Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Primary Human Osteoblasts in Response to 25-Hydroxyvitamin D3, 1,25-Dihydroxyvitamin D3 and 24R,25-Dihydroxyvitamin D3

  • Karen van der Meijden,

    Affiliation Department of Internal Medicine/Endocrinology, VU University Medical Center, Research Institute MOVE, Amsterdam, The Netherlands

  • Paul Lips,

    Affiliation Department of Internal Medicine/Endocrinology, VU University Medical Center, Research Institute MOVE, Amsterdam, The Netherlands

  • Marjolein van Driel,

    Affiliation Department of Internal Medicine/Endocrinology, Erasmus Medical Center, Rotterdam, The Netherlands

  • Annemieke C. Heijboer,

    Affiliation Department of Clinical Chemistry, VU University Medical Center, Research Institute MOVE, Amsterdam, The Netherlands

  • Engelbert A. J. M. Schulten,

    Affiliation Department of Oral and Maxillofacial Surgery/Oral Pathology, VU University Medical Center, Academic Centre for Dentistry Amsterdam, Amsterdam, The Netherlands

  • Martin den Heijer,

    Affiliation Department of Internal Medicine/Endocrinology, VU University Medical Center, Research Institute MOVE, Amsterdam, The Netherlands

  • Nathalie Bravenboer

    n.bravenboer@vumc.nl

    Affiliation Department of Clinical Chemistry, VU University Medical Center, Research Institute MOVE, Amsterdam, The Netherlands

Primary Human Osteoblasts in Response to 25-Hydroxyvitamin D3, 1,25-Dihydroxyvitamin D3 and 24R,25-Dihydroxyvitamin D3

  • Karen van der Meijden, 
  • Paul Lips, 
  • Marjolein van Driel, 
  • Annemieke C. Heijboer, 
  • Engelbert A. J. M. Schulten, 
  • Martin den Heijer, 
  • Nathalie Bravenboer
PLOS
x

Abstract

The most biologically active metabolite 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) has well known direct effects on osteoblast growth and differentiation in vitro. The precursor 25-hydroxyvitamin D3 (25(OH)D3) can affect osteoblast function via conversion to 1,25(OH)2D3, however, it is largely unknown whether 25(OH)D3 can affect primary osteoblast function on its own. Furthermore, 25(OH)D3 is not only converted to 1,25(OH)2D3, but also to 24R,25-dihydroxyvitamin D3 (24R,25(OH)2D3) which may have bioactivity as well. Therefore we used a primary human osteoblast model to examine whether 25(OH)D3 itself can affect osteoblast function using CYP27B1 silencing and to investigate whether 24R,25(OH)2D3 can affect osteoblast function. We showed that primary human osteoblasts responded to both 25(OH)D3 and 1,25(OH)2D3 by reducing their proliferation and enhancing their differentiation by the increase of alkaline phosphatase, osteocalcin and osteopontin expression. Osteoblasts expressed CYP27B1 and CYP24 and synthesized 1,25(OH)2D3 and 24R,25(OH)2D3 dose-dependently. Silencing of CYP27B1 resulted in a decline of 1,25(OH)2D3 synthesis, but we observed no significant differences in mRNA levels of differentiation markers in CYP27B1-silenced cells compared to control cells after treatment with 25(OH)D3. We demonstrated that 24R,25(OH)2D3 increased mRNA levels of alkaline phosphatase, osteocalcin and osteopontin. In addition, 24R,25(OH)2D3 strongly increased CYP24 mRNA. In conclusion, the vitamin D metabolites 25(OH)D3, 1,25(OH)2D3 and 24R,25(OH)2D3 can affect osteoblast differentiation directly or indirectly. We showed that primary human osteoblasts not only respond to 1,25(OH)2D3, but also to 24R,25(OH)2D3 by enhancing osteoblast differentiation. This suggests that 25(OH)D3 can affect osteoblast differentiation via conversion to the active metabolite 1,25(OH)2D3, but also via conversion to 24R,25(OH)2D3. Whether 25(OH)D3 has direct actions on osteoblast function needs further investigation.

Introduction

Vitamin D deficiency, a common condition in the elderly population, has been associated to numerous skeletal health problems. Vitamin D deficiency causes a decrease of calcium absorption from the intestines and secondary hyperparathyroidism which leads to bone loss, osteoporosis and mineralization defects in the long term [1]. Vitamin D status is determined by the measurement of the metabolite 25-hydroxyvitamin D3 (25(OH)D3) [2], which is the major circulating form of vitamin D. The metabolite 25(OH)D3 is metabolized in the kidney by the enzyme 1α-hydroxylase (CYP27B1) into the biologically most active metabolite 1,25-dihydroxyvitamin D3 (1,25(OH)2D3) [3], which is the classical pathway for vitamin D activation. Both 25(OH)D3 and 1,25(OH)2D3 are metabolized by the enzyme 24-hydroxylase (CYP24), responsible for the first step in the inactivation process, to respectively 24R,25-dihydroxyvitamin D3 (24R,25(OH)2D3) and 1,24R,25-trihydroxyvitamin D3 (1,24R,25(OH)3D3) [4]. In addition, alternative pathways for vitamin D activation have been described, and one of such is the CYP11A1-mediated pathway [5]. This pathway for activation of vitamin D has been demonstrated in placentas ex utero, adrenal glands ex vivo and in cultured epidermal keratinocytes and colonic Caco-2 cells [6], [7]. Hydroxyvitamin D derivatives synthesized by the action of CYP11A1 not only act on the vitamin D receptor (VDR), but also on the retinoic acid related receptors α and γ (RORα and RORγ) [8].

The metabolite 1,25(OH)2D3 exerts its function by binding to the VDR which is present in numerous tissues, including bone tissue [3]. Bone formation is affected by 1,25(OH)2D3 both in an indirect and direct manner. Indirect effects of 1,25(OH)2D3 occur through stimulation of intestinal calcium absorption required for the maintenance of normal serum calcium levels and bone mineralization [3]. Direct effects of 1,25(OH)2D3 on osteoblasts have been demonstrated in vitro. These in vitro studies show that 1,25(OH)2D3 decreases osteoblast proliferation and stimulates osteoblast differentiation by increasing collagen type I synthesis and by secreting several non-collagenous proteins, for example osteocalcin and osteopontin [9]. The metabolite 1,25(OH)2D3 also increases the alkaline phosphatase (ALP) activity and the mineralization of bone matrix synthesized by human osteoblasts [10][12].

