either no BrdU, no anti-BrdU antibody or no secondary antibody did not generate any fluorescence.
Alizarin red staining. ARS was used to assess mineral matrix deposition in osteogenically-supplemented cultures on day 14 of incubation. ARS showed a 24 and four-fold increase in mineral matrix production in β-Gly supplemented Saos-2 and hMSC cultures respectively (Fig. 7). Asc and Dex did not appear to affect mineral deposition in either cell line. However, there was an increased ARS concentration in β-Gly-negative hMSCs compared with in β-Gly-negative Saos-2.
Gene expression analysis. The expression of proliferation markers and osteoblast-associated genes was normalised to the expression of YWHAZ and determined every two days of incubation with the different combinations of osteogenic supplements (Figs. 8,9). Saos-2 cells demonstrated a gradual decrease in MKI67 (Fig. 8a) and PCNA (Fig. 8b)   www.nature.com/scientificreports/ combinations of Asc, Dex and β-Gly was not identified. hMSCs showed a similar trend for MKI67 expression (Fig. 9a) but a more marked decrease in the mRNA synthesis of PCNA (Fig. 9b) after 4 days of osteogenic supplementation than Saos-2 (Fig. 8b). RUNX2 expression was significantly upregulated by the presence of osteogenic  (Fig. 1), hMSCs did not show a significant change in numbers with the addition of β-Gly. Nor did Asc and/or Dex affect hMSCs numbers compared with the control. Blue and red bars represent cultures without and with β-Gly respectively. n = 3. www.nature.com/scientificreports/ supplements in Saos-2 cells and increased during the incubation period relative to negative controls (Fig. 8d). In contrast, the RUNX2 activity increased only at day 8 of the experiment and thereafter in osteogenically supplemented hMSCs (Fig. 9d). ALP expression was upregulated in the Dex, Asc, β-Gly supplemented hMSCs (Fig. 9c) and peaked on day 8 of incubation. Saos-2 showed an upregulation of OCN in the presence of β-Gly (Fig. 8e) at day 8 of the experiment. PHEX, an osteocyte-associated gene was significantly activated in all Saos-2 cultures containing osteogenic supplements at day 8 of incubation (Fig. 8f). No significant trend in PHEX expression was observed in hMSCs ( Fig. 9f) at any stage of the experiment.

