25-Hydroxy- and 1α,25-Dihydroxycholecalciferol Have Greater Potencies than 25-Hydroxy- and 1α,25-Dihydroxyergocalciferol in Modulating Cultured Human and Mouse Osteoblast Activities

Despite differences in the phamacokinetics of 25-hydroxycholecalciferol (25(OH)D3) and 25-hydroxyergocalciferol (25(OH)D2) in man, the effects of these and their 1α-hydroxylated forms (1,25(OH)2D3 and 1,25(OH)2D2) on cellular activity of vitamin D-responsive cells have hardly been compared. We studied differences in the effects of these metabolites on cell number, gene transcription, protein expression and mineralisation of cultured human bone marrow-derived stromal cells (hBMSC) and rapidly mineralising mouse 2T3 osteoblasts. 50–1000 nM 25(OH) and 0.05–10 nM 1,25(OH)2 metabolites were used. At high concentrations, 25(OH)D2/D3 and 1,25(OH)2D2/D3 suppressed cell number in both human and mouse cells. The suppression was greater with cholecalciferol (D3) metabolites than with those of ergocalciferol (D2). In both cell types, 25(OH)D2 and 25(OH)D3 increased the expression of osteopontin, osteocalcin, collagen-1, receptor activator of nuclear factor kappa-B ligand, vitamin D receptor, CYP24A1 and CYP27B1 genes. Whereas there was little or no difference between the effects of 25(OH)D2 and 25(OH)D3 in hBMSCs, differences were observed in the magnitude of the effects of these metabolites on the expression of most studied genes in 2T3 cells. Alkaline phosphatase (ALP) activity was increased by 25(OH)D2/D3 and 1,25(OH)2D2/D3 in hBMSC and 2T3 cells, and the increase was greater with the D3 metabolites at high concentrations. In hBMSCs, mineralisation was also increased by 25(OH)D2/D3 and 1,25(OH)2D2/D3 at high concentrations, with D3 metabolites exerting a greater influence. In 2T3 cells, the effects of these compounds on mineralisation were stimulatory at low concentrations and inhibitory when high concentrations were used. The suppression at high concentrations was greater with the D3 metabolites. These findings suggest that there are differences in the effects of 25-hydroxy and 1α,25(OH)2 metabolites of D3 and D2 on human preosteoblasts and mouse osteoblasts, with the D3 metabolites being more potent in suppressing cell number, increasing ALP activity and influencing mineralisation.

Patients with vit D deficiency carry a risk of reduced bone mineral density and developing osteoporosis. With the recognition that large portions of populations living at high latitudes are deficient in vit D [16][17][18], a need to identify and treat the deficiency has increasing been realised. For the treatment, it has been generally assumed that the two prescribed forms of vit D, cholecalciferol (D 3 ) and ergocalciferol (D 2 ), have equal potencies. However, this has been questioned [19,20]. A few studies in humans have shown that lower circulating concentrations of 25(OH)D are achieved as a result of the administration of D 2 compared with D 3 [21,22]. Other studies have shown a lower binding to vit D-binding protein and a higher clearance rate for D 2 metabolites compared with those of D 3 [23,24], and differences in the effects of active D 2 and D 3 metabolites on plasma calcium and bone mineral content have been observed [25,26]. However, very few studies have compared potencies of 25(OH) and 1,25(OH) 2 metabolites of D 2 and D 3 at the cellular level. Although 1,25(OH) 2 D 2 and 1,25(OH) 2 D 3 were shown in separate studies to have equal affinities for VDR in chick intestine, rat intestine, porcine kidney, human breast cancer cells and HeLa cells [27][28][29], 1,25(OH) 2 D 3 was able to up-regulate rat intestinal VDR more than 1,25(OH) 2 D 2 did [30], and differences in the effects of 1α-cholecalciferol and 1α-ergocalciferol were shown in rat intestinal calcium absorption, and osteoclast numbers and activity [31,32]. As for osteoblasts, to the best of our knowledge, no study has compared their response to 25(OH) or 1,25(OH) 2 metabolites of D 2 and D 3 . Therefore, we have undertaken a comparison of the effects and potencies of 25(OH) and 1,25(OH) 2 metabolites of D 2 and D 3 on osteoblast cell number, cellular activity and mineralisation. For this, we have used two very distinct cell types: primary human bone marrowderived stromal cells (hBMSC) and mouse 2T3 osteoblastic cell line. Human BMSCs are at the initial stage of differentiation, and require osteogenic factors and time to develop into pre-and full osteoblasts. Based on our unpublished data, these cells require a period of 4 weeks to become responsive and to begin to mineralise, whereas murine 2T3 cells are primary murine osteoblasts that have been immortalised using SV40 large T antigen on the BMP2 promoter and that respond rapidly to matrix maturation and mineralisation signals.