While the effects of 1,25(OH)2D3 on human osteoblasts are well-known, fewer studies have focused on the response of human osteoblasts to the precursor 25(OH)D3. Van Driel et al [12] have shown that 25(OH)D3 increases the ALP activity, the osteocalcin expression, and the early phase of mineralization in the human SV-HFO cell line. In primary osteoblasts, 25(OH)D3 inhibits the proliferation, stimulates the expression of osteocalcin and osteopontin, and increases the mineralization [13]. The actions of 25(OH)D3 on human osteoblasts are thought to take place after its conversion to 1,25(OH)2D3, since osteoblasts express 1α-hydroxylase and are capable of synthesizing 1,25(OH)2D3 from 25(OH)D3 [12][14]. Locally synthesized 1,25(OH)2D3 is thought to act in an autocrine or paracrine manner to regulate osteoblast proliferation and differentiation [12], [13]. However, it is largely unknown whether, in addition to the effects of 25(OH)D3 that occur via hydroxylation to 1,25(OH)2D3, 25(OH)D3 can affect primary osteoblast function on its own.

In addition to 1α-hydroxylase, osteoblasts express 24-hydroxylase [12], [13] and have the capability to synthesize 24R,25(OH)2D3 from 25(OH)D3 [14]. The metabolite 24R,25(OH)2D3 was originally thought to be inactive, however, several in vivo and in vitro studies support 24R,25(OH)2D3 bioactivity in bone tissue. In chickens, 24R,25(OH)2D3 in combination with 1,25(OH)2D3 treatment promotes fracture healing [15]. In addition, CYP24 knockout mice demonstrate a delayed fracture healing [16]. In vitro, 24R,25(OH)2D3 has positive actions on SV-HFO osteoblast differentiation by increasing ALP activity, osteocalcin secretion and matrix mineralization [17]. These findings suggest that primary human osteoblasts not only respond to the active metabolite 1,25(OH)2D3 but also to 24R,25(OH)2D3.

The aim of this research was to determine the effects of 25(OH)D3 on primary human osteoblast proliferation and differentiation, compared to 1,25(OH)2D3. To examine whether these effects of 25(OH)D3 occur through hydroxylation to 1,25(OH)2D3 we silenced CYP27B1 expression. However, osteoblasts synthesize not only 1,25(OH)2D3 from 25(OH)D3, but also 24R,25(OH)2D3 from 25(OH)D3. Therefore we hypothesized that the effects of 25(OH)D3 not only occurred through conversion to 1,25(OH)2D3 but also to 24R,25(OH)2D3.

Materials and Methods

Primary human osteoblast culture

Primary human osteoblasts were isolated from redundant trabecular bone fragments obtained from healthy donors undergoing pre-implant bony reconstruction of the mandible or maxilla with autologous bone from the anterior iliac crest. The donor group consisted of 11 males and 12 females with a mean age of 49.3±18.6 years. The protocol was approved by the Medical Ethical Review Board of the VU University Medical Center, Amsterdam, the Netherlands, and all donors gave their written informed consent.

A modification of the methods of Beresford and Marie [18], [19] was used. Shortly, the trabecular bone fragments were minced into small pieces and washed extensively with phosphate buffered saline (PBS). The bone pieces were treated with 2 mg/ml collagenase type II (300 U/mg; Worthington Biochemical Corporation, Lakewood, NJ, USA) for two hours in a shaking waterbath at 37°C. The pieces were placed in culture flasks with Dulbecco’s Modified Eagle Medium: Nutrient Mixture F-12 (DMEM/F12; GIBCO, Life technologies) supplemented with 10% Fetal Clone I (HyClone, Thermo Fisher Scientific), 100 U/ml penicillin and 100 µg/ml streptomycin (GIBCO, Life technologies), 1.25 µg/ml fungizone (GIBCO, Life technologies) and incubated at 37°C and 5% CO2. Medium was changed twice a week until cells reached confluence.

Primary human osteoblast treatments

The vitamin D metabolites 25(OH)D3, 1,25(OH)2D3 and 24R,25(OH)2D3 were obtained from Sigma-Aldrich. Primary human osteoblasts were treated with or without different vitamin D concentrations as indicated in the figure legends.

To enable differentiation, primary human osteoblasts were cultured in osteogenic medium. Osteogenic medium consisted of complete medium with 10 mmol/L β-glycerophosphate (Sigma-Aldrich), 10 nmol/L dexamethasone (Sigma-Aldrich) and 50 µg/ml ascorbic acid (Sigma-Aldrich).

All experiments were performed in complete medium with 5% Fetal Clone I unless otherwise stated and all conditions, including treated and control groups, contained 0.1% ethanol.

Proliferation

Primary human osteoblasts of the first passage were plated out in a 96 wells plate at a density of 4.000 cells/well. After 24 hours cells were exposed to medium with 25(OH)D3 (0, 100, 200 or 400 nmol/L) or 1,25(OH)2D3 (0, 1, 10 or 100 nmol/L). Medium was replaced every 3 days by complete medium with or without 25(OH)D3 or 1,25(OH)2D3. The proliferation of primary human osteoblasts was measured at day 3 and 6 using the XTT Cell Proliferation Kit (Roche Diagnostics) according to the manufacturer’s protocol. Briefly, cells were incubated with the XTT solution at 37°C, whereby the viable cells formed an orange formazan dye by cleaving the yellow tetrazolium salt XTT. After 2 hours the orange formazan solution was quantified by a photospectrometer (Berthold Technologies) at 450 nm.

Differentiation

Primary human osteoblasts of the first or second passage were seeded into a 12 wells plate at a cell density of 40.000 cells/well. Cells were allowed to attach to the well for 24 hours before medium was changed to osteogenic medium with 25(OH)D3 (0 or 400 nmol/L) or 1,25(OH)2D3 (0 or 100 nmol/L). Medium was replaced every 3 or 4 days by complete medium with or without 25(OH)D3 or 1,25(OH)2D3. Culture medium was collected at day 3, 7, 10 and 14 of the differentiation culture and cell lysates were prepared for the measurement of osteoblast markers.

