Facial trauma or tumor surgery in the head and face area often lead to massive destruction of the facial skeleton 1. The reconstruction of damaged or lost bone is a clinical challenge in modern reconstructive surgery. The repair of bone defects still poses a significant problem for many clinicians. In the early decades of bone reconstruction surgeons used artificial tissue substitutes containing metals, ceramics, and polymers to maintain skeletal function 2. These artificial materials have facilitated surgeons to restore the form and – to some extent – the function of defective bones. Nevertheless, these artificial materials have specific disadvantages, and thus encouraged surgeons to develop alternative approaches including cell-based devices. Transplantation of autografts is a frequently used treatment strategy in routine clinical practice and has gained the "gold standard" in bone reconstructive surgery, despite donor site morbidity and donor shortage 3.

Modern cell-based bone reconstruction techniques may offer new therapeutic opportunities for the repair of bone damaged by disease or injury. Generally, the combination of scaffolds, bioactive factors, and living cells provides a surgically implantable product for use in tissue regeneration and functional restoration 45. Numerous attempts were undertaken with various success to restore bone defects by various biomaterials alone 678910 or in combination with bioactive cytokines such as bone morphogenetic protein (BMP)-7, BMP-2 or BMP-2-mutants 1112. Cell-based strategies in bone tissue engineering use different cell sources including autologous cells as well as allogenic and xenogenic cells 13141516. There are some reports that use totipotential embryonic stem cells in tissue engineering of bone 1718.

Embryonic stem cells (ESCs) are routinely derived from the inner cell mass of blastocysts and represent pluripotential embryonic precursor cells that give rise to all cell types in the developing organism. ESCs have historically been maintained in co-culture with mitotically inactive fibroblasts 192021. This co-culture system is unnecessary if the medium is supplemented with leukemia inhibitory factor (LIF) 2223. In the absence of LIF embryonic stem cells will differentiate into a morphologically mixed cell population expressing features of endoderm and mesoderm lineages 24. By definition ESCs have the potential to differentiate into osteogenic cells under selective culture conditions. Specifically, it has been shown by various investigators that ESCs can differentiate into osteogenic cells under selective culture conditions 171825. However, it is unclear which medium is most suitable to initiate osteogenic differentiation. BMP-2 and a mixture of dexamethasone, ascorbic acid and β-glycerophosphate (DAG) are good candidates 1925. Thus, we examined the time-dependent expression of the osteoblastic markers osteopontin 26, collagen I 27, alkaline phosphatase 28, and osteocalcin 29 in ESC cells.

In ESC cultures supplemented with DAG we found the largest deposition of calcium phosphate-containing minerals, as judged by von Kossa staining (Fig. 1). ESCs cultured in the presence of BMP-2 exhibited less mineralization, and there were no signs of mineral deposition in unstimulated control cells or cells stimulated with LIF.

In order to assess the differentiation of ESCs cultured under different conditions, we used the hematopoetic stem cell marker CD34. Only in ESC cultures without any additional stimulus (ESCs without LIF) the expected amplicon appeared in agarose-gel electrophoresis. ESCs which were differentiated with BMP-2 or DAG have downregulated this marker (Fig. 2).

Next the kinetics of gene expression in ESCs during differentiation and matrix formation were evaluated. The values were plotted as a multiple of the expression in the control group (ESCs with LIF). Expression of osteopontin was reduced in ESC treated with LIF as compared to all other samples (without LIF, with BMP-2 or with DAG). The low level of osteopontin mRNA synthesis persisted in the presence of DAG, and in BMP-2-treated cells showed a steep increase after 2.5 weeks of culture. ESCs without LIF showed similar expression rates as the DAG group (Fig. 3).

Starting at day 9 of culture, a strong increase in collagen I expression was recorded in the DAG culture, which was paralleled to a lesser extent by the collagen expression in the BMP-2-treated cells. After three weeks of culture the expression level of collagen I mRNA was similar in all groups of the differentially treated cells (Fig. 4). Only the DAG culture showed a second but smaller increase at day 23.

The transcription of mRNA coding for alkaline phosphatase was slightly increased in cells stimulated with BMP-2. ESCs exposed to DAG did not significantly differ from the control culture (Fig. 5). Expression of osteocalcin, which is regarded as a highly specific marker for osteoblasts, demonstrated showed a double-peaked curve in the DAG-treated cells. ESCs cultured in DAG-supplemented medium showed a prominent peak after 9 days and a second peak beginning at day 17. The first increase was also seen in ESCs cultured in the presence of BMP-2 or the absence of LIF. Interestingly, in all the differentially treated cells a second peak of osteocalcin transcription was observed 7 days later (Fig. 6). All three ESC cultures showed similar expression pattern of the hematopoetic stem cell marker CD34 (Fig. 7).