Cell culture
Human BMSCs were derived from discarded bone material from patients undergoing total hip replacement. The use of bone material was approved by the University of Oxford Musculoskeletal Biobank Ethical Committee in compliance with Human Tissue Act ethical guidelines, and was after obtaining written, informed consent from patients [33]. Human BMSCs were cultured and studied at passage 1-2 in α-Modified Eagle's Minimum Essential Medium (αMEM, Lonza), supplemented with 10% heat-inactivated foetal bovine serum (HI-FBS; Lonza, UK), L-glutamine (2mM), ribonucleosides, penicillin and streptomycin (100U/ml).
Murine 2T3 pre-osteoblast cells were kind donations from Professor Stephen E. Harris (University of Texas, San Antonio, USA) and have been characterized previously [34]. These cells were isolated and cloned from transgenic mice, containing SV-40 large T antigen, driven by a BMP-2 promoter. When treated with osteogenic supplements, these cells mineralize and form bone matrix, expressing osteoblastic markers. The cells were plated at 1 x 10 6 per T-150 flasks in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% heat-inactivated foetal bovine serum (HI-FBS; Lonza, UK), penicillin (100 U/ml), streptomycin (100 U/ml) and L-glutamine (Invitrogen, 2 mM) in a standard humidified incubator at 37˚C and 5% CO 2 . 25(OH)D 2 , 1,25(OH) 2 D 2 , 25(OH)D 3 and 1,25(OH) 2 D 3 (Isosciences, CertiMass™) were dissolved in absolute ethanol at 10 −3 M concentration as a stock solution, and stored in light-protected glass vials at -80˚C. Working dilutions of 25(OH)D 2 and 25(OH)D 3 were evaluated by liquid chromatography-mass spectrometry. All sera used for tissue culture were routinely assessed for endogenous levels of 25(OH)D 2 and 25(OH)D 3 .
To induce hBMSCs differentiation, these cells were plated in αMEM plus 10% HI-FBS to reach confluency, then supplemented with osteogenic medium (OSM, 10% HI-FBS, 10 mM β-glycerol 2-phosphate disodium salt and 50 μg/ml L-ascorbic acid (Sigma Aldrich, UK) and the addition of 10 nM dexamethasone. To promote differentiation of mouse 2T3 cells at confluency, the medium was changed to OSM. Cells were cultured in replicates in OSM in the presence of absence of various concentrations of 25(OH)D 2 /D 3 (50-1000 nM) or 1,25(OH) 2 D 2 /D 3 (0.05-10 nM) in ethanol as vehicle (total added volume was <1% of the culture medium volume). OSM media containing the vit D metabolites were changed every other day prior to RNA and protein extraction or until mineralisation was assessed.

Cell number
Mouse 2T3 and hBMSCs were seeded in multiple 96-well plates at 5x10 3 cell density per well (DMEM-10% with HI-FBS, and αMEM plus 10% HI-FBS, respectively) in replicates in the presence or absence of different concentrations of 25(OH)D 2 /D 3 (50-1000 nM) or 1,25 (OH) 2 D 2 /D 3 (0.05-1 nM) for 24 h. After 21h, 10% MTS tetrazolium CellTiter 1 One Solution Reagent (Promega) was added to the wells and incubated for the last 3h under standard tissue culture conditions. Plates were removed and growth was assayed at 490 nm using a SPECTRAmax plate reader (Molecular DEVICES, USA).