Discussion
The diverse in vitro models used in osteoblast research create challenges with comparability of data between studies 14,15 . Osteosarcoma cell lines represent an affordable and straightforward option for researchers, however different cell origins can generate varying osteogenic gene expression profiles and phenotypes. The latter can produce variable data between studies and influence potential interpretation. The present study evaluated the response of an osteosarcoma cell line, Saos-2 to Asc, Dex and β-Gly used to induce osteogenic cell behaviour. An immortalised multipotent hMSC line was used for comparison of responses to osteogenic supplements by Saos-2 that may have been influenced by its tumour origin. It is crucial to emphasize that although the hMSCs did not originate from malignant tissue, they were immortalized through telomerase induction using transformation with SV40 Large T antigen 16 . This process resulted in a cellular phenotype that is more physiologically representative and "normal" when compared with Saos-2 cells. However, it is important to acknowledge that the cell line underwent certain modifications to achieve this state. Previously it has been reported that 6,7 Asc and/or Dex supplementation did not affect cell numbers in osteosarcoma cell lines including MG63 and Saos-3 suggesting either an anti-proliferative or toxic effect of β-Gly. However, in the present study Saos-2 cells demonstrated a decrease in cell number in the presence of β-Gly (Fig. 1) which was not observed in hMSCs (Fig. 2). β-Gly is used to promote mineralisation of the ECM by providing organic phosphate ions 5 . Inorganic phosphate concentrations of 5-7 mM in culture medium have been reported to induce apoptosis in osteoblasts 18 , possibly mediated by the mitochondrial damage via hyperpolarisation of the electrochemical gradient across the inner mitochondrial membrane and a consequent release of excess reactive oxygen species 19,20 However, it has also been indicated that organic phosphates, such as β-Gly were not harmful to cells at similar concentrations 21 . Furthermore, osteoblasts are suggested to be adapted to the elevated phosphate concentrations in comparison with other cell types arising from involvement in bone-remodelling 18 . In the present study, the LDH assay confirmed that β-Gly did not change cell viability and therefore was not cytotoxic to Saos-2 at 9.3 mM (Fig. 3). www.nature.com/scientificreports/ Interestingly, β-Gly appeared to decrease proliferation of Saos-2 cells but not hMSCs as shown by the BrdU staining (Figs. 4, 5 and 6). A decrease in cell proliferation rate and arrest in the G 0 phase are expected events at the later stages of the cell cycle 22 , which would explain the decrease in cell numbers in the later phases of incubation of Saos-2 cells. Mature cell phenotypes can enter a non-dividing, G 0 state via either terminal differentiation, senescence or quiescence 23 , expressing low to non-detectable levels of MKI67 and PCNA mRNA 24 MKI67 gene codes for Ki-67, a protein involved in the formation of the perichromosomal layer necessary for chromosome condensation in mitosis 25 . It is expressed in G 1 , S, G 2 and M phases of the cell cycle 26 . Whereas PCNA, typically transcribed in G 1 and S 27 , codes for an accessory protein for DNA polymerase alpha 26 . Hence, the increased activity of MKI67 and PCNA were used as indicators of cell proliferation in addition to BrdU staining. A significantly lower MKI67 expression level characteristic of G 0 was observed in the later stages of the experiment in all osteogenic conditions both in the Saos-2 cells and hMSCs (Figs. 8A and 9A). Furthermore, the decrease in PCNA expression after 2 and 4 days in all Saos-2 and hMSCs cultures respectively suggested the cell cycle arrest in G 0 or exit from G 1 and S. The reduction in transcription of both genes in all osteogenic conditions including those not containing β-Gly was possibly caused by quiescence typically arising from nutrient and/or oxygen deprivation or the build-up of toxic metabolites with increased cell numbers in longer cell cultures 28 However, a marked decrease of BrdU-stained nuclei observed only in β-Gly-exposed Saos-2 (Fig. 4C, F) implied a unique role for β-Gly in the cell cycle arrest of Saos-2. Decreased BrdU uptake showed that fewer Saos-2 cells were in S phase after 4 days of incubation that corresponded with a reduction in proliferation. A possible cause of the latter is the cell cycle arrest triggered by terminal differentiation of Saos-2.
Terminal differentiation leads to a permanent cell cycle exit in most cell types. Terminally differentiated osteoblasts become either lining cells, osteocytes or undergo apoptosis 29 . It is necessary to highlight the mature osteoblastic phenotype of Saos-2 relative to the earlier phase of osteogenic differentiation of hMSCs as the latter affects proliferation kinetics in cells. Coelho and Fernandes 11 reported decreased proliferation of human bone marrow cells in the presence of β-Gly after 42 days and linked it with progression of osteogenic differentiation. Interestingly in the present study, PHEX, an osteocyte associated marker was upregulated in osteogenically . Saos-2 exposed to β-Gly demonstrated a lower proportion of BrdU-stained nuclei than Saos-2 controls, 3.1% and 35.6% respectively. β-Gly-positive and β-Gly-negative hMSCs cultures did not show a significant difference in the proportion of BrdU labelled nuclei-23.7% and 25% respectively. n = 3.
Scientific Reports | (2023) 13:14472 | https://doi.org/10.1038/s41598-023-40835-w www.nature.com/scientificreports/ induced Saos-2 cultures, unlike hMSCs further confirming the phenotypic difference of the two in the presence of β-Gly. However, Saos-2 cells continued to express PCNA and MKI67 as well as synthesising mineralised matrix which are not typical of terminally differentiated cells. Although, it is worth noting that osteocytes are notoriously difficult to maintain in vitro 30 , so terminal differentiation should not be excluded as the cause of the anti-proliferative effect of β-Gly on Saos-2.
RUNX2 is an important regulator of osteogenic differentiation 31 and ALP has often been described as an early osteogenic marker 32 . Although, since ALP proteins are not unique to bone, it may be argued that the ALP expression level is inadequate for addressing osteoblast differentiation. Saos-2 cells showed no particularly defined patterns of RUNX2 or ALP expression depending on the osteogenic supplements provided and no significant difference was detected between treated and control cells. (Fig. 8C, D). This suggested that Asc, Dex and β-Gly did not affect early osteoblastic differentiation in Saos-2 cells as expected due to the relatively "mature" phenotype. This contrasted with the hMSCs where an upregulation of RUNX2 and ALP expression was observed in cells subjected to Asc and/or Dex and/or β-Gly promoting early osteoblastic differentiation (Fig. 9C, D). Expression of OCN was increased in Saos-2 in the presence of β-Gly (Fig. 8E). Increased OCN transcription indicated β-Gly interaction with the later stages of osteogenesis, whilst the activity of OCN in hMSCs did not appear to be affected exclusively by β-Gly and was upregulated in all supplemented cultures after day 10 (Fig. 9E). This implied a transition of hMSCs into a later stage of osteogenic differentiation during the course of the experiment similarly to healthy osteoblasts described in the literature 29 . These findings suggested a potential distinct role of β-Gly in later osteogenic differentiation of Saos-2 cells not detected in hMSCs. Based on this data, it was likely β-Gly affected Saos-2 proliferation via interacting with molecular pathways in the late osteogenesis.

Conclusion
β-Gly was shown to have a unique anti-proliferative effect on Saos-2 cells, which was not observed in a less differentiated hMSC line. Such a response may be mediated via interaction with late osteogenic differentiation of Saos-2. The derivation of Saos-2 from an osteosarcoma may have affected the response to osteogenic supplementation resulting in altered proliferation kinetics. It is important these findings are considered during selection of a model for osteogenesis in vitro.     Figure 9. Heatmaps for the gene expression levels by hMSCs incubated in osteogenic media over 14 days relative to the expression of YWHAZ. The data from each osteogenic condition were statistically analysed in comparison with control cultures at the corresponding time point. Where significant statistical differences existed, these are indicated as P values. Like Saos-2 cells, hMSCs showed downregulation in the expression of MKI67 (a) and PCNA (b) for all culture conditions. ALP (c) expression was upregulated in all osteogenic conditions, with the highest increase observed with Dex, Asc, and β-Gly. The upregulation peaked on day 8 of incubation. RUNX2 expression increased only in hMSCs after 8 days of incubation (d). A less marked increase in transcription of OCN and PHEX was detected in β-Gly supplemented hMSCs after 8 days (e, f) than in Saos-2. n = 3. The heatmaps were created using GraphPad Prism, Version 9.5.0 (https:// www. graph pad. com/ updat es/ prism-950-relea se-notes).

Data availability
The data created during this research are openly available from the Open Scientific Framework repository at https:// osf. io/ t6skh/.