Procollagen type I aminoterminal propeptide (P1NP) was measured in culture medium using the UniQ PINP radioimmunoassay (Orion Diagnostica). The interassay variation was <8% over the whole concentration range.

ALP activity was measured in cell lysate that was made by scraping the cells in PBS-0.1% triton [12], and by sonificating of the lysate two times for 30 seconds at 50 Hz. ALP activity was measured by the ALP IFCC liquid assay (Roche Diagnostics), performed on a Modular analyzer (Roche Diagnostics). ALP activity was adjusted for total protein, measured by the BCA protein assay (Thermo Fisher Scientific) according to the manufacturer’s protocol.

Osteocalcin was measured in culture medium using an enzyme immunoassay (Biosource). Interassay variation was 15% at a level of 0.5 nmol/L, 8% at a level of 2 nmol/L and 9% at a level of 8 nmol/L.

RNA isolation and real time RT-PCR

For RNA experiments primary human osteblasts of the first or second passage were seeded into a 12 wells plate at a cell density of 40.000 cells/well. Medium was changed after 24 hours and primary human osteoblasts were treated with different vitamin D metabolites as indicated in the figure legends. Total RNA isolation of primary osteoblasts was performed using the RNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol. For removing residual DNA amounts an additional on-column DNA treatment was accomplished during the RNA isolation procedure. Total RNA concentration was measured with the Nanodrop spectrophotometer (Nanodrop Technologies).

RNA was reverse transcribed from 100 ng total RNA in a 20 µl reaction mixture containing 5 mmol/L MgCl2 (Eurogentec), 1X RT buffer (Promega), 1 mmol/L dATP, 1 mmol/L dCTP, 1 mmol/L dGTP, 1 mmol/L dTTP (Roche Diagnostics), 1 mmol/L Betaïne, 10 ng/µl random primer, 0.4 U/µl RNAsin (Promega) and 5 U/µl M-MLV RT-enzym (Promega). The PCR reaction of total 25 µl contained 3 µl cDNA, 300 nmol/L reverse and forward primer (table 1) and SYBR Green Supermix (Bio-Rad). The PCR was performed on an iCycler iQ Real-Time PCR Detection System (Bio-Rad): 3 minutes at 95°C, 40 cycli consisting of 15 seconds at 95°C and 1 minute at 60°C. The relative gene expression was calculated by the 2−ΔCt method and TATA binding protein (TBP) was used as housekeeping gene.

siRNA transfection

Silencing RNA was carried out to suppress CYP27B1 mRNA. Knockdown was performed using CYP27B1 SMART pool and the negative control ON-TARGET plus SMART pool (Thermo Fisher Scientific). Primary human osteoblasts of the first passage were electroporated with the Microporator Pipetype Electroporation System (Digital Bio, Hopkinton, MA, USA) using 1 pulse of 1200 V for 40 ms. After electroporation, 100.000 cells were seeded in a 24 wells plate in DMEM/F12 with 10% fetal clone I. Two days after electroporation of the cells, total RNA was isolated to determine CYP27B1 knockdown. Four days after the electroporation treatment, cells were incubated in complete medium with 25(OH)D3 (0 or 400 nmol/L) for 3 days. Complete medium was collected and stored at −20°C until 1,25(OH)2D3, 25(OH)D3 and 24R,25(OH)2D3 measurements. Cells were lysed and stored at −80°C until total RNA isolation.

1,25(OH)2D3, 25(OH)D3 and 24R,25(OH)2D3 measurements

Primary human osteoblasts were seeded into a 6 wells plate with a cell density of 500.000 cells/well. High bone cell density was used to raise the 1,25(OH)2D3 concentrations above the detection levels. After 24 hours, cells were incubated in medium consisting of DMEM/F12, 0.2% BSA, 100 U/ml penicillin, 100 µg/ml streptomycin, 1.25 µg/ml fungizone and 25(OH)D3 (0, 100, 200, 400 or 1.000 nmol/L). Medium was collected after 24 hours exposure of osteoblasts to 25(OH)D3.

The metabolite 1,25(OH)2D3 was measured in non-conditioned and conditioned medium using a radioimmunoassay (IDS). Cross reactivity with 25(OH)D3 and 24R,25(OH)2D3 was 0.1% and <0.01% respectively. Intra-assay variation was 8% at a level of 25 pmol/L and 9% at a level of 70 pmol/L, and interassay variation was 11% at a concentration of 25 and 70 pmol/L.

The metabolites 25(OH)D3 and 24R,25(OH)2D3 were analyzed in non-conditioned and conditioned medium using a liquid chromatography-tandem mass spectrometry (LC-MS/MS) method. Briefly, samples were incubated with deuterated internal vitamin D standards (d6–25(OH)D3 and d6-24R,25(OH)2D3) and protein-precipitated using acetonitrile. Supernatant was, after PTAD derivatization, purified using a Symbiosis online solid phase extraction (SPE) system (Spark Holland, Emmen, the Netherlands), followed by detection with a Quattro Premier XE tandem mass spectrometer (Waters Corp., Milford, MA). Intra-assay variation of 25(OH)D3 was 9.6%, 6.0% and 8.5% at a level of 58, 191 and 516 nmol/L, respectively. Intra-assay variation of 24R,25(OH)2D3 was 5.4% and 9.1% at a level of 46 and 150 nmol/L, respectively.

Statistical analyses

Data were presented as mean ± SEM. Differences between 2 groups were assessed using Wilcoxon signed rank test. Differences between 3 or more groups were assessed using Friedman test followed by Dunn’s post hoc test. A p-value<0.05 was considered to be significant (*p<0.05, **p<0.01, ***p<0.001).

Results

Effects of 1,25(OH)2D3 and 25(OH)D3 on osteoblast proliferation

Primary human osteoblasts were cultured in the presence of 1,25(OH)2D3 or 25(OH)D3 for 6 days to compare the effects of these metabolites on the proliferation. Both 1,25(OH)2D3 (Fig. 1A) and 25(OH)D3 (Fig. 1B) significantly decreased the proliferation after 3 and 6 days of treatment. The reduction of the proportion of viable cells was found to be 28% (p<0.01) and 47% (p<0.01) in the presence of 100 nmol/L 1,25(OH)2D3 compared to control cultures at day 3 and 6 respectively. The metabolite 25(OH)D3 decreased the proliferation of primary human osteoblasts after 3 and 6 days of treatment at a concentration of 400 nmol/L. The reduction of the proportion of viable cells was 12% (p<0.01) and 28% (p<0.05) at day 3 and 6 respectively compared to control cultures.

thumbnail
Figure 1. Effects of 1,25(OH)2D3 and 25(OH)D3 on primary human osteoblast proliferation.