ALP activity
Human BMSCs (1-5x10 4 ) and mouse 2T3 cells (1x10 4 ) were re-plated in replicates into 24/48 well plates in αMEM-10% HI-FBS and DMEM plus 10% HI-FBS culture media, respectively, to reach confluence, after which differentiation was induced by OSM in the presence or absence of vit D metabolites. For hBMSCs, the culture was for 7 days, during which the cells were treated every other day with either 100-1000 nM 25(OH)D 2 /D 3 or 0.1-10 nM 1,25(OH) 2 D 2 /D 3 in OSM. For 2T3 osteoblasts, the cells were cultured for 24 h, during which they were treated once with 100-500 nM 25(OH)D 2 /D 3 or 0.1-0.5 nM 1,25(OH) 2 D 2 /D 3 in OSM. Upon the completion of the treatment periods, cells were washed with Dulbeccos's phosphate-buffered saline (DPBS) and lysed in 100 μl radioimmunoprecipitation assay buffer containing protease inhibitor (RIPA-PI; Sigma). Cell lysates were sonicated for 15 sec and total protein content was measured by bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, UK). For each treated condition, ALP activities (pmol substrate/μg protein) of cell lysates were determined fluorometrically using fluorescent substrate 4-methylumbelliferyl phosphate (4-MUP; Sigma) as previously described [35]. Briefly, 10 μl of each cell lysate was incubated with 100 μl of freshly made 4MUP (0.16 mM in 50 mM Tris pH 6.0) substrate in white 96-well plates (NUNC) for 45 min in the dark at 37˚C, followed by the addition of 100 μl of 0.6M Na 2 CO 3 to stop the reaction. Fluorescence was measured at 360 nm excitation, 450 nm emission and 435 nm cut-off wavelength using a BMG Optima FluoSTAR plate reader (BMG LABTECH, Germany). ALP activities were calculated from standard curves of 0-20,000 pmol 4-methylumbelliferone (4MU), and values were normalised against the total protein content of each relevant well or treatment.
concentrations of 25(OH)D 2 /D 3 over 14 days. Mouse 2T3 cells were cultured in DMEM supplemented with 10% HI-FBS in 4-well plates at 1x10 5 cell density until full confluence, followed by the addition of OSM in the presence or absence of 200 nM 25(OH)D 2 /D 3 . Total RNA from each treatment was extracted by RNeasy Mini Kit (Qiagen, UK) according to the manufacturer's instructions. Nucleic acid concentrations were measured by NanoDrop ND-1000 spectrophotometer (Thermo Scientific, UK) at 260 nm. The absorbance ratios at 260/280 and 260/230 were used to detect any protein or organic carryover. Samples with both 260/280 and 260/230 ratios of !2 were used for further analysis. The integrity of total RNA purified with Qiagen RNeasy kit was also assessed using 1.5% agarose gel electrophoresis and ethidium bromide staining.
A total of 2 μg from each extract was treated with RNase-free DNase I (Thermo Scientific, UK) for 30 min at 37˚C. Removal of genomic DNA was terminated by further heat inactivation with 100mM EDTA at 65˚C for 10 mins. First strand cDNA synthesis of template RNA extracts was performed using a Veriti 96 well Thermal Cycler (Applied Biosystems, UK) using BIO-RAD iScript Reverse Transcription Supermix in a final reaction volume of 40 μl according to the manufacturer's instruction.
RT-PCR was performed using a ViiA7 system (Applied Biosytems, UK) with commercially available lyophilized Quanititect Qiagen primers; col1, opn, ocn, opg, rankl, vdr, cyp27b1, cyp24a1. One μl of one-twentieth dilutions of templates were used in a total volume of 10 μl reaction in 384 well plates (Applied Biosystems, UK) by 2-step cycling (polymerase activation at 95˚C for 2 min, 40 cycles of template denaturation at 95˚C for 5 sec, primer annealing and extension at 65˚C for 30 sec) using SYBR green Master Mix (BIOLINE; SensiFAST SYBR Lo-Rox Kit). The C t values for treated samples were normalised to housekeeping genes gapdh and 18S and the relative expressions were calculated using ΔΔ C t with amplification and accuracy of 98-100%.

Mineralisation
Human BMSCs were cultured at 1x10 4 in 96-well plates in αMEM-10% HI-FBS, and mouse 2T3 cells at 0.1-1x10 4 in 96-, 48-or 24-well plates in replicates in DMEM-10%HI-FBS to reach confluency. Media at this time point (day zero) was replaced with OSM with or without vit D metabolites. For hBMSCs, OSM containing 500-1000 nM 25(OH)D 2 /D 3 or 1-10 nM 1,25 (OH) 2 D 2 /D 3 was used, and for 2T3 cells, OSM contained 100-500 nM 25(OH)D 2 /D 3 or 01-0.5 nM 1,25(OH) 2 D 2 /D 3 . The media were changed every other day up to 21 days for hBMSCs and up to 8 days for 2T3 cells prior to mineralisation assays. Upon completion of treatments, cell cultures were stopped, media removed, cells were washed with DPBS, fixed with 70% ethanol, air-dried and stained with 1.5% alizarin red dye solution (Sigma-Aldrich, UK; pH 4.1). Plates were washed with 70% ethanol and deionised water, air dried and staining were extracted by 10% acetic acid over orbit shaker, neutralised with 10% ammonium hydroxide, transferred into 96-well plates and measured colourimetrically at 550 nm.

