Stimulation of osteoblast activity by homocysteine

Homocysteine (HCY) has recently been linked to fragility fractures. Moreover, HCY activates osteoclasts. Little is known about the effect of HCY on activity of human osteoblasts (OBs). We hypothesized that HCY decreases the activity of OBs. Osteoblasts obtained from tra-becular human bone specimens of eight donors were cultured with conditioned medium. Culture medium was adjusted to 0, 100, 500, 1000 and 2000 μM HCY. After 14 days alkaline phosphatase (AP) activity, pro-collagen type I N-terminal peptide (PINP) and osteocalcin (OC) secretion in the supernatant were measured. After 20 days the formation of mineralized matrix was analyzed. HCY-stimulated AP activity gradually (100 μM HCY: 118%, P= 0.006; 500 μM HCY: 125%, P < 0.001). At 1000 and 2000 μM HCY the increase of AP activity was reversible (1000 μM HCY: 106%, P= 0.317; 2000 μM HCY: 102%, P < 0.737). The PINP secretion was also stimulated by HCY reaching a maximum of 260 ± 154 μg/l at 500 μmol/l versus 205 ± 94 μ,g/l in controls. After 20 days of culture the formation of bone matrix was increased at 100 and 500 μM HCY. OC secretion was not significantly changed. The results of the present study consistently demonstrate a moderate stimulation of primary human OB activity by increasing concentrations of HCY. However, the magnitude of this effect seems to be less pronounced than recent observations on primary human osteoclasts, suggesting a dysbalance between OBs and osteoclasts in favour of osteoclasts

diet, a 2% homocystine-enriched diet or a control diet for 3 months. Mean plasma HCY was approximately 30 mol/l in the methioninegroup, 50 mol/l in the homocystine-group and <10 mol/l in controls. At the end of the treatment period, the hyperhomocysteinemic animals exhibited a reduction of bone strength (axial compression of the femoral neck until fracture) by up to 40% and a drastic removal of trabecular bone (up to 90%). However, the underlying mechanisms remain obscure. Several studies on human beings observed a correlation between HCY and circulating concentrations of biochemical bone resorption markers, such as desoxypyridinoline cross-links (DPD) [16,17] and C-terminal telopeptides of collagen type I (ICTP) [18]. First experimental data from our group [19] and Koh et al. [20] demonstrate a stimulation of human osteoclasts activity by increasing concentrations of HCY. In addition, extracellular collagen cross-linking, which is important for bone stability, might be disturbed [20][21][22]. However, little is known about the effect of HCY on osteoblast (OB) activity. In hyperhomocysteinemic animals circulating osteocalcin (OC) concentrations have been found to be 40% lower than in controls [15], indicating a reduced OB activity. A reduction of OC expression by HCY could be confirmed in cell culture experiments performed with MC3T3-E1 preosteoblastic cells [23]. Another recent cell culture study with human bone marrow stromal cells demonstrated an enhanced apoptosis by HCY stimulation [24]. Contrary to these experimental observations, in human beings HCY positively correlates with the circulating concentration of the OB activity marker OC [16].
Based on existing data, we hypothesized that HCY decreases the activity of primary human OB. Accordingly, we analyzed the influence of increasing HCY concentrations on the activity of primary human OB in vitro.

PINP secretion in the supernatant was measured with a chemiluminescence immunoassay (Roche Diagnostics) on an Elecsys 2010 automated analyzer (Roche Diagnostics). This sandwich assay uses two monoclonal anti PINP antibodies. The first anti-PINP antibody is bound to biotin and fixes PINP of the sample to streptavidin coated micro particles (solid phase). The second antibody is marked with a Tris(2,2'bipyridyl)ruthenium(II) complex.
Magnetic forces bind the micro particles to the surface of an electrode. After washing, the chemiluminescence emission is induced electrically and measured by use of a photomultiplier. Intra-and inter-assay imprecision of this assay are 2.1 and 2.4% at 270 g/l, and 2.9 and 3.7% at 800 g/l.

OC, another well-established bone formation marker, was quantified with the N-MID OC assay (Roche Diagnostics) on an Elecsys 2010 automated analyzer (Roche Diagnostics). This sandwich assay uses two monoclonal antibodies against the 1-43 fragment of OC and detects intact OC as well as the stable 1-43 fragment. The first anti-OC antibody is bound to biotin and fixes OC of the sample to streptavidin-coated micro particles (solid phase).
The second antibody is marked with a Tris(2,2'bipyridyl)ruthenium(II) complex. Magnetic forces bind the micro particles to the surface of an electrode. After washing, the chemiluminescence emission is induced electrically and measured by use of a photomultiplier. Intra-and inter-assay imprecision of this assay are 4.0 and 6.5% at 15.5 g/l.

Mineralization assay
For the mineralization assay OB were transferred at a density of 10,000 cells cm 2 to a 48 multi-well plate and were cultured for 20 days in ␣-MEM supplemented with 10% FCS, 50 g/ml ascorbic acid (Sigma-Aldrich), 10 mM ␤-glycerolphosphate (Sigma-Aldrich) and 10

Statistical analysis
The descriptive statistics provide data as means ± SD. Means were compared by use of a student's t-test (comparison of two means) or a one-way ANOVA (more than two means) with an LSD (least significant difference) post hoc test. Cultures without addition of HCY were used as controls.