Osteoblasts were cultured in the presence of 0, 1, 10 or 100 nM 1,25(OH)2D3 (A) and 0, 100, 200 or 400 nM 25(OH)D3 (B) and the proliferation was quantified at day 3 and 6. Results (mean ± SEM) are expressed as treatment versus control ratios (time-point 0 was set at 1.0) using cells from 4 (A) or 7 (B) different donors. Results were analysed using Friedman test followed by Dunn’s post hoc test for each timepoint (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g001

Effects of 1,25(OH)2D3 and 25(OH)D3 on osteoblast differentiation

Primary human osteoblasts were cultured in osteogenic medium containing 1,25(OH)2D3 or 25(OH)D3 for 14 days to compare the effects of these metabolites on the differentiation. Both 1,25(OH)2D3 (Fig. 2A) and 25(OH)D3 (Fig. 2B) stimulated the ALP activity during differentiation. The metabolite 1,25(OH)2D3 increased ALP activity at day 3 (332%; p<0.05) and 10 (238%; p<0.05) compared to control cultures. The metabolite 25(OH)D3 increased ALP activity at day 3 (369%; p<0.05), 7 (326%; p<0.05) and 14 (146%; p<0.05) compared to control cultures. P1NP secretion was decreased by 1,25(OH)2D3 (Fig. 2C) at day 10 (47%; p<0.05) and 14 (65%; p<0.05) compared to control cultures, but P1NP secretion was not significantly affected by 25(OH)D3 (Fig. 2D). Osteocalcin secretion was markedly enhanced by both 1,25(OH)2D3 (Fig. 2E) and 25(OH)D3 (Fig. 2F). The metabolite 1,25(OH)2D3 stimulated the secretion at day 3 (p<0.05), whereas 25(OH)D3 increased the secretion at day 10 (p<0.05).

thumbnail
Figure 2. Effects of 1,25(OH)2D3 and 25(OH)D3 on ALP activity, P1NP and osteocalcin secretion by primary human osteoblast.

Osteoblasts were cultured in the presence of 0 or 100 nM 1,25(OH)2D3 and 0 or 400 nM 25(OH)D3 and ALP activity (A and B respectively), P1NP (C and D respectively) and osteocalcin secretion (E and F respectively) were measured at day 3, 7, 10 and 14 of the differentiation. Results are expressed as mean ± SEM using cells from 5 different donors. Results were analyzed using Wilcoxon signed rank test for each timepoint (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g002

Figure 3 demonstrates effects of 1,25(OH)2D3 or 25(OH)D3 on mRNA levels of genes involved in primary human osteoblast differentiation. ALP mRNA levels (Fig. 3A) were stimulated by both metabolites. The metabolite 1,25(OH)2D3 increased ALP mRNA levels at a concentration of 100 nmol/L (203%; p<0.01) and 25(OH)D3 increased ALP mRNA at a concentration of 200 nmol/L (191%; p<0.05) and 400 nmol/L (209%; p<0.05). Significant effects of 1,25(OH)2D3 or 25(OH)D3 on COL1α1 mRNA levels (Fig. 3B) were not observed. Osteocalcin mRNA was increased by 10 nmol/L (2147%; p<0.05) and 100 nmol/L 1,25(OH)2D3 (3289%; p<0.01) and by 200 nmol/L (2100%; p<0.01) and 400 nmol/L 25(OH)D3 (2102%; p<0.01). Osteopontin mRNA levels (Fig. 3D) were only significantly increased by 25(OH)D3 at a concentration of 400 nmol/L (314%; p<0.05).

thumbnail
Figure 3. Effects of 1,25(OH)2D3 and 25(OH)D3 on mRNA levels of genes involved in primary human osteoblast differentiation.

Osteoblasts were cultured in the presence of 0, 1, 10 and 100 nM 1,25(OH)2D3 or 0, 100, 200 or 400 nM 25(OH)D3 for 10 days in osteogenic medium and mRNA levels of ALP (A), COL1α1 (B), osteocalcin (C) and osteopontin (D) were determined. Results (mean ± SEM) are expressed as treatment versus control ratios (control was set at 1.0; dashed line) using cells from 5 different donors. Results were analysed using Friedman test followed by Dunn’s post hoc test (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g003

Effects of 1,25(OH)2D3 and 25(OH)D3 on VDR, CYP27B1 and CYP24 mRNA levels in primary human osteoblasts

Proliferation and differentiation experiments showed that primary human osteoblasts were able to respond to 1,25(OH)2D3 and 25(OH)D3. Therefore we examined effects of both metabolites on mRNA levels of VDR, and metabolizing enzymes, CYP27B1 and CYP24. VDR mRNA levels (Fig. 4A) increased in the presence of 100 nmol/L 1,25(OH)2D3 (162%; p<0.01) and 400 nmol/L 25(OH)D3 (149%; p<0.05). CYP27B1 mRNA (Fig. 4B) did not respond to either 1,25(OH)2D3 or 25(OH)D3. CYP24 mRNA levels (Fig. 4C) increased dose-dependently in response to 1,25(OH)2D3 at a concentration of 10 nmol/L (p<0.05) and 100 nmol/L (p<0.001). In response to 25(OH)D3, a significant increase of CYP24 mRNA was found at a concentration of 400 nmol/L (p<0.01).

thumbnail
Figure 4. Effects of 1,25(OH)2D3 and 25(OH)D3 on VDR, CYP27B1 and CYP24 mRNA levels in primary human osteoblast.