Statistical analysis
All experiments were carried out at least in triplicates and the mean ± SEM was calculated. Statistical analyses were carried out using SPSS version 11.0 for windows (SPSS Inc., Chicago, IL, USA). Effects of treatments were compared by Kruskal-Wallis one-way analysis of variance (ANOVA), with differences between treatments assessed using Bonferroni error protection for multiple comparisons. A P value of <0.05 was regarded to indicate a significant difference.

Human BMS and mouse cell numbers
Human BMSC numbers decreased in response to 25(OH)D 2 at doses of !500 nM (and in response to 25(OH)D 3 at doses of !200 nM (P<0.01 and P<0.001, respectively; Fig 1A). 25 (OH)D 3 was more potent than D 2 at concentrations of !200 nM in human primary cells (P<0.01; Fig 1A). Only a small decrease in hBMSC numbers as a result of treatment with the highest dose of 1,25(OH) 2 D 3 (1 nM) was observed (P<0.05; Fig 1B).
When mouse 2T3 osteoblasts were exposed to 25(OH)D 2 , cell numbers were reduced at doses of 500 nM and 1000 nM (P<0.05 and P<0.01, respectively; Fig 1C). A decrease in 2T3 cell numbers was also observed with 25(OH)D 3 at doses of 500 nM and 1000 nM (P<0.001 for both concentrations; Fig 1C). The magnitude of the decrease was greater with 25(OH)D 3  Fig 1C). In 2T3 cells, whereas 1,25(OH) 2 D 3 at !0.5 nM concentrations reduced cell numbers significantly (P<0.001; Fig 1D), 1,25(OH) 2 D 2 had no such effects, and, if anything, showed a tendency to increase cell numbers only at 0.2 nM (P<0.01; Fig 1D).
In 2T3 cells treated with 25(OH)D 2 or 25(OH)D 3 , there were moderate increases in VDR, CYP27B1, OPN and OCN at 200 nM ( Fig 4I). However, a reduction of these proteins was observed with 500 nM of both 25(OH)D 2 and 25(OH)D 3 , with the latter being more potent in this reduction (Fig 4I). There were dose-dependent increases in VDR, OPN and OCN protein expression in response to 1,25(OH) 2 D 2 and 1,25(OH) 2 D 3 , where 0.2-0.5 nM 1,25(OH) 2 D 2 elicited the greatest response (Fig 4J).

Fig 2. Comparative effects of 25(OH)D 2 , 25(OH)D 3 , 1,25(OH) 2 D 2 and 1,25(OH) 2 D 3 on ALP enzyme activity.
Human BMSCs were grown in αMEM-10% FBS for 4 days until confluency (day 0) and treated with the D 2 or D 3 metabolites every other day (days 0, 2, 4 and 6). Human BMSC cultures were stopped at day 7, and ALP activity was determined (panels A&B). 2T3 cells were grown in 10% DMEM-FBS for 24 h until full confluency (day 0), and were treated once with either 100-500 nM 25(OH)D 2 /D 3 (panels C-E) or 0.1-0.5 nM 1,25(OH) 2 D 2 /D 3 (panels F&H) in differentiation media. Cultures were stopped at 12 or 24 h, and ALP activity was measured. ALP activities were normalised to total protein content. Mean ± SEM enzymatic activity values from triplicate experiments have been presented as the amount of 4-methylumbelliferone (4MU) generated after 45 min at 37˚C. # P<0.001 for comparisons between the treatments and the control (vehicle); * P<0.05 and ** P<0.01 for comparisons between the D 2 and D 3 metabolites. When hBMSCs were cultured in osteogenic media for 21 days and treated with the vit D metabolites every other day, no mineralisation was observed at concentrations of 500 nM of 25 (OH)D 2 or 25(OH)D 3 (Fig 5A). Only at a concentration of 1000 nM 25(OH)D 2 and D 3 was mineralisation significantly increased in these cells (P<0.05 and P<0.001, respectively), and the effects were more pronounced with the 25(OH)D 3 (P<0.001; Fig 5A and 5C). Quantified alizarin red staining revealed that 1-10 nM 1,25(OH) 2 D 3 was more potent in mineralizing hBMSCs than 1,25(OH) 2 D 2 (P<0.05; Fig 5B and 5D).