AP activity
AP activity exhibited a noticeable inter-individual variation, which was mainly due to differences in growth behaviour. After

(B) Mean (95% confidence interval) concentration of pro-collagen type I N-terminal peptide in the supernatant. (C) Mean (95% confidence interval) concentration of osteocalcin in the supernatant. P: P-value of one-way ANOVA representing the variation of the group means around the overall mean. (*)P < 0.10 (LSD post hoc test) versus control cultures without HCY, *P < 0.05 (LSD post hoc test) versus control cultures without HCY and **P < 0.001 (LSD post hoc test) versus control cultures without HCY.
HCY stimulated AP activity gradually, reaching a maximum of 125% at 500 mol/l (Fig. 1A). At higher HCY concentrations the increase in AP was reversible, which is probably due to toxic effects of HCY.

PINP and OC secretion
In order to confirm the AP results we measured PINP and OC secretion in the supernatant. The mean PINP concentration in the supernatant of control cultures was 205 ± 94 g/l. HCY concentrations of 100, 500 and 1000 mol/l HCY induced a gradual increase of PINP with a maximum of 260 ± 154 g/l at 500 mol/l (Fig. 1B). This corresponds to an increase of 27%, which is comparable with the increase of AP.
In cultures without HCY addition mean OC was 7.4 ± 0.8 g/l, indicating a low variability between donors. In contrast to AP and PINP, the addition of HCY to the culture medium did not change OC secretion (Fig. 1C).

Formation of mineralized matrix
As shown in Fig. 2, addition of HCY results in an increased formation of mineralized matrix. This effect was already evident at 100 mol/l. At higher HCY concentrations this effect was reversible.

Discussion
The main finding of this study is a moderate stimulation of primary human OB activity by HCY, as indicated by a stimulation of AP activity, PINP secretion and mineralized matrix formation. The maximum increase of OB activity was about 25% and could be observed at a HCY concentration of 500 mol/l. The strength of our results is due to a high number of repetitions, cells from different donors and the use of well-standardized quantitative measures.

and OC is mainly expressed during the matrix mineralization (starting around day 20). Consequently, the OB activity markers used in this study reflect different aspects of OB function. Since the supernatants were sampled at day 14, the missing effects regarding OC can possibly be explained by the later onset of OC expression during the mineralization phase. However, AP and PINP are two well-established markers of bone formation, reflecting OB proliferation and matrix synthesis. Both markers consistently showed a HCY-induced stimulation of OB activity. The results obtained with AP and PINP could be confirmed by the nodule formation assay.
Since experimental studies regarding HCY and bone formation are rare, comparison of our results with other studies is difficult. Our own animal study [15] and a cell culture study performed with MC3T3-E1 preosteoblastic cells [23]  Osteoporosis is characterized by a dysbalance between OB and osteoclasts [29]. Previous studies by Koh et al. and ourselves consistently demonstrated a strong stimulation of osteoclasts by HCY [19,20]. The maximum increase of tartrate-resistant acid phosphatase (TRAP) activity was about 50%. In recent cell culture experiments we could confirm a HCY-induced stimulation of TRAP activity by 50%. Moreover, we demonstrated a increase in bone resorption activity up to 400% [30]. The stimulation of OB activity in the present study showed a maximum of 25%, indicating that the HCY-induced stimulation of OB activity is less pronounced than those observed for osteoclasts. Existing data suggests a dysbalance between OB and osteoclasts in favour of osteoclasts as a major pathomechanism for HCY-induced bone loss and reduced bone quality in vivo [15].
The main reasons for HHCY in adults are deficiencies of folate, vitamin B12 and B6 as well as an impaired renal function. Current results from our group showed that B-vitamin deficiencies also stimulate OC activity and lead to elevated HCY concentrations in the supernatant [30]. Bone resorption activity increased up to 200%. This observation is another strong hint for a mainly osteoclast driven deterioration of bone by HHCY. In addition, Kim et al. showed a suppression of OB activity in the presence of low vitamin B12 concentrations.
Another potential mechanism involved in HHCY-related reduction of bone quality is a disturbed cross-linking of collagen fibrils [21]. In a first study on human beings, Saito et al. compared 25 female fracture cases with 25 post-mortem controls and found higher circulating HCY concentrations (~2 mol/l) and significantly lower enzymatic cross-links in the bone tissue [31]. However, these results need to be confirmed by others. The potential clinical impact of HHCY for bone health can be deduced from a current large-scale intervention trial by Sato et al. [32]. In a population of stroke patients, a group at high risk for fragility fractures, a 2-year HCY lowering treatment by supplementation of folate and vitamin B12 resulted in a 75% decrease of hip fractures and overall fractures.
In conclusion, the results of the present study consistently demonstrate a moderate stimulation of primary human OB activity by increasing concentrations of HCY. The results are based on a high number of repetitions, cells from several donors and wellstandardized analytic tools. However, the magnitude of this effect seems to be less pronounced than recent observations on primary human osteoclasts, suggesting a dysbalance between OB and osteoclasts in favour of osteoclasts. Future studies need to clarify if the effects obtained by stimulation with HCY can be confirmed by decreasing concentrations of folate, vitamin B12 and B6.