Currently, there are many efforts to establish cell-based strategies in bone tissue engineering. ESCs are one of many different cell populations, which are being tested for their feasibility for these treatment options. The purpose of this investigation was to determine which supplements in culture medium are most suitable to initiate osteogenic differentiation in ESC cultures. In addition, we investigated the kinetics of gene expression during in vitro differentiation.

The results of our microscopical analysis revealed that ESCs cultured in the presence of DAG show by far the highest extent of mineralisation as determined by the occurrence of calcium-phosphate-containing crystals. With respect to extracellular matrix maturation and mineral deposition as crucial steps in the osteogenic cascade 34, DAG seems to be the most promising supplement for inducing osteogenic differentiation in ESCs. In accordance with our microscopical results, a strong increase of collagen I expression was observed at day 11 in the DAG-treated cells. Stimulation with BMP-2 also increased collagen synthesis. Expression of osteocalcin mRNA followed a different pattern and appeared as a double-peaked curve, when ESCs were supplemented with osteogenic agents (DAG or BMP-2). However, the peak induction of osteocalcin mRNA in the BMP-2-treated cells was lower and delayed as compared to DAG-exposed cells. Taken together, these results support the use of DAG as a potent agent for inducing in vitro differentiation of ESCs into osteoblast-like cells.

There are only few reports addressing osteogenic differentiation of ESCs published in the literature so far 18253435. In agreement with these results we describe here that mineralisation is microscopically evident as early as two weeks of culture. Buttery and co-workers also used DAG as a culture supplement and found that mineralisation was detectable when dexamethasone was added only at day 14 or later 35. By following this protocol the differentiation process was delayed as compared to the findings in our ESC cultures. While Buttery used only microscopical methods for studying osteogenic differentiation, zur Nieden and colleagues performed also gene expression analyses for osteogenic markers 34. With respect to the time-course of gene expression with an early increase of collagen I and a later increase of osteocalcin transcripts, their data are comparable to our findings as shown above. Unlike to the findings of zur Nieden and colleagues, an early peak of osteocalcin expression and a minor increase of osteopontin were found in the presented study. The differences could be explained by different concentrations of supplements used for cell differentiation. Zur Nieden et al. used 1,25-OH vitamin D₃ instead of dexamethasone. According to Zhang et al. vitamin D₃ increases osteopontin expression in osteoblasts and inhibits expression of osteocalcin 36. Chaudhry and co-workers replaced dexamethasone with retinoid acid, which was found to be an inductor of mineralization in three-dimensional scaffolds 25. Notably, alkaline phosphatase was constitutively expressed at high levels in undifferentiated cells 37. In this experimental setting the mineralisation process was delayed and was detectable only after day 21. Treatment with DAG appeared to be equal or even superior to BMP-2 stimulation regarding the induction of osteogenic differentiation in ESCs. Other authors have used BMP-2 in combination with osteogenic supplements for this purpose 1838.

An advantage of using ESCs instead of tissue-derived progenitor cells is that ESCs are immortal and could potentially provide an unlimited supply of differentiated osteoblast and osteoprogenitor cells for transplantation. In contrast to embryonic cells, the proliferative, self-renewal and differentiation capacity of cells derived from adult tissues generally decreases with age 3940. One major challenge pointing to the use of ESCs lies in overcoming immunological rejection from the transplant recipient. Interestingly, Burt and colleagues performed ESC transplantation in major histocompatibility complex (MHC)-mismatched mice without clinical or histological evidence of graft-versus-host disease (GVHD) 21. In addition, recent data indicate that ESCs may allow for a low-risk induction of tolerance not requiring any immunosuppression 41.

In conclusion, ESCs differentiate into osteoblast-like cells in vitro when stimulated with DAG and showed a time-dependent induction of osteogenic markers. Thus, stimulation with these agents is suitable to generate a promising cell population used for bone tissue engineering.

The authors declare that they have no competing interests.

JH conceived the study, calculated the statistics and drafted the manuscript, KB carried out the cell culture and the gene expression analysis, HW helped to perform and evaluate the histological investigations, RD, CN, MO, NK and UM participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.