Discussion
To the best of our knowledge, this is the first study that has compared the effects of 25(OH) or 1,25(OH) 2 metabolites of D 2 and D 3 on osteoblasts. Unlike previous studies that have employed pharmacologically high concentrations of 25(OH) and 1,25(OH) 2 metabolites, we used concentrations that were generally much lower and more physiological [4,7,9,15]. Differences in the effects of these metabolites on osteoblastic cells from human and mouse that were also very distinct in their stage of differentiation were studied. Most previous studies have shown the effects of vit D metabolites on osteoblast proliferation to be inhibitory [14,36,37].
In hBMSCs, 25(OH)D 3 had a greater inhibitory effect on cell number than 25(OH)D 2 , and with 1,25(OH) 2 metabolites, an inhibition was seen only with 1,25(OH) 2 D 3 at a concentration of 1000 pmol/L. We also found both 25(OH)D and 1,25(OH) 2 D metabolites to inhibit increasing 2T3 cell number, and the suppressive effects to be greater with the D 3 metabolites. A recent study on human osteoblasts have shown 1,25(OH) 2 D 3 metabolite to decrease cell proliferation, but only at a pharmacologically high concentration of 100 nmol/L administered over 3 days [15]. Our data show that over a period of 24 h, some stasis in cell number, albeit small, appear to be present at lower vit D metabolite concentrations, particularly with those of D 3 . Such antiproliferative effects are achieved by cell cycle arrest, with the effects of 25(OH)D presumably mediated through the activity of CYP27B1 and the generation of 1,25(OH) 2 [38].
During the differentiation of hBMSCs into osteoblasts, their VDR expression increases [39]. Our observations on VDR gene and protein expression lend support to this, particularly in relation to stimulation by 25(OH)D metabolites, with 25(OH)D 3 having a greater effect on the VDR protein. Activated VDR acts as a transcription factor for osteoblast signature genes, Comparison of the Effects of Vitamin D2 and D3 Metabolites on Human and Mouse Osteoblasts and in hBMSCs, active vit D metabolites stimulate the manufacture of bone matrix proteins that include Col1, OCN and OPN, and increase ALP production that promotes mineralisation [7,13,40,41]. Our data in hBMSCs are consistent with such previous findings, with both 25 (OH)D metabolites increasing the transcription of col1, ocn and opn matrix protein genes, as well as that of opn, rankl, and cyp24a1 and cyp27b1. Our data showed a greater effect of 25 (OH)D 3 than 25(OH)D 2 in increasing the manufacture of ALP, OCN, CYP24A1 and CY27B1 proteins.
It is unclear why 1,25(OH) 2 metabolites did not affect hBMSC cell number or stimulate VDR, CYP24A1 and CYP27B1 as much as 25(OH) metabolites did. Whether this may be related to higher 1,25(OH) 2 D concentrations produced at cellular level from nM range of 25 (OH)D used is unknown, although some previous findings seem to suggest this [38]. An increase in CYP27B1 increases 1,25(OH) 2 D, and CYP24A1 converts 25(OH)D and, in particular, 1,25(OH) 2 D metabolites into their 24-hydroxylated forms [42]. It has been shown that both 24,25(OH) 2 D and particularly 1,24,25(OH) 3 D are biologically active and exert stimulatory effects on vit D-responsive cells [43,44], but the exact influence of these metabolites on osteoblasts remains unknown.
Unlike the scenario in hBMSCs, in 2T3 cells 25(OH)D 3 was a more potent stimulus to vdr and ocn gene expression than 25(OH)D 2 , but both VDR and OCN proteins were higher with 1,25(OH) 2 D 2 compared with 1,25(OH)D 3 . As with previous reports [38], we also found 25 (OH) metabolites to increase cyp27B1 gene expression in 2T3 cells, but in contrast to human cells, this was greater with 25(OH)D 2 . In contrast to hBMSCs, 2T3 cells produced large amounts of CYP27B1 enzyme in response to 1,25(OH) 2 D 2 . Furthermore, unlike hBMSCs, mouse cells expressed cyp24A1 gene only in response to 25(OH)D 3 and not 25(OH)D 2 . Whereas the influence of CYP27B1 expression is consistent with the cell number and mineralisation effects of vit D as described above, the physiological effects of stimulation of CYP24A1 is uncertain.