Osteoblasts were cultured in the presence of 0, 1, 10 and 100 nM 1,25(OH)2D3 or 0, 100, 200 and 400 nM 25(OH)D3 for 24 hours and mRNA levels of VDR (A), CYP27B1 (B) and CYP24 (C) were determined. Results (mean ± SEM) are expressed as treatment versus control ratios (control was set at 1.0; dashed line) using cells from 5 or 6 different donors. Results were analysed using Friedman test followed by Dunn’s post hoc test (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g004

Synthesis of 1,25(OH)2D3 and 24R,25(OH)2D3 by primary human osteoblasts

Primary human osteoblasts were cultured in the presence of increasing concentrations of 25(OH)D3 to study the conversion to 1,25(OH)D3 and 24R,25(OH)2D3. After 24 hours incubation with 100, 200, 400 and 1.000 nmol/L 25(OH)D3, levels of 25(OH)D3 were strongly reduced to respectively 16%, 20%, 29% and 33% of non-conditioned values (Fig. 5A). The metabolite 1,25(OH)2D3 (Fig. 5B) was produced in a dose-dependent manner after 25(OH)D3 treatment. Mean concentrations of 1,25(OH)2D3 in medium were 8.8, 41.7, 62.3, 125.6 and 197.3 pmol/L after 24 hours incubation of cells with respectively 0, 100, 200, 400 and 1.000 nmol/L 25(OH)D3. In non-conditioned medium, 1,25(OH)2D3 concentrations ranging from 3.3–60.8 pmol/L were measured. The metabolite 24R,25(OH)2D3 (Fig. 5C) was also produced in a dose-dependent manner after 25(OH)D3 treatment. Mean concentrations of 24R,25(OH)2D3 in medium were <3, 16.1, 45.3, 70.2 and 105.4 nmol/L after 24 hours incubation of cells with respectively 0, 100, 200, 400 and 1.000 nmol/L 25(OH)D3. In non-conditioned medium, 24R,25(OH)2D3 was not detected (<3 nmol/L).

thumbnail
Figure 5. Synthesis of 1,25(OH)2D3 and 24R,25(OH)2D3 by primary human osteoblasts.

Osteoblasts were cultured in the presence of 0, 100, 200, 400 and 1.000 nM 25(OH)D3 for 24 hours and 25(OH)D3 (A) 1,25(OH)2D3 (B) and 24R,25(OH)2D3 (C) levels were measured in non-conditioned and conditioned culture medium. Results are expressed as mean ± SEM using cells from 3 different donors.

https://doi.org/10.1371/journal.pone.0110283.g005

Effects of 25(OH)D3 on mRNA levels of genes involved in primary human osteoblast differentiation after CYP27B1 silencing

Silencing of CYP27B1 gene expression was used to examine whether 25(OH)D3 can directly act on osteoblast function. Treatment with CYP27B1 siRNA resulted in a 58% reduction of CYP27B1 mRNA compared to the control culture (p<0.05) (Fig. 6A). After 25(OH)D3 treatment, the reduction of CYP27B1 mRNA resulted in a decreased 1,25(OH)2D3 synthesis of 30% (p<0.05) compared to the control culture (Fig. 6B). Levels of 24R,25(OH)2D3 (Fig. 6C) and 25(OH)D3 (Fig. 6D) did not change in silenced and control cultures. After 72 hours of 25(OH)D3 treatment, the reduction of CYP27B1 mRNA was still 62% in the absence of 25(OH)D3 and 45% in the presence of 25(OH)D3 (Fig. 6E). No significant differences were seen between mRNA levels of CYP24 (Fig. 6F), VDR (Fig. 6G), ALP (Fig. 6 H), osteocalcin (Fig. 6I) and osteopontin (Fig. 6J) in control and CYP27B1-silenced cells that were exposed to 25(OH)D3.

thumbnail
Figure 6. Effects of 25(OH)D3 on mRNA levels of genes involved in primary human osteoblast differentiation after CYP27B1 silencing.

CYP27B1-silenced and control cells were incubated in the presence of 0 or 400 nM 25(OH)D3 for 3 days. CYP27B1 knock down was determined before 25(OH)D3 treatment (A). After 72 hours incubation with 25(OH)D3, we examined levels of 1,25(OH)2D3 (B), 24R,25(OH)2D3 (C) and 25(OH)D3 (D), and mRNA levels of CYP27B1 (E), CYP24 (F), VDR (G), ALP (H), osteocalcin (I) and osteopontin (J) in CYP27B1-silenced and control cells. Results are expressed as mean ± SEM using cells from 5 different donors. Results were analysed using Friedman test followed by Dunn’s post hoc test (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g006

Effects of 24R,25(OH)2D3 on osteoblast differentiation

In addition to 1,25(OH)2D3, we showed that osteoblasts are able to synthesize 24R,25(OH)2D3 from the precursor 25(OH)D3. To examine whether 24R,25(OH)2D3 can act on osteoblast differentiation, primary human osteoblasts were cultured in the presence of 24R,25(OH)2D3. After 72 hours, 24R,25(OH)2D3 did not affect COL1α1 mRNA levels (Fig. 7A), but 400 nmol/L 24R,25(OH)2D3 increased ALP (137%; p<0.05) (Fig. 7B), osteocalcin (6182%; p<0.01) (Fig. 7C) and osteopontin (387%; p<0.05) (Fig. 7D) mRNA levels.

thumbnail
Figure 7. Effects of 24R,25(OH)2D3 on mRNA levels of genes involved in osteoblast differentiation.

Osteoblasts were cultured in the presence of 0, 100, 200 or 400 nM 24R,25(OH)2D3 and mRNA levels of COL1α1 (A), ALP (B), osteocalcin (C) and osteopontin (D) were determined after 72 hours. Results (mean ± SEM) are expressed as treatment versus control ratios (control was set at 1.0; dashed line) using cells from 4 different donors. Results were analysed using Friedman test followed by Dunn’s post hoc test (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g007

Effects of 24R,25(OH)2D3 on VDR, CYP27B1 and CYP24 mRNA levels in primary human osteoblasts

We did not observe effects of 24R,25(OH)2D3 on mRNA levels of VDR (Fig. 8A) and CYP27B1 (Fig. 8B). The metabolite 24R,25(OH)2D3 highly induced CYP24 mRNA (Fig. 8C) levels in cells treated with 400 nmol/L 24R,25(OH)2D3 (p<0.001).

thumbnail
Figure 8. Effects of 24R,25(OH)2D3 on VDR, CYP27B1 and CYP24 mRNA levels in primary human osteoblasts.