Both 25(OH)D 2 and 25(OH)D 3 greatly increased rankl gene transcription in human and mouse cells. The two metabolites modestly suppressed opg gene transcription in hBMSCs, but had inconsistent effects in 2T3 cells. RANKL and OPG exert opposing effects on osteoclastogenesis and osteoclast activation, and a pattern of protein expression similar to that of gene expression observed in human cells here would be a stimulus to osteoblast activation [45][46][47]. With regards mineralisation as a crucial clinical endpoint of vit D effect in relation to bone mineral density, hBMSC's mineralisation increased with vit D metabolites. This was in accord with our ALP data, and these effects tended to be greater with 25(OH)D 3 and 1,25(OH)D 3 at high concentrations. These add further support to a prevailing general notion that D 2 treatment has less effect than D 3 administration in man.
However, previous data in some murine osteoblasts have shown that, unlike its effects on human cells, vit D metabolites appear to inhibit differentiation and mineralisation, although studies on murine ALP and Col1 show some conflicting data [9,[48][49][50][51]. Our observations on 2T3 mineralisation indicate a bimodal concentration effect of vit D metabolites in these cells. Thus, there was an increase in mineralisation at low vit D metabolites concentrations, but the Comparison of the Effects of Vitamin D2 and D3 Metabolites on Human and Mouse Osteoblasts effects were reversed and became suppressive as concentrations increased. These data may suggest that some of the inhibitory effects on mineralisation observed in some previous studies may have been due to the use of very high, supra-physiological doses of these metabolites in those studies. At high concentrations, the suppressive effects of vit D metabolites on 2T3 cell mineralisation was greater with 25(OH)D 3 and 1,25(OH) 2 D 3 than the corresponding D 2 compounds. Two further murine cell lines, MBA 15.4 and Mc3T3 E1, displayed similar dose responses (data not shown), with enhanced differentiation over doses 100-200 nM and repression at !500 nM.
These findings indicate that cell activity can change significantly over time under the influence of active vit D metabolites. In that respect, two different concentrations of a vit D metabolite may be able to bring about two patterns of change in cell activity that are out of synch with one another during their time course. For this reason, it is possible that some of the conflicting differences in the expression of a particular protein seen with D 3 and D 2 metabolites may be secondary to a greater potency of one metabolite that imposes a shift in the time course of change in that protein. The data on such a protein can then become conflicting when there is both an increase and a decrease in protein expression over a time course. Such apparent conflicts may be possible in our study as we assessed gene expression over long periods of 1 and 14 days of treatment of 2T3 and hBMS cells, respectively. Furthermore, the effects of vit D metabolites on osteoblasts are known to be influenced by other factors such as calcium concentrations [52]. How these other factors change over time is unknown, but they can potentially add further complexity to the dynamics of the effects of vit D metabolites that may possess different potencies.
Overall, the data from this study suggest that in general vit D 3 metabolites appear to elicit a greater influence on both hBMSC and mouse 2T3 cells. These greater effects appear not to be universal, but seem to be more pronounced in some major aspects of osteoblast function, particularly in changing cell number and in increasing ALP and mineralisation. The only greater influences of 25(OH)D 2 that we found were on OCN expression in hBMSCs and on CYP27B1 expression in 2T3 calls. Assuming that the effects of 25(OH)D are mostly mediated through the generation of its α-hydroxylated compound by CYP27B1, as well as the metabolism of 25 (OH)D and 1,25(OH) 2 D by CYP24A1, consideration must be given to the possibility that the velocity and the Km of these enzymes may differ significantly for 25(OH)D 2 and 25(OH)D 3 .
Furthermore, some of the differences in the effects of vit D 3 metabolites on osteoblast proliferation and function appear to be also time-and cell-type-dependent [10]. The same may be true for the effects of D 2 metabolites. As with others, we have observed that the direction and the effects of vit D metabolites can differ sufficiently in the two osteoblastic cell types studied so as to preclude the use of mouse model to derive knowledge about human cells. In some respects, the data on mouse osteoblasts may in fact be even misleading in predicting the behaviour of human cells. However, human osteosarcoma cell lines display similarly diverse responses, with neither MG63 nor SaOS2 cell lines responding to the same dose and range of vitamin D as primary hBMSCs with either ALP activity or mineralisation (S1 Fig in supporting  information). In conclusion, our data on osteoblasts add significant weight to some previous findings that D 2 may not have the same potency and may not elicit the same physiological effect as D 3 . If corroborated by further future studies, these findings would underline a need to assess differences in long-term physiological effects of the two vitamins in human subjects. Evidence gathered from such studies may prove to have implications for prescribing D 2 preparations in clinical practice.