Osteoblasts were cultured in the presence of 0, 100, 200 or 400 nM 24R,25(OH)2D3 and mRNA levels of VDR (A), CYP27B1 (B) and CYP24 (C) were determined after 72 hours. Results (mean ± SEM) are expressed as treatment versus control ratios (control was set at 1.0; dashed line) using cells from 4 different donors. Results were analysed using Friedman test followed by Dunn’s post hoc test (*p<0.05, **p<0.01, ***p<0.001).

https://doi.org/10.1371/journal.pone.0110283.g008

Discussion

This in vitro study shows the response of primary human osteoblasts to 25(OH)D3, 1,25(OH)2D3 and 24R,25(OH)2D3. Primary human osteoblasts responded to 25(OH)D3 by reducing their proliferation and enhancing their differentiation, similarly to 1,25(OH)2D3. We hypothesized that these 25(OH)D3 actions on osteoblast function occurred not only through hydroxylation to 1,25(OH)2D3, but possibly also through hydroxylation to 24R,25(OH)2D3. We could demonstrate that primary human osteoblasts expressed CYP27B1 and CYP24 and were capable to synthesize respectively 1,25(OH)2D3 as well as 24R,25(OH)2D3 from 25(OH)D3. Moreover, we showed that 24R,25(OH)2D3 increased mRNA levels of genes involved in primary human osteoblast differentiation.

The prohormone 25(OH)D3 has comparable effects to 1,25(OH)2D3 on growth and differentiation of primary osteoblasts. The metabolites 25(OH)D3 and 1,25(OH)2D3 reduced osteoblast proliferation and stimulated the differentiation as shown by increasing ALP activity (mRNA and protein) and osteocalcin secretion (mRNA and protein). Our study confirms previous studies in human osteoblastic cell lines and primary osteoblasts [12], [13], [20].

The effects of 25(OH)D3 on proliferation and differentiation likely occur through hydroxylation to 1,25(OH)2D3 [12], [13], since we and others demonstrated that osteoblasts are able to synthesize the active metabolite 1,25(OH)2D3 after exposure to the precursor 25(OH)D3 [12][14]. The consideration that effects of 25(OH)D3 occur through conversion to 1,25(OH)2D3 is supported by several in vitro blocking studies. In CYP27B1-silenced HOS cells, a human osteoblast cell line, it has been shown that exposure to 25(OH)D3 leads to a decline of osteonectin and CYP24 mRNA expression compared to control cells [21]. In human marrow stromal cells differentiated to osteoblasts, CYP27B1 is reported to be necessary for the antiproliferative and prodifferentiation effects of 25(OH)D3 [20]. Furthermore, in SV-HFO osteoblasts, ketoconazole almost complete blocked the effects of 25(OH)D3 on osteocalcin mRNA levels [12].

In our study, CYP27B1-silencing resulted in a decline of 1,25(OH)2D3 synthesis by primary osteoblasts. Despite this reduction, no significant differences in mRNA levels of differentiation markers were seen in CYP27B1-silenced cells compared to control cells after treatment with 25(OH)D3. It is likely that CYP27B1-silenced cells produced sufficient 1,25(OH)2D3 to induce a response. It is also possible that 25(OH)D3 affected osteoblast function through hydroxylation to 24R,25(OH)2D3. Levels of 24R,25(OH)2D3 were present in control and silenced cultures. Moreover, we showed that osteoblast cultures exposed to 24R,25(OH)2D3 had increased mRNA levels of ALP, osteocalcin and osteopontin. These results indicate a role for 24R,25(OH)2D3 in osteoblast differentiation. This is in line with previous research in the human osteoblast cell line SV-HFO in which 24R,25(OH)2D3 stimulated ALP activity and osteocalcin secretion by binding to the VDR [17]. Our results are also supported by a study in human mesenchymal stem cells, in which 24R,25(OH)2D3 enhances the osteoblastic differentiation by increasing ALP activity, osteocalcin mRNA levels and calcium mineralization of matrix [22]. In addition, our results are supported by a study in primary human osteoblasts that found increased osteocalcin production after 24R,25(OH)2D3 treatment [23]. However, due to incomplete CYP27B1 knockdown, 24R,25(OH)2D3 effects may be caused by 1α-hydroxylation to 1,24R,25(OH)3D3. The strong reduction of 25(OH)D3 levels in medium supports the idea that also other metabolites than 1,25(OH)2D3 and 24R,25(OH)2D3 are formed, for example 1,24R,25(OH)3D3. The metabolite 1,24R,25(OH)3D3 is able to enhance ALP activity, osteocalcin production and mineralization by SV-HFO osteoblasts [17]. In addition, 1,24R,25(OH)3D3 is even more potent than 24R,25(OH)2D3 [17].

In addition to the actions of 24R,25(OH)2D3 on mRNA levels of differentiation genes, 24R,25(OH)2D3 was also able to markedly enhance mRNA levels of CYP24. This may result in a higher production of 24R,25(OH)2D3 which suggests that 24R,25(OH)2D3 has the ability to regulate its own synthesis in a positive way (positive feedback). This is not in line with research performed in human mesenchymal stem cells differentiated to osteoblasts [22]. These osteoblasts decrease their CYP24 mRNA levels in response to 24R,25(OH)2D3 (at a concentration of 10 nmol/L) [22]. Added concentrations of 24R,25(OH)2D3 may explain the opposite results, since our results were obtained by using high 24R,25(OH)2D3 concentrations (400 nmol/L). Furthermore, we showed that 24R,25(OH)2D3, as well as 25(OH)D3 and 1,25(OH)2D3, had no effect on CYP27B1 mRNA levels. In human mesenchymal stem cells differentiated to osteoblasts, 24R,25(OH)2D3 (at a concentration of 10 nmol/L) decreases CYP27B1 mRNA and 1,25(OH)2D3 synthesis [22].

Actions of 24R,25(OH)2D3 can take place by activating the nuclear VDR [17], although the binding affinity of 24R,25(OH)2D3 to the VDR is 100 times less than 1,25(OH)2D3 [24]. In our study, 24R,25(OH)2D3 did not affect VDR mRNA, while 25(OH)D3 and 1,25(OH)2D3 increased VDR mRNA. Effects of 24R,25(OH)2D3 effects may be cell and concentration dependent, because at lower concentrations (10 nmol/L), 24R,25(OH)2D3 can decrease VDR mRNA and protein in human mesenchymal stem cells differentiated to osteoblasts [22].

The metabolite 25(OH)D3 itself may also be capable to activate the VDR and to subsequently affect mRNA levels of differentiation genes. This is supported by the in vivo study of Rowling et al [25] that supraphysiological levels of 25(OH)D3 can affect calcium and bone metabolism in the absence of its hydroxylation to 1,25(OH)2D3. In this study CYP27B1 knock out mice were fed a diet high in cholecalciferol which prevented hypocalcemia and almost rescued skeletal growth [25]. Several in vitro studies also support the hypothesis that 25(OH)D3 has direct effects on cells. Curtis [22] showed that 25(OH)D3 stimulates osteoblast mineralization in the presence of the cytochrome P450 inhibitor ketoconazole. Lou et al [26] showed that 25(OH)D3 is an agonistic VDR ligand and has direct inhibitory effects on proliferation in human LNCaP prostate cancer cells. In bovine parathyroid cells, 25(OH)D3 suppressed PTH secretion while 1α-hydroxylase was inhibited by clotrimazole. Therefore further studies are needed to clarify whether 25(OH)D3 can directly affect primary human osteoblasts. Although 25(OH)D3 may activate the VDR, the binding affinity to the VDR is less for 25(OH)D3 compared to 1,25(OH)2D3. Bouillon [24] reported that the binding affinity for 25(OH)D3 to the VDR is 50 times less than 1,25(OH)2D3.

In bone tissue, 25(OH)D3 metabolism may be beneficial since it is thought that locally synthesized 1,25(OH)2D3 supports osteoblast differentiation and matrix mineralization [12], [13], [27]. Serum 25(OH)D3 levels serve as substrate for local 25(OH)D3 metabolism and an adequate vitamin D status may therefore be essential [28]. In addition, low 25(OH)D3 serum levels in the range of deficiency may be a limiting factor for the synthesis of 1,25(OH)2D3 [29] and 24R,25(OH)2D3, and may result in reduced osteoblast differentiation and thereby a reduction of bone strength.

A limitation of this study is that complete blocking of the 1,25(OH)2D3 synthesis was not achieved in the RNA-silencing experiments. Therefore the question whether 25(OH)D3 itself is able to affect osteoblast function, can not be answered. Additional research is needed to achieve completely blocking of 1,25(OH)2D3 synthesis, for example studies with osteoblasts isolated from bone from CYP27B1 knock out mice. Furthermore, a critical point in our primary osteoblast cell culture model is the use of relatively high concentrations of 25(OH)D3, 1,25(OH)2D3 and 24R,25(OH)2D3 compared to normal serum levels in humans. Our concentrations of vitamin D metabolites were based on other studies in literature [12], [13], but effects of physiological levels of vitamin D metabolites on bone formation may be different. Lastly, in non-conditioned medium relatively high 1,25(OH)2D3 levels were measured. These levels of 1,25(OH)2D3 levels in non-conditioned medium are probably caused by cross-reactivity with 25(OH)D3 because of the high doses of 25(OH)D3 used in this study. However, our study clearly showed increased 1,25(OH)2D3 concentrations in conditioned medium compared to non-conditioned medium, which demonstrates the synthesis of 1,25(OH)2D3 by osteoblasts.

In conclusion, the vitamin D metabolites 25(OH)D3, 1,25(OH)2D3 and 24R,25(OH)2D3 can affect osteoblast differentiation directly or indirectly. The metabolite 25(OH)D3 is converted to both 1,25(OH)2D3 and 24R,25(OH)2D3, as demonstrated by measurements in culture medium. We showed that primary human osteoblasts not only respond to 1,25(OH)2D3, but also to 24R,25(OH)2D3 by enhancing the differentiation. This suggests that 25(OH)D3 can affect osteoblast differentiation via conversion to the active metabolite 1,25(OH)2D3, but also via conversion to 24R,25(OH)2D3 (direct or indirect via 1,24R,25(OH)3D3). Whether 25(OH)D3 has direct actions on osteoblast differentiation needs further investigation.

Acknowledgments

We thank Huib van Essen for his valuable technical advice. We thank Niek Dirks for performing 24R,25(OH)2D3 and 25(OH)D3 measurements. We also thank the technicians of the department of Clinical Chemistry of the VU University Medical Center for performing P1NP, ALP, osteocalcin and 1,25(OH)2D3 measurements.

Author Contributions

Conceived and designed the experiments: KM NB PL. Performed the experiments: KM. Analyzed the data: KM NB PL. Contributed reagents/materials/analysis tools: KM EAJMS ACH NB. Contributed to the writing of the manuscript: KM PL MD ACH EAJMS MH NB.

References

  1. 1. Lips P, van Schoor NM (2011) The effect of vitamin D on bone and osteoporosis. Best Pract Res Clin Endocrinol Metab 25: 585–591.
  2. 2. Zerwekh JE (2008) Blood biomarkers of vitamin D status. Am J Clin Nutr 87: 1087S–1091S. 87/4/1087S [pii].
  3. 3. Lips P (2006) Vitamin D physiology. Prog Biophys Mol Biol 92: 4–8 S0079-6107(06)00008-3 [pii];10.1016/j.pbiomolbio.2006.02.016 [doi].
  4. 4. St-Arnaud R (2011) CYP24A1: Structure, Function, and Physiological Role. In: Feldman D, Pike JW, Adams JS, editors. Vitamin D (Third edition). San Diego: Academic Press. 43–56.
  5. 5. Slominski AT, Kim TK, Li W, Yi AK, Postlethwaite A, et al. (2014) The role of CYP11A1 in the production of vitamin D metabolites and their role in the regulation of epidermal functions. J Steroid Biochem Mol Biol 144PA: 28–39 S0960-0760(13)00205-7 [pii];10.1016/j.jsbmb.2013.10.012 [doi].
  6. 6. Slominski AT, Kim TK, Shehabi HZ, Semak I, Tang EK, et al. (2012) In vivo evidence for a novel pathway of vitamin D3 metabolism initiated by P450scc and modified by CYP27B1. FASEB J 26: 3901–3915 fj.12-208975 [pii];10.1096/fj.12-208975 [doi].
  7. 7. Slominski AT, Kim TK, Shehabi HZ, Tang EK, Benson HA, et al. (2014) In vivo production of novel vitamin D2 hydroxy-derivatives by human placentas, epidermal keratinocytes, Caco-2 colon cells and the adrenal gland. Mol Cell Endocrinol 383: 181–192 S0303-7207(13)00527-3 [pii];10.1016/j.mce.2013.12.012 [doi].
  8. 8. Slominski AT, Kim TK, Takeda Y, Janjetovic Z, Brozyna AA, et al. (2014) RORα and RORγ are expressed in human skin and serve as receptors for endogenously produced noncalcemic 20-hydroxy- and 20,23-dihydroxyvitamin D. FASEB J. 28: 2775–2789 fj.13-242040 [pii];10.1096/fj.13-242040 [doi].
  9. 9. van Driel M, Pols HA, van Leeuwen JP (2004) Osteoblast differentiation and control by vitamin D and vitamin D metabolites. Curr Pharm Des 10: 2535–2555.
  10. 10. Beresford JN, Gallagher JA, Russell RG (1986) 1,25-Dihydroxyvitamin D3 and human bone-derived cells in vitro: effects on alkaline phosphatase, type I collagen and proliferation. Endocrinology 119: 1776–1785.
  11. 11. Chen FP, Lee N, Wang KC, Soong YK, Huang KE (2002) Effect of estrogen and 1α,25(OH)2- vitamin D3 on the activity and growth of human primary osteoblast-like cells in vitro. Fertil Steril 77: 1038–1043.
  12. 12. van Driel M, Koedam M, Buurman CJ, Hewison M, Chiba H, et al. (2006) Evidence for auto/paracrine actions of vitamin D in bone: 1α-hydroxylase expression and activity in human bone cells. FASEB J 20: 2417–2419.
  13. 13. Atkins GJ, Anderson PH, Findlay DM, Welldon KJ, Vincent C, et al. (2007) Metabolism of vitamin D3 in human osteoblasts: evidence for autocrine and paracrine activities of 1α,25-dihydroxyvitamin D3. Bone 40: 1517–1528.
  14. 14. Howard GA, Turner RT, Sherrard DJ, Baylink DJ (1981) Human bone cells in culture metabolize 25-hydroxyvitamin D3 to 1,25-dihydroxyvitamin D3 and 24,25-dihydroxyvitamin D3. J Biol Chem 256: 7738–7740.
  15. 15. Seo EG, Einhorn TA, Norman AW (1997) 24R,25-dihydroxyvitamin D3: an essential vitamin D3 metabolite for both normal bone integrity and healing of tibial fracture in chicks. Endocrinology 138: 3864–3872 10.1210/endo.138.9.5398 [doi].
  16. 16. St-Arnaud R (2010) CYP24A1-deficient mice as a tool to uncover a biological activity for vitamin D metabolites hydroxylated at position 24. J Steroid Biochem Mol Biol 121: 254–256 S0960-0760(10)00044-0 [pii];10.1016/j.jsbmb.2010.02.002 [doi].
  17. 17. van Driel M, Koedam M, Buurman CJ, Roelse M, Weyts F, et al. (2006) Evidence that both 1α,25-dihydroxyvitamin D3 and 24-hydroxylated D3 enhance human osteoblast differentiation and mineralization. J Cell Biochem 99: 922–935.
  18. 18. Beresford JN, Gallagher JA, Gowen M, McGuire MKB, Poser JW, et al. (1983) Human bone cells in culture: a novel system for the investigation of bone cell metabolism. Clinical Sciences 64: 33–39.
  19. 19. Marie PJ, Lomri A, Sabbagh A, Basle M (1989) Culture and behavior of osteoblastic cells isolated from normal trabecular bone surfaces. In Vitro Cell Dev Biol 25: 373–380.
  20. 20. Geng S, Zhou S, Glowacki J (2011) Effects of 25-hydroxyvitamin D3 on proliferation and osteoblast differentiation of human marrow stromal cells require CYP27B1/1α-hydroxylase. J Bone Miner Res 26: 1145–1153.
  21. 21. Anderson PH, Atkins GJ, Findlay DM, Oloughlin PD, Welldon K, et al. (2007) RNAi-mediated silencing of CYP27B1 abolishes 1,25(OH)2D3 synthesis and reduces osteocalcin and CYP24 mRNA expression in human osteosarcoma (HOS) cells. J Steroid Biochem Mol Biol 103: 601–605.
  22. 22. Curtis KM, Aenlle KK, Roos BA, Howard GA (2014) 24R,25-dihydroxyvitamin D3 promotes the osteoblastic differentiation of human mesenchymal stem cells. Mol Endocrinol 28: 644–658 10.1210/me.2013-1241 [doi].
  23. 23. Yamamoto T, Ozono K, Shima M, Yamaoka K, Okada S (1998) 24R,25-dihydroxyvitamin D3 increases cyclic GMP contents, leading to an enhancement of osteocalcin synthesis by 1,25-dihydroxyvitamin D3 in cultured human osteoblastic cells. Exp Cell Res 244: 71–76. S0014-4827(98)94189-1 [pii]; 10.1006/excr.1998.4189 [doi].
  24. 24. Bouillon R, Okamura WH, Norman AW (1995) Structure-function relationships in the vitamin D endocrine system. Endocr Rev 16: 200–257 10.1210/edrv-16-2-200 [doi].
  25. 25. Rowling MJ, Gliniak C, Welsh J, Fleet JC (2007) High dietary vitamin D prevents hypocalcemia and osteomalacia in CYP27B1 knockout mice. J Nutr 137: 2608–2615. 137/12/2608 [pii].
  26. 26. Lou YR, Molnar F, Perakyla M, Qiao S, Kalueff AV, et al.. (2010) 25-Hydroxyvitamin D3 is an agonistic vitamin D receptor ligand. J Steroid Biochem Mol Biol 118: 162–170. S0960-0760(09)00282-9 [pii]; 10.1016/j.jsbmb.2009.11.011 [doi].
  27. 27. Anderson PH, Atkins GJ (2008) The skeleton as an intracrine organ for vitamin D metabolism. Mol Aspects Med 29: 397–406.
  28. 28. Morris HA (2014) Vitamin D activities for health outcomes. Ann Lab Med 34: 181–186 10.3343/alm.2014.34.3.181 [doi].
  29. 29. Peterlik M, Cross HS (2005) Vitamin D and calcium deficits predispose for multiple chronic diseases. Eur J Clin Invest 35: 290–304. ECI1487 [pii]; 10.1111/j.1365-2362.2005.01487.x